155714Skris 255714Skris OpenSSL CHANGES 355714Skris _______________ 455714Skris 5285330Sjkim Changes between 1.0.1o and 1.0.1p [9 Jul 2015] 6285330Sjkim 7285330Sjkim *) Alternate chains certificate forgery 8285330Sjkim 9285330Sjkim During certificate verfification, OpenSSL will attempt to find an 10285330Sjkim alternative certificate chain if the first attempt to build such a chain 11285330Sjkim fails. An error in the implementation of this logic can mean that an 12285330Sjkim attacker could cause certain checks on untrusted certificates to be 13285330Sjkim bypassed, such as the CA flag, enabling them to use a valid leaf 14285330Sjkim certificate to act as a CA and "issue" an invalid certificate. 15285330Sjkim 16285330Sjkim This issue was reported to OpenSSL by Adam Langley/David Benjamin 17285330Sjkim (Google/BoringSSL). 18285330Sjkim [Matt Caswell] 19285330Sjkim 20284330Sjkim Changes between 1.0.1n and 1.0.1o [12 Jun 2015] 21284330Sjkim 22284330Sjkim *) Fix HMAC ABI incompatibility. The previous version introduced an ABI 23284330Sjkim incompatibility in the handling of HMAC. The previous ABI has now been 24284330Sjkim restored. 25284330Sjkim 26284285Sjkim Changes between 1.0.1m and 1.0.1n [11 Jun 2015] 27284285Sjkim 28284285Sjkim *) Malformed ECParameters causes infinite loop 29284285Sjkim 30284285Sjkim When processing an ECParameters structure OpenSSL enters an infinite loop 31284285Sjkim if the curve specified is over a specially malformed binary polynomial 32284285Sjkim field. 33284285Sjkim 34284285Sjkim This can be used to perform denial of service against any 35284285Sjkim system which processes public keys, certificate requests or 36284285Sjkim certificates. This includes TLS clients and TLS servers with 37284285Sjkim client authentication enabled. 38284285Sjkim 39284285Sjkim This issue was reported to OpenSSL by Joseph Barr-Pixton. 40284285Sjkim (CVE-2015-1788) 41284285Sjkim [Andy Polyakov] 42284285Sjkim 43284285Sjkim *) Exploitable out-of-bounds read in X509_cmp_time 44284285Sjkim 45284285Sjkim X509_cmp_time does not properly check the length of the ASN1_TIME 46284285Sjkim string and can read a few bytes out of bounds. In addition, 47284285Sjkim X509_cmp_time accepts an arbitrary number of fractional seconds in the 48284285Sjkim time string. 49284285Sjkim 50284285Sjkim An attacker can use this to craft malformed certificates and CRLs of 51284285Sjkim various sizes and potentially cause a segmentation fault, resulting in 52284285Sjkim a DoS on applications that verify certificates or CRLs. TLS clients 53284285Sjkim that verify CRLs are affected. TLS clients and servers with client 54284285Sjkim authentication enabled may be affected if they use custom verification 55284285Sjkim callbacks. 56284285Sjkim 57284285Sjkim This issue was reported to OpenSSL by Robert Swiecki (Google), and 58284285Sjkim independently by Hanno B�ck. 59284285Sjkim (CVE-2015-1789) 60284285Sjkim [Emilia K�sper] 61284285Sjkim 62284285Sjkim *) PKCS7 crash with missing EnvelopedContent 63284285Sjkim 64284285Sjkim The PKCS#7 parsing code does not handle missing inner EncryptedContent 65284285Sjkim correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs 66284285Sjkim with missing content and trigger a NULL pointer dereference on parsing. 67284285Sjkim 68284285Sjkim Applications that decrypt PKCS#7 data or otherwise parse PKCS#7 69284285Sjkim structures from untrusted sources are affected. OpenSSL clients and 70284285Sjkim servers are not affected. 71284285Sjkim 72284285Sjkim This issue was reported to OpenSSL by Michal Zalewski (Google). 73284285Sjkim (CVE-2015-1790) 74284285Sjkim [Emilia K�sper] 75284285Sjkim 76284285Sjkim *) CMS verify infinite loop with unknown hash function 77284285Sjkim 78284285Sjkim When verifying a signedData message the CMS code can enter an infinite loop 79284285Sjkim if presented with an unknown hash function OID. This can be used to perform 80284285Sjkim denial of service against any system which verifies signedData messages using 81284285Sjkim the CMS code. 82284285Sjkim This issue was reported to OpenSSL by Johannes Bauer. 83284285Sjkim (CVE-2015-1792) 84284285Sjkim [Stephen Henson] 85284285Sjkim 86284285Sjkim *) Race condition handling NewSessionTicket 87284285Sjkim 88284285Sjkim If a NewSessionTicket is received by a multi-threaded client when attempting to 89284285Sjkim reuse a previous ticket then a race condition can occur potentially leading to 90284285Sjkim a double free of the ticket data. 91284285Sjkim (CVE-2015-1791) 92284285Sjkim [Matt Caswell] 93284285Sjkim 94284285Sjkim *) Reject DH handshakes with parameters shorter than 768 bits. 95284285Sjkim [Kurt Roeckx and Emilia Kasper] 96284285Sjkim 97280304Sjkim Changes between 1.0.1l and 1.0.1m [19 Mar 2015] 98280304Sjkim 99280304Sjkim *) Segmentation fault in ASN1_TYPE_cmp fix 100280304Sjkim 101280304Sjkim The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is 102280304Sjkim made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check 103280304Sjkim certificate signature algorithm consistency this can be used to crash any 104280304Sjkim certificate verification operation and exploited in a DoS attack. Any 105280304Sjkim application which performs certificate verification is vulnerable including 106280304Sjkim OpenSSL clients and servers which enable client authentication. 107280304Sjkim (CVE-2015-0286) 108280304Sjkim [Stephen Henson] 109280304Sjkim 110280304Sjkim *) ASN.1 structure reuse memory corruption fix 111280304Sjkim 112280304Sjkim Reusing a structure in ASN.1 parsing may allow an attacker to cause 113280304Sjkim memory corruption via an invalid write. Such reuse is and has been 114280304Sjkim strongly discouraged and is believed to be rare. 115280304Sjkim 116280304Sjkim Applications that parse structures containing CHOICE or ANY DEFINED BY 117280304Sjkim components may be affected. Certificate parsing (d2i_X509 and related 118280304Sjkim functions) are however not affected. OpenSSL clients and servers are 119280304Sjkim not affected. 120280304Sjkim (CVE-2015-0287) 121280304Sjkim [Stephen Henson] 122280304Sjkim 123280304Sjkim *) PKCS7 NULL pointer dereferences fix 124280304Sjkim 125280304Sjkim The PKCS#7 parsing code does not handle missing outer ContentInfo 126280304Sjkim correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with 127280304Sjkim missing content and trigger a NULL pointer dereference on parsing. 128280304Sjkim 129280304Sjkim Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or 130280304Sjkim otherwise parse PKCS#7 structures from untrusted sources are 131280304Sjkim affected. OpenSSL clients and servers are not affected. 132280304Sjkim 133280304Sjkim This issue was reported to OpenSSL by Michal Zalewski (Google). 134280304Sjkim (CVE-2015-0289) 135280304Sjkim [Emilia K�sper] 136280304Sjkim 137280304Sjkim *) DoS via reachable assert in SSLv2 servers fix 138280304Sjkim 139280304Sjkim A malicious client can trigger an OPENSSL_assert (i.e., an abort) in 140280304Sjkim servers that both support SSLv2 and enable export cipher suites by sending 141280304Sjkim a specially crafted SSLv2 CLIENT-MASTER-KEY message. 142280304Sjkim 143280304Sjkim This issue was discovered by Sean Burford (Google) and Emilia K�sper 144280304Sjkim (OpenSSL development team). 145280304Sjkim (CVE-2015-0293) 146280304Sjkim [Emilia K�sper] 147280304Sjkim 148280304Sjkim *) Use After Free following d2i_ECPrivatekey error fix 149280304Sjkim 150280304Sjkim A malformed EC private key file consumed via the d2i_ECPrivateKey function 151280304Sjkim could cause a use after free condition. This, in turn, could cause a double 152280304Sjkim free in several private key parsing functions (such as d2i_PrivateKey 153280304Sjkim or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption 154280304Sjkim for applications that receive EC private keys from untrusted 155280304Sjkim sources. This scenario is considered rare. 156280304Sjkim 157280304Sjkim This issue was discovered by the BoringSSL project and fixed in their 158280304Sjkim commit 517073cd4b. 159280304Sjkim (CVE-2015-0209) 160280304Sjkim [Matt Caswell] 161280304Sjkim 162280304Sjkim *) X509_to_X509_REQ NULL pointer deref fix 163280304Sjkim 164280304Sjkim The function X509_to_X509_REQ will crash with a NULL pointer dereference if 165280304Sjkim the certificate key is invalid. This function is rarely used in practice. 166280304Sjkim 167280304Sjkim This issue was discovered by Brian Carpenter. 168280304Sjkim (CVE-2015-0288) 169280304Sjkim [Stephen Henson] 170280304Sjkim 171280304Sjkim *) Removed the export ciphers from the DEFAULT ciphers 172280304Sjkim [Kurt Roeckx] 173280304Sjkim 174277597Sjkim Changes between 1.0.1k and 1.0.1l [15 Jan 2015] 175277597Sjkim 176277597Sjkim *) Build fixes for the Windows and OpenVMS platforms 177277597Sjkim [Matt Caswell and Richard Levitte] 178277597Sjkim 179276864Sjkim Changes between 1.0.1j and 1.0.1k [8 Jan 2015] 180276864Sjkim 181276864Sjkim *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS 182276864Sjkim message can cause a segmentation fault in OpenSSL due to a NULL pointer 183276864Sjkim dereference. This could lead to a Denial Of Service attack. Thanks to 184276864Sjkim Markus Stenberg of Cisco Systems, Inc. for reporting this issue. 185276864Sjkim (CVE-2014-3571) 186276864Sjkim [Steve Henson] 187276864Sjkim 188276864Sjkim *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the 189276864Sjkim dtls1_buffer_record function under certain conditions. In particular this 190276864Sjkim could occur if an attacker sent repeated DTLS records with the same 191276864Sjkim sequence number but for the next epoch. The memory leak could be exploited 192276864Sjkim by an attacker in a Denial of Service attack through memory exhaustion. 193276864Sjkim Thanks to Chris Mueller for reporting this issue. 194276864Sjkim (CVE-2015-0206) 195276864Sjkim [Matt Caswell] 196276864Sjkim 197276864Sjkim *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is 198276864Sjkim built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl 199276864Sjkim method would be set to NULL which could later result in a NULL pointer 200276864Sjkim dereference. Thanks to Frank Schmirler for reporting this issue. 201276864Sjkim (CVE-2014-3569) 202276864Sjkim [Kurt Roeckx] 203276864Sjkim 204276864Sjkim *) Abort handshake if server key exchange message is omitted for ephemeral 205276864Sjkim ECDH ciphersuites. 206276864Sjkim 207276864Sjkim Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for 208276864Sjkim reporting this issue. 209276864Sjkim (CVE-2014-3572) 210276864Sjkim [Steve Henson] 211276864Sjkim 212276864Sjkim *) Remove non-export ephemeral RSA code on client and server. This code 213276864Sjkim violated the TLS standard by allowing the use of temporary RSA keys in 214276864Sjkim non-export ciphersuites and could be used by a server to effectively 215276864Sjkim downgrade the RSA key length used to a value smaller than the server 216276864Sjkim certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at 217276864Sjkim INRIA or reporting this issue. 218276864Sjkim (CVE-2015-0204) 219276864Sjkim [Steve Henson] 220276864Sjkim 221276864Sjkim *) Fixed issue where DH client certificates are accepted without verification. 222276864Sjkim An OpenSSL server will accept a DH certificate for client authentication 223276864Sjkim without the certificate verify message. This effectively allows a client to 224276864Sjkim authenticate without the use of a private key. This only affects servers 225276864Sjkim which trust a client certificate authority which issues certificates 226276864Sjkim containing DH keys: these are extremely rare and hardly ever encountered. 227276864Sjkim Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting 228276864Sjkim this issue. 229276864Sjkim (CVE-2015-0205) 230276864Sjkim [Steve Henson] 231276864Sjkim 232276864Sjkim *) Ensure that the session ID context of an SSL is updated when its 233276864Sjkim SSL_CTX is updated via SSL_set_SSL_CTX. 234276864Sjkim 235276864Sjkim The session ID context is typically set from the parent SSL_CTX, 236276864Sjkim and can vary with the CTX. 237276864Sjkim [Adam Langley] 238276864Sjkim 239276864Sjkim *) Fix various certificate fingerprint issues. 240276864Sjkim 241276864Sjkim By using non-DER or invalid encodings outside the signed portion of a 242276864Sjkim certificate the fingerprint can be changed without breaking the signature. 243276864Sjkim Although no details of the signed portion of the certificate can be changed 244276864Sjkim this can cause problems with some applications: e.g. those using the 245276864Sjkim certificate fingerprint for blacklists. 246276864Sjkim 247276864Sjkim 1. Reject signatures with non zero unused bits. 248276864Sjkim 249276864Sjkim If the BIT STRING containing the signature has non zero unused bits reject 250276864Sjkim the signature. All current signature algorithms require zero unused bits. 251276864Sjkim 252276864Sjkim 2. Check certificate algorithm consistency. 253276864Sjkim 254276864Sjkim Check the AlgorithmIdentifier inside TBS matches the one in the 255276864Sjkim certificate signature. NB: this will result in signature failure 256276864Sjkim errors for some broken certificates. 257276864Sjkim 258276864Sjkim Thanks to Konrad Kraszewski from Google for reporting this issue. 259276864Sjkim 260276864Sjkim 3. Check DSA/ECDSA signatures use DER. 261276864Sjkim 262276864Sjkim Reencode DSA/ECDSA signatures and compare with the original received 263276864Sjkim signature. Return an error if there is a mismatch. 264276864Sjkim 265276864Sjkim This will reject various cases including garbage after signature 266276864Sjkim (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS 267276864Sjkim program for discovering this case) and use of BER or invalid ASN.1 INTEGERs 268276864Sjkim (negative or with leading zeroes). 269276864Sjkim 270276864Sjkim Further analysis was conducted and fixes were developed by Stephen Henson 271276864Sjkim of the OpenSSL core team. 272276864Sjkim 273276864Sjkim (CVE-2014-8275) 274276864Sjkim [Steve Henson] 275276864Sjkim 276276864Sjkim *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect 277276864Sjkim results on some platforms, including x86_64. This bug occurs at random 278276864Sjkim with a very low probability, and is not known to be exploitable in any 279276864Sjkim way, though its exact impact is difficult to determine. Thanks to Pieter 280276864Sjkim Wuille (Blockstream) who reported this issue and also suggested an initial 281276864Sjkim fix. Further analysis was conducted by the OpenSSL development team and 282276864Sjkim Adam Langley of Google. The final fix was developed by Andy Polyakov of 283276864Sjkim the OpenSSL core team. 284276864Sjkim (CVE-2014-3570) 285276864Sjkim [Andy Polyakov] 286276864Sjkim 287276864Sjkim *) Do not resume sessions on the server if the negotiated protocol 288276864Sjkim version does not match the session's version. Resuming with a different 289276864Sjkim version, while not strictly forbidden by the RFC, is of questionable 290276864Sjkim sanity and breaks all known clients. 291276864Sjkim [David Benjamin, Emilia K�sper] 292276864Sjkim 293276864Sjkim *) Tighten handling of the ChangeCipherSpec (CCS) message: reject 294276864Sjkim early CCS messages during renegotiation. (Note that because 295276864Sjkim renegotiation is encrypted, this early CCS was not exploitable.) 296276864Sjkim [Emilia K�sper] 297276864Sjkim 298276864Sjkim *) Tighten client-side session ticket handling during renegotiation: 299276864Sjkim ensure that the client only accepts a session ticket if the server sends 300276864Sjkim the extension anew in the ServerHello. Previously, a TLS client would 301276864Sjkim reuse the old extension state and thus accept a session ticket if one was 302276864Sjkim announced in the initial ServerHello. 303276864Sjkim 304276864Sjkim Similarly, ensure that the client requires a session ticket if one 305276864Sjkim was advertised in the ServerHello. Previously, a TLS client would 306276864Sjkim ignore a missing NewSessionTicket message. 307276864Sjkim [Emilia K�sper] 308276864Sjkim 309273149Sjkim Changes between 1.0.1i and 1.0.1j [15 Oct 2014] 310273149Sjkim 311273149Sjkim *) SRTP Memory Leak. 312273149Sjkim 313273149Sjkim A flaw in the DTLS SRTP extension parsing code allows an attacker, who 314273149Sjkim sends a carefully crafted handshake message, to cause OpenSSL to fail 315273149Sjkim to free up to 64k of memory causing a memory leak. This could be 316273149Sjkim exploited in a Denial Of Service attack. This issue affects OpenSSL 317273149Sjkim 1.0.1 server implementations for both SSL/TLS and DTLS regardless of 318273149Sjkim whether SRTP is used or configured. Implementations of OpenSSL that 319273149Sjkim have been compiled with OPENSSL_NO_SRTP defined are not affected. 320273149Sjkim 321273149Sjkim The fix was developed by the OpenSSL team. 322273149Sjkim (CVE-2014-3513) 323273149Sjkim [OpenSSL team] 324273149Sjkim 325273149Sjkim *) Session Ticket Memory Leak. 326273149Sjkim 327273149Sjkim When an OpenSSL SSL/TLS/DTLS server receives a session ticket the 328273149Sjkim integrity of that ticket is first verified. In the event of a session 329273149Sjkim ticket integrity check failing, OpenSSL will fail to free memory 330273149Sjkim causing a memory leak. By sending a large number of invalid session 331273149Sjkim tickets an attacker could exploit this issue in a Denial Of Service 332273149Sjkim attack. 333273149Sjkim (CVE-2014-3567) 334273149Sjkim [Steve Henson] 335273149Sjkim 336273149Sjkim *) Build option no-ssl3 is incomplete. 337273149Sjkim 338273149Sjkim When OpenSSL is configured with "no-ssl3" as a build option, servers 339273149Sjkim could accept and complete a SSL 3.0 handshake, and clients could be 340273149Sjkim configured to send them. 341273149Sjkim (CVE-2014-3568) 342273149Sjkim [Akamai and the OpenSSL team] 343273149Sjkim 344273149Sjkim *) Add support for TLS_FALLBACK_SCSV. 345273149Sjkim Client applications doing fallback retries should call 346273149Sjkim SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV). 347273149Sjkim (CVE-2014-3566) 348273149Sjkim [Adam Langley, Bodo Moeller] 349273149Sjkim 350273149Sjkim *) Add additional DigestInfo checks. 351273149Sjkim 352273149Sjkim Reencode DigestInto in DER and check against the original when 353273149Sjkim verifying RSA signature: this will reject any improperly encoded 354273149Sjkim DigestInfo structures. 355273149Sjkim 356273149Sjkim Note: this is a precautionary measure and no attacks are currently known. 357273149Sjkim 358273149Sjkim [Steve Henson] 359273149Sjkim 360269686Sjkim Changes between 1.0.1h and 1.0.1i [6 Aug 2014] 361269686Sjkim 362269686Sjkim *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the 363269686Sjkim SRP code can be overrun an internal buffer. Add sanity check that 364269686Sjkim g, A, B < N to SRP code. 365269686Sjkim 366269686Sjkim Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC 367269686Sjkim Group for discovering this issue. 368269686Sjkim (CVE-2014-3512) 369269686Sjkim [Steve Henson] 370269686Sjkim 371269686Sjkim *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate 372269686Sjkim TLS 1.0 instead of higher protocol versions when the ClientHello message 373269686Sjkim is badly fragmented. This allows a man-in-the-middle attacker to force a 374269686Sjkim downgrade to TLS 1.0 even if both the server and the client support a 375269686Sjkim higher protocol version, by modifying the client's TLS records. 376269686Sjkim 377269686Sjkim Thanks to David Benjamin and Adam Langley (Google) for discovering and 378269686Sjkim researching this issue. 379269686Sjkim (CVE-2014-3511) 380269686Sjkim [David Benjamin] 381269686Sjkim 382269686Sjkim *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject 383269686Sjkim to a denial of service attack. A malicious server can crash the client 384269686Sjkim with a null pointer dereference (read) by specifying an anonymous (EC)DH 385269686Sjkim ciphersuite and sending carefully crafted handshake messages. 386269686Sjkim 387269686Sjkim Thanks to Felix Gr�bert (Google) for discovering and researching this 388269686Sjkim issue. 389269686Sjkim (CVE-2014-3510) 390269686Sjkim [Emilia K�sper] 391269686Sjkim 392269686Sjkim *) By sending carefully crafted DTLS packets an attacker could cause openssl 393269686Sjkim to leak memory. This can be exploited through a Denial of Service attack. 394269686Sjkim Thanks to Adam Langley for discovering and researching this issue. 395269686Sjkim (CVE-2014-3507) 396269686Sjkim [Adam Langley] 397269686Sjkim 398269686Sjkim *) An attacker can force openssl to consume large amounts of memory whilst 399269686Sjkim processing DTLS handshake messages. This can be exploited through a 400269686Sjkim Denial of Service attack. 401269686Sjkim Thanks to Adam Langley for discovering and researching this issue. 402269686Sjkim (CVE-2014-3506) 403269686Sjkim [Adam Langley] 404269686Sjkim 405269686Sjkim *) An attacker can force an error condition which causes openssl to crash 406269686Sjkim whilst processing DTLS packets due to memory being freed twice. This 407269686Sjkim can be exploited through a Denial of Service attack. 408269686Sjkim Thanks to Adam Langley and Wan-Teh Chang for discovering and researching 409269686Sjkim this issue. 410269686Sjkim (CVE-2014-3505) 411269686Sjkim [Adam Langley] 412269686Sjkim 413269686Sjkim *) If a multithreaded client connects to a malicious server using a resumed 414269686Sjkim session and the server sends an ec point format extension it could write 415269686Sjkim up to 255 bytes to freed memory. 416269686Sjkim 417269686Sjkim Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this 418269686Sjkim issue. 419269686Sjkim (CVE-2014-3509) 420269686Sjkim [Gabor Tyukasz] 421269686Sjkim 422269686Sjkim *) A malicious server can crash an OpenSSL client with a null pointer 423269686Sjkim dereference (read) by specifying an SRP ciphersuite even though it was not 424269686Sjkim properly negotiated with the client. This can be exploited through a 425269686Sjkim Denial of Service attack. 426269686Sjkim 427269686Sjkim Thanks to Joonas Kuorilehto and Riku Hietam�ki (Codenomicon) for 428269686Sjkim discovering and researching this issue. 429269686Sjkim (CVE-2014-5139) 430269686Sjkim [Steve Henson] 431269686Sjkim 432269686Sjkim *) A flaw in OBJ_obj2txt may cause pretty printing functions such as 433269686Sjkim X509_name_oneline, X509_name_print_ex et al. to leak some information 434269686Sjkim from the stack. Applications may be affected if they echo pretty printing 435269686Sjkim output to the attacker. 436269686Sjkim 437269686Sjkim Thanks to Ivan Fratric (Google) for discovering this issue. 438269686Sjkim (CVE-2014-3508) 439269686Sjkim [Emilia K�sper, and Steve Henson] 440269686Sjkim 441269686Sjkim *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.) 442269686Sjkim for corner cases. (Certain input points at infinity could lead to 443269686Sjkim bogus results, with non-infinity inputs mapped to infinity too.) 444269686Sjkim [Bodo Moeller] 445269686Sjkim 446267258Sjkim Changes between 1.0.1g and 1.0.1h [5 Jun 2014] 447267258Sjkim 448267258Sjkim *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted 449267258Sjkim handshake can force the use of weak keying material in OpenSSL 450267258Sjkim SSL/TLS clients and servers. 451267258Sjkim 452267258Sjkim Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and 453267258Sjkim researching this issue. (CVE-2014-0224) 454267258Sjkim [KIKUCHI Masashi, Steve Henson] 455267258Sjkim 456267258Sjkim *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an 457267258Sjkim OpenSSL DTLS client the code can be made to recurse eventually crashing 458267258Sjkim in a DoS attack. 459267258Sjkim 460267258Sjkim Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue. 461267258Sjkim (CVE-2014-0221) 462267258Sjkim [Imre Rad, Steve Henson] 463267258Sjkim 464267258Sjkim *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can 465267258Sjkim be triggered by sending invalid DTLS fragments to an OpenSSL DTLS 466267258Sjkim client or server. This is potentially exploitable to run arbitrary 467267258Sjkim code on a vulnerable client or server. 468267258Sjkim 469267258Sjkim Thanks to J�ri Aedla for reporting this issue. (CVE-2014-0195) 470267258Sjkim [J�ri Aedla, Steve Henson] 471267258Sjkim 472267258Sjkim *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites 473267258Sjkim are subject to a denial of service attack. 474267258Sjkim 475267258Sjkim Thanks to Felix Gr�bert and Ivan Fratric at Google for discovering 476267258Sjkim this issue. (CVE-2014-3470) 477267258Sjkim [Felix Gr�bert, Ivan Fratric, Steve Henson] 478267258Sjkim 479267258Sjkim *) Harmonize version and its documentation. -f flag is used to display 480267258Sjkim compilation flags. 481267258Sjkim [mancha <mancha1@zoho.com>] 482267258Sjkim 483267258Sjkim *) Fix eckey_priv_encode so it immediately returns an error upon a failure 484267258Sjkim in i2d_ECPrivateKey. 485267258Sjkim [mancha <mancha1@zoho.com>] 486267258Sjkim 487267258Sjkim *) Fix some double frees. These are not thought to be exploitable. 488267258Sjkim [mancha <mancha1@zoho.com>] 489267258Sjkim 490264331Sjkim Changes between 1.0.1f and 1.0.1g [7 Apr 2014] 491264331Sjkim 492264331Sjkim *) A missing bounds check in the handling of the TLS heartbeat extension 493264331Sjkim can be used to reveal up to 64k of memory to a connected client or 494264331Sjkim server. 495264331Sjkim 496264331Sjkim Thanks for Neel Mehta of Google Security for discovering this bug and to 497264331Sjkim Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for 498264331Sjkim preparing the fix (CVE-2014-0160) 499264331Sjkim [Adam Langley, Bodo Moeller] 500264331Sjkim 501264331Sjkim *) Fix for the attack described in the paper "Recovering OpenSSL 502264331Sjkim ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack" 503264331Sjkim by Yuval Yarom and Naomi Benger. Details can be obtained from: 504264331Sjkim http://eprint.iacr.org/2014/140 505264331Sjkim 506264331Sjkim Thanks to Yuval Yarom and Naomi Benger for discovering this 507264331Sjkim flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076) 508264331Sjkim [Yuval Yarom and Naomi Benger] 509264331Sjkim 510264331Sjkim *) TLS pad extension: draft-agl-tls-padding-03 511264331Sjkim 512264331Sjkim Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the 513264331Sjkim TLS client Hello record length value would otherwise be > 255 and 514264331Sjkim less that 512 pad with a dummy extension containing zeroes so it 515264331Sjkim is at least 512 bytes long. 516264331Sjkim 517264331Sjkim [Adam Langley, Steve Henson] 518264331Sjkim 519264331Sjkim Changes between 1.0.1e and 1.0.1f [6 Jan 2014] 520264331Sjkim 521264331Sjkim *) Fix for TLS record tampering bug. A carefully crafted invalid 522264331Sjkim handshake could crash OpenSSL with a NULL pointer exception. 523264331Sjkim Thanks to Anton Johansson for reporting this issues. 524264331Sjkim (CVE-2013-4353) 525264331Sjkim 526264331Sjkim *) Keep original DTLS digest and encryption contexts in retransmission 527264331Sjkim structures so we can use the previous session parameters if they need 528264331Sjkim to be resent. (CVE-2013-6450) 529264331Sjkim [Steve Henson] 530264331Sjkim 531264331Sjkim *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which 532264331Sjkim avoids preferring ECDHE-ECDSA ciphers when the client appears to be 533264331Sjkim Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for 534264331Sjkim several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug 535264331Sjkim is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing 536264331Sjkim 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer. 537264331Sjkim [Rob Stradling, Adam Langley] 538264331Sjkim 539246772Sjkim Changes between 1.0.1d and 1.0.1e [11 Feb 2013] 540246772Sjkim 541264331Sjkim *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI 542264331Sjkim supporting platforms or when small records were transferred. 543264331Sjkim [Andy Polyakov, Steve Henson] 544246772Sjkim 545246772Sjkim Changes between 1.0.1c and 1.0.1d [5 Feb 2013] 546246772Sjkim 547246772Sjkim *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time. 548246772Sjkim 549246772Sjkim This addresses the flaw in CBC record processing discovered by 550246772Sjkim Nadhem Alfardan and Kenny Paterson. Details of this attack can be found 551246772Sjkim at: http://www.isg.rhul.ac.uk/tls/ 552246772Sjkim 553246772Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 554246772Sjkim Security Group at Royal Holloway, University of London 555246772Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and 556246772Sjkim Emilia K�sper for the initial patch. 557246772Sjkim (CVE-2013-0169) 558246772Sjkim [Emilia K�sper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson] 559246772Sjkim 560246772Sjkim *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode 561246772Sjkim ciphersuites which can be exploited in a denial of service attack. 562246772Sjkim Thanks go to and to Adam Langley <agl@chromium.org> for discovering 563246772Sjkim and detecting this bug and to Wolfgang Ettlinger 564246772Sjkim <wolfgang.ettlinger@gmail.com> for independently discovering this issue. 565246772Sjkim (CVE-2012-2686) 566246772Sjkim [Adam Langley] 567246772Sjkim 568246772Sjkim *) Return an error when checking OCSP signatures when key is NULL. 569246772Sjkim This fixes a DoS attack. (CVE-2013-0166) 570246772Sjkim [Steve Henson] 571246772Sjkim 572246772Sjkim *) Make openssl verify return errors. 573246772Sjkim [Chris Palmer <palmer@google.com> and Ben Laurie] 574246772Sjkim 575246772Sjkim *) Call OCSP Stapling callback after ciphersuite has been chosen, so 576246772Sjkim the right response is stapled. Also change SSL_get_certificate() 577246772Sjkim so it returns the certificate actually sent. 578246772Sjkim See http://rt.openssl.org/Ticket/Display.html?id=2836. 579246772Sjkim [Rob Stradling <rob.stradling@comodo.com>] 580246772Sjkim 581246772Sjkim *) Fix possible deadlock when decoding public keys. 582246772Sjkim [Steve Henson] 583246772Sjkim 584246772Sjkim *) Don't use TLS 1.0 record version number in initial client hello 585246772Sjkim if renegotiating. 586246772Sjkim [Steve Henson] 587246772Sjkim 588238405Sjkim Changes between 1.0.1b and 1.0.1c [10 May 2012] 589237657Sjkim 590238405Sjkim *) Sanity check record length before skipping explicit IV in TLS 591238405Sjkim 1.2, 1.1 and DTLS to fix DoS attack. 592237657Sjkim 593237657Sjkim Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic 594237657Sjkim fuzzing as a service testing platform. 595237657Sjkim (CVE-2012-2333) 596237657Sjkim [Steve Henson] 597237657Sjkim 598237657Sjkim *) Initialise tkeylen properly when encrypting CMS messages. 599237657Sjkim Thanks to Solar Designer of Openwall for reporting this issue. 600237657Sjkim [Steve Henson] 601237657Sjkim 602238405Sjkim *) In FIPS mode don't try to use composite ciphers as they are not 603238405Sjkim approved. 604238405Sjkim [Steve Henson] 605237657Sjkim 606238405Sjkim Changes between 1.0.1a and 1.0.1b [26 Apr 2012] 607237657Sjkim 608238405Sjkim *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and 609238405Sjkim 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately 610238405Sjkim mean any application compiled against OpenSSL 1.0.0 headers setting 611238405Sjkim SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng 612238405Sjkim TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to 613238405Sjkim 0x10000000L Any application which was previously compiled against 614238405Sjkim OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1 615238405Sjkim will need to be recompiled as a result. Letting be results in 616238405Sjkim inability to disable specifically TLS 1.1 and in client context, 617238405Sjkim in unlike event, limit maximum offered version to TLS 1.0 [see below]. 618238405Sjkim [Steve Henson] 619237657Sjkim 620238405Sjkim *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not 621238405Sjkim disable just protocol X, but all protocols above X *if* there are 622238405Sjkim protocols *below* X still enabled. In more practical terms it means 623238405Sjkim that if application wants to disable TLS1.0 in favor of TLS1.1 and 624238405Sjkim above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass 625238405Sjkim SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to 626238405Sjkim client side. 627238405Sjkim [Andy Polyakov] 628238405Sjkim 629238405Sjkim Changes between 1.0.1 and 1.0.1a [19 Apr 2012] 630238405Sjkim 631237657Sjkim *) Check for potentially exploitable overflows in asn1_d2i_read_bio 632237657Sjkim BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer 633237657Sjkim in CRYPTO_realloc_clean. 634237657Sjkim 635237657Sjkim Thanks to Tavis Ormandy, Google Security Team, for discovering this 636237657Sjkim issue and to Adam Langley <agl@chromium.org> for fixing it. 637237657Sjkim (CVE-2012-2110) 638237657Sjkim [Adam Langley (Google), Tavis Ormandy, Google Security Team] 639237657Sjkim 640238405Sjkim *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections. 641238405Sjkim [Adam Langley] 642237657Sjkim 643238405Sjkim *) Workarounds for some broken servers that "hang" if a client hello 644238405Sjkim record length exceeds 255 bytes. 645238405Sjkim 646238405Sjkim 1. Do not use record version number > TLS 1.0 in initial client 647238405Sjkim hello: some (but not all) hanging servers will now work. 648238405Sjkim 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate 649238405Sjkim the number of ciphers sent in the client hello. This should be 650238405Sjkim set to an even number, such as 50, for example by passing: 651238405Sjkim -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure. 652238405Sjkim Most broken servers should now work. 653238405Sjkim 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable 654238405Sjkim TLS 1.2 client support entirely. 655238405Sjkim [Steve Henson] 656238405Sjkim 657238405Sjkim *) Fix SEGV in Vector Permutation AES module observed in OpenSSH. 658238405Sjkim [Andy Polyakov] 659238405Sjkim 660238405Sjkim Changes between 1.0.0h and 1.0.1 [14 Mar 2012] 661238405Sjkim 662238405Sjkim *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET 663238405Sjkim STRING form instead of a DigestInfo. 664238405Sjkim [Steve Henson] 665238405Sjkim 666238405Sjkim *) The format used for MDC2 RSA signatures is inconsistent between EVP 667238405Sjkim and the RSA_sign/RSA_verify functions. This was made more apparent when 668238405Sjkim OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular 669238405Sjkim those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect 670238405Sjkim the correct format in RSA_verify so both forms transparently work. 671238405Sjkim [Steve Henson] 672238405Sjkim 673238405Sjkim *) Some servers which support TLS 1.0 can choke if we initially indicate 674238405Sjkim support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA 675238405Sjkim encrypted premaster secret. As a workaround use the maximum pemitted 676238405Sjkim client version in client hello, this should keep such servers happy 677238405Sjkim and still work with previous versions of OpenSSL. 678238405Sjkim [Steve Henson] 679238405Sjkim 680238405Sjkim *) Add support for TLS/DTLS heartbeats. 681238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>] 682238405Sjkim 683238405Sjkim *) Add support for SCTP. 684238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>] 685238405Sjkim 686238405Sjkim *) Improved PRNG seeding for VOS. 687238405Sjkim [Paul Green <Paul.Green@stratus.com>] 688238405Sjkim 689238405Sjkim *) Extensive assembler packs updates, most notably: 690238405Sjkim 691238405Sjkim - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support; 692238405Sjkim - x86[_64]: SSSE3 support (SHA1, vector-permutation AES); 693238405Sjkim - x86_64: bit-sliced AES implementation; 694238405Sjkim - ARM: NEON support, contemporary platforms optimizations; 695238405Sjkim - s390x: z196 support; 696238405Sjkim - *: GHASH and GF(2^m) multiplication implementations; 697238405Sjkim 698238405Sjkim [Andy Polyakov] 699238405Sjkim 700238405Sjkim *) Make TLS-SRP code conformant with RFC 5054 API cleanup 701238405Sjkim (removal of unnecessary code) 702238405Sjkim [Peter Sylvester <peter.sylvester@edelweb.fr>] 703238405Sjkim 704238405Sjkim *) Add TLS key material exporter from RFC 5705. 705238405Sjkim [Eric Rescorla] 706238405Sjkim 707238405Sjkim *) Add DTLS-SRTP negotiation from RFC 5764. 708238405Sjkim [Eric Rescorla] 709238405Sjkim 710238405Sjkim *) Add Next Protocol Negotiation, 711238405Sjkim http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be 712238405Sjkim disabled with a no-npn flag to config or Configure. Code donated 713238405Sjkim by Google. 714238405Sjkim [Adam Langley <agl@google.com> and Ben Laurie] 715238405Sjkim 716238405Sjkim *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224, 717238405Sjkim NIST-P256, NIST-P521, with constant-time single point multiplication on 718238405Sjkim typical inputs. Compiler support for the nonstandard type __uint128_t is 719238405Sjkim required to use this (present in gcc 4.4 and later, for 64-bit builds). 720238405Sjkim Code made available under Apache License version 2.0. 721238405Sjkim 722238405Sjkim Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command 723238405Sjkim line to include this in your build of OpenSSL, and run "make depend" (or 724238405Sjkim "make update"). This enables the following EC_METHODs: 725238405Sjkim 726238405Sjkim EC_GFp_nistp224_method() 727238405Sjkim EC_GFp_nistp256_method() 728238405Sjkim EC_GFp_nistp521_method() 729238405Sjkim 730238405Sjkim EC_GROUP_new_by_curve_name() will automatically use these (while 731238405Sjkim EC_GROUP_new_curve_GFp() currently prefers the more flexible 732238405Sjkim implementations). 733238405Sjkim [Emilia K�sper, Adam Langley, Bodo Moeller (Google)] 734238405Sjkim 735238405Sjkim *) Use type ossl_ssize_t instad of ssize_t which isn't available on 736238405Sjkim all platforms. Move ssize_t definition from e_os.h to the public 737238405Sjkim header file e_os2.h as it now appears in public header file cms.h 738238405Sjkim [Steve Henson] 739238405Sjkim 740238405Sjkim *) New -sigopt option to the ca, req and x509 utilities. Additional 741238405Sjkim signature parameters can be passed using this option and in 742238405Sjkim particular PSS. 743238405Sjkim [Steve Henson] 744238405Sjkim 745238405Sjkim *) Add RSA PSS signing function. This will generate and set the 746238405Sjkim appropriate AlgorithmIdentifiers for PSS based on those in the 747238405Sjkim corresponding EVP_MD_CTX structure. No application support yet. 748238405Sjkim [Steve Henson] 749238405Sjkim 750238405Sjkim *) Support for companion algorithm specific ASN1 signing routines. 751238405Sjkim New function ASN1_item_sign_ctx() signs a pre-initialised 752238405Sjkim EVP_MD_CTX structure and sets AlgorithmIdentifiers based on 753238405Sjkim the appropriate parameters. 754238405Sjkim [Steve Henson] 755238405Sjkim 756238405Sjkim *) Add new algorithm specific ASN1 verification initialisation function 757238405Sjkim to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1 758238405Sjkim handling will be the same no matter what EVP_PKEY_METHOD is used. 759238405Sjkim Add a PSS handler to support verification of PSS signatures: checked 760238405Sjkim against a number of sample certificates. 761238405Sjkim [Steve Henson] 762238405Sjkim 763238405Sjkim *) Add signature printing for PSS. Add PSS OIDs. 764238405Sjkim [Steve Henson, Martin Kaiser <lists@kaiser.cx>] 765238405Sjkim 766238405Sjkim *) Add algorithm specific signature printing. An individual ASN1 method 767238405Sjkim can now print out signatures instead of the standard hex dump. 768238405Sjkim 769238405Sjkim More complex signatures (e.g. PSS) can print out more meaningful 770238405Sjkim information. Include DSA version that prints out the signature 771238405Sjkim parameters r, s. 772238405Sjkim [Steve Henson] 773238405Sjkim 774238405Sjkim *) Password based recipient info support for CMS library: implementing 775238405Sjkim RFC3211. 776238405Sjkim [Steve Henson] 777238405Sjkim 778238405Sjkim *) Split password based encryption into PBES2 and PBKDF2 functions. This 779238405Sjkim neatly separates the code into cipher and PBE sections and is required 780238405Sjkim for some algorithms that split PBES2 into separate pieces (such as 781238405Sjkim password based CMS). 782238405Sjkim [Steve Henson] 783238405Sjkim 784238405Sjkim *) Session-handling fixes: 785238405Sjkim - Fix handling of connections that are resuming with a session ID, 786238405Sjkim but also support Session Tickets. 787238405Sjkim - Fix a bug that suppressed issuing of a new ticket if the client 788238405Sjkim presented a ticket with an expired session. 789238405Sjkim - Try to set the ticket lifetime hint to something reasonable. 790238405Sjkim - Make tickets shorter by excluding irrelevant information. 791238405Sjkim - On the client side, don't ignore renewed tickets. 792238405Sjkim [Adam Langley, Bodo Moeller (Google)] 793238405Sjkim 794238405Sjkim *) Fix PSK session representation. 795238405Sjkim [Bodo Moeller] 796238405Sjkim 797238405Sjkim *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations. 798238405Sjkim 799238405Sjkim This work was sponsored by Intel. 800238405Sjkim [Andy Polyakov] 801238405Sjkim 802238405Sjkim *) Add GCM support to TLS library. Some custom code is needed to split 803238405Sjkim the IV between the fixed (from PRF) and explicit (from TLS record) 804238405Sjkim portions. This adds all GCM ciphersuites supported by RFC5288 and 805238405Sjkim RFC5289. Generalise some AES* cipherstrings to inlclude GCM and 806238405Sjkim add a special AESGCM string for GCM only. 807238405Sjkim [Steve Henson] 808238405Sjkim 809238405Sjkim *) Expand range of ctrls for AES GCM. Permit setting invocation 810238405Sjkim field on decrypt and retrieval of invocation field only on encrypt. 811238405Sjkim [Steve Henson] 812238405Sjkim 813238405Sjkim *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support. 814238405Sjkim As required by RFC5289 these ciphersuites cannot be used if for 815238405Sjkim versions of TLS earlier than 1.2. 816238405Sjkim [Steve Henson] 817238405Sjkim 818238405Sjkim *) For FIPS capable OpenSSL interpret a NULL default public key method 819238405Sjkim as unset and return the appopriate default but do *not* set the default. 820238405Sjkim This means we can return the appopriate method in applications that 821238405Sjkim swicth between FIPS and non-FIPS modes. 822238405Sjkim [Steve Henson] 823238405Sjkim 824238405Sjkim *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an 825238405Sjkim ENGINE is used then we cannot handle that in the FIPS module so we 826238405Sjkim keep original code iff non-FIPS operations are allowed. 827238405Sjkim [Steve Henson] 828238405Sjkim 829238405Sjkim *) Add -attime option to openssl utilities. 830238405Sjkim [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson] 831238405Sjkim 832238405Sjkim *) Redirect DSA and DH operations to FIPS module in FIPS mode. 833238405Sjkim [Steve Henson] 834238405Sjkim 835238405Sjkim *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use 836238405Sjkim FIPS EC methods unconditionally for now. 837238405Sjkim [Steve Henson] 838238405Sjkim 839238405Sjkim *) New build option no-ec2m to disable characteristic 2 code. 840238405Sjkim [Steve Henson] 841238405Sjkim 842238405Sjkim *) Backport libcrypto audit of return value checking from 1.1.0-dev; not 843238405Sjkim all cases can be covered as some introduce binary incompatibilities. 844238405Sjkim [Steve Henson] 845238405Sjkim 846238405Sjkim *) Redirect RSA operations to FIPS module including keygen, 847238405Sjkim encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods. 848238405Sjkim [Steve Henson] 849238405Sjkim 850238405Sjkim *) Add similar low level API blocking to ciphers. 851238405Sjkim [Steve Henson] 852238405Sjkim 853238405Sjkim *) Low level digest APIs are not approved in FIPS mode: any attempt 854238405Sjkim to use these will cause a fatal error. Applications that *really* want 855238405Sjkim to use them can use the private_* version instead. 856238405Sjkim [Steve Henson] 857238405Sjkim 858238405Sjkim *) Redirect cipher operations to FIPS module for FIPS builds. 859238405Sjkim [Steve Henson] 860238405Sjkim 861238405Sjkim *) Redirect digest operations to FIPS module for FIPS builds. 862238405Sjkim [Steve Henson] 863238405Sjkim 864238405Sjkim *) Update build system to add "fips" flag which will link in fipscanister.o 865238405Sjkim for static and shared library builds embedding a signature if needed. 866238405Sjkim [Steve Henson] 867238405Sjkim 868238405Sjkim *) Output TLS supported curves in preference order instead of numerical 869238405Sjkim order. This is currently hardcoded for the highest order curves first. 870238405Sjkim This should be configurable so applications can judge speed vs strength. 871238405Sjkim [Steve Henson] 872238405Sjkim 873238405Sjkim *) Add TLS v1.2 server support for client authentication. 874238405Sjkim [Steve Henson] 875238405Sjkim 876238405Sjkim *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers 877238405Sjkim and enable MD5. 878238405Sjkim [Steve Henson] 879238405Sjkim 880238405Sjkim *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying 881238405Sjkim FIPS modules versions. 882238405Sjkim [Steve Henson] 883238405Sjkim 884238405Sjkim *) Add TLS v1.2 client side support for client authentication. Keep cache 885238405Sjkim of handshake records longer as we don't know the hash algorithm to use 886238405Sjkim until after the certificate request message is received. 887238405Sjkim [Steve Henson] 888238405Sjkim 889238405Sjkim *) Initial TLS v1.2 client support. Add a default signature algorithms 890238405Sjkim extension including all the algorithms we support. Parse new signature 891238405Sjkim format in client key exchange. Relax some ECC signing restrictions for 892238405Sjkim TLS v1.2 as indicated in RFC5246. 893238405Sjkim [Steve Henson] 894238405Sjkim 895238405Sjkim *) Add server support for TLS v1.2 signature algorithms extension. Switch 896238405Sjkim to new signature format when needed using client digest preference. 897238405Sjkim All server ciphersuites should now work correctly in TLS v1.2. No client 898238405Sjkim support yet and no support for client certificates. 899238405Sjkim [Steve Henson] 900238405Sjkim 901238405Sjkim *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch 902238405Sjkim to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based 903238405Sjkim ciphersuites. At present only RSA key exchange ciphersuites work with 904238405Sjkim TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete 905238405Sjkim SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods 906238405Sjkim and version checking. 907238405Sjkim [Steve Henson] 908238405Sjkim 909238405Sjkim *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled 910238405Sjkim with this defined it will not be affected by any changes to ssl internal 911238405Sjkim structures. Add several utility functions to allow openssl application 912238405Sjkim to work with OPENSSL_NO_SSL_INTERN defined. 913238405Sjkim [Steve Henson] 914238405Sjkim 915238405Sjkim *) Add SRP support. 916238405Sjkim [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie] 917238405Sjkim 918238405Sjkim *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id. 919238405Sjkim [Steve Henson] 920238405Sjkim 921238405Sjkim *) Permit abbreviated handshakes when renegotiating using the function 922238405Sjkim SSL_renegotiate_abbreviated(). 923238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>] 924238405Sjkim 925238405Sjkim *) Add call to ENGINE_register_all_complete() to 926238405Sjkim ENGINE_load_builtin_engines(), so some implementations get used 927238405Sjkim automatically instead of needing explicit application support. 928238405Sjkim [Steve Henson] 929238405Sjkim 930238405Sjkim *) Add support for TLS key exporter as described in RFC5705. 931238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson] 932238405Sjkim 933238405Sjkim *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only 934238405Sjkim a few changes are required: 935238405Sjkim 936238405Sjkim Add SSL_OP_NO_TLSv1_1 flag. 937238405Sjkim Add TLSv1_1 methods. 938238405Sjkim Update version checking logic to handle version 1.1. 939238405Sjkim Add explicit IV handling (ported from DTLS code). 940238405Sjkim Add command line options to s_client/s_server. 941238405Sjkim [Steve Henson] 942238405Sjkim 943238405Sjkim Changes between 1.0.0g and 1.0.0h [12 Mar 2012] 944238405Sjkim 945237657Sjkim *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness 946237657Sjkim in CMS and PKCS7 code. When RSA decryption fails use a random key for 947237657Sjkim content decryption and always return the same error. Note: this attack 948237657Sjkim needs on average 2^20 messages so it only affects automated senders. The 949237657Sjkim old behaviour can be reenabled in the CMS code by setting the 950237657Sjkim CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where 951237657Sjkim an MMA defence is not necessary. 952237657Sjkim Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering 953237657Sjkim this issue. (CVE-2012-0884) 954237657Sjkim [Steve Henson] 955237657Sjkim 956237657Sjkim *) Fix CVE-2011-4619: make sure we really are receiving a 957237657Sjkim client hello before rejecting multiple SGC restarts. Thanks to 958237657Sjkim Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug. 959237657Sjkim [Steve Henson] 960237657Sjkim 961238405Sjkim Changes between 1.0.0f and 1.0.0g [18 Jan 2012] 962238405Sjkim 963238405Sjkim *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109. 964238405Sjkim Thanks to Antonio Martin, Enterprise Secure Access Research and 965238405Sjkim Development, Cisco Systems, Inc. for discovering this bug and 966238405Sjkim preparing a fix. (CVE-2012-0050) 967238405Sjkim [Antonio Martin] 968238405Sjkim 969238405Sjkim Changes between 1.0.0e and 1.0.0f [4 Jan 2012] 970238405Sjkim 971238405Sjkim *) Nadhem Alfardan and Kenny Paterson have discovered an extension 972238405Sjkim of the Vaudenay padding oracle attack on CBC mode encryption 973238405Sjkim which enables an efficient plaintext recovery attack against 974238405Sjkim the OpenSSL implementation of DTLS. Their attack exploits timing 975238405Sjkim differences arising during decryption processing. A research 976238405Sjkim paper describing this attack can be found at: 977238405Sjkim http://www.isg.rhul.ac.uk/~kp/dtls.pdf 978238405Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 979238405Sjkim Security Group at Royal Holloway, University of London 980238405Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann 981238405Sjkim <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de> 982238405Sjkim for preparing the fix. (CVE-2011-4108) 983238405Sjkim [Robin Seggelmann, Michael Tuexen] 984238405Sjkim 985238405Sjkim *) Clear bytes used for block padding of SSL 3.0 records. 986238405Sjkim (CVE-2011-4576) 987238405Sjkim [Adam Langley (Google)] 988238405Sjkim 989238405Sjkim *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George 990238405Sjkim Kadianakis <desnacked@gmail.com> for discovering this issue and 991238405Sjkim Adam Langley for preparing the fix. (CVE-2011-4619) 992238405Sjkim [Adam Langley (Google)] 993238405Sjkim 994238405Sjkim *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027) 995238405Sjkim [Andrey Kulikov <amdeich@gmail.com>] 996238405Sjkim 997238405Sjkim *) Prevent malformed RFC3779 data triggering an assertion failure. 998238405Sjkim Thanks to Andrew Chi, BBN Technologies, for discovering the flaw 999238405Sjkim and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577) 1000238405Sjkim [Rob Austein <sra@hactrn.net>] 1001238405Sjkim 1002238405Sjkim *) Improved PRNG seeding for VOS. 1003238405Sjkim [Paul Green <Paul.Green@stratus.com>] 1004238405Sjkim 1005238405Sjkim *) Fix ssl_ciph.c set-up race. 1006238405Sjkim [Adam Langley (Google)] 1007238405Sjkim 1008238405Sjkim *) Fix spurious failures in ecdsatest.c. 1009238405Sjkim [Emilia K�sper (Google)] 1010238405Sjkim 1011238405Sjkim *) Fix the BIO_f_buffer() implementation (which was mixing different 1012238405Sjkim interpretations of the '..._len' fields). 1013238405Sjkim [Adam Langley (Google)] 1014238405Sjkim 1015238405Sjkim *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than 1016238405Sjkim BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent 1017238405Sjkim threads won't reuse the same blinding coefficients. 1018238405Sjkim 1019238405Sjkim This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING 1020238405Sjkim lock to call BN_BLINDING_invert_ex, and avoids one use of 1021238405Sjkim BN_BLINDING_update for each BN_BLINDING structure (previously, 1022238405Sjkim the last update always remained unused). 1023238405Sjkim [Emilia K�sper (Google)] 1024238405Sjkim 1025238405Sjkim *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf. 1026238405Sjkim [Bob Buckholz (Google)] 1027238405Sjkim 1028238405Sjkim Changes between 1.0.0d and 1.0.0e [6 Sep 2011] 1029238405Sjkim 1030238405Sjkim *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted 1031238405Sjkim by initialising X509_STORE_CTX properly. (CVE-2011-3207) 1032238405Sjkim [Kaspar Brand <ossl@velox.ch>] 1033238405Sjkim 1034238405Sjkim *) Fix SSL memory handling for (EC)DH ciphersuites, in particular 1035238405Sjkim for multi-threaded use of ECDH. (CVE-2011-3210) 1036238405Sjkim [Adam Langley (Google)] 1037238405Sjkim 1038238405Sjkim *) Fix x509_name_ex_d2i memory leak on bad inputs. 1039238405Sjkim [Bodo Moeller] 1040238405Sjkim 1041238405Sjkim *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check 1042238405Sjkim signature public key algorithm by using OID xref utilities instead. 1043238405Sjkim Before this you could only use some ECC ciphersuites with SHA1 only. 1044238405Sjkim [Steve Henson] 1045238405Sjkim 1046238405Sjkim *) Add protection against ECDSA timing attacks as mentioned in the paper 1047238405Sjkim by Billy Bob Brumley and Nicola Tuveri, see: 1048238405Sjkim 1049238405Sjkim http://eprint.iacr.org/2011/232.pdf 1050238405Sjkim 1051238405Sjkim [Billy Bob Brumley and Nicola Tuveri] 1052238405Sjkim 1053238405Sjkim Changes between 1.0.0c and 1.0.0d [8 Feb 2011] 1054238405Sjkim 1055238405Sjkim *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014 1056238405Sjkim [Neel Mehta, Adam Langley, Bodo Moeller (Google)] 1057238405Sjkim 1058238405Sjkim *) Fix bug in string printing code: if *any* escaping is enabled we must 1059238405Sjkim escape the escape character (backslash) or the resulting string is 1060238405Sjkim ambiguous. 1061238405Sjkim [Steve Henson] 1062238405Sjkim 1063238405Sjkim Changes between 1.0.0b and 1.0.0c [2 Dec 2010] 1064238405Sjkim 1065238405Sjkim *) Disable code workaround for ancient and obsolete Netscape browsers 1066238405Sjkim and servers: an attacker can use it in a ciphersuite downgrade attack. 1067238405Sjkim Thanks to Martin Rex for discovering this bug. CVE-2010-4180 1068238405Sjkim [Steve Henson] 1069238405Sjkim 1070238405Sjkim *) Fixed J-PAKE implementation error, originally discovered by 1071238405Sjkim Sebastien Martini, further info and confirmation from Stefan 1072238405Sjkim Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252 1073238405Sjkim [Ben Laurie] 1074238405Sjkim 1075238405Sjkim Changes between 1.0.0a and 1.0.0b [16 Nov 2010] 1076238405Sjkim 1077238405Sjkim *) Fix extension code to avoid race conditions which can result in a buffer 1078238405Sjkim overrun vulnerability: resumed sessions must not be modified as they can 1079238405Sjkim be shared by multiple threads. CVE-2010-3864 1080238405Sjkim [Steve Henson] 1081238405Sjkim 1082238405Sjkim *) Fix WIN32 build system to correctly link an ENGINE directory into 1083238405Sjkim a DLL. 1084238405Sjkim [Steve Henson] 1085238405Sjkim 1086238405Sjkim Changes between 1.0.0 and 1.0.0a [01 Jun 2010] 1087238405Sjkim 1088238405Sjkim *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover 1089238405Sjkim (CVE-2010-1633) 1090238405Sjkim [Steve Henson, Peter-Michael Hager <hager@dortmund.net>] 1091238405Sjkim 1092238405Sjkim Changes between 0.9.8n and 1.0.0 [29 Mar 2010] 1093238405Sjkim 1094238405Sjkim *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher 1095238405Sjkim context. The operation can be customised via the ctrl mechanism in 1096238405Sjkim case ENGINEs want to include additional functionality. 1097238405Sjkim [Steve Henson] 1098238405Sjkim 1099238405Sjkim *) Tolerate yet another broken PKCS#8 key format: private key value negative. 1100238405Sjkim [Steve Henson] 1101238405Sjkim 1102238405Sjkim *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to 1103238405Sjkim output hashes compatible with older versions of OpenSSL. 1104238405Sjkim [Willy Weisz <weisz@vcpc.univie.ac.at>] 1105238405Sjkim 1106238405Sjkim *) Fix compression algorithm handling: if resuming a session use the 1107238405Sjkim compression algorithm of the resumed session instead of determining 1108238405Sjkim it from client hello again. Don't allow server to change algorithm. 1109238405Sjkim [Steve Henson] 1110238405Sjkim 1111238405Sjkim *) Add load_crls() function to apps tidying load_certs() too. Add option 1112238405Sjkim to verify utility to allow additional CRLs to be included. 1113238405Sjkim [Steve Henson] 1114238405Sjkim 1115238405Sjkim *) Update OCSP request code to permit adding custom headers to the request: 1116238405Sjkim some responders need this. 1117238405Sjkim [Steve Henson] 1118238405Sjkim 1119238405Sjkim *) The function EVP_PKEY_sign() returns <=0 on error: check return code 1120238405Sjkim correctly. 1121238405Sjkim [Julia Lawall <julia@diku.dk>] 1122238405Sjkim 1123238405Sjkim *) Update verify callback code in apps/s_cb.c and apps/verify.c, it 1124238405Sjkim needlessly dereferenced structures, used obsolete functions and 1125238405Sjkim didn't handle all updated verify codes correctly. 1126238405Sjkim [Steve Henson] 1127238405Sjkim 1128238405Sjkim *) Disable MD2 in the default configuration. 1129238405Sjkim [Steve Henson] 1130238405Sjkim 1131238405Sjkim *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to 1132238405Sjkim indicate the initial BIO being pushed or popped. This makes it possible 1133238405Sjkim to determine whether the BIO is the one explicitly called or as a result 1134238405Sjkim of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so 1135238405Sjkim it handles reference counts correctly and doesn't zero out the I/O bio 1136238405Sjkim when it is not being explicitly popped. WARNING: applications which 1137238405Sjkim included workarounds for the old buggy behaviour will need to be modified 1138238405Sjkim or they could free up already freed BIOs. 1139238405Sjkim [Steve Henson] 1140238405Sjkim 1141238405Sjkim *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni 1142238405Sjkim renaming to all platforms (within the 0.9.8 branch, this was 1143238405Sjkim done conditionally on Netware platforms to avoid a name clash). 1144238405Sjkim [Guenter <lists@gknw.net>] 1145238405Sjkim 1146238405Sjkim *) Add ECDHE and PSK support to DTLS. 1147238405Sjkim [Michael Tuexen <tuexen@fh-muenster.de>] 1148238405Sjkim 1149238405Sjkim *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't 1150238405Sjkim be used on C++. 1151238405Sjkim [Steve Henson] 1152238405Sjkim 1153238405Sjkim *) Add "missing" function EVP_MD_flags() (without this the only way to 1154238405Sjkim retrieve a digest flags is by accessing the structure directly. Update 1155238405Sjkim EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest 1156238405Sjkim or cipher is registered as in the "from" argument. Print out all 1157238405Sjkim registered digests in the dgst usage message instead of manually 1158238405Sjkim attempting to work them out. 1159238405Sjkim [Steve Henson] 1160238405Sjkim 1161238405Sjkim *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello: 1162238405Sjkim this allows the use of compression and extensions. Change default cipher 1163238405Sjkim string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2 1164238405Sjkim by default unless an application cipher string requests it. 1165238405Sjkim [Steve Henson] 1166238405Sjkim 1167238405Sjkim *) Alter match criteria in PKCS12_parse(). It used to try to use local 1168238405Sjkim key ids to find matching certificates and keys but some PKCS#12 files 1169238405Sjkim don't follow the (somewhat unwritten) rules and this strategy fails. 1170238405Sjkim Now just gather all certificates together and the first private key 1171238405Sjkim then look for the first certificate that matches the key. 1172238405Sjkim [Steve Henson] 1173238405Sjkim 1174238405Sjkim *) Support use of registered digest and cipher names for dgst and cipher 1175238405Sjkim commands instead of having to add each one as a special case. So now 1176238405Sjkim you can do: 1177238405Sjkim 1178238405Sjkim openssl sha256 foo 1179238405Sjkim 1180238405Sjkim as well as: 1181238405Sjkim 1182238405Sjkim openssl dgst -sha256 foo 1183238405Sjkim 1184238405Sjkim and this works for ENGINE based algorithms too. 1185238405Sjkim 1186238405Sjkim [Steve Henson] 1187238405Sjkim 1188238405Sjkim *) Update Gost ENGINE to support parameter files. 1189238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>] 1190238405Sjkim 1191238405Sjkim *) Support GeneralizedTime in ca utility. 1192238405Sjkim [Oliver Martin <oliver@volatilevoid.net>, Steve Henson] 1193238405Sjkim 1194238405Sjkim *) Enhance the hash format used for certificate directory links. The new 1195238405Sjkim form uses the canonical encoding (meaning equivalent names will work 1196238405Sjkim even if they aren't identical) and uses SHA1 instead of MD5. This form 1197238405Sjkim is incompatible with the older format and as a result c_rehash should 1198238405Sjkim be used to rebuild symbolic links. 1199238405Sjkim [Steve Henson] 1200238405Sjkim 1201238405Sjkim *) Make PKCS#8 the default write format for private keys, replacing the 1202238405Sjkim traditional format. This form is standardised, more secure and doesn't 1203238405Sjkim include an implicit MD5 dependency. 1204238405Sjkim [Steve Henson] 1205238405Sjkim 1206238405Sjkim *) Add a $gcc_devteam_warn option to Configure. The idea is that any code 1207238405Sjkim committed to OpenSSL should pass this lot as a minimum. 1208238405Sjkim [Steve Henson] 1209238405Sjkim 1210238405Sjkim *) Add session ticket override functionality for use by EAP-FAST. 1211238405Sjkim [Jouni Malinen <j@w1.fi>] 1212238405Sjkim 1213238405Sjkim *) Modify HMAC functions to return a value. Since these can be implemented 1214238405Sjkim in an ENGINE errors can occur. 1215238405Sjkim [Steve Henson] 1216238405Sjkim 1217238405Sjkim *) Type-checked OBJ_bsearch_ex. 1218238405Sjkim [Ben Laurie] 1219238405Sjkim 1220238405Sjkim *) Type-checked OBJ_bsearch. Also some constification necessitated 1221238405Sjkim by type-checking. Still to come: TXT_DB, bsearch(?), 1222238405Sjkim OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING, 1223238405Sjkim CONF_VALUE. 1224238405Sjkim [Ben Laurie] 1225238405Sjkim 1226238405Sjkim *) New function OPENSSL_gmtime_adj() to add a specific number of days and 1227238405Sjkim seconds to a tm structure directly, instead of going through OS 1228238405Sjkim specific date routines. This avoids any issues with OS routines such 1229238405Sjkim as the year 2038 bug. New *_adj() functions for ASN1 time structures 1230238405Sjkim and X509_time_adj_ex() to cover the extended range. The existing 1231238405Sjkim X509_time_adj() is still usable and will no longer have any date issues. 1232238405Sjkim [Steve Henson] 1233238405Sjkim 1234238405Sjkim *) Delta CRL support. New use deltas option which will attempt to locate 1235238405Sjkim and search any appropriate delta CRLs available. 1236238405Sjkim 1237238405Sjkim This work was sponsored by Google. 1238238405Sjkim [Steve Henson] 1239238405Sjkim 1240238405Sjkim *) Support for CRLs partitioned by reason code. Reorganise CRL processing 1241238405Sjkim code and add additional score elements. Validate alternate CRL paths 1242238405Sjkim as part of the CRL checking and indicate a new error "CRL path validation 1243238405Sjkim error" in this case. Applications wanting additional details can use 1244238405Sjkim the verify callback and check the new "parent" field. If this is not 1245238405Sjkim NULL CRL path validation is taking place. Existing applications wont 1246238405Sjkim see this because it requires extended CRL support which is off by 1247238405Sjkim default. 1248238405Sjkim 1249238405Sjkim This work was sponsored by Google. 1250238405Sjkim [Steve Henson] 1251238405Sjkim 1252238405Sjkim *) Support for freshest CRL extension. 1253238405Sjkim 1254238405Sjkim This work was sponsored by Google. 1255238405Sjkim [Steve Henson] 1256238405Sjkim 1257238405Sjkim *) Initial indirect CRL support. Currently only supported in the CRLs 1258238405Sjkim passed directly and not via lookup. Process certificate issuer 1259238405Sjkim CRL entry extension and lookup CRL entries by bother issuer name 1260238405Sjkim and serial number. Check and process CRL issuer entry in IDP extension. 1261238405Sjkim 1262238405Sjkim This work was sponsored by Google. 1263238405Sjkim [Steve Henson] 1264238405Sjkim 1265238405Sjkim *) Add support for distinct certificate and CRL paths. The CRL issuer 1266238405Sjkim certificate is validated separately in this case. Only enabled if 1267238405Sjkim an extended CRL support flag is set: this flag will enable additional 1268238405Sjkim CRL functionality in future. 1269238405Sjkim 1270238405Sjkim This work was sponsored by Google. 1271238405Sjkim [Steve Henson] 1272238405Sjkim 1273238405Sjkim *) Add support for policy mappings extension. 1274238405Sjkim 1275238405Sjkim This work was sponsored by Google. 1276238405Sjkim [Steve Henson] 1277238405Sjkim 1278238405Sjkim *) Fixes to pathlength constraint, self issued certificate handling, 1279238405Sjkim policy processing to align with RFC3280 and PKITS tests. 1280238405Sjkim 1281238405Sjkim This work was sponsored by Google. 1282238405Sjkim [Steve Henson] 1283238405Sjkim 1284238405Sjkim *) Support for name constraints certificate extension. DN, email, DNS 1285238405Sjkim and URI types are currently supported. 1286238405Sjkim 1287238405Sjkim This work was sponsored by Google. 1288238405Sjkim [Steve Henson] 1289238405Sjkim 1290238405Sjkim *) To cater for systems that provide a pointer-based thread ID rather 1291238405Sjkim than numeric, deprecate the current numeric thread ID mechanism and 1292238405Sjkim replace it with a structure and associated callback type. This 1293238405Sjkim mechanism allows a numeric "hash" to be extracted from a thread ID in 1294238405Sjkim either case, and on platforms where pointers are larger than 'long', 1295238405Sjkim mixing is done to help ensure the numeric 'hash' is usable even if it 1296238405Sjkim can't be guaranteed unique. The default mechanism is to use "&errno" 1297238405Sjkim as a pointer-based thread ID to distinguish between threads. 1298238405Sjkim 1299238405Sjkim Applications that want to provide their own thread IDs should now use 1300238405Sjkim CRYPTO_THREADID_set_callback() to register a callback that will call 1301238405Sjkim either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer(). 1302238405Sjkim 1303238405Sjkim Note that ERR_remove_state() is now deprecated, because it is tied 1304238405Sjkim to the assumption that thread IDs are numeric. ERR_remove_state(0) 1305238405Sjkim to free the current thread's error state should be replaced by 1306238405Sjkim ERR_remove_thread_state(NULL). 1307238405Sjkim 1308238405Sjkim (This new approach replaces the functions CRYPTO_set_idptr_callback(), 1309238405Sjkim CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in 1310238405Sjkim OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an 1311238405Sjkim application was previously providing a numeric thread callback that 1312238405Sjkim was inappropriate for distinguishing threads, then uniqueness might 1313238405Sjkim have been obtained with &errno that happened immediately in the 1314238405Sjkim intermediate development versions of OpenSSL; this is no longer the 1315238405Sjkim case, the numeric thread callback will now override the automatic use 1316238405Sjkim of &errno.) 1317238405Sjkim [Geoff Thorpe, with help from Bodo Moeller] 1318238405Sjkim 1319238405Sjkim *) Initial support for different CRL issuing certificates. This covers a 1320238405Sjkim simple case where the self issued certificates in the chain exist and 1321238405Sjkim the real CRL issuer is higher in the existing chain. 1322238405Sjkim 1323238405Sjkim This work was sponsored by Google. 1324238405Sjkim [Steve Henson] 1325238405Sjkim 1326238405Sjkim *) Removed effectively defunct crypto/store from the build. 1327238405Sjkim [Ben Laurie] 1328238405Sjkim 1329238405Sjkim *) Revamp of STACK to provide stronger type-checking. Still to come: 1330238405Sjkim TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE, 1331238405Sjkim ASN1_STRING, CONF_VALUE. 1332238405Sjkim [Ben Laurie] 1333238405Sjkim 1334238405Sjkim *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer 1335238405Sjkim RAM on SSL connections. This option can save about 34k per idle SSL. 1336238405Sjkim [Nick Mathewson] 1337238405Sjkim 1338238405Sjkim *) Revamp of LHASH to provide stronger type-checking. Still to come: 1339238405Sjkim STACK, TXT_DB, bsearch, qsort. 1340238405Sjkim [Ben Laurie] 1341238405Sjkim 1342238405Sjkim *) Initial support for Cryptographic Message Syntax (aka CMS) based 1343238405Sjkim on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility, 1344238405Sjkim support for data, signedData, compressedData, digestedData and 1345238405Sjkim encryptedData, envelopedData types included. Scripts to check against 1346238405Sjkim RFC4134 examples draft and interop and consistency checks of many 1347238405Sjkim content types and variants. 1348238405Sjkim [Steve Henson] 1349238405Sjkim 1350238405Sjkim *) Add options to enc utility to support use of zlib compression BIO. 1351238405Sjkim [Steve Henson] 1352238405Sjkim 1353238405Sjkim *) Extend mk1mf to support importing of options and assembly language 1354238405Sjkim files from Configure script, currently only included in VC-WIN32. 1355238405Sjkim The assembly language rules can now optionally generate the source 1356238405Sjkim files from the associated perl scripts. 1357238405Sjkim [Steve Henson] 1358238405Sjkim 1359238405Sjkim *) Implement remaining functionality needed to support GOST ciphersuites. 1360238405Sjkim Interop testing has been performed using CryptoPro implementations. 1361238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>] 1362238405Sjkim 1363238405Sjkim *) s390x assembler pack. 1364238405Sjkim [Andy Polyakov] 1365238405Sjkim 1366238405Sjkim *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU 1367238405Sjkim "family." 1368238405Sjkim [Andy Polyakov] 1369238405Sjkim 1370238405Sjkim *) Implement Opaque PRF Input TLS extension as specified in 1371238405Sjkim draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an 1372238405Sjkim official specification yet and no extension type assignment by 1373238405Sjkim IANA exists, this extension (for now) will have to be explicitly 1374238405Sjkim enabled when building OpenSSL by providing the extension number 1375238405Sjkim to use. For example, specify an option 1376238405Sjkim 1377238405Sjkim -DTLSEXT_TYPE_opaque_prf_input=0x9527 1378238405Sjkim 1379238405Sjkim to the "config" or "Configure" script to enable the extension, 1380238405Sjkim assuming extension number 0x9527 (which is a completely arbitrary 1381238405Sjkim and unofficial assignment based on the MD5 hash of the Internet 1382238405Sjkim Draft). Note that by doing so, you potentially lose 1383238405Sjkim interoperability with other TLS implementations since these might 1384238405Sjkim be using the same extension number for other purposes. 1385238405Sjkim 1386238405Sjkim SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the 1387238405Sjkim opaque PRF input value to use in the handshake. This will create 1388238405Sjkim an interal copy of the length-'len' string at 'src', and will 1389238405Sjkim return non-zero for success. 1390238405Sjkim 1391238405Sjkim To get more control and flexibility, provide a callback function 1392238405Sjkim by using 1393238405Sjkim 1394238405Sjkim SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb) 1395238405Sjkim SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg) 1396238405Sjkim 1397238405Sjkim where 1398238405Sjkim 1399238405Sjkim int (*cb)(SSL *, void *peerinput, size_t len, void *arg); 1400238405Sjkim void *arg; 1401238405Sjkim 1402238405Sjkim Callback function 'cb' will be called in handshakes, and is 1403238405Sjkim expected to use SSL_set_tlsext_opaque_prf_input() as appropriate. 1404238405Sjkim Argument 'arg' is for application purposes (the value as given to 1405238405Sjkim SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly 1406238405Sjkim be provided to the callback function). The callback function 1407238405Sjkim has to return non-zero to report success: usually 1 to use opaque 1408238405Sjkim PRF input just if possible, or 2 to enforce use of the opaque PRF 1409238405Sjkim input. In the latter case, the library will abort the handshake 1410238405Sjkim if opaque PRF input is not successfully negotiated. 1411238405Sjkim 1412238405Sjkim Arguments 'peerinput' and 'len' given to the callback function 1413238405Sjkim will always be NULL and 0 in the case of a client. A server will 1414238405Sjkim see the client's opaque PRF input through these variables if 1415238405Sjkim available (NULL and 0 otherwise). Note that if the server 1416238405Sjkim provides an opaque PRF input, the length must be the same as the 1417238405Sjkim length of the client's opaque PRF input. 1418238405Sjkim 1419238405Sjkim Note that the callback function will only be called when creating 1420238405Sjkim a new session (session resumption can resume whatever was 1421238405Sjkim previously negotiated), and will not be called in SSL 2.0 1422238405Sjkim handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or 1423238405Sjkim SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended 1424238405Sjkim for applications that need to enforce opaque PRF input. 1425238405Sjkim 1426238405Sjkim [Bodo Moeller] 1427238405Sjkim 1428238405Sjkim *) Update ssl code to support digests other than SHA1+MD5 for handshake 1429238405Sjkim MAC. 1430238405Sjkim 1431238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>] 1432238405Sjkim 1433238405Sjkim *) Add RFC4507 support to OpenSSL. This includes the corrections in 1434238405Sjkim RFC4507bis. The encrypted ticket format is an encrypted encoded 1435238405Sjkim SSL_SESSION structure, that way new session features are automatically 1436238405Sjkim supported. 1437238405Sjkim 1438238405Sjkim If a client application caches session in an SSL_SESSION structure 1439238405Sjkim support is transparent because tickets are now stored in the encoded 1440238405Sjkim SSL_SESSION. 1441238405Sjkim 1442238405Sjkim The SSL_CTX structure automatically generates keys for ticket 1443238405Sjkim protection in servers so again support should be possible 1444238405Sjkim with no application modification. 1445238405Sjkim 1446238405Sjkim If a client or server wishes to disable RFC4507 support then the option 1447238405Sjkim SSL_OP_NO_TICKET can be set. 1448238405Sjkim 1449238405Sjkim Add a TLS extension debugging callback to allow the contents of any client 1450238405Sjkim or server extensions to be examined. 1451238405Sjkim 1452238405Sjkim This work was sponsored by Google. 1453238405Sjkim [Steve Henson] 1454238405Sjkim 1455238405Sjkim *) Final changes to avoid use of pointer pointer casts in OpenSSL. 1456238405Sjkim OpenSSL should now compile cleanly on gcc 4.2 1457238405Sjkim [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson] 1458238405Sjkim 1459238405Sjkim *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC 1460238405Sjkim support including streaming MAC support: this is required for GOST 1461238405Sjkim ciphersuite support. 1462238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson] 1463238405Sjkim 1464238405Sjkim *) Add option -stream to use PKCS#7 streaming in smime utility. New 1465238405Sjkim function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream() 1466238405Sjkim to output in BER and PEM format. 1467238405Sjkim [Steve Henson] 1468238405Sjkim 1469238405Sjkim *) Experimental support for use of HMAC via EVP_PKEY interface. This 1470238405Sjkim allows HMAC to be handled via the EVP_DigestSign*() interface. The 1471238405Sjkim EVP_PKEY "key" in this case is the HMAC key, potentially allowing 1472238405Sjkim ENGINE support for HMAC keys which are unextractable. New -mac and 1473238405Sjkim -macopt options to dgst utility. 1474238405Sjkim [Steve Henson] 1475238405Sjkim 1476238405Sjkim *) New option -sigopt to dgst utility. Update dgst to use 1477238405Sjkim EVP_Digest{Sign,Verify}*. These two changes make it possible to use 1478238405Sjkim alternative signing paramaters such as X9.31 or PSS in the dgst 1479238405Sjkim utility. 1480238405Sjkim [Steve Henson] 1481238405Sjkim 1482238405Sjkim *) Change ssl_cipher_apply_rule(), the internal function that does 1483238405Sjkim the work each time a ciphersuite string requests enabling 1484238405Sjkim ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or 1485238405Sjkim removing ("!foo+bar") a class of ciphersuites: Now it maintains 1486238405Sjkim the order of disabled ciphersuites such that those ciphersuites 1487238405Sjkim that most recently went from enabled to disabled not only stay 1488238405Sjkim in order with respect to each other, but also have higher priority 1489238405Sjkim than other disabled ciphersuites the next time ciphersuites are 1490238405Sjkim enabled again. 1491238405Sjkim 1492238405Sjkim This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable 1493238405Sjkim the same ciphersuites as with "HIGH" alone, but in a specific 1494238405Sjkim order where the PSK ciphersuites come first (since they are the 1495238405Sjkim most recently disabled ciphersuites when "HIGH" is parsed). 1496238405Sjkim 1497238405Sjkim Also, change ssl_create_cipher_list() (using this new 1498238405Sjkim funcionality) such that between otherwise identical 1499238405Sjkim cihpersuites, ephemeral ECDH is preferred over ephemeral DH in 1500238405Sjkim the default order. 1501238405Sjkim [Bodo Moeller] 1502238405Sjkim 1503238405Sjkim *) Change ssl_create_cipher_list() so that it automatically 1504238405Sjkim arranges the ciphersuites in reasonable order before starting 1505238405Sjkim to process the rule string. Thus, the definition for "DEFAULT" 1506238405Sjkim (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but 1507238405Sjkim remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH". 1508238405Sjkim This makes it much easier to arrive at a reasonable default order 1509238405Sjkim in applications for which anonymous ciphers are OK (meaning 1510238405Sjkim that you can't actually use DEFAULT). 1511238405Sjkim [Bodo Moeller; suggested by Victor Duchovni] 1512238405Sjkim 1513238405Sjkim *) Split the SSL/TLS algorithm mask (as used for ciphersuite string 1514238405Sjkim processing) into multiple integers instead of setting 1515238405Sjkim "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK", 1516238405Sjkim "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer. 1517238405Sjkim (These masks as well as the individual bit definitions are hidden 1518238405Sjkim away into the non-exported interface ssl/ssl_locl.h, so this 1519238405Sjkim change to the definition of the SSL_CIPHER structure shouldn't 1520238405Sjkim affect applications.) This give us more bits for each of these 1521238405Sjkim categories, so there is no longer a need to coagulate AES128 and 1522238405Sjkim AES256 into a single algorithm bit, and to coagulate Camellia128 1523238405Sjkim and Camellia256 into a single algorithm bit, which has led to all 1524238405Sjkim kinds of kludges. 1525238405Sjkim 1526238405Sjkim Thus, among other things, the kludge introduced in 0.9.7m and 1527238405Sjkim 0.9.8e for masking out AES256 independently of AES128 or masking 1528238405Sjkim out Camellia256 independently of AES256 is not needed here in 0.9.9. 1529238405Sjkim 1530238405Sjkim With the change, we also introduce new ciphersuite aliases that 1531238405Sjkim so far were missing: "AES128", "AES256", "CAMELLIA128", and 1532238405Sjkim "CAMELLIA256". 1533238405Sjkim [Bodo Moeller] 1534238405Sjkim 1535238405Sjkim *) Add support for dsa-with-SHA224 and dsa-with-SHA256. 1536238405Sjkim Use the leftmost N bytes of the signature input if the input is 1537238405Sjkim larger than the prime q (with N being the size in bytes of q). 1538238405Sjkim [Nils Larsch] 1539238405Sjkim 1540238405Sjkim *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses 1541238405Sjkim it yet and it is largely untested. 1542238405Sjkim [Steve Henson] 1543238405Sjkim 1544238405Sjkim *) Add support for the ecdsa-with-SHA224/256/384/512 signature types. 1545238405Sjkim [Nils Larsch] 1546238405Sjkim 1547238405Sjkim *) Initial incomplete changes to avoid need for function casts in OpenSSL 1548238405Sjkim some compilers (gcc 4.2 and later) reject their use. Safestack is 1549238405Sjkim reimplemented. Update ASN1 to avoid use of legacy functions. 1550238405Sjkim [Steve Henson] 1551238405Sjkim 1552238405Sjkim *) Win32/64 targets are linked with Winsock2. 1553238405Sjkim [Andy Polyakov] 1554238405Sjkim 1555238405Sjkim *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected 1556238405Sjkim to external functions. This can be used to increase CRL handling 1557238405Sjkim efficiency especially when CRLs are very large by (for example) storing 1558238405Sjkim the CRL revoked certificates in a database. 1559238405Sjkim [Steve Henson] 1560238405Sjkim 1561238405Sjkim *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so 1562238405Sjkim new CRLs added to a directory can be used. New command line option 1563238405Sjkim -verify_return_error to s_client and s_server. This causes real errors 1564238405Sjkim to be returned by the verify callback instead of carrying on no matter 1565238405Sjkim what. This reflects the way a "real world" verify callback would behave. 1566238405Sjkim [Steve Henson] 1567238405Sjkim 1568238405Sjkim *) GOST engine, supporting several GOST algorithms and public key formats. 1569238405Sjkim Kindly donated by Cryptocom. 1570238405Sjkim [Cryptocom] 1571238405Sjkim 1572238405Sjkim *) Partial support for Issuing Distribution Point CRL extension. CRLs 1573238405Sjkim partitioned by DP are handled but no indirect CRL or reason partitioning 1574238405Sjkim (yet). Complete overhaul of CRL handling: now the most suitable CRL is 1575238405Sjkim selected via a scoring technique which handles IDP and AKID in CRLs. 1576238405Sjkim [Steve Henson] 1577238405Sjkim 1578238405Sjkim *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which 1579238405Sjkim will ultimately be used for all verify operations: this will remove the 1580238405Sjkim X509_STORE dependency on certificate verification and allow alternative 1581238405Sjkim lookup methods. X509_STORE based implementations of these two callbacks. 1582238405Sjkim [Steve Henson] 1583238405Sjkim 1584238405Sjkim *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names. 1585238405Sjkim Modify get_crl() to find a valid (unexpired) CRL if possible. 1586238405Sjkim [Steve Henson] 1587238405Sjkim 1588238405Sjkim *) New function X509_CRL_match() to check if two CRLs are identical. Normally 1589238405Sjkim this would be called X509_CRL_cmp() but that name is already used by 1590238405Sjkim a function that just compares CRL issuer names. Cache several CRL 1591238405Sjkim extensions in X509_CRL structure and cache CRLDP in X509. 1592238405Sjkim [Steve Henson] 1593238405Sjkim 1594238405Sjkim *) Store a "canonical" representation of X509_NAME structure (ASN1 Name) 1595238405Sjkim this maps equivalent X509_NAME structures into a consistent structure. 1596238405Sjkim Name comparison can then be performed rapidly using memcmp(). 1597238405Sjkim [Steve Henson] 1598238405Sjkim 1599238405Sjkim *) Non-blocking OCSP request processing. Add -timeout option to ocsp 1600238405Sjkim utility. 1601238405Sjkim [Steve Henson] 1602238405Sjkim 1603238405Sjkim *) Allow digests to supply their own micalg string for S/MIME type using 1604238405Sjkim the ctrl EVP_MD_CTRL_MICALG. 1605238405Sjkim [Steve Henson] 1606238405Sjkim 1607238405Sjkim *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the 1608238405Sjkim EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN 1609238405Sjkim ctrl. It can then customise the structure before and/or after signing 1610238405Sjkim if necessary. 1611238405Sjkim [Steve Henson] 1612238405Sjkim 1613238405Sjkim *) New function OBJ_add_sigid() to allow application defined signature OIDs 1614238405Sjkim to be added to OpenSSLs internal tables. New function OBJ_sigid_free() 1615238405Sjkim to free up any added signature OIDs. 1616238405Sjkim [Steve Henson] 1617238405Sjkim 1618238405Sjkim *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(), 1619238405Sjkim EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal 1620238405Sjkim digest and cipher tables. New options added to openssl utility: 1621238405Sjkim list-message-digest-algorithms and list-cipher-algorithms. 1622238405Sjkim [Steve Henson] 1623238405Sjkim 1624238405Sjkim *) Change the array representation of binary polynomials: the list 1625238405Sjkim of degrees of non-zero coefficients is now terminated with -1. 1626238405Sjkim Previously it was terminated with 0, which was also part of the 1627238405Sjkim value; thus, the array representation was not applicable to 1628238405Sjkim polynomials where t^0 has coefficient zero. This change makes 1629238405Sjkim the array representation useful in a more general context. 1630238405Sjkim [Douglas Stebila] 1631238405Sjkim 1632238405Sjkim *) Various modifications and fixes to SSL/TLS cipher string 1633238405Sjkim handling. For ECC, the code now distinguishes between fixed ECDH 1634238405Sjkim with RSA certificates on the one hand and with ECDSA certificates 1635238405Sjkim on the other hand, since these are separate ciphersuites. The 1636238405Sjkim unused code for Fortezza ciphersuites has been removed. 1637238405Sjkim 1638238405Sjkim For consistency with EDH, ephemeral ECDH is now called "EECDH" 1639238405Sjkim (not "ECDHE"). For consistency with the code for DH 1640238405Sjkim certificates, use of ECDH certificates is now considered ECDH 1641238405Sjkim authentication, not RSA or ECDSA authentication (the latter is 1642238405Sjkim merely the CA's signing algorithm and not actively used in the 1643238405Sjkim protocol). 1644238405Sjkim 1645238405Sjkim The temporary ciphersuite alias "ECCdraft" is no longer 1646238405Sjkim available, and ECC ciphersuites are no longer excluded from "ALL" 1647238405Sjkim and "DEFAULT". The following aliases now exist for RFC 4492 1648238405Sjkim ciphersuites, most of these by analogy with the DH case: 1649238405Sjkim 1650238405Sjkim kECDHr - ECDH cert, signed with RSA 1651238405Sjkim kECDHe - ECDH cert, signed with ECDSA 1652238405Sjkim kECDH - ECDH cert (signed with either RSA or ECDSA) 1653238405Sjkim kEECDH - ephemeral ECDH 1654238405Sjkim ECDH - ECDH cert or ephemeral ECDH 1655238405Sjkim 1656238405Sjkim aECDH - ECDH cert 1657238405Sjkim aECDSA - ECDSA cert 1658238405Sjkim ECDSA - ECDSA cert 1659238405Sjkim 1660238405Sjkim AECDH - anonymous ECDH 1661238405Sjkim EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH") 1662238405Sjkim 1663238405Sjkim [Bodo Moeller] 1664238405Sjkim 1665238405Sjkim *) Add additional S/MIME capabilities for AES and GOST ciphers if supported. 1666238405Sjkim Use correct micalg parameters depending on digest(s) in signed message. 1667238405Sjkim [Steve Henson] 1668238405Sjkim 1669238405Sjkim *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process 1670238405Sjkim an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code. 1671238405Sjkim [Steve Henson] 1672238405Sjkim 1673238405Sjkim *) Initial engine support for EVP_PKEY_METHOD. New functions to permit 1674238405Sjkim an engine to register a method. Add ENGINE lookups for methods and 1675238405Sjkim functional reference processing. 1676238405Sjkim [Steve Henson] 1677238405Sjkim 1678238405Sjkim *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of 1679238405Sjkim EVP_{Sign,Verify}* which allow an application to customise the signature 1680238405Sjkim process. 1681238405Sjkim [Steve Henson] 1682238405Sjkim 1683238405Sjkim *) New -resign option to smime utility. This adds one or more signers 1684238405Sjkim to an existing PKCS#7 signedData structure. Also -md option to use an 1685238405Sjkim alternative message digest algorithm for signing. 1686238405Sjkim [Steve Henson] 1687238405Sjkim 1688238405Sjkim *) Tidy up PKCS#7 routines and add new functions to make it easier to 1689238405Sjkim create PKCS7 structures containing multiple signers. Update smime 1690238405Sjkim application to support multiple signers. 1691238405Sjkim [Steve Henson] 1692238405Sjkim 1693238405Sjkim *) New -macalg option to pkcs12 utility to allow setting of an alternative 1694238405Sjkim digest MAC. 1695238405Sjkim [Steve Henson] 1696238405Sjkim 1697238405Sjkim *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC. 1698238405Sjkim Reorganize PBE internals to lookup from a static table using NIDs, 1699238405Sjkim add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl: 1700238405Sjkim EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative 1701238405Sjkim PRF which will be automatically used with PBES2. 1702238405Sjkim [Steve Henson] 1703238405Sjkim 1704238405Sjkim *) Replace the algorithm specific calls to generate keys in "req" with the 1705238405Sjkim new API. 1706238405Sjkim [Steve Henson] 1707238405Sjkim 1708238405Sjkim *) Update PKCS#7 enveloped data routines to use new API. This is now 1709238405Sjkim supported by any public key method supporting the encrypt operation. A 1710238405Sjkim ctrl is added to allow the public key algorithm to examine or modify 1711238405Sjkim the PKCS#7 RecipientInfo structure if it needs to: for RSA this is 1712238405Sjkim a no op. 1713238405Sjkim [Steve Henson] 1714238405Sjkim 1715238405Sjkim *) Add a ctrl to asn1 method to allow a public key algorithm to express 1716238405Sjkim a default digest type to use. In most cases this will be SHA1 but some 1717238405Sjkim algorithms (such as GOST) need to specify an alternative digest. The 1718238405Sjkim return value indicates how strong the prefernce is 1 means optional and 1719238405Sjkim 2 is mandatory (that is it is the only supported type). Modify 1720238405Sjkim ASN1_item_sign() to accept a NULL digest argument to indicate it should 1721238405Sjkim use the default md. Update openssl utilities to use the default digest 1722238405Sjkim type for signing if it is not explicitly indicated. 1723238405Sjkim [Steve Henson] 1724238405Sjkim 1725238405Sjkim *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New 1726238405Sjkim EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant 1727238405Sjkim signing method from the key type. This effectively removes the link 1728238405Sjkim between digests and public key types. 1729238405Sjkim [Steve Henson] 1730238405Sjkim 1731238405Sjkim *) Add an OID cross reference table and utility functions. Its purpose is to 1732238405Sjkim translate between signature OIDs such as SHA1WithrsaEncryption and SHA1, 1733238405Sjkim rsaEncryption. This will allow some of the algorithm specific hackery 1734238405Sjkim needed to use the correct OID to be removed. 1735238405Sjkim [Steve Henson] 1736238405Sjkim 1737238405Sjkim *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO 1738238405Sjkim structures for PKCS7_sign(). They are now set up by the relevant public 1739238405Sjkim key ASN1 method. 1740238405Sjkim [Steve Henson] 1741238405Sjkim 1742238405Sjkim *) Add provisional EC pkey method with support for ECDSA and ECDH. 1743238405Sjkim [Steve Henson] 1744238405Sjkim 1745238405Sjkim *) Add support for key derivation (agreement) in the API, DH method and 1746238405Sjkim pkeyutl. 1747238405Sjkim [Steve Henson] 1748238405Sjkim 1749238405Sjkim *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support 1750238405Sjkim public and private key formats. As a side effect these add additional 1751238405Sjkim command line functionality not previously available: DSA signatures can be 1752238405Sjkim generated and verified using pkeyutl and DH key support and generation in 1753238405Sjkim pkey, genpkey. 1754238405Sjkim [Steve Henson] 1755238405Sjkim 1756238405Sjkim *) BeOS support. 1757238405Sjkim [Oliver Tappe <zooey@hirschkaefer.de>] 1758238405Sjkim 1759238405Sjkim *) New make target "install_html_docs" installs HTML renditions of the 1760238405Sjkim manual pages. 1761238405Sjkim [Oliver Tappe <zooey@hirschkaefer.de>] 1762238405Sjkim 1763238405Sjkim *) New utility "genpkey" this is analagous to "genrsa" etc except it can 1764238405Sjkim generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to 1765238405Sjkim support key and parameter generation and add initial key generation 1766238405Sjkim functionality for RSA. 1767238405Sjkim [Steve Henson] 1768238405Sjkim 1769238405Sjkim *) Add functions for main EVP_PKEY_method operations. The undocumented 1770238405Sjkim functions EVP_PKEY_{encrypt,decrypt} have been renamed to 1771238405Sjkim EVP_PKEY_{encrypt,decrypt}_old. 1772238405Sjkim [Steve Henson] 1773238405Sjkim 1774238405Sjkim *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public 1775238405Sjkim key API, doesn't do much yet. 1776238405Sjkim [Steve Henson] 1777238405Sjkim 1778238405Sjkim *) New function EVP_PKEY_asn1_get0_info() to retrieve information about 1779238405Sjkim public key algorithms. New option to openssl utility: 1780238405Sjkim "list-public-key-algorithms" to print out info. 1781238405Sjkim [Steve Henson] 1782238405Sjkim 1783238405Sjkim *) Implement the Supported Elliptic Curves Extension for 1784238405Sjkim ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 1785238405Sjkim [Douglas Stebila] 1786238405Sjkim 1787238405Sjkim *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or 1788238405Sjkim EVP_CIPHER structures to avoid later problems in EVP_cleanup(). 1789238405Sjkim [Steve Henson] 1790238405Sjkim 1791238405Sjkim *) New utilities pkey and pkeyparam. These are similar to algorithm specific 1792238405Sjkim utilities such as rsa, dsa, dsaparam etc except they process any key 1793238405Sjkim type. 1794238405Sjkim [Steve Henson] 1795238405Sjkim 1796238405Sjkim *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New 1797238405Sjkim functions EVP_PKEY_print_public(), EVP_PKEY_print_private(), 1798238405Sjkim EVP_PKEY_print_param() to print public key data from an EVP_PKEY 1799238405Sjkim structure. 1800238405Sjkim [Steve Henson] 1801238405Sjkim 1802238405Sjkim *) Initial support for pluggable public key ASN1. 1803238405Sjkim De-spaghettify the public key ASN1 handling. Move public and private 1804238405Sjkim key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate 1805238405Sjkim algorithm specific handling to a single module within the relevant 1806238405Sjkim algorithm directory. Add functions to allow (near) opaque processing 1807238405Sjkim of public and private key structures. 1808238405Sjkim [Steve Henson] 1809238405Sjkim 1810238405Sjkim *) Implement the Supported Point Formats Extension for 1811238405Sjkim ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 1812238405Sjkim [Douglas Stebila] 1813238405Sjkim 1814238405Sjkim *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members 1815238405Sjkim for the psk identity [hint] and the psk callback functions to the 1816238405Sjkim SSL_SESSION, SSL and SSL_CTX structure. 1817238405Sjkim 1818238405Sjkim New ciphersuites: 1819238405Sjkim PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA, 1820238405Sjkim PSK-AES256-CBC-SHA 1821238405Sjkim 1822238405Sjkim New functions: 1823238405Sjkim SSL_CTX_use_psk_identity_hint 1824238405Sjkim SSL_get_psk_identity_hint 1825238405Sjkim SSL_get_psk_identity 1826238405Sjkim SSL_use_psk_identity_hint 1827238405Sjkim 1828238405Sjkim [Mika Kousa and Pasi Eronen of Nokia Corporation] 1829238405Sjkim 1830238405Sjkim *) Add RFC 3161 compliant time stamp request creation, response generation 1831238405Sjkim and response verification functionality. 1832238405Sjkim [Zolt�n Gl�zik <zglozik@opentsa.org>, The OpenTSA Project] 1833238405Sjkim 1834238405Sjkim *) Add initial support for TLS extensions, specifically for the server_name 1835238405Sjkim extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 1836238405Sjkim have new members for a host name. The SSL data structure has an 1837238405Sjkim additional member SSL_CTX *initial_ctx so that new sessions can be 1838238405Sjkim stored in that context to allow for session resumption, even after the 1839238405Sjkim SSL has been switched to a new SSL_CTX in reaction to a client's 1840238405Sjkim server_name extension. 1841238405Sjkim 1842238405Sjkim New functions (subject to change): 1843238405Sjkim 1844238405Sjkim SSL_get_servername() 1845238405Sjkim SSL_get_servername_type() 1846238405Sjkim SSL_set_SSL_CTX() 1847238405Sjkim 1848238405Sjkim New CTRL codes and macros (subject to change): 1849238405Sjkim 1850238405Sjkim SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 1851238405Sjkim - SSL_CTX_set_tlsext_servername_callback() 1852238405Sjkim SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 1853238405Sjkim - SSL_CTX_set_tlsext_servername_arg() 1854238405Sjkim SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 1855238405Sjkim 1856238405Sjkim openssl s_client has a new '-servername ...' option. 1857238405Sjkim 1858238405Sjkim openssl s_server has new options '-servername_host ...', '-cert2 ...', 1859238405Sjkim '-key2 ...', '-servername_fatal' (subject to change). This allows 1860238405Sjkim testing the HostName extension for a specific single host name ('-cert' 1861238405Sjkim and '-key' remain fallbacks for handshakes without HostName 1862238405Sjkim negotiation). If the unrecogninzed_name alert has to be sent, this by 1863238405Sjkim default is a warning; it becomes fatal with the '-servername_fatal' 1864238405Sjkim option. 1865238405Sjkim 1866238405Sjkim [Peter Sylvester, Remy Allais, Christophe Renou] 1867238405Sjkim 1868238405Sjkim *) Whirlpool hash implementation is added. 1869238405Sjkim [Andy Polyakov] 1870238405Sjkim 1871238405Sjkim *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to 1872238405Sjkim bn(64,32). Because of instruction set limitations it doesn't have 1873238405Sjkim any negative impact on performance. This was done mostly in order 1874238405Sjkim to make it possible to share assembler modules, such as bn_mul_mont 1875238405Sjkim implementations, between 32- and 64-bit builds without hassle. 1876238405Sjkim [Andy Polyakov] 1877238405Sjkim 1878238405Sjkim *) Move code previously exiled into file crypto/ec/ec2_smpt.c 1879238405Sjkim to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP 1880238405Sjkim macro. 1881238405Sjkim [Bodo Moeller] 1882238405Sjkim 1883238405Sjkim *) New candidate for BIGNUM assembler implementation, bn_mul_mont, 1884238405Sjkim dedicated Montgomery multiplication procedure, is introduced. 1885238405Sjkim BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher 1886238405Sjkim "64-bit" performance on certain 32-bit targets. 1887238405Sjkim [Andy Polyakov] 1888238405Sjkim 1889238405Sjkim *) New option SSL_OP_NO_COMP to disable use of compression selectively 1890238405Sjkim in SSL structures. New SSL ctrl to set maximum send fragment size. 1891238405Sjkim Save memory by seeting the I/O buffer sizes dynamically instead of 1892238405Sjkim using the maximum available value. 1893238405Sjkim [Steve Henson] 1894238405Sjkim 1895238405Sjkim *) New option -V for 'openssl ciphers'. This prints the ciphersuite code 1896238405Sjkim in addition to the text details. 1897238405Sjkim [Bodo Moeller] 1898238405Sjkim 1899238405Sjkim *) Very, very preliminary EXPERIMENTAL support for printing of general 1900238405Sjkim ASN1 structures. This currently produces rather ugly output and doesn't 1901238405Sjkim handle several customised structures at all. 1902238405Sjkim [Steve Henson] 1903238405Sjkim 1904238405Sjkim *) Integrated support for PVK file format and some related formats such 1905238405Sjkim as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support 1906238405Sjkim these in the 'rsa' and 'dsa' utilities. 1907238405Sjkim [Steve Henson] 1908238405Sjkim 1909238405Sjkim *) Support for PKCS#1 RSAPublicKey format on rsa utility command line. 1910238405Sjkim [Steve Henson] 1911238405Sjkim 1912238405Sjkim *) Remove the ancient ASN1_METHOD code. This was only ever used in one 1913238405Sjkim place for the (very old) "NETSCAPE" format certificates which are now 1914238405Sjkim handled using new ASN1 code equivalents. 1915238405Sjkim [Steve Henson] 1916238405Sjkim 1917238405Sjkim *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD 1918238405Sjkim pointer and make the SSL_METHOD parameter in SSL_CTX_new, 1919238405Sjkim SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'. 1920238405Sjkim [Nils Larsch] 1921238405Sjkim 1922238405Sjkim *) Modify CRL distribution points extension code to print out previously 1923238405Sjkim unsupported fields. Enhance extension setting code to allow setting of 1924238405Sjkim all fields. 1925238405Sjkim [Steve Henson] 1926238405Sjkim 1927238405Sjkim *) Add print and set support for Issuing Distribution Point CRL extension. 1928238405Sjkim [Steve Henson] 1929238405Sjkim 1930238405Sjkim *) Change 'Configure' script to enable Camellia by default. 1931238405Sjkim [NTT] 1932238405Sjkim 1933206046Ssimon Changes between 0.9.8m and 0.9.8n [24 Mar 2010] 1934206046Ssimon 1935206046Ssimon *) When rejecting SSL/TLS records due to an incorrect version number, never 1936206046Ssimon update s->server with a new major version number. As of 1937206046Ssimon - OpenSSL 0.9.8m if 'short' is a 16-bit type, 1938206046Ssimon - OpenSSL 0.9.8f if 'short' is longer than 16 bits, 1939206046Ssimon the previous behavior could result in a read attempt at NULL when 1940206046Ssimon receiving specific incorrect SSL/TLS records once record payload 1941206046Ssimon protection is active. (CVE-2010-0740) 1942206046Ssimon [Bodo Moeller, Adam Langley <agl@chromium.org>] 1943206046Ssimon 1944206046Ssimon *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL 1945206046Ssimon could be crashed if the relevant tables were not present (e.g. chrooted). 1946206046Ssimon [Tomas Hoger <thoger@redhat.com>] 1947206046Ssimon 1948205128Ssimon Changes between 0.9.8l and 0.9.8m [25 Feb 2010] 1949205128Ssimon 1950205128Ssimon *) Always check bn_wexpend() return values for failure. (CVE-2009-3245) 1951205128Ssimon [Martin Olsson, Neel Mehta] 1952205128Ssimon 1953205128Ssimon *) Fix X509_STORE locking: Every 'objs' access requires a lock (to 1954205128Ssimon accommodate for stack sorting, always a write lock!). 1955205128Ssimon [Bodo Moeller] 1956205128Ssimon 1957205128Ssimon *) On some versions of WIN32 Heap32Next is very slow. This can cause 1958205128Ssimon excessive delays in the RAND_poll(): over a minute. As a workaround 1959205128Ssimon include a time check in the inner Heap32Next loop too. 1960205128Ssimon [Steve Henson] 1961205128Ssimon 1962205128Ssimon *) The code that handled flushing of data in SSL/TLS originally used the 1963205128Ssimon BIO_CTRL_INFO ctrl to see if any data was pending first. This caused 1964205128Ssimon the problem outlined in PR#1949. The fix suggested there however can 1965205128Ssimon trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions 1966205128Ssimon of Apache). So instead simplify the code to flush unconditionally. 1967205128Ssimon This should be fine since flushing with no data to flush is a no op. 1968205128Ssimon [Steve Henson] 1969205128Ssimon 1970205128Ssimon *) Handle TLS versions 2.0 and later properly and correctly use the 1971205128Ssimon highest version of TLS/SSL supported. Although TLS >= 2.0 is some way 1972205128Ssimon off ancient servers have a habit of sticking around for a while... 1973205128Ssimon [Steve Henson] 1974205128Ssimon 1975205128Ssimon *) Modify compression code so it frees up structures without using the 1976205128Ssimon ex_data callbacks. This works around a problem where some applications 1977205128Ssimon call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when 1978205128Ssimon restarting) then use compression (e.g. SSL with compression) later. 1979205128Ssimon This results in significant per-connection memory leaks and 1980205128Ssimon has caused some security issues including CVE-2008-1678 and 1981205128Ssimon CVE-2009-4355. 1982205128Ssimon [Steve Henson] 1983205128Ssimon 1984205128Ssimon *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't 1985205128Ssimon change when encrypting or decrypting. 1986205128Ssimon [Bodo Moeller] 1987205128Ssimon 1988205128Ssimon *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to 1989205128Ssimon connect and renegotiate with servers which do not support RI. 1990205128Ssimon Until RI is more widely deployed this option is enabled by default. 1991205128Ssimon [Steve Henson] 1992205128Ssimon 1993205128Ssimon *) Add "missing" ssl ctrls to clear options and mode. 1994205128Ssimon [Steve Henson] 1995205128Ssimon 1996205128Ssimon *) If client attempts to renegotiate and doesn't support RI respond with 1997205128Ssimon a no_renegotiation alert as required by RFC5746. Some renegotiating 1998205128Ssimon TLS clients will continue a connection gracefully when they receive 1999205128Ssimon the alert. Unfortunately OpenSSL mishandled this alert and would hang 2000205128Ssimon waiting for a server hello which it will never receive. Now we treat a 2001205128Ssimon received no_renegotiation alert as a fatal error. This is because 2002205128Ssimon applications requesting a renegotiation might well expect it to succeed 2003205128Ssimon and would have no code in place to handle the server denying it so the 2004205128Ssimon only safe thing to do is to terminate the connection. 2005205128Ssimon [Steve Henson] 2006205128Ssimon 2007205128Ssimon *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if 2008205128Ssimon peer supports secure renegotiation and 0 otherwise. Print out peer 2009205128Ssimon renegotiation support in s_client/s_server. 2010205128Ssimon [Steve Henson] 2011205128Ssimon 2012205128Ssimon *) Replace the highly broken and deprecated SPKAC certification method with 2013205128Ssimon the updated NID creation version. This should correctly handle UTF8. 2014205128Ssimon [Steve Henson] 2015205128Ssimon 2016205128Ssimon *) Implement RFC5746. Re-enable renegotiation but require the extension 2017205128Ssimon as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION 2018205128Ssimon turns out to be a bad idea. It has been replaced by 2019205128Ssimon SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with 2020205128Ssimon SSL_CTX_set_options(). This is really not recommended unless you 2021205128Ssimon know what you are doing. 2022205128Ssimon [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson] 2023205128Ssimon 2024205128Ssimon *) Fixes to stateless session resumption handling. Use initial_ctx when 2025205128Ssimon issuing and attempting to decrypt tickets in case it has changed during 2026205128Ssimon servername handling. Use a non-zero length session ID when attempting 2027205128Ssimon stateless session resumption: this makes it possible to determine if 2028205128Ssimon a resumption has occurred immediately after receiving server hello 2029205128Ssimon (several places in OpenSSL subtly assume this) instead of later in 2030205128Ssimon the handshake. 2031205128Ssimon [Steve Henson] 2032205128Ssimon 2033205128Ssimon *) The functions ENGINE_ctrl(), OPENSSL_isservice(), 2034205128Ssimon CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error 2035205128Ssimon fixes for a few places where the return code is not checked 2036205128Ssimon correctly. 2037205128Ssimon [Julia Lawall <julia@diku.dk>] 2038205128Ssimon 2039205128Ssimon *) Add --strict-warnings option to Configure script to include devteam 2040205128Ssimon warnings in other configurations. 2041205128Ssimon [Steve Henson] 2042205128Ssimon 2043205128Ssimon *) Add support for --libdir option and LIBDIR variable in makefiles. This 2044205128Ssimon makes it possible to install openssl libraries in locations which 2045205128Ssimon have names other than "lib", for example "/usr/lib64" which some 2046205128Ssimon systems need. 2047205128Ssimon [Steve Henson, based on patch from Jeremy Utley] 2048205128Ssimon 2049205128Ssimon *) Don't allow the use of leading 0x80 in OIDs. This is a violation of 2050205128Ssimon X690 8.9.12 and can produce some misleading textual output of OIDs. 2051205128Ssimon [Steve Henson, reported by Dan Kaminsky] 2052205128Ssimon 2053205128Ssimon *) Delete MD2 from algorithm tables. This follows the recommendation in 2054205128Ssimon several standards that it is not used in new applications due to 2055205128Ssimon several cryptographic weaknesses. For binary compatibility reasons 2056205128Ssimon the MD2 API is still compiled in by default. 2057205128Ssimon [Steve Henson] 2058205128Ssimon 2059205128Ssimon *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved 2060205128Ssimon and restored. 2061205128Ssimon [Steve Henson] 2062205128Ssimon 2063205128Ssimon *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and 2064205128Ssimon OPENSSL_asc2uni conditionally on Netware platforms to avoid a name 2065205128Ssimon clash. 2066205128Ssimon [Guenter <lists@gknw.net>] 2067205128Ssimon 2068205128Ssimon *) Fix the server certificate chain building code to use X509_verify_cert(), 2069205128Ssimon it used to have an ad-hoc builder which was unable to cope with anything 2070205128Ssimon other than a simple chain. 2071205128Ssimon [David Woodhouse <dwmw2@infradead.org>, Steve Henson] 2072205128Ssimon 2073205128Ssimon *) Don't check self signed certificate signatures in X509_verify_cert() 2074205128Ssimon by default (a flag can override this): it just wastes time without 2075205128Ssimon adding any security. As a useful side effect self signed root CAs 2076205128Ssimon with non-FIPS digests are now usable in FIPS mode. 2077205128Ssimon [Steve Henson] 2078205128Ssimon 2079205128Ssimon *) In dtls1_process_out_of_seq_message() the check if the current message 2080205128Ssimon is already buffered was missing. For every new message was memory 2081205128Ssimon allocated, allowing an attacker to perform an denial of service attack 2082205128Ssimon with sending out of seq handshake messages until there is no memory 2083205128Ssimon left. Additionally every future messege was buffered, even if the 2084205128Ssimon sequence number made no sense and would be part of another handshake. 2085205128Ssimon So only messages with sequence numbers less than 10 in advance will be 2086205128Ssimon buffered. (CVE-2009-1378) 2087205128Ssimon [Robin Seggelmann, discovered by Daniel Mentz] 2088205128Ssimon 2089205128Ssimon *) Records are buffered if they arrive with a future epoch to be 2090205128Ssimon processed after finishing the corresponding handshake. There is 2091205128Ssimon currently no limitation to this buffer allowing an attacker to perform 2092205128Ssimon a DOS attack with sending records with future epochs until there is no 2093205128Ssimon memory left. This patch adds the pqueue_size() function to detemine 2094205128Ssimon the size of a buffer and limits the record buffer to 100 entries. 2095205128Ssimon (CVE-2009-1377) 2096205128Ssimon [Robin Seggelmann, discovered by Daniel Mentz] 2097205128Ssimon 2098205128Ssimon *) Keep a copy of frag->msg_header.frag_len so it can be used after the 2099205128Ssimon parent structure is freed. (CVE-2009-1379) 2100205128Ssimon [Daniel Mentz] 2101205128Ssimon 2102205128Ssimon *) Handle non-blocking I/O properly in SSL_shutdown() call. 2103205128Ssimon [Darryl Miles <darryl-mailinglists@netbauds.net>] 2104205128Ssimon 2105205128Ssimon *) Add 2.5.4.* OIDs 2106205128Ssimon [Ilya O. <vrghost@gmail.com>] 2107205128Ssimon 2108205128Ssimon Changes between 0.9.8k and 0.9.8l [5 Nov 2009] 2109205128Ssimon 2110205128Ssimon *) Disable renegotiation completely - this fixes a severe security 2111205128Ssimon problem (CVE-2009-3555) at the cost of breaking all 2112205128Ssimon renegotiation. Renegotiation can be re-enabled by setting 2113205128Ssimon SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at 2114205128Ssimon run-time. This is really not recommended unless you know what 2115205128Ssimon you're doing. 2116205128Ssimon [Ben Laurie] 2117205128Ssimon 2118194206Ssimon Changes between 0.9.8j and 0.9.8k [25 Mar 2009] 2119167612Ssimon 2120194206Ssimon *) Don't set val to NULL when freeing up structures, it is freed up by 2121194206Ssimon underlying code. If sizeof(void *) > sizeof(long) this can result in 2122194206Ssimon zeroing past the valid field. (CVE-2009-0789) 2123194206Ssimon [Paolo Ganci <Paolo.Ganci@AdNovum.CH>] 2124194206Ssimon 2125194206Ssimon *) Fix bug where return value of CMS_SignerInfo_verify_content() was not 2126194206Ssimon checked correctly. This would allow some invalid signed attributes to 2127194206Ssimon appear to verify correctly. (CVE-2009-0591) 2128194206Ssimon [Ivan Nestlerode <inestlerode@us.ibm.com>] 2129194206Ssimon 2130194206Ssimon *) Reject UniversalString and BMPString types with invalid lengths. This 2131194206Ssimon prevents a crash in ASN1_STRING_print_ex() which assumes the strings have 2132194206Ssimon a legal length. (CVE-2009-0590) 2133194206Ssimon [Steve Henson] 2134194206Ssimon 2135194206Ssimon *) Set S/MIME signing as the default purpose rather than setting it 2136194206Ssimon unconditionally. This allows applications to override it at the store 2137194206Ssimon level. 2138194206Ssimon [Steve Henson] 2139194206Ssimon 2140194206Ssimon *) Permit restricted recursion of ASN1 strings. This is needed in practice 2141194206Ssimon to handle some structures. 2142194206Ssimon [Steve Henson] 2143194206Ssimon 2144194206Ssimon *) Improve efficiency of mem_gets: don't search whole buffer each time 2145194206Ssimon for a '\n' 2146194206Ssimon [Jeremy Shapiro <jnshapir@us.ibm.com>] 2147194206Ssimon 2148194206Ssimon *) New -hex option for openssl rand. 2149194206Ssimon [Matthieu Herrb] 2150194206Ssimon 2151194206Ssimon *) Print out UTF8String and NumericString when parsing ASN1. 2152194206Ssimon [Steve Henson] 2153194206Ssimon 2154194206Ssimon *) Support NumericString type for name components. 2155194206Ssimon [Steve Henson] 2156194206Ssimon 2157194206Ssimon *) Allow CC in the environment to override the automatically chosen 2158194206Ssimon compiler. Note that nothing is done to ensure flags work with the 2159194206Ssimon chosen compiler. 2160194206Ssimon [Ben Laurie] 2161194206Ssimon 2162194206Ssimon Changes between 0.9.8i and 0.9.8j [07 Jan 2009] 2163194206Ssimon 2164194206Ssimon *) Properly check EVP_VerifyFinal() and similar return values 2165194206Ssimon (CVE-2008-5077). 2166194206Ssimon [Ben Laurie, Bodo Moeller, Google Security Team] 2167194206Ssimon 2168194206Ssimon *) Enable TLS extensions by default. 2169194206Ssimon [Ben Laurie] 2170194206Ssimon 2171194206Ssimon *) Allow the CHIL engine to be loaded, whether the application is 2172194206Ssimon multithreaded or not. (This does not release the developer from the 2173194206Ssimon obligation to set up the dynamic locking callbacks.) 2174194206Ssimon [Sander Temme <sander@temme.net>] 2175194206Ssimon 2176194206Ssimon *) Use correct exit code if there is an error in dgst command. 2177194206Ssimon [Steve Henson; problem pointed out by Roland Dirlewanger] 2178194206Ssimon 2179194206Ssimon *) Tweak Configure so that you need to say "experimental-jpake" to enable 2180194206Ssimon JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications. 2181194206Ssimon [Bodo Moeller] 2182194206Ssimon 2183194206Ssimon *) Add experimental JPAKE support, including demo authentication in 2184194206Ssimon s_client and s_server. 2185194206Ssimon [Ben Laurie] 2186194206Ssimon 2187194206Ssimon *) Set the comparison function in v3_addr_canonize(). 2188194206Ssimon [Rob Austein <sra@hactrn.net>] 2189194206Ssimon 2190194206Ssimon *) Add support for XMPP STARTTLS in s_client. 2191194206Ssimon [Philip Paeps <philip@freebsd.org>] 2192194206Ssimon 2193194206Ssimon *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior 2194194206Ssimon to ensure that even with this option, only ciphersuites in the 2195194206Ssimon server's preference list will be accepted. (Note that the option 2196194206Ssimon applies only when resuming a session, so the earlier behavior was 2197194206Ssimon just about the algorithm choice for symmetric cryptography.) 2198194206Ssimon [Bodo Moeller] 2199194206Ssimon 2200194206Ssimon Changes between 0.9.8h and 0.9.8i [15 Sep 2008] 2201194206Ssimon 2202205128Ssimon *) Fix NULL pointer dereference if a DTLS server received 2203205128Ssimon ChangeCipherSpec as first record (CVE-2009-1386). 2204205128Ssimon [PR #1679] 2205205128Ssimon 2206194206Ssimon *) Fix a state transitition in s3_srvr.c and d1_srvr.c 2207194206Ssimon (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...). 2208194206Ssimon [Nagendra Modadugu] 2209194206Ssimon 2210194206Ssimon *) The fix in 0.9.8c that supposedly got rid of unsafe 2211194206Ssimon double-checked locking was incomplete for RSA blinding, 2212194206Ssimon addressing just one layer of what turns out to have been 2213194206Ssimon doubly unsafe triple-checked locking. 2214194206Ssimon 2215194206Ssimon So now fix this for real by retiring the MONT_HELPER macro 2216194206Ssimon in crypto/rsa/rsa_eay.c. 2217194206Ssimon 2218194206Ssimon [Bodo Moeller; problem pointed out by Marius Schilder] 2219194206Ssimon 2220194206Ssimon *) Various precautionary measures: 2221194206Ssimon 2222194206Ssimon - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h). 2223194206Ssimon 2224194206Ssimon - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c). 2225194206Ssimon (NB: This would require knowledge of the secret session ticket key 2226194206Ssimon to exploit, in which case you'd be SOL either way.) 2227194206Ssimon 2228194206Ssimon - Change bn_nist.c so that it will properly handle input BIGNUMs 2229194206Ssimon outside the expected range. 2230194206Ssimon 2231194206Ssimon - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG 2232194206Ssimon builds. 2233194206Ssimon 2234194206Ssimon [Neel Mehta, Bodo Moeller] 2235194206Ssimon 2236194206Ssimon *) Allow engines to be "soft loaded" - i.e. optionally don't die if 2237194206Ssimon the load fails. Useful for distros. 2238194206Ssimon [Ben Laurie and the FreeBSD team] 2239194206Ssimon 2240194206Ssimon *) Add support for Local Machine Keyset attribute in PKCS#12 files. 2241194206Ssimon [Steve Henson] 2242194206Ssimon 2243194206Ssimon *) Fix BN_GF2m_mod_arr() top-bit cleanup code. 2244194206Ssimon [Huang Ying] 2245194206Ssimon 2246194206Ssimon *) Expand ENGINE to support engine supplied SSL client certificate functions. 2247194206Ssimon 2248194206Ssimon This work was sponsored by Logica. 2249194206Ssimon [Steve Henson] 2250194206Ssimon 2251194206Ssimon *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows 2252194206Ssimon keystores. Support for SSL/TLS client authentication too. 2253194206Ssimon Not compiled unless enable-capieng specified to Configure. 2254194206Ssimon 2255194206Ssimon This work was sponsored by Logica. 2256194206Ssimon [Steve Henson] 2257194206Ssimon 2258194206Ssimon *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using 2259194206Ssimon ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain 2260194206Ssimon attribute creation routines such as certifcate requests and PKCS#12 2261194206Ssimon files. 2262194206Ssimon [Steve Henson] 2263194206Ssimon 2264194206Ssimon Changes between 0.9.8g and 0.9.8h [28 May 2008] 2265194206Ssimon 2266194206Ssimon *) Fix flaw if 'Server Key exchange message' is omitted from a TLS 2267194206Ssimon handshake which could lead to a cilent crash as found using the 2268194206Ssimon Codenomicon TLS test suite (CVE-2008-1672) 2269194206Ssimon [Steve Henson, Mark Cox] 2270194206Ssimon 2271194206Ssimon *) Fix double free in TLS server name extensions which could lead to 2272194206Ssimon a remote crash found by Codenomicon TLS test suite (CVE-2008-0891) 2273194206Ssimon [Joe Orton] 2274194206Ssimon 2275194206Ssimon *) Clear error queue in SSL_CTX_use_certificate_chain_file() 2276194206Ssimon 2277194206Ssimon Clear the error queue to ensure that error entries left from 2278194206Ssimon older function calls do not interfere with the correct operation. 2279194206Ssimon [Lutz Jaenicke, Erik de Castro Lopo] 2280194206Ssimon 2281194206Ssimon *) Remove root CA certificates of commercial CAs: 2282194206Ssimon 2283194206Ssimon The OpenSSL project does not recommend any specific CA and does not 2284194206Ssimon have any policy with respect to including or excluding any CA. 2285194206Ssimon Therefore it does not make any sense to ship an arbitrary selection 2286194206Ssimon of root CA certificates with the OpenSSL software. 2287194206Ssimon [Lutz Jaenicke] 2288194206Ssimon 2289194206Ssimon *) RSA OAEP patches to fix two separate invalid memory reads. 2290194206Ssimon The first one involves inputs when 'lzero' is greater than 2291194206Ssimon 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes 2292194206Ssimon before the beginning of from). The second one involves inputs where 2293194206Ssimon the 'db' section contains nothing but zeroes (there is a one-byte 2294194206Ssimon invalid read after the end of 'db'). 2295194206Ssimon [Ivan Nestlerode <inestlerode@us.ibm.com>] 2296194206Ssimon 2297194206Ssimon *) Partial backport from 0.9.9-dev: 2298194206Ssimon 2299194206Ssimon Introduce bn_mul_mont (dedicated Montgomery multiplication 2300194206Ssimon procedure) as a candidate for BIGNUM assembler implementation. 2301194206Ssimon While 0.9.9-dev uses assembler for various architectures, only 2302194206Ssimon x86_64 is available by default here in the 0.9.8 branch, and 2303194206Ssimon 32-bit x86 is available through a compile-time setting. 2304194206Ssimon 2305194206Ssimon To try the 32-bit x86 assembler implementation, use Configure 2306194206Ssimon option "enable-montasm" (which exists only for this backport). 2307194206Ssimon 2308194206Ssimon As "enable-montasm" for 32-bit x86 disclaims code stability 2309194206Ssimon anyway, in this constellation we activate additional code 2310194206Ssimon backported from 0.9.9-dev for further performance improvements, 2311194206Ssimon namely BN_from_montgomery_word. (To enable this otherwise, 2312194206Ssimon e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".) 2313194206Ssimon 2314194206Ssimon [Andy Polyakov (backport partially by Bodo Moeller)] 2315194206Ssimon 2316194206Ssimon *) Add TLS session ticket callback. This allows an application to set 2317194206Ssimon TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed 2318194206Ssimon values. This is useful for key rollover for example where several key 2319194206Ssimon sets may exist with different names. 2320194206Ssimon [Steve Henson] 2321194206Ssimon 2322194206Ssimon *) Reverse ENGINE-internal logic for caching default ENGINE handles. 2323194206Ssimon This was broken until now in 0.9.8 releases, such that the only way 2324194206Ssimon a registered ENGINE could be used (assuming it initialises 2325194206Ssimon successfully on the host) was to explicitly set it as the default 2326194206Ssimon for the relevant algorithms. This is in contradiction with 0.9.7 2327194206Ssimon behaviour and the documentation. With this fix, when an ENGINE is 2328194206Ssimon registered into a given algorithm's table of implementations, the 2329194206Ssimon 'uptodate' flag is reset so that auto-discovery will be used next 2330194206Ssimon time a new context for that algorithm attempts to select an 2331194206Ssimon implementation. 2332194206Ssimon [Ian Lister (tweaked by Geoff Thorpe)] 2333194206Ssimon 2334194206Ssimon *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9 2335194206Ssimon implemention in the following ways: 2336194206Ssimon 2337194206Ssimon Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be 2338194206Ssimon hard coded. 2339194206Ssimon 2340194206Ssimon Lack of BER streaming support means one pass streaming processing is 2341194206Ssimon only supported if data is detached: setting the streaming flag is 2342194206Ssimon ignored for embedded content. 2343194206Ssimon 2344194206Ssimon CMS support is disabled by default and must be explicitly enabled 2345194206Ssimon with the enable-cms configuration option. 2346194206Ssimon [Steve Henson] 2347194206Ssimon 2348194206Ssimon *) Update the GMP engine glue to do direct copies between BIGNUM and 2349194206Ssimon mpz_t when openssl and GMP use the same limb size. Otherwise the 2350194206Ssimon existing "conversion via a text string export" trick is still used. 2351194206Ssimon [Paul Sheer <paulsheer@gmail.com>] 2352194206Ssimon 2353194206Ssimon *) Zlib compression BIO. This is a filter BIO which compressed and 2354194206Ssimon uncompresses any data passed through it. 2355194206Ssimon [Steve Henson] 2356194206Ssimon 2357194206Ssimon *) Add AES_wrap_key() and AES_unwrap_key() functions to implement 2358194206Ssimon RFC3394 compatible AES key wrapping. 2359194206Ssimon [Steve Henson] 2360194206Ssimon 2361194206Ssimon *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0(): 2362194206Ssimon sets string data without copying. X509_ALGOR_set0() and 2363194206Ssimon X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier) 2364194206Ssimon data. Attribute function X509at_get0_data_by_OBJ(): retrieves data 2365194206Ssimon from an X509_ATTRIBUTE structure optionally checking it occurs only 2366194206Ssimon once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied 2367194206Ssimon data. 2368194206Ssimon [Steve Henson] 2369194206Ssimon 2370194206Ssimon *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set() 2371194206Ssimon to get the expected BN_FLG_CONSTTIME behavior. 2372194206Ssimon [Bodo Moeller (Google)] 2373194206Ssimon 2374194206Ssimon *) Netware support: 2375194206Ssimon 2376194206Ssimon - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets 2377194206Ssimon - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT) 2378194206Ssimon - added some more tests to do_tests.pl 2379194206Ssimon - fixed RunningProcess usage so that it works with newer LIBC NDKs too 2380194206Ssimon - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency 2381194206Ssimon - added new Configure targets netware-clib-bsdsock, netware-clib-gcc, 2382194206Ssimon netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc 2383194206Ssimon - various changes to netware.pl to enable gcc-cross builds on Win32 2384194206Ssimon platform 2385194206Ssimon - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD) 2386194206Ssimon - various changes to fix missing prototype warnings 2387194206Ssimon - fixed x86nasm.pl to create correct asm files for NASM COFF output 2388194206Ssimon - added AES, WHIRLPOOL and CPUID assembler code to build files 2389194206Ssimon - added missing AES assembler make rules to mk1mf.pl 2390194206Ssimon - fixed order of includes in apps/ocsp.c so that e_os.h settings apply 2391194206Ssimon [Guenter Knauf <eflash@gmx.net>] 2392194206Ssimon 2393194206Ssimon *) Implement certificate status request TLS extension defined in RFC3546. 2394194206Ssimon A client can set the appropriate parameters and receive the encoded 2395194206Ssimon OCSP response via a callback. A server can query the supplied parameters 2396194206Ssimon and set the encoded OCSP response in the callback. Add simplified examples 2397194206Ssimon to s_client and s_server. 2398194206Ssimon [Steve Henson] 2399194206Ssimon 2400194206Ssimon Changes between 0.9.8f and 0.9.8g [19 Oct 2007] 2401194206Ssimon 2402194206Ssimon *) Fix various bugs: 2403194206Ssimon + Binary incompatibility of ssl_ctx_st structure 2404194206Ssimon + DTLS interoperation with non-compliant servers 2405194206Ssimon + Don't call get_session_cb() without proposed session 2406194206Ssimon + Fix ia64 assembler code 2407194206Ssimon [Andy Polyakov, Steve Henson] 2408194206Ssimon 2409194206Ssimon Changes between 0.9.8e and 0.9.8f [11 Oct 2007] 2410194206Ssimon 2411194206Ssimon *) DTLS Handshake overhaul. There were longstanding issues with 2412194206Ssimon OpenSSL DTLS implementation, which were making it impossible for 2413194206Ssimon RFC 4347 compliant client to communicate with OpenSSL server. 2414194206Ssimon Unfortunately just fixing these incompatibilities would "cut off" 2415194206Ssimon pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e 2416194206Ssimon server keeps tolerating non RFC compliant syntax. The opposite is 2417194206Ssimon not true, 0.9.8f client can not communicate with earlier server. 2418194206Ssimon This update even addresses CVE-2007-4995. 2419194206Ssimon [Andy Polyakov] 2420194206Ssimon 2421194206Ssimon *) Changes to avoid need for function casts in OpenSSL: some compilers 2422194206Ssimon (gcc 4.2 and later) reject their use. 2423194206Ssimon [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>, 2424194206Ssimon Steve Henson] 2425194206Ssimon 2426194206Ssimon *) Add RFC4507 support to OpenSSL. This includes the corrections in 2427194206Ssimon RFC4507bis. The encrypted ticket format is an encrypted encoded 2428194206Ssimon SSL_SESSION structure, that way new session features are automatically 2429194206Ssimon supported. 2430194206Ssimon 2431194206Ssimon If a client application caches session in an SSL_SESSION structure 2432194206Ssimon support is transparent because tickets are now stored in the encoded 2433194206Ssimon SSL_SESSION. 2434194206Ssimon 2435194206Ssimon The SSL_CTX structure automatically generates keys for ticket 2436194206Ssimon protection in servers so again support should be possible 2437194206Ssimon with no application modification. 2438194206Ssimon 2439194206Ssimon If a client or server wishes to disable RFC4507 support then the option 2440194206Ssimon SSL_OP_NO_TICKET can be set. 2441194206Ssimon 2442194206Ssimon Add a TLS extension debugging callback to allow the contents of any client 2443194206Ssimon or server extensions to be examined. 2444194206Ssimon 2445194206Ssimon This work was sponsored by Google. 2446194206Ssimon [Steve Henson] 2447194206Ssimon 2448194206Ssimon *) Add initial support for TLS extensions, specifically for the server_name 2449194206Ssimon extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 2450194206Ssimon have new members for a host name. The SSL data structure has an 2451194206Ssimon additional member SSL_CTX *initial_ctx so that new sessions can be 2452194206Ssimon stored in that context to allow for session resumption, even after the 2453194206Ssimon SSL has been switched to a new SSL_CTX in reaction to a client's 2454194206Ssimon server_name extension. 2455194206Ssimon 2456194206Ssimon New functions (subject to change): 2457194206Ssimon 2458194206Ssimon SSL_get_servername() 2459194206Ssimon SSL_get_servername_type() 2460194206Ssimon SSL_set_SSL_CTX() 2461194206Ssimon 2462194206Ssimon New CTRL codes and macros (subject to change): 2463194206Ssimon 2464194206Ssimon SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 2465194206Ssimon - SSL_CTX_set_tlsext_servername_callback() 2466194206Ssimon SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 2467194206Ssimon - SSL_CTX_set_tlsext_servername_arg() 2468194206Ssimon SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 2469194206Ssimon 2470194206Ssimon openssl s_client has a new '-servername ...' option. 2471194206Ssimon 2472194206Ssimon openssl s_server has new options '-servername_host ...', '-cert2 ...', 2473194206Ssimon '-key2 ...', '-servername_fatal' (subject to change). This allows 2474194206Ssimon testing the HostName extension for a specific single host name ('-cert' 2475194206Ssimon and '-key' remain fallbacks for handshakes without HostName 2476194206Ssimon negotiation). If the unrecogninzed_name alert has to be sent, this by 2477194206Ssimon default is a warning; it becomes fatal with the '-servername_fatal' 2478194206Ssimon option. 2479194206Ssimon 2480194206Ssimon [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson] 2481194206Ssimon 2482194206Ssimon *) Add AES and SSE2 assembly language support to VC++ build. 2483194206Ssimon [Steve Henson] 2484194206Ssimon 2485194206Ssimon *) Mitigate attack on final subtraction in Montgomery reduction. 2486194206Ssimon [Andy Polyakov] 2487194206Ssimon 2488194206Ssimon *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0 2489194206Ssimon (which previously caused an internal error). 2490194206Ssimon [Bodo Moeller] 2491194206Ssimon 2492194206Ssimon *) Squeeze another 10% out of IGE mode when in != out. 2493194206Ssimon [Ben Laurie] 2494194206Ssimon 2495194206Ssimon *) AES IGE mode speedup. 2496194206Ssimon [Dean Gaudet (Google)] 2497194206Ssimon 2498194206Ssimon *) Add the Korean symmetric 128-bit cipher SEED (see 2499194206Ssimon http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and 2500194206Ssimon add SEED ciphersuites from RFC 4162: 2501194206Ssimon 2502194206Ssimon TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA" 2503194206Ssimon TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA" 2504194206Ssimon TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA" 2505194206Ssimon TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA" 2506194206Ssimon 2507194206Ssimon To minimize changes between patchlevels in the OpenSSL 0.9.8 2508194206Ssimon series, SEED remains excluded from compilation unless OpenSSL 2509194206Ssimon is configured with 'enable-seed'. 2510194206Ssimon [KISA, Bodo Moeller] 2511194206Ssimon 2512194206Ssimon *) Mitigate branch prediction attacks, which can be practical if a 2513194206Ssimon single processor is shared, allowing a spy process to extract 2514194206Ssimon information. For detailed background information, see 2515194206Ssimon http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron, 2516194206Ssimon J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL 2517194206Ssimon and Necessary Software Countermeasures"). The core of the change 2518194206Ssimon are new versions BN_div_no_branch() and 2519194206Ssimon BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(), 2520194206Ssimon respectively, which are slower, but avoid the security-relevant 2521194206Ssimon conditional branches. These are automatically called by BN_div() 2522194206Ssimon and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one 2523194206Ssimon of the input BIGNUMs. Also, BN_is_bit_set() has been changed to 2524194206Ssimon remove a conditional branch. 2525194206Ssimon 2526194206Ssimon BN_FLG_CONSTTIME is the new name for the previous 2527194206Ssimon BN_FLG_EXP_CONSTTIME flag, since it now affects more than just 2528194206Ssimon modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag 2529194206Ssimon in the exponent causes BN_mod_exp_mont() to use the alternative 2530194206Ssimon implementation in BN_mod_exp_mont_consttime().) The old name 2531194206Ssimon remains as a deprecated alias. 2532194206Ssimon 2533194206Ssimon Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general 2534194206Ssimon RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses 2535194206Ssimon constant-time implementations for more than just exponentiation. 2536194206Ssimon Here too the old name is kept as a deprecated alias. 2537194206Ssimon 2538194206Ssimon BN_BLINDING_new() will now use BN_dup() for the modulus so that 2539194206Ssimon the BN_BLINDING structure gets an independent copy of the 2540194206Ssimon modulus. This means that the previous "BIGNUM *m" argument to 2541194206Ssimon BN_BLINDING_new() and to BN_BLINDING_create_param() now 2542194206Ssimon essentially becomes "const BIGNUM *m", although we can't actually 2543194206Ssimon change this in the header file before 0.9.9. It allows 2544194206Ssimon RSA_setup_blinding() to use BN_with_flags() on the modulus to 2545194206Ssimon enable BN_FLG_CONSTTIME. 2546194206Ssimon 2547194206Ssimon [Matthew D Wood (Intel Corp)] 2548194206Ssimon 2549194206Ssimon *) In the SSL/TLS server implementation, be strict about session ID 2550194206Ssimon context matching (which matters if an application uses a single 2551194206Ssimon external cache for different purposes). Previously, 2552194206Ssimon out-of-context reuse was forbidden only if SSL_VERIFY_PEER was 2553194206Ssimon set. This did ensure strict client verification, but meant that, 2554194206Ssimon with applications using a single external cache for quite 2555194206Ssimon different requirements, clients could circumvent ciphersuite 2556194206Ssimon restrictions for a given session ID context by starting a session 2557194206Ssimon in a different context. 2558194206Ssimon [Bodo Moeller] 2559194206Ssimon 2560167612Ssimon *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 2561167612Ssimon a ciphersuite string such as "DEFAULT:RSA" cannot enable 2562167612Ssimon authentication-only ciphersuites. 2563167612Ssimon [Bodo Moeller] 2564167612Ssimon 2565194206Ssimon *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was 2566194206Ssimon not complete and could lead to a possible single byte overflow 2567194206Ssimon (CVE-2007-5135) [Ben Laurie] 2568194206Ssimon 2569194206Ssimon Changes between 0.9.8d and 0.9.8e [23 Feb 2007] 2570194206Ssimon 2571167612Ssimon *) Since AES128 and AES256 (and similarly Camellia128 and 2572167612Ssimon Camellia256) share a single mask bit in the logic of 2573167612Ssimon ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 2574167612Ssimon kludge to work properly if AES128 is available and AES256 isn't 2575167612Ssimon (or if Camellia128 is available and Camellia256 isn't). 2576167612Ssimon [Victor Duchovni] 2577167612Ssimon 2578167612Ssimon *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c 2579167612Ssimon (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters): 2580167612Ssimon When a point or a seed is encoded in a BIT STRING, we need to 2581167612Ssimon prevent the removal of trailing zero bits to get the proper DER 2582167612Ssimon encoding. (By default, crypto/asn1/a_bitstr.c assumes the case 2583167612Ssimon of a NamedBitList, for which trailing 0 bits need to be removed.) 2584167612Ssimon [Bodo Moeller] 2585167612Ssimon 2586167612Ssimon *) Have SSL/TLS server implementation tolerate "mismatched" record 2587167612Ssimon protocol version while receiving ClientHello even if the 2588167612Ssimon ClientHello is fragmented. (The server can't insist on the 2589167612Ssimon particular protocol version it has chosen before the ServerHello 2590167612Ssimon message has informed the client about his choice.) 2591167612Ssimon [Bodo Moeller] 2592167612Ssimon 2593167612Ssimon *) Add RFC 3779 support. 2594167612Ssimon [Rob Austein for ARIN, Ben Laurie] 2595167612Ssimon 2596167612Ssimon *) Load error codes if they are not already present instead of using a 2597167612Ssimon static variable. This allows them to be cleanly unloaded and reloaded. 2598167612Ssimon Improve header file function name parsing. 2599167612Ssimon [Steve Henson] 2600167612Ssimon 2601167612Ssimon *) extend SMTP and IMAP protocol emulation in s_client to use EHLO 2602167612Ssimon or CAPABILITY handshake as required by RFCs. 2603167612Ssimon [Goetz Babin-Ebell] 2604167612Ssimon 2605162911Ssimon Changes between 0.9.8c and 0.9.8d [28 Sep 2006] 2606162911Ssimon 2607162911Ssimon *) Introduce limits to prevent malicious keys being able to 2608162911Ssimon cause a denial of service. (CVE-2006-2940) 2609162911Ssimon [Steve Henson, Bodo Moeller] 2610162911Ssimon 2611162911Ssimon *) Fix ASN.1 parsing of certain invalid structures that can result 2612162911Ssimon in a denial of service. (CVE-2006-2937) [Steve Henson] 2613162911Ssimon 2614162911Ssimon *) Fix buffer overflow in SSL_get_shared_ciphers() function. 2615162911Ssimon (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 2616162911Ssimon 2617162911Ssimon *) Fix SSL client code which could crash if connecting to a 2618162911Ssimon malicious SSLv2 server. (CVE-2006-4343) 2619162911Ssimon [Tavis Ormandy and Will Drewry, Google Security Team] 2620162911Ssimon 2621162911Ssimon *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites 2622162911Ssimon match only those. Before that, "AES256-SHA" would be interpreted 2623162911Ssimon as a pattern and match "AES128-SHA" too (since AES128-SHA got 2624162911Ssimon the same strength classification in 0.9.7h) as we currently only 2625162911Ssimon have a single AES bit in the ciphersuite description bitmap. 2626162911Ssimon That change, however, also applied to ciphersuite strings such as 2627162911Ssimon "RC4-MD5" that intentionally matched multiple ciphersuites -- 2628162911Ssimon namely, SSL 2.0 ciphersuites in addition to the more common ones 2629162911Ssimon from SSL 3.0/TLS 1.0. 2630162911Ssimon 2631162911Ssimon So we change the selection algorithm again: Naming an explicit 2632162911Ssimon ciphersuite selects this one ciphersuite, and any other similar 2633162911Ssimon ciphersuite (same bitmap) from *other* protocol versions. 2634162911Ssimon Thus, "RC4-MD5" again will properly select both the SSL 2.0 2635162911Ssimon ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite. 2636162911Ssimon 2637162911Ssimon Since SSL 2.0 does not have any ciphersuites for which the 2638162911Ssimon 128/256 bit distinction would be relevant, this works for now. 2639162911Ssimon The proper fix will be to use different bits for AES128 and 2640162911Ssimon AES256, which would have avoided the problems from the beginning; 2641162911Ssimon however, bits are scarce, so we can only do this in a new release 2642162911Ssimon (not just a patchlevel) when we can change the SSL_CIPHER 2643162911Ssimon definition to split the single 'unsigned long mask' bitmap into 2644162911Ssimon multiple values to extend the available space. 2645162911Ssimon 2646162911Ssimon [Bodo Moeller] 2647162911Ssimon 2648162911Ssimon Changes between 0.9.8b and 0.9.8c [05 Sep 2006] 2649162911Ssimon 2650162911Ssimon *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 2651162911Ssimon (CVE-2006-4339) [Ben Laurie and Google Security Team] 2652162911Ssimon 2653162911Ssimon *) Add AES IGE and biIGE modes. 2654162911Ssimon [Ben Laurie] 2655162911Ssimon 2656162911Ssimon *) Change the Unix randomness entropy gathering to use poll() when 2657162911Ssimon possible instead of select(), since the latter has some 2658162911Ssimon undesirable limitations. 2659162911Ssimon [Darryl Miles via Richard Levitte and Bodo Moeller] 2660162911Ssimon 2661162911Ssimon *) Disable "ECCdraft" ciphersuites more thoroughly. Now special 2662162911Ssimon treatment in ssl/ssl_ciph.s makes sure that these ciphersuites 2663162911Ssimon cannot be implicitly activated as part of, e.g., the "AES" alias. 2664162911Ssimon However, please upgrade to OpenSSL 0.9.9[-dev] for 2665162911Ssimon non-experimental use of the ECC ciphersuites to get TLS extension 2666162911Ssimon support, which is required for curve and point format negotiation 2667162911Ssimon to avoid potential handshake problems. 2668162911Ssimon [Bodo Moeller] 2669162911Ssimon 2670162911Ssimon *) Disable rogue ciphersuites: 2671162911Ssimon 2672162911Ssimon - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 2673162911Ssimon - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 2674162911Ssimon - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 2675162911Ssimon 2676162911Ssimon The latter two were purportedly from 2677162911Ssimon draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 2678162911Ssimon appear there. 2679162911Ssimon 2680167612Ssimon Also deactivate the remaining ciphersuites from 2681162911Ssimon draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 2682162911Ssimon unofficial, and the ID has long expired. 2683162911Ssimon [Bodo Moeller] 2684162911Ssimon 2685162911Ssimon *) Fix RSA blinding Heisenbug (problems sometimes occured on 2686162911Ssimon dual-core machines) and other potential thread-safety issues. 2687162911Ssimon [Bodo Moeller] 2688162911Ssimon 2689162911Ssimon *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key 2690162911Ssimon versions), which is now available for royalty-free use 2691162911Ssimon (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html). 2692162911Ssimon Also, add Camellia TLS ciphersuites from RFC 4132. 2693162911Ssimon 2694162911Ssimon To minimize changes between patchlevels in the OpenSSL 0.9.8 2695162911Ssimon series, Camellia remains excluded from compilation unless OpenSSL 2696162911Ssimon is configured with 'enable-camellia'. 2697162911Ssimon [NTT] 2698162911Ssimon 2699162911Ssimon *) Disable the padding bug check when compression is in use. The padding 2700162911Ssimon bug check assumes the first packet is of even length, this is not 2701162911Ssimon necessarily true if compresssion is enabled and can result in false 2702162911Ssimon positives causing handshake failure. The actual bug test is ancient 2703162911Ssimon code so it is hoped that implementations will either have fixed it by 2704162911Ssimon now or any which still have the bug do not support compression. 2705162911Ssimon [Steve Henson] 2706162911Ssimon 2707160814Ssimon Changes between 0.9.8a and 0.9.8b [04 May 2006] 2708160814Ssimon 2709160814Ssimon *) When applying a cipher rule check to see if string match is an explicit 2710160814Ssimon cipher suite and only match that one cipher suite if it is. 2711160814Ssimon [Steve Henson] 2712160814Ssimon 2713160814Ssimon *) Link in manifests for VC++ if needed. 2714160814Ssimon [Austin Ziegler <halostatue@gmail.com>] 2715160814Ssimon 2716160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 2717160814Ssimon draft-ietf-tls-ecc-12.txt with proposed changes (but without 2718160814Ssimon TLS extensions, which are supported starting with the 0.9.9 2719160814Ssimon branch, not in the OpenSSL 0.9.8 branch). 2720160814Ssimon [Douglas Stebila] 2721160814Ssimon 2722160814Ssimon *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support 2723160814Ssimon opaque EVP_CIPHER_CTX handling. 2724160814Ssimon [Steve Henson] 2725160814Ssimon 2726160814Ssimon *) Fixes and enhancements to zlib compression code. We now only use 2727160814Ssimon "zlib1.dll" and use the default __cdecl calling convention on Win32 2728160814Ssimon to conform with the standards mentioned here: 2729160814Ssimon http://www.zlib.net/DLL_FAQ.txt 2730160814Ssimon Static zlib linking now works on Windows and the new --with-zlib-include 2731160814Ssimon --with-zlib-lib options to Configure can be used to supply the location 2732160814Ssimon of the headers and library. Gracefully handle case where zlib library 2733160814Ssimon can't be loaded. 2734160814Ssimon [Steve Henson] 2735160814Ssimon 2736160814Ssimon *) Several fixes and enhancements to the OID generation code. The old code 2737160814Ssimon sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't 2738160814Ssimon handle numbers larger than ULONG_MAX, truncated printing and had a 2739160814Ssimon non standard OBJ_obj2txt() behaviour. 2740160814Ssimon [Steve Henson] 2741160814Ssimon 2742160814Ssimon *) Add support for building of engines under engine/ as shared libraries 2743160814Ssimon under VC++ build system. 2744160814Ssimon [Steve Henson] 2745160814Ssimon 2746160814Ssimon *) Corrected the numerous bugs in the Win32 path splitter in DSO. 2747160814Ssimon Hopefully, we will not see any false combination of paths any more. 2748160814Ssimon [Richard Levitte] 2749160814Ssimon 2750160814Ssimon Changes between 0.9.8 and 0.9.8a [11 Oct 2005] 2751160814Ssimon 2752160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 2753160814Ssimon (part of SSL_OP_ALL). This option used to disable the 2754160814Ssimon countermeasure against man-in-the-middle protocol-version 2755160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 2756160814Ssimon idea. (CVE-2005-2969) 2757160814Ssimon 2758160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 2759160814Ssimon for Information Security, National Institute of Advanced Industrial 2760160814Ssimon Science and Technology [AIST], Japan)] 2761160814Ssimon 2762160814Ssimon *) Add two function to clear and return the verify parameter flags. 2763160814Ssimon [Steve Henson] 2764160814Ssimon 2765160814Ssimon *) Keep cipherlists sorted in the source instead of sorting them at 2766160814Ssimon runtime, thus removing the need for a lock. 2767160814Ssimon [Nils Larsch] 2768160814Ssimon 2769160814Ssimon *) Avoid some small subgroup attacks in Diffie-Hellman. 2770160814Ssimon [Nick Mathewson and Ben Laurie] 2771160814Ssimon 2772160814Ssimon *) Add functions for well-known primes. 2773160814Ssimon [Nick Mathewson] 2774160814Ssimon 2775160814Ssimon *) Extended Windows CE support. 2776160814Ssimon [Satoshi Nakamura and Andy Polyakov] 2777160814Ssimon 2778160814Ssimon *) Initialize SSL_METHOD structures at compile time instead of during 2779160814Ssimon runtime, thus removing the need for a lock. 2780160814Ssimon [Steve Henson] 2781160814Ssimon 2782160814Ssimon *) Make PKCS7_decrypt() work even if no certificate is supplied by 2783160814Ssimon attempting to decrypt each encrypted key in turn. Add support to 2784160814Ssimon smime utility. 2785160814Ssimon [Steve Henson] 2786160814Ssimon 2787160814Ssimon Changes between 0.9.7h and 0.9.8 [05 Jul 2005] 2788160814Ssimon 2789162911Ssimon [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after 2790162911Ssimon OpenSSL 0.9.8.] 2791162911Ssimon 2792160814Ssimon *) Add libcrypto.pc and libssl.pc for those who feel they need them. 2793160814Ssimon [Richard Levitte] 2794160814Ssimon 2795160814Ssimon *) Change CA.sh and CA.pl so they don't bundle the CSR and the private 2796160814Ssimon key into the same file any more. 2797160814Ssimon [Richard Levitte] 2798160814Ssimon 2799160814Ssimon *) Add initial support for Win64, both IA64 and AMD64/x64 flavors. 2800160814Ssimon [Andy Polyakov] 2801160814Ssimon 2802160814Ssimon *) Add -utf8 command line and config file option to 'ca'. 2803160814Ssimon [Stefan <stf@udoma.org] 2804160814Ssimon 2805160814Ssimon *) Removed the macro des_crypt(), as it seems to conflict with some 2806160814Ssimon libraries. Use DES_crypt(). 2807160814Ssimon [Richard Levitte] 2808160814Ssimon 2809160814Ssimon *) Correct naming of the 'chil' and '4758cca' ENGINEs. This 2810160814Ssimon involves renaming the source and generated shared-libs for 2811160814Ssimon both. The engines will accept the corrected or legacy ids 2812160814Ssimon ('ncipher' and '4758_cca' respectively) when binding. NB, 2813160814Ssimon this only applies when building 'shared'. 2814160814Ssimon [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe] 2815160814Ssimon 2816160814Ssimon *) Add attribute functions to EVP_PKEY structure. Modify 2817160814Ssimon PKCS12_create() to recognize a CSP name attribute and 2818160814Ssimon use it. Make -CSP option work again in pkcs12 utility. 2819160814Ssimon [Steve Henson] 2820160814Ssimon 2821160814Ssimon *) Add new functionality to the bn blinding code: 2822160814Ssimon - automatic re-creation of the BN_BLINDING parameters after 2823160814Ssimon a fixed number of uses (currently 32) 2824160814Ssimon - add new function for parameter creation 2825160814Ssimon - introduce flags to control the update behaviour of the 2826160814Ssimon BN_BLINDING parameters 2827160814Ssimon - hide BN_BLINDING structure 2828160814Ssimon Add a second BN_BLINDING slot to the RSA structure to improve 2829160814Ssimon performance when a single RSA object is shared among several 2830160814Ssimon threads. 2831160814Ssimon [Nils Larsch] 2832160814Ssimon 2833160814Ssimon *) Add support for DTLS. 2834160814Ssimon [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie] 2835160814Ssimon 2836160814Ssimon *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1) 2837160814Ssimon to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file() 2838160814Ssimon [Walter Goulet] 2839160814Ssimon 2840160814Ssimon *) Remove buggy and incompletet DH cert support from 2841160814Ssimon ssl/ssl_rsa.c and ssl/s3_both.c 2842160814Ssimon [Nils Larsch] 2843160814Ssimon 2844160814Ssimon *) Use SHA-1 instead of MD5 as the default digest algorithm for 2845160814Ssimon the apps/openssl applications. 2846160814Ssimon [Nils Larsch] 2847160814Ssimon 2848160814Ssimon *) Compile clean with "-Wall -Wmissing-prototypes 2849160814Ssimon -Wstrict-prototypes -Wmissing-declarations -Werror". Currently 2850160814Ssimon DEBUG_SAFESTACK must also be set. 2851160814Ssimon [Ben Laurie] 2852160814Ssimon 2853160814Ssimon *) Change ./Configure so that certain algorithms can be disabled by default. 2854160814Ssimon The new counterpiece to "no-xxx" is "enable-xxx". 2855160814Ssimon 2856160814Ssimon The patented RC5 and MDC2 algorithms will now be disabled unless 2857160814Ssimon "enable-rc5" and "enable-mdc2", respectively, are specified. 2858160814Ssimon 2859160814Ssimon (IDEA remains enabled despite being patented. This is because IDEA 2860160814Ssimon is frequently required for interoperability, and there is no license 2861160814Ssimon fee for non-commercial use. As before, "no-idea" can be used to 2862160814Ssimon avoid this algorithm.) 2863160814Ssimon 2864160814Ssimon [Bodo Moeller] 2865160814Ssimon 2866160814Ssimon *) Add processing of proxy certificates (see RFC 3820). This work was 2867160814Ssimon sponsored by KTH (The Royal Institute of Technology in Stockholm) and 2868160814Ssimon EGEE (Enabling Grids for E-science in Europe). 2869160814Ssimon [Richard Levitte] 2870160814Ssimon 2871160814Ssimon *) RC4 performance overhaul on modern architectures/implementations, such 2872160814Ssimon as Intel P4, IA-64 and AMD64. 2873160814Ssimon [Andy Polyakov] 2874160814Ssimon 2875160814Ssimon *) New utility extract-section.pl. This can be used specify an alternative 2876160814Ssimon section number in a pod file instead of having to treat each file as 2877160814Ssimon a separate case in Makefile. This can be done by adding two lines to the 2878160814Ssimon pod file: 2879160814Ssimon 2880160814Ssimon =for comment openssl_section:XXX 2881160814Ssimon 2882160814Ssimon The blank line is mandatory. 2883160814Ssimon 2884160814Ssimon [Steve Henson] 2885160814Ssimon 2886160814Ssimon *) New arguments -certform, -keyform and -pass for s_client and s_server 2887160814Ssimon to allow alternative format key and certificate files and passphrase 2888160814Ssimon sources. 2889160814Ssimon [Steve Henson] 2890160814Ssimon 2891160814Ssimon *) New structure X509_VERIFY_PARAM which combines current verify parameters, 2892160814Ssimon update associated structures and add various utility functions. 2893160814Ssimon 2894160814Ssimon Add new policy related verify parameters, include policy checking in 2895160814Ssimon standard verify code. Enhance 'smime' application with extra parameters 2896160814Ssimon to support policy checking and print out. 2897160814Ssimon [Steve Henson] 2898160814Ssimon 2899160814Ssimon *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3 2900160814Ssimon Nehemiah processors. These extensions support AES encryption in hardware 2901160814Ssimon as well as RNG (though RNG support is currently disabled). 2902160814Ssimon [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov] 2903160814Ssimon 2904160814Ssimon *) Deprecate BN_[get|set]_params() functions (they were ignored internally). 2905160814Ssimon [Geoff Thorpe] 2906160814Ssimon 2907160814Ssimon *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented. 2908160814Ssimon [Andy Polyakov and a number of other people] 2909160814Ssimon 2910160814Ssimon *) Improved PowerPC platform support. Most notably BIGNUM assembler 2911160814Ssimon implementation contributed by IBM. 2912160814Ssimon [Suresh Chari, Peter Waltenberg, Andy Polyakov] 2913160814Ssimon 2914160814Ssimon *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public 2915160814Ssimon exponent rather than 'unsigned long'. There is a corresponding change to 2916160814Ssimon the new 'rsa_keygen' element of the RSA_METHOD structure. 2917160814Ssimon [Jelte Jansen, Geoff Thorpe] 2918160814Ssimon 2919160814Ssimon *) Functionality for creating the initial serial number file is now 2920160814Ssimon moved from CA.pl to the 'ca' utility with a new option -create_serial. 2921160814Ssimon 2922160814Ssimon (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial 2923160814Ssimon number file to 1, which is bound to cause problems. To avoid 2924160814Ssimon the problems while respecting compatibility between different 0.9.7 2925160814Ssimon patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in 2926160814Ssimon CA.pl for serial number initialization. With the new release 0.9.8, 2927160814Ssimon we can fix the problem directly in the 'ca' utility.) 2928160814Ssimon [Steve Henson] 2929160814Ssimon 2930160814Ssimon *) Reduced header interdepencies by declaring more opaque objects in 2931160814Ssimon ossl_typ.h. As a consequence, including some headers (eg. engine.h) will 2932160814Ssimon give fewer recursive includes, which could break lazy source code - so 2933160814Ssimon this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always, 2934160814Ssimon developers should define this symbol when building and using openssl to 2935160814Ssimon ensure they track the recommended behaviour, interfaces, [etc], but 2936160814Ssimon backwards-compatible behaviour prevails when this isn't defined. 2937160814Ssimon [Geoff Thorpe] 2938160814Ssimon 2939160814Ssimon *) New function X509_POLICY_NODE_print() which prints out policy nodes. 2940160814Ssimon [Steve Henson] 2941160814Ssimon 2942160814Ssimon *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality. 2943160814Ssimon This will generate a random key of the appropriate length based on the 2944160814Ssimon cipher context. The EVP_CIPHER can provide its own random key generation 2945160814Ssimon routine to support keys of a specific form. This is used in the des and 2946160814Ssimon 3des routines to generate a key of the correct parity. Update S/MIME 2947160814Ssimon code to use new functions and hence generate correct parity DES keys. 2948160814Ssimon Add EVP_CHECK_DES_KEY #define to return an error if the key is not 2949160814Ssimon valid (weak or incorrect parity). 2950160814Ssimon [Steve Henson] 2951160814Ssimon 2952160814Ssimon *) Add a local set of CRLs that can be used by X509_verify_cert() as well 2953160814Ssimon as looking them up. This is useful when the verified structure may contain 2954160814Ssimon CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs 2955160814Ssimon present unless the new PKCS7_NO_CRL flag is asserted. 2956160814Ssimon [Steve Henson] 2957160814Ssimon 2958160814Ssimon *) Extend ASN1 oid configuration module. It now additionally accepts the 2959160814Ssimon syntax: 2960160814Ssimon 2961160814Ssimon shortName = some long name, 1.2.3.4 2962160814Ssimon [Steve Henson] 2963160814Ssimon 2964160814Ssimon *) Reimplemented the BN_CTX implementation. There is now no more static 2965160814Ssimon limitation on the number of variables it can handle nor the depth of the 2966160814Ssimon "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack 2967160814Ssimon information can now expand as required, and rather than having a single 2968160814Ssimon static array of bignums, BN_CTX now uses a linked-list of such arrays 2969160814Ssimon allowing it to expand on demand whilst maintaining the usefulness of 2970160814Ssimon BN_CTX's "bundling". 2971160814Ssimon [Geoff Thorpe] 2972160814Ssimon 2973160814Ssimon *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD 2974160814Ssimon to allow all RSA operations to function using a single BN_CTX. 2975160814Ssimon [Geoff Thorpe] 2976160814Ssimon 2977160814Ssimon *) Preliminary support for certificate policy evaluation and checking. This 2978160814Ssimon is initially intended to pass the tests outlined in "Conformance Testing 2979160814Ssimon of Relying Party Client Certificate Path Processing Logic" v1.07. 2980160814Ssimon [Steve Henson] 2981160814Ssimon 2982160814Ssimon *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and 2983160814Ssimon remained unused and not that useful. A variety of other little bignum 2984160814Ssimon tweaks and fixes have also been made continuing on from the audit (see 2985160814Ssimon below). 2986160814Ssimon [Geoff Thorpe] 2987160814Ssimon 2988160814Ssimon *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with 2989160814Ssimon associated ASN1, EVP and SSL functions and old ASN1 macros. 2990160814Ssimon [Richard Levitte] 2991160814Ssimon 2992160814Ssimon *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results, 2993160814Ssimon and this should never fail. So the return value from the use of 2994160814Ssimon BN_set_word() (which can fail due to needless expansion) is now deprecated; 2995160814Ssimon if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro. 2996160814Ssimon [Geoff Thorpe] 2997160814Ssimon 2998160814Ssimon *) BN_CTX_get() should return zero-valued bignums, providing the same 2999160814Ssimon initialised value as BN_new(). 3000160814Ssimon [Geoff Thorpe, suggested by Ulf M�ller] 3001160814Ssimon 3002160814Ssimon *) Support for inhibitAnyPolicy certificate extension. 3003160814Ssimon [Steve Henson] 3004160814Ssimon 3005160814Ssimon *) An audit of the BIGNUM code is underway, for which debugging code is 3006160814Ssimon enabled when BN_DEBUG is defined. This makes stricter enforcements on what 3007160814Ssimon is considered valid when processing BIGNUMs, and causes execution to 3008160814Ssimon assert() when a problem is discovered. If BN_DEBUG_RAND is defined, 3009160814Ssimon further steps are taken to deliberately pollute unused data in BIGNUM 3010160814Ssimon structures to try and expose faulty code further on. For now, openssl will 3011160814Ssimon (in its default mode of operation) continue to tolerate the inconsistent 3012160814Ssimon forms that it has tolerated in the past, but authors and packagers should 3013160814Ssimon consider trying openssl and their own applications when compiled with 3014160814Ssimon these debugging symbols defined. It will help highlight potential bugs in 3015160814Ssimon their own code, and will improve the test coverage for OpenSSL itself. At 3016160814Ssimon some point, these tighter rules will become openssl's default to improve 3017160814Ssimon maintainability, though the assert()s and other overheads will remain only 3018160814Ssimon in debugging configurations. See bn.h for more details. 3019160814Ssimon [Geoff Thorpe, Nils Larsch, Ulf M�ller] 3020160814Ssimon 3021160814Ssimon *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure 3022160814Ssimon that can only be obtained through BN_CTX_new() (which implicitly 3023160814Ssimon initialises it). The presence of this function only made it possible 3024160814Ssimon to overwrite an existing structure (and cause memory leaks). 3025160814Ssimon [Geoff Thorpe] 3026160814Ssimon 3027160814Ssimon *) Because of the callback-based approach for implementing LHASH as a 3028160814Ssimon template type, lh_insert() adds opaque objects to hash-tables and 3029160814Ssimon lh_doall() or lh_doall_arg() are typically used with a destructor callback 3030160814Ssimon to clean up those corresponding objects before destroying the hash table 3031160814Ssimon (and losing the object pointers). So some over-zealous constifications in 3032160814Ssimon LHASH have been relaxed so that lh_insert() does not take (nor store) the 3033160814Ssimon objects as "const" and the lh_doall[_arg] callback wrappers are not 3034160814Ssimon prototyped to have "const" restrictions on the object pointers they are 3035160814Ssimon given (and so aren't required to cast them away any more). 3036160814Ssimon [Geoff Thorpe] 3037160814Ssimon 3038160814Ssimon *) The tmdiff.h API was so ugly and minimal that our own timing utility 3039160814Ssimon (speed) prefers to use its own implementation. The two implementations 3040160814Ssimon haven't been consolidated as yet (volunteers?) but the tmdiff API has had 3041160814Ssimon its object type properly exposed (MS_TM) instead of casting to/from "char 3042160814Ssimon *". This may still change yet if someone realises MS_TM and "ms_time_***" 3043160814Ssimon aren't necessarily the greatest nomenclatures - but this is what was used 3044160814Ssimon internally to the implementation so I've used that for now. 3045160814Ssimon [Geoff Thorpe] 3046160814Ssimon 3047160814Ssimon *) Ensure that deprecated functions do not get compiled when 3048160814Ssimon OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of 3049160814Ssimon the self-tests were still using deprecated key-generation functions so 3050160814Ssimon these have been updated also. 3051160814Ssimon [Geoff Thorpe] 3052160814Ssimon 3053160814Ssimon *) Reorganise PKCS#7 code to separate the digest location functionality 3054160814Ssimon into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest(). 3055160814Ssimon New function PKCS7_set_digest() to set the digest type for PKCS#7 3056160814Ssimon digestedData type. Add additional code to correctly generate the 3057160814Ssimon digestedData type and add support for this type in PKCS7 initialization 3058160814Ssimon functions. 3059160814Ssimon [Steve Henson] 3060160814Ssimon 3061160814Ssimon *) New function PKCS7_set0_type_other() this initializes a PKCS7 3062160814Ssimon structure of type "other". 3063160814Ssimon [Steve Henson] 3064160814Ssimon 3065160814Ssimon *) Fix prime generation loop in crypto/bn/bn_prime.pl by making 3066160814Ssimon sure the loop does correctly stop and breaking ("division by zero") 3067160814Ssimon modulus operations are not performed. The (pre-generated) prime 3068160814Ssimon table crypto/bn/bn_prime.h was already correct, but it could not be 3069160814Ssimon re-generated on some platforms because of the "division by zero" 3070160814Ssimon situation in the script. 3071160814Ssimon [Ralf S. Engelschall] 3072160814Ssimon 3073160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 3074160814Ssimon draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with 3075160814Ssimon SHA-1 now is only used for "small" curves (where the 3076160814Ssimon representation of a field element takes up to 24 bytes); for 3077160814Ssimon larger curves, the field element resulting from ECDH is directly 3078160814Ssimon used as premaster secret. 3079160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 3080160814Ssimon 3081160814Ssimon *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2 3082160814Ssimon curve secp160r1 to the tests. 3083160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 3084160814Ssimon 3085160814Ssimon *) Add the possibility to load symbols globally with DSO. 3086160814Ssimon [G�tz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte] 3087160814Ssimon 3088160814Ssimon *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better 3089160814Ssimon control of the error stack. 3090160814Ssimon [Richard Levitte] 3091160814Ssimon 3092160814Ssimon *) Add support for STORE in ENGINE. 3093160814Ssimon [Richard Levitte] 3094160814Ssimon 3095160814Ssimon *) Add the STORE type. The intention is to provide a common interface 3096160814Ssimon to certificate and key stores, be they simple file-based stores, or 3097160814Ssimon HSM-type store, or LDAP stores, or... 3098160814Ssimon NOTE: The code is currently UNTESTED and isn't really used anywhere. 3099160814Ssimon [Richard Levitte] 3100160814Ssimon 3101160814Ssimon *) Add a generic structure called OPENSSL_ITEM. This can be used to 3102160814Ssimon pass a list of arguments to any function as well as provide a way 3103160814Ssimon for a function to pass data back to the caller. 3104160814Ssimon [Richard Levitte] 3105160814Ssimon 3106160814Ssimon *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup() 3107160814Ssimon works like BUF_strdup() but can be used to duplicate a portion of 3108160814Ssimon a string. The copy gets NUL-terminated. BUF_memdup() duplicates 3109160814Ssimon a memory area. 3110160814Ssimon [Richard Levitte] 3111160814Ssimon 3112160814Ssimon *) Add the function sk_find_ex() which works like sk_find(), but will 3113160814Ssimon return an index to an element even if an exact match couldn't be 3114160814Ssimon found. The index is guaranteed to point at the element where the 3115160814Ssimon searched-for key would be inserted to preserve sorting order. 3116160814Ssimon [Richard Levitte] 3117160814Ssimon 3118160814Ssimon *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but 3119160814Ssimon takes an extra flags argument for optional functionality. Currently, 3120160814Ssimon the following flags are defined: 3121160814Ssimon 3122160814Ssimon OBJ_BSEARCH_VALUE_ON_NOMATCH 3123160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 3124160814Ssimon element where the comparing function returns a negative or zero 3125160814Ssimon number. 3126160814Ssimon 3127160814Ssimon OBJ_BSEARCH_FIRST_VALUE_ON_MATCH 3128160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 3129160814Ssimon element where the comparing function returns zero. This is useful 3130160814Ssimon if there are more than one element where the comparing function 3131160814Ssimon returns zero. 3132160814Ssimon [Richard Levitte] 3133160814Ssimon 3134160814Ssimon *) Make it possible to create self-signed certificates with 'openssl ca' 3135160814Ssimon in such a way that the self-signed certificate becomes part of the 3136160814Ssimon CA database and uses the same mechanisms for serial number generation 3137160814Ssimon as all other certificate signing. The new flag '-selfsign' enables 3138160814Ssimon this functionality. Adapt CA.sh and CA.pl.in. 3139160814Ssimon [Richard Levitte] 3140160814Ssimon 3141160814Ssimon *) Add functionality to check the public key of a certificate request 3142160814Ssimon against a given private. This is useful to check that a certificate 3143160814Ssimon request can be signed by that key (self-signing). 3144160814Ssimon [Richard Levitte] 3145160814Ssimon 3146160814Ssimon *) Make it possible to have multiple active certificates with the same 3147160814Ssimon subject in the CA index file. This is done only if the keyword 3148160814Ssimon 'unique_subject' is set to 'no' in the main CA section (default 3149160814Ssimon if 'CA_default') of the configuration file. The value is saved 3150160814Ssimon with the database itself in a separate index attribute file, 3151160814Ssimon named like the index file with '.attr' appended to the name. 3152160814Ssimon [Richard Levitte] 3153160814Ssimon 3154160814Ssimon *) Generate muti valued AVAs using '+' notation in config files for 3155160814Ssimon req and dirName. 3156160814Ssimon [Steve Henson] 3157160814Ssimon 3158160814Ssimon *) Support for nameConstraints certificate extension. 3159160814Ssimon [Steve Henson] 3160160814Ssimon 3161160814Ssimon *) Support for policyConstraints certificate extension. 3162160814Ssimon [Steve Henson] 3163160814Ssimon 3164160814Ssimon *) Support for policyMappings certificate extension. 3165160814Ssimon [Steve Henson] 3166160814Ssimon 3167160814Ssimon *) Make sure the default DSA_METHOD implementation only uses its 3168160814Ssimon dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL, 3169160814Ssimon and change its own handlers to be NULL so as to remove unnecessary 3170160814Ssimon indirection. This lets alternative implementations fallback to the 3171160814Ssimon default implementation more easily. 3172160814Ssimon [Geoff Thorpe] 3173160814Ssimon 3174160814Ssimon *) Support for directoryName in GeneralName related extensions 3175160814Ssimon in config files. 3176160814Ssimon [Steve Henson] 3177160814Ssimon 3178160814Ssimon *) Make it possible to link applications using Makefile.shared. 3179160814Ssimon Make that possible even when linking against static libraries! 3180160814Ssimon [Richard Levitte] 3181160814Ssimon 3182160814Ssimon *) Support for single pass processing for S/MIME signing. This now 3183160814Ssimon means that S/MIME signing can be done from a pipe, in addition 3184160814Ssimon cleartext signing (multipart/signed type) is effectively streaming 3185160814Ssimon and the signed data does not need to be all held in memory. 3186160814Ssimon 3187160814Ssimon This is done with a new flag PKCS7_STREAM. When this flag is set 3188160814Ssimon PKCS7_sign() only initializes the PKCS7 structure and the actual signing 3189160814Ssimon is done after the data is output (and digests calculated) in 3190160814Ssimon SMIME_write_PKCS7(). 3191160814Ssimon [Steve Henson] 3192160814Ssimon 3193160814Ssimon *) Add full support for -rpath/-R, both in shared libraries and 3194160814Ssimon applications, at least on the platforms where it's known how 3195160814Ssimon to do it. 3196160814Ssimon [Richard Levitte] 3197160814Ssimon 3198160814Ssimon *) In crypto/ec/ec_mult.c, implement fast point multiplication with 3199160814Ssimon precomputation, based on wNAF splitting: EC_GROUP_precompute_mult() 3200160814Ssimon will now compute a table of multiples of the generator that 3201160814Ssimon makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul() 3202160814Ssimon faster (notably in the case of a single point multiplication, 3203160814Ssimon scalar * generator). 3204160814Ssimon [Nils Larsch, Bodo Moeller] 3205160814Ssimon 3206160814Ssimon *) IPv6 support for certificate extensions. The various extensions 3207160814Ssimon which use the IP:a.b.c.d can now take IPv6 addresses using the 3208160814Ssimon formats of RFC1884 2.2 . IPv6 addresses are now also displayed 3209160814Ssimon correctly. 3210160814Ssimon [Steve Henson] 3211160814Ssimon 3212160814Ssimon *) Added an ENGINE that implements RSA by performing private key 3213160814Ssimon exponentiations with the GMP library. The conversions to and from 3214160814Ssimon GMP's mpz_t format aren't optimised nor are any montgomery forms 3215160814Ssimon cached, and on x86 it appears OpenSSL's own performance has caught up. 3216160814Ssimon However there are likely to be other architectures where GMP could 3217160814Ssimon provide a boost. This ENGINE is not built in by default, but it can be 3218160814Ssimon specified at Configure time and should be accompanied by the necessary 3219160814Ssimon linker additions, eg; 3220160814Ssimon ./config -DOPENSSL_USE_GMP -lgmp 3221160814Ssimon [Geoff Thorpe] 3222160814Ssimon 3223160814Ssimon *) "openssl engine" will not display ENGINE/DSO load failure errors when 3224160814Ssimon testing availability of engines with "-t" - the old behaviour is 3225160814Ssimon produced by increasing the feature's verbosity with "-tt". 3226160814Ssimon [Geoff Thorpe] 3227160814Ssimon 3228160814Ssimon *) ECDSA routines: under certain error conditions uninitialized BN objects 3229160814Ssimon could be freed. Solution: make sure initialization is performed early 3230160814Ssimon enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de> 3231160814Ssimon via PR#459) 3232160814Ssimon [Lutz Jaenicke] 3233160814Ssimon 3234160814Ssimon *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD 3235160814Ssimon and DH_METHOD (eg. by ENGINE implementations) to override the normal 3236160814Ssimon software implementations. For DSA and DH, parameter generation can 3237160814Ssimon also be overriden by providing the appropriate method callbacks. 3238160814Ssimon [Geoff Thorpe] 3239160814Ssimon 3240160814Ssimon *) Change the "progress" mechanism used in key-generation and 3241160814Ssimon primality testing to functions that take a new BN_GENCB pointer in 3242160814Ssimon place of callback/argument pairs. The new API functions have "_ex" 3243160814Ssimon postfixes and the older functions are reimplemented as wrappers for 3244160814Ssimon the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide 3245160814Ssimon declarations of the old functions to help (graceful) attempts to 3246160814Ssimon migrate to the new functions. Also, the new key-generation API 3247160814Ssimon functions operate on a caller-supplied key-structure and return 3248160814Ssimon success/failure rather than returning a key or NULL - this is to 3249160814Ssimon help make "keygen" another member function of RSA_METHOD etc. 3250160814Ssimon 3251160814Ssimon Example for using the new callback interface: 3252160814Ssimon 3253160814Ssimon int (*my_callback)(int a, int b, BN_GENCB *cb) = ...; 3254160814Ssimon void *my_arg = ...; 3255160814Ssimon BN_GENCB my_cb; 3256160814Ssimon 3257160814Ssimon BN_GENCB_set(&my_cb, my_callback, my_arg); 3258160814Ssimon 3259160814Ssimon return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb); 3260160814Ssimon /* For the meaning of a, b in calls to my_callback(), see the 3261160814Ssimon * documentation of the function that calls the callback. 3262160814Ssimon * cb will point to my_cb; my_arg can be retrieved as cb->arg. 3263160814Ssimon * my_callback should return 1 if it wants BN_is_prime_ex() 3264160814Ssimon * to continue, or 0 to stop. 3265160814Ssimon */ 3266160814Ssimon 3267160814Ssimon [Geoff Thorpe] 3268160814Ssimon 3269160814Ssimon *) Change the ZLIB compression method to be stateful, and make it 3270160814Ssimon available to TLS with the number defined in 3271160814Ssimon draft-ietf-tls-compression-04.txt. 3272160814Ssimon [Richard Levitte] 3273160814Ssimon 3274160814Ssimon *) Add the ASN.1 structures and functions for CertificatePair, which 3275160814Ssimon is defined as follows (according to X.509_4thEditionDraftV6.pdf): 3276160814Ssimon 3277160814Ssimon CertificatePair ::= SEQUENCE { 3278160814Ssimon forward [0] Certificate OPTIONAL, 3279160814Ssimon reverse [1] Certificate OPTIONAL, 3280160814Ssimon -- at least one of the pair shall be present -- } 3281160814Ssimon 3282160814Ssimon Also implement the PEM functions to read and write certificate 3283160814Ssimon pairs, and defined the PEM tag as "CERTIFICATE PAIR". 3284160814Ssimon 3285160814Ssimon This needed to be defined, mostly for the sake of the LDAP 3286160814Ssimon attribute crossCertificatePair, but may prove useful elsewhere as 3287160814Ssimon well. 3288160814Ssimon [Richard Levitte] 3289160814Ssimon 3290160814Ssimon *) Make it possible to inhibit symlinking of shared libraries in 3291160814Ssimon Makefile.shared, for Cygwin's sake. 3292160814Ssimon [Richard Levitte] 3293160814Ssimon 3294160814Ssimon *) Extend the BIGNUM API by creating a function 3295160814Ssimon void BN_set_negative(BIGNUM *a, int neg); 3296160814Ssimon and a macro that behave like 3297160814Ssimon int BN_is_negative(const BIGNUM *a); 3298160814Ssimon 3299160814Ssimon to avoid the need to access 'a->neg' directly in applications. 3300160814Ssimon [Nils Larsch] 3301160814Ssimon 3302160814Ssimon *) Implement fast modular reduction for pseudo-Mersenne primes 3303160814Ssimon used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c). 3304160814Ssimon EC_GROUP_new_curve_GFp() will now automatically use this 3305160814Ssimon if applicable. 3306160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3307160814Ssimon 3308160814Ssimon *) Add new lock type (CRYPTO_LOCK_BN). 3309160814Ssimon [Bodo Moeller] 3310160814Ssimon 3311160814Ssimon *) Change the ENGINE framework to automatically load engines 3312160814Ssimon dynamically from specific directories unless they could be 3313160814Ssimon found to already be built in or loaded. Move all the 3314160814Ssimon current engines except for the cryptodev one to a new 3315160814Ssimon directory engines/. 3316160814Ssimon The engines in engines/ are built as shared libraries if 3317160814Ssimon the "shared" options was given to ./Configure or ./config. 3318160814Ssimon Otherwise, they are inserted in libcrypto.a. 3319160814Ssimon /usr/local/ssl/engines is the default directory for dynamic 3320160814Ssimon engines, but that can be overriden at configure time through 3321160814Ssimon the usual use of --prefix and/or --openssldir, and at run 3322160814Ssimon time with the environment variable OPENSSL_ENGINES. 3323160814Ssimon [Geoff Thorpe and Richard Levitte] 3324160814Ssimon 3325160814Ssimon *) Add Makefile.shared, a helper makefile to build shared 3326160814Ssimon libraries. Addapt Makefile.org. 3327160814Ssimon [Richard Levitte] 3328160814Ssimon 3329160814Ssimon *) Add version info to Win32 DLLs. 3330160814Ssimon [Peter 'Luna' Runestig" <peter@runestig.com>] 3331160814Ssimon 3332160814Ssimon *) Add new 'medium level' PKCS#12 API. Certificates and keys 3333160814Ssimon can be added using this API to created arbitrary PKCS#12 3334160814Ssimon files while avoiding the low level API. 3335160814Ssimon 3336160814Ssimon New options to PKCS12_create(), key or cert can be NULL and 3337160814Ssimon will then be omitted from the output file. The encryption 3338160814Ssimon algorithm NIDs can be set to -1 for no encryption, the mac 3339160814Ssimon iteration count can be set to 0 to omit the mac. 3340160814Ssimon 3341160814Ssimon Enhance pkcs12 utility by making the -nokeys and -nocerts 3342160814Ssimon options work when creating a PKCS#12 file. New option -nomac 3343160814Ssimon to omit the mac, NONE can be set for an encryption algorithm. 3344160814Ssimon New code is modified to use the enhanced PKCS12_create() 3345160814Ssimon instead of the low level API. 3346160814Ssimon [Steve Henson] 3347160814Ssimon 3348160814Ssimon *) Extend ASN1 encoder to support indefinite length constructed 3349160814Ssimon encoding. This can output sequences tags and octet strings in 3350160814Ssimon this form. Modify pk7_asn1.c to support indefinite length 3351160814Ssimon encoding. This is experimental and needs additional code to 3352160814Ssimon be useful, such as an ASN1 bio and some enhanced streaming 3353160814Ssimon PKCS#7 code. 3354160814Ssimon 3355160814Ssimon Extend template encode functionality so that tagging is passed 3356160814Ssimon down to the template encoder. 3357160814Ssimon [Steve Henson] 3358160814Ssimon 3359160814Ssimon *) Let 'openssl req' fail if an argument to '-newkey' is not 3360160814Ssimon recognized instead of using RSA as a default. 3361160814Ssimon [Bodo Moeller] 3362160814Ssimon 3363160814Ssimon *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt. 3364160814Ssimon As these are not official, they are not included in "ALL"; 3365160814Ssimon the "ECCdraft" ciphersuite group alias can be used to select them. 3366160814Ssimon [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)] 3367160814Ssimon 3368160814Ssimon *) Add ECDH engine support. 3369160814Ssimon [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)] 3370160814Ssimon 3371160814Ssimon *) Add ECDH in new directory crypto/ecdh/. 3372160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 3373160814Ssimon 3374160814Ssimon *) Let BN_rand_range() abort with an error after 100 iterations 3375160814Ssimon without success (which indicates a broken PRNG). 3376160814Ssimon [Bodo Moeller] 3377160814Ssimon 3378160814Ssimon *) Change BN_mod_sqrt() so that it verifies that the input value 3379160814Ssimon is really the square of the return value. (Previously, 3380160814Ssimon BN_mod_sqrt would show GIGO behaviour.) 3381160814Ssimon [Bodo Moeller] 3382160814Ssimon 3383160814Ssimon *) Add named elliptic curves over binary fields from X9.62, SECG, 3384160814Ssimon and WAP/WTLS; add OIDs that were still missing. 3385160814Ssimon 3386160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3387160814Ssimon (Sun Microsystems Laboratories)] 3388160814Ssimon 3389160814Ssimon *) Extend the EC library for elliptic curves over binary fields 3390160814Ssimon (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/). 3391160814Ssimon New EC_METHOD: 3392160814Ssimon 3393160814Ssimon EC_GF2m_simple_method 3394160814Ssimon 3395160814Ssimon New API functions: 3396160814Ssimon 3397160814Ssimon EC_GROUP_new_curve_GF2m 3398160814Ssimon EC_GROUP_set_curve_GF2m 3399160814Ssimon EC_GROUP_get_curve_GF2m 3400160814Ssimon EC_POINT_set_affine_coordinates_GF2m 3401160814Ssimon EC_POINT_get_affine_coordinates_GF2m 3402160814Ssimon EC_POINT_set_compressed_coordinates_GF2m 3403160814Ssimon 3404160814Ssimon Point compression for binary fields is disabled by default for 3405160814Ssimon patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to 3406160814Ssimon enable it). 3407160814Ssimon 3408160814Ssimon As binary polynomials are represented as BIGNUMs, various members 3409160814Ssimon of the EC_GROUP and EC_POINT data structures can be shared 3410160814Ssimon between the implementations for prime fields and binary fields; 3411160814Ssimon the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m) 3412160814Ssimon are essentially identical to their ..._GFp counterparts. 3413160814Ssimon (For simplicity, the '..._GFp' prefix has been dropped from 3414160814Ssimon various internal method names.) 3415160814Ssimon 3416160814Ssimon An internal 'field_div' method (similar to 'field_mul' and 3417160814Ssimon 'field_sqr') has been added; this is used only for binary fields. 3418160814Ssimon 3419160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3420160814Ssimon (Sun Microsystems Laboratories)] 3421160814Ssimon 3422160814Ssimon *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult() 3423160814Ssimon through methods ('mul', 'precompute_mult'). 3424160814Ssimon 3425160814Ssimon The generic implementations (now internally called 'ec_wNAF_mul' 3426160814Ssimon and 'ec_wNAF_precomputed_mult') remain the default if these 3427160814Ssimon methods are undefined. 3428160814Ssimon 3429160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3430160814Ssimon (Sun Microsystems Laboratories)] 3431160814Ssimon 3432160814Ssimon *) New function EC_GROUP_get_degree, which is defined through 3433160814Ssimon EC_METHOD. For curves over prime fields, this returns the bit 3434160814Ssimon length of the modulus. 3435160814Ssimon 3436160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3437160814Ssimon (Sun Microsystems Laboratories)] 3438160814Ssimon 3439160814Ssimon *) New functions EC_GROUP_dup, EC_POINT_dup. 3440160814Ssimon (These simply call ..._new and ..._copy). 3441160814Ssimon 3442160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3443160814Ssimon (Sun Microsystems Laboratories)] 3444160814Ssimon 3445160814Ssimon *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c. 3446160814Ssimon Polynomials are represented as BIGNUMs (where the sign bit is not 3447160814Ssimon used) in the following functions [macros]: 3448160814Ssimon 3449160814Ssimon BN_GF2m_add 3450160814Ssimon BN_GF2m_sub [= BN_GF2m_add] 3451160814Ssimon BN_GF2m_mod [wrapper for BN_GF2m_mod_arr] 3452160814Ssimon BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr] 3453160814Ssimon BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr] 3454160814Ssimon BN_GF2m_mod_inv 3455160814Ssimon BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr] 3456160814Ssimon BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr] 3457160814Ssimon BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr] 3458160814Ssimon BN_GF2m_cmp [= BN_ucmp] 3459160814Ssimon 3460160814Ssimon (Note that only the 'mod' functions are actually for fields GF(2^m). 3461160814Ssimon BN_GF2m_add() is misnomer, but this is for the sake of consistency.) 3462160814Ssimon 3463160814Ssimon For some functions, an the irreducible polynomial defining a 3464160814Ssimon field can be given as an 'unsigned int[]' with strictly 3465160814Ssimon decreasing elements giving the indices of those bits that are set; 3466160814Ssimon i.e., p[] represents the polynomial 3467160814Ssimon f(t) = t^p[0] + t^p[1] + ... + t^p[k] 3468160814Ssimon where 3469160814Ssimon p[0] > p[1] > ... > p[k] = 0. 3470160814Ssimon This applies to the following functions: 3471160814Ssimon 3472160814Ssimon BN_GF2m_mod_arr 3473160814Ssimon BN_GF2m_mod_mul_arr 3474160814Ssimon BN_GF2m_mod_sqr_arr 3475160814Ssimon BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv] 3476160814Ssimon BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div] 3477160814Ssimon BN_GF2m_mod_exp_arr 3478160814Ssimon BN_GF2m_mod_sqrt_arr 3479160814Ssimon BN_GF2m_mod_solve_quad_arr 3480160814Ssimon BN_GF2m_poly2arr 3481160814Ssimon BN_GF2m_arr2poly 3482160814Ssimon 3483160814Ssimon Conversion can be performed by the following functions: 3484160814Ssimon 3485160814Ssimon BN_GF2m_poly2arr 3486160814Ssimon BN_GF2m_arr2poly 3487160814Ssimon 3488160814Ssimon bntest.c has additional tests for binary polynomial arithmetic. 3489160814Ssimon 3490160814Ssimon Two implementations for BN_GF2m_mod_div() are available. 3491160814Ssimon The default algorithm simply uses BN_GF2m_mod_inv() and 3492160814Ssimon BN_GF2m_mod_mul(). The alternative algorithm is compiled in only 3493160814Ssimon if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the 3494160814Ssimon copyright notice in crypto/bn/bn_gf2m.c before enabling it). 3495160814Ssimon 3496160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3497160814Ssimon (Sun Microsystems Laboratories)] 3498160814Ssimon 3499160814Ssimon *) Add new error code 'ERR_R_DISABLED' that can be used when some 3500160814Ssimon functionality is disabled at compile-time. 3501160814Ssimon [Douglas Stebila <douglas.stebila@sun.com>] 3502160814Ssimon 3503160814Ssimon *) Change default behaviour of 'openssl asn1parse' so that more 3504160814Ssimon information is visible when viewing, e.g., a certificate: 3505160814Ssimon 3506160814Ssimon Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump' 3507160814Ssimon mode the content of non-printable OCTET STRINGs is output in a 3508160814Ssimon style similar to INTEGERs, but with '[HEX DUMP]' prepended to 3509160814Ssimon avoid the appearance of a printable string. 3510160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3511160814Ssimon 3512160814Ssimon *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access 3513160814Ssimon functions 3514160814Ssimon EC_GROUP_set_asn1_flag() 3515160814Ssimon EC_GROUP_get_asn1_flag() 3516160814Ssimon EC_GROUP_set_point_conversion_form() 3517160814Ssimon EC_GROUP_get_point_conversion_form() 3518160814Ssimon These control ASN1 encoding details: 3519160814Ssimon - Curves (i.e., groups) are encoded explicitly unless asn1_flag 3520160814Ssimon has been set to OPENSSL_EC_NAMED_CURVE. 3521160814Ssimon - Points are encoded in uncompressed form by default; options for 3522160814Ssimon asn1_for are as for point2oct, namely 3523160814Ssimon POINT_CONVERSION_COMPRESSED 3524160814Ssimon POINT_CONVERSION_UNCOMPRESSED 3525160814Ssimon POINT_CONVERSION_HYBRID 3526160814Ssimon 3527160814Ssimon Also add 'seed' and 'seed_len' members to EC_GROUP with access 3528160814Ssimon functions 3529160814Ssimon EC_GROUP_set_seed() 3530160814Ssimon EC_GROUP_get0_seed() 3531160814Ssimon EC_GROUP_get_seed_len() 3532160814Ssimon This is used only for ASN1 purposes (so far). 3533160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3534160814Ssimon 3535160814Ssimon *) Add 'field_type' member to EC_METHOD, which holds the NID 3536160814Ssimon of the appropriate field type OID. The new function 3537160814Ssimon EC_METHOD_get_field_type() returns this value. 3538160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3539160814Ssimon 3540160814Ssimon *) Add functions 3541160814Ssimon EC_POINT_point2bn() 3542160814Ssimon EC_POINT_bn2point() 3543160814Ssimon EC_POINT_point2hex() 3544160814Ssimon EC_POINT_hex2point() 3545160814Ssimon providing useful interfaces to EC_POINT_point2oct() and 3546160814Ssimon EC_POINT_oct2point(). 3547160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3548160814Ssimon 3549160814Ssimon *) Change internals of the EC library so that the functions 3550160814Ssimon EC_GROUP_set_generator() 3551160814Ssimon EC_GROUP_get_generator() 3552160814Ssimon EC_GROUP_get_order() 3553160814Ssimon EC_GROUP_get_cofactor() 3554160814Ssimon are implemented directly in crypto/ec/ec_lib.c and not dispatched 3555160814Ssimon to methods, which would lead to unnecessary code duplication when 3556160814Ssimon adding different types of curves. 3557160814Ssimon [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller] 3558160814Ssimon 3559160814Ssimon *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM 3560160814Ssimon arithmetic, and such that modified wNAFs are generated 3561160814Ssimon (which avoid length expansion in many cases). 3562160814Ssimon [Bodo Moeller] 3563160814Ssimon 3564160814Ssimon *) Add a function EC_GROUP_check_discriminant() (defined via 3565160814Ssimon EC_METHOD) that verifies that the curve discriminant is non-zero. 3566160814Ssimon 3567160814Ssimon Add a function EC_GROUP_check() that makes some sanity tests 3568160814Ssimon on a EC_GROUP, its generator and order. This includes 3569160814Ssimon EC_GROUP_check_discriminant(). 3570160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3571160814Ssimon 3572160814Ssimon *) Add ECDSA in new directory crypto/ecdsa/. 3573160814Ssimon 3574160814Ssimon Add applications 'openssl ecparam' and 'openssl ecdsa' 3575160814Ssimon (these are based on 'openssl dsaparam' and 'openssl dsa'). 3576160814Ssimon 3577160814Ssimon ECDSA support is also included in various other files across the 3578160814Ssimon library. Most notably, 3579160814Ssimon - 'openssl req' now has a '-newkey ecdsa:file' option; 3580160814Ssimon - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA; 3581160814Ssimon - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and 3582160814Ssimon d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make 3583160814Ssimon them suitable for ECDSA where domain parameters must be 3584160814Ssimon extracted before the specific public key; 3585160814Ssimon - ECDSA engine support has been added. 3586160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3587160814Ssimon 3588160814Ssimon *) Include some named elliptic curves, and add OIDs from X9.62, 3589160814Ssimon SECG, and WAP/WTLS. Each curve can be obtained from the new 3590160814Ssimon function 3591160814Ssimon EC_GROUP_new_by_curve_name(), 3592160814Ssimon and the list of available named curves can be obtained with 3593160814Ssimon EC_get_builtin_curves(). 3594160814Ssimon Also add a 'curve_name' member to EC_GROUP objects, which can be 3595160814Ssimon accessed via 3596160814Ssimon EC_GROUP_set_curve_name() 3597160814Ssimon EC_GROUP_get_curve_name() 3598160814Ssimon [Nils Larsch <larsch@trustcenter.de, Bodo Moeller] 3599160814Ssimon 3600160814Ssimon *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 3601160814Ssimon was actually never needed) and in BN_mul(). The removal in BN_mul() 3602160814Ssimon required a small change in bn_mul_part_recursive() and the addition 3603160814Ssimon of the functions bn_cmp_part_words(), bn_sub_part_words() and 3604160814Ssimon bn_add_part_words(), which do the same thing as bn_cmp_words(), 3605160814Ssimon bn_sub_words() and bn_add_words() except they take arrays with 3606160814Ssimon differing sizes. 3607160814Ssimon [Richard Levitte] 3608160814Ssimon 3609194206Ssimon Changes between 0.9.7l and 0.9.7m [23 Feb 2007] 3610194206Ssimon 3611167612Ssimon *) Cleanse PEM buffers before freeing them since they may contain 3612167612Ssimon sensitive data. 3613167612Ssimon [Benjamin Bennett <ben@psc.edu>] 3614167612Ssimon 3615167612Ssimon *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 3616167612Ssimon a ciphersuite string such as "DEFAULT:RSA" cannot enable 3617167612Ssimon authentication-only ciphersuites. 3618167612Ssimon [Bodo Moeller] 3619167612Ssimon 3620167612Ssimon *) Since AES128 and AES256 share a single mask bit in the logic of 3621167612Ssimon ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 3622167612Ssimon kludge to work properly if AES128 is available and AES256 isn't. 3623167612Ssimon [Victor Duchovni] 3624167612Ssimon 3625194206Ssimon *) Expand security boundary to match 1.1.1 module. 3626194206Ssimon [Steve Henson] 3627194206Ssimon 3628194206Ssimon *) Remove redundant features: hash file source, editing of test vectors 3629194206Ssimon modify fipsld to use external fips_premain.c signature. 3630194206Ssimon [Steve Henson] 3631194206Ssimon 3632194206Ssimon *) New perl script mkfipsscr.pl to create shell scripts or batch files to 3633194206Ssimon run algorithm test programs. 3634194206Ssimon [Steve Henson] 3635194206Ssimon 3636194206Ssimon *) Make algorithm test programs more tolerant of whitespace. 3637194206Ssimon [Steve Henson] 3638194206Ssimon 3639167612Ssimon *) Have SSL/TLS server implementation tolerate "mismatched" record 3640167612Ssimon protocol version while receiving ClientHello even if the 3641167612Ssimon ClientHello is fragmented. (The server can't insist on the 3642167612Ssimon particular protocol version it has chosen before the ServerHello 3643167612Ssimon message has informed the client about his choice.) 3644167612Ssimon [Bodo Moeller] 3645167612Ssimon 3646167612Ssimon *) Load error codes if they are not already present instead of using a 3647167612Ssimon static variable. This allows them to be cleanly unloaded and reloaded. 3648167612Ssimon [Steve Henson] 3649167612Ssimon 3650167612Ssimon Changes between 0.9.7k and 0.9.7l [28 Sep 2006] 3651167612Ssimon 3652167612Ssimon *) Introduce limits to prevent malicious keys being able to 3653167612Ssimon cause a denial of service. (CVE-2006-2940) 3654167612Ssimon [Steve Henson, Bodo Moeller] 3655167612Ssimon 3656167612Ssimon *) Fix ASN.1 parsing of certain invalid structures that can result 3657167612Ssimon in a denial of service. (CVE-2006-2937) [Steve Henson] 3658167612Ssimon 3659167612Ssimon *) Fix buffer overflow in SSL_get_shared_ciphers() function. 3660167612Ssimon (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 3661167612Ssimon 3662167612Ssimon *) Fix SSL client code which could crash if connecting to a 3663167612Ssimon malicious SSLv2 server. (CVE-2006-4343) 3664167612Ssimon [Tavis Ormandy and Will Drewry, Google Security Team] 3665167612Ssimon 3666162911Ssimon *) Change ciphersuite string processing so that an explicit 3667162911Ssimon ciphersuite selects this one ciphersuite (so that "AES256-SHA" 3668162911Ssimon will no longer include "AES128-SHA"), and any other similar 3669162911Ssimon ciphersuite (same bitmap) from *other* protocol versions (so that 3670162911Ssimon "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the 3671162911Ssimon SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining 3672162911Ssimon changes from 0.9.8b and 0.9.8d. 3673162911Ssimon [Bodo Moeller] 3674162911Ssimon 3675162911Ssimon Changes between 0.9.7j and 0.9.7k [05 Sep 2006] 3676162911Ssimon 3677162911Ssimon *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 3678162911Ssimon (CVE-2006-4339) [Ben Laurie and Google Security Team] 3679162911Ssimon 3680162911Ssimon *) Change the Unix randomness entropy gathering to use poll() when 3681162911Ssimon possible instead of select(), since the latter has some 3682162911Ssimon undesirable limitations. 3683162911Ssimon [Darryl Miles via Richard Levitte and Bodo Moeller] 3684162911Ssimon 3685162911Ssimon *) Disable rogue ciphersuites: 3686162911Ssimon 3687162911Ssimon - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 3688162911Ssimon - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 3689162911Ssimon - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 3690162911Ssimon 3691162911Ssimon The latter two were purportedly from 3692162911Ssimon draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 3693162911Ssimon appear there. 3694162911Ssimon 3695162911Ssimon Also deactive the remaining ciphersuites from 3696162911Ssimon draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 3697162911Ssimon unofficial, and the ID has long expired. 3698162911Ssimon [Bodo Moeller] 3699162911Ssimon 3700162911Ssimon *) Fix RSA blinding Heisenbug (problems sometimes occured on 3701162911Ssimon dual-core machines) and other potential thread-safety issues. 3702162911Ssimon [Bodo Moeller] 3703162911Ssimon 3704162911Ssimon Changes between 0.9.7i and 0.9.7j [04 May 2006] 3705162911Ssimon 3706162911Ssimon *) Adapt fipsld and the build system to link against the validated FIPS 3707162911Ssimon module in FIPS mode. 3708162911Ssimon [Steve Henson] 3709162911Ssimon 3710162911Ssimon *) Fixes for VC++ 2005 build under Windows. 3711162911Ssimon [Steve Henson] 3712162911Ssimon 3713162911Ssimon *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make 3714162911Ssimon from a Windows bash shell such as MSYS. It is autodetected from the 3715162911Ssimon "config" script when run from a VC++ environment. Modify standard VC++ 3716162911Ssimon build to use fipscanister.o from the GNU make build. 3717162911Ssimon [Steve Henson] 3718162911Ssimon 3719160814Ssimon Changes between 0.9.7h and 0.9.7i [14 Oct 2005] 3720160814Ssimon 3721160814Ssimon *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS. 3722160814Ssimon The value now differs depending on if you build for FIPS or not. 3723160814Ssimon BEWARE! A program linked with a shared FIPSed libcrypto can't be 3724160814Ssimon safely run with a non-FIPSed libcrypto, as it may crash because of 3725160814Ssimon the difference induced by this change. 3726160814Ssimon [Andy Polyakov] 3727160814Ssimon 3728160814Ssimon Changes between 0.9.7g and 0.9.7h [11 Oct 2005] 3729160814Ssimon 3730160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 3731160814Ssimon (part of SSL_OP_ALL). This option used to disable the 3732160814Ssimon countermeasure against man-in-the-middle protocol-version 3733160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 3734160814Ssimon idea. (CVE-2005-2969) 3735160814Ssimon 3736160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 3737160814Ssimon for Information Security, National Institute of Advanced Industrial 3738160814Ssimon Science and Technology [AIST], Japan)] 3739160814Ssimon 3740160814Ssimon *) Minimal support for X9.31 signatures and PSS padding modes. This is 3741160814Ssimon mainly for FIPS compliance and not fully integrated at this stage. 3742160814Ssimon [Steve Henson] 3743160814Ssimon 3744160814Ssimon *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform 3745160814Ssimon the exponentiation using a fixed-length exponent. (Otherwise, 3746160814Ssimon the information leaked through timing could expose the secret key 3747160814Ssimon after many signatures; cf. Bleichenbacher's attack on DSA with 3748160814Ssimon biased k.) 3749160814Ssimon [Bodo Moeller] 3750160814Ssimon 3751160814Ssimon *) Make a new fixed-window mod_exp implementation the default for 3752160814Ssimon RSA, DSA, and DH private-key operations so that the sequence of 3753160814Ssimon squares and multiplies and the memory access pattern are 3754160814Ssimon independent of the particular secret key. This will mitigate 3755160814Ssimon cache-timing and potential related attacks. 3756160814Ssimon 3757160814Ssimon BN_mod_exp_mont_consttime() is the new exponentiation implementation, 3758160814Ssimon and this is automatically used by BN_mod_exp_mont() if the new flag 3759160814Ssimon BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH 3760160814Ssimon will use this BN flag for private exponents unless the flag 3761160814Ssimon RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or 3762160814Ssimon DH_FLAG_NO_EXP_CONSTTIME, respectively, is set. 3763160814Ssimon 3764160814Ssimon [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller] 3765160814Ssimon 3766160814Ssimon *) Change the client implementation for SSLv23_method() and 3767160814Ssimon SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0 3768160814Ssimon Client Hello message format if the SSL_OP_NO_SSLv2 option is set. 3769160814Ssimon (Previously, the SSL 2.0 backwards compatible Client Hello 3770160814Ssimon message format would be used even with SSL_OP_NO_SSLv2.) 3771160814Ssimon [Bodo Moeller] 3772160814Ssimon 3773160814Ssimon *) Add support for smime-type MIME parameter in S/MIME messages which some 3774160814Ssimon clients need. 3775160814Ssimon [Steve Henson] 3776160814Ssimon 3777160814Ssimon *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in 3778160814Ssimon a threadsafe manner. Modify rsa code to use new function and add calls 3779160814Ssimon to dsa and dh code (which had race conditions before). 3780160814Ssimon [Steve Henson] 3781160814Ssimon 3782160814Ssimon *) Include the fixed error library code in the C error file definitions 3783160814Ssimon instead of fixing them up at runtime. This keeps the error code 3784160814Ssimon structures constant. 3785160814Ssimon [Steve Henson] 3786160814Ssimon 3787160814Ssimon Changes between 0.9.7f and 0.9.7g [11 Apr 2005] 3788160814Ssimon 3789160814Ssimon [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after 3790160814Ssimon OpenSSL 0.9.8.] 3791160814Ssimon 3792160814Ssimon *) Fixes for newer kerberos headers. NB: the casts are needed because 3793160814Ssimon the 'length' field is signed on one version and unsigned on another 3794160814Ssimon with no (?) obvious way to tell the difference, without these VC++ 3795160814Ssimon complains. Also the "definition" of FAR (blank) is no longer included 3796160814Ssimon nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up 3797160814Ssimon some needed definitions. 3798160814Ssimon [Steve Henson] 3799160814Ssimon 3800160814Ssimon *) Undo Cygwin change. 3801160814Ssimon [Ulf M�ller] 3802160814Ssimon 3803160814Ssimon *) Added support for proxy certificates according to RFC 3820. 3804160814Ssimon Because they may be a security thread to unaware applications, 3805160814Ssimon they must be explicitely allowed in run-time. See 3806160814Ssimon docs/HOWTO/proxy_certificates.txt for further information. 3807160814Ssimon [Richard Levitte] 3808160814Ssimon 3809160814Ssimon Changes between 0.9.7e and 0.9.7f [22 Mar 2005] 3810160814Ssimon 3811160814Ssimon *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating 3812160814Ssimon server and client random values. Previously 3813160814Ssimon (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in 3814160814Ssimon less random data when sizeof(time_t) > 4 (some 64 bit platforms). 3815160814Ssimon 3816160814Ssimon This change has negligible security impact because: 3817160814Ssimon 3818160814Ssimon 1. Server and client random values still have 24 bytes of pseudo random 3819160814Ssimon data. 3820160814Ssimon 3821160814Ssimon 2. Server and client random values are sent in the clear in the initial 3822160814Ssimon handshake. 3823160814Ssimon 3824160814Ssimon 3. The master secret is derived using the premaster secret (48 bytes in 3825160814Ssimon size for static RSA ciphersuites) as well as client server and random 3826160814Ssimon values. 3827160814Ssimon 3828160814Ssimon The OpenSSL team would like to thank the UK NISCC for bringing this issue 3829160814Ssimon to our attention. 3830160814Ssimon 3831160814Ssimon [Stephen Henson, reported by UK NISCC] 3832160814Ssimon 3833160814Ssimon *) Use Windows randomness collection on Cygwin. 3834160814Ssimon [Ulf M�ller] 3835160814Ssimon 3836160814Ssimon *) Fix hang in EGD/PRNGD query when communication socket is closed 3837160814Ssimon prematurely by EGD/PRNGD. 3838160814Ssimon [Darren Tucker <dtucker@zip.com.au> via Lutz J�nicke, resolves #1014] 3839160814Ssimon 3840160814Ssimon *) Prompt for pass phrases when appropriate for PKCS12 input format. 3841160814Ssimon [Steve Henson] 3842160814Ssimon 3843160814Ssimon *) Back-port of selected performance improvements from development 3844160814Ssimon branch, as well as improved support for PowerPC platforms. 3845160814Ssimon [Andy Polyakov] 3846160814Ssimon 3847160814Ssimon *) Add lots of checks for memory allocation failure, error codes to indicate 3848160814Ssimon failure and freeing up memory if a failure occurs. 3849160814Ssimon [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson] 3850160814Ssimon 3851160814Ssimon *) Add new -passin argument to dgst. 3852160814Ssimon [Steve Henson] 3853160814Ssimon 3854160814Ssimon *) Perform some character comparisons of different types in X509_NAME_cmp: 3855160814Ssimon this is needed for some certificates that reencode DNs into UTF8Strings 3856160814Ssimon (in violation of RFC3280) and can't or wont issue name rollover 3857160814Ssimon certificates. 3858160814Ssimon [Steve Henson] 3859160814Ssimon 3860160814Ssimon *) Make an explicit check during certificate validation to see that 3861160814Ssimon the CA setting in each certificate on the chain is correct. As a 3862160814Ssimon side effect always do the following basic checks on extensions, 3863160814Ssimon not just when there's an associated purpose to the check: 3864160814Ssimon 3865160814Ssimon - if there is an unhandled critical extension (unless the user 3866160814Ssimon has chosen to ignore this fault) 3867160814Ssimon - if the path length has been exceeded (if one is set at all) 3868160814Ssimon - that certain extensions fit the associated purpose (if one has 3869160814Ssimon been given) 3870160814Ssimon [Richard Levitte] 3871160814Ssimon 3872142425Snectar Changes between 0.9.7d and 0.9.7e [25 Oct 2004] 3873142425Snectar 3874142425Snectar *) Avoid a race condition when CRLs are checked in a multi threaded 3875142425Snectar environment. This would happen due to the reordering of the revoked 3876142425Snectar entries during signature checking and serial number lookup. Now the 3877142425Snectar encoding is cached and the serial number sort performed under a lock. 3878142425Snectar Add new STACK function sk_is_sorted(). 3879142425Snectar [Steve Henson] 3880142425Snectar 3881142425Snectar *) Add Delta CRL to the extension code. 3882142425Snectar [Steve Henson] 3883142425Snectar 3884142425Snectar *) Various fixes to s3_pkt.c so alerts are sent properly. 3885142425Snectar [David Holmes <d.holmes@f5.com>] 3886142425Snectar 3887142425Snectar *) Reduce the chances of duplicate issuer name and serial numbers (in 3888142425Snectar violation of RFC3280) using the OpenSSL certificate creation utilities. 3889142425Snectar This is done by creating a random 64 bit value for the initial serial 3890142425Snectar number when a serial number file is created or when a self signed 3891142425Snectar certificate is created using 'openssl req -x509'. The initial serial 3892142425Snectar number file is created using 'openssl x509 -next_serial' in CA.pl 3893142425Snectar rather than being initialized to 1. 3894142425Snectar [Steve Henson] 3895142425Snectar 3896127128Snectar Changes between 0.9.7c and 0.9.7d [17 Mar 2004] 3897127128Snectar 3898127128Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 3899160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 3900127128Snectar [Joe Orton, Steve Henson] 3901127128Snectar 3902127128Snectar *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites 3903160814Ssimon (CVE-2004-0112) 3904127128Snectar [Joe Orton, Steve Henson] 3905127128Snectar 3906127128Snectar *) Make it possible to have multiple active certificates with the same 3907127128Snectar subject in the CA index file. This is done only if the keyword 3908127128Snectar 'unique_subject' is set to 'no' in the main CA section (default 3909127128Snectar if 'CA_default') of the configuration file. The value is saved 3910127128Snectar with the database itself in a separate index attribute file, 3911127128Snectar named like the index file with '.attr' appended to the name. 3912127128Snectar [Richard Levitte] 3913127128Snectar 3914127128Snectar *) X509 verify fixes. Disable broken certificate workarounds when 3915127128Snectar X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if 3916127128Snectar keyUsage extension present. Don't accept CRLs with unhandled critical 3917127128Snectar extensions: since verify currently doesn't process CRL extensions this 3918127128Snectar rejects a CRL with *any* critical extensions. Add new verify error codes 3919127128Snectar for these cases. 3920127128Snectar [Steve Henson] 3921127128Snectar 3922127128Snectar *) When creating an OCSP nonce use an OCTET STRING inside the extnValue. 3923127128Snectar A clarification of RFC2560 will require the use of OCTET STRINGs and 3924127128Snectar some implementations cannot handle the current raw format. Since OpenSSL 3925127128Snectar copies and compares OCSP nonces as opaque blobs without any attempt at 3926127128Snectar parsing them this should not create any compatibility issues. 3927127128Snectar [Steve Henson] 3928127128Snectar 3929127128Snectar *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when 3930127128Snectar calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without 3931127128Snectar this HMAC (and other) operations are several times slower than OpenSSL 3932127128Snectar < 0.9.7. 3933127128Snectar [Steve Henson] 3934127128Snectar 3935127128Snectar *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex(). 3936127128Snectar [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 3937127128Snectar 3938127128Snectar *) Use the correct content when signing type "other". 3939127128Snectar [Steve Henson] 3940127128Snectar 3941120631Snectar Changes between 0.9.7b and 0.9.7c [30 Sep 2003] 3942120631Snectar 3943120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 3944120631Snectar 3945120631Snectar Stop out of bounds reads in the ASN1 code when presented with 3946160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 3947120631Snectar 3948160814Ssimon Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545). 3949120631Snectar 3950120631Snectar If verify callback ignores invalid public key errors don't try to check 3951120631Snectar certificate signature with the NULL public key. 3952120631Snectar 3953120631Snectar [Steve Henson] 3954120631Snectar 3955120631Snectar *) New -ignore_err option in ocsp application to stop the server 3956120631Snectar exiting on the first error in a request. 3957120631Snectar [Steve Henson] 3958120631Snectar 3959120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 3960120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 3961120631Snectar specifications. 3962120631Snectar [Steve Henson] 3963120631Snectar 3964120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 3965120631Snectar extra data after the compression methods not only for TLS 1.0 3966120631Snectar but also for SSL 3.0 (as required by the specification). 3967120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 3968120631Snectar 3969120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 3970120631Snectar when it's 512 *bits* long, not 512 bytes. 3971120631Snectar [Richard Levitte] 3972120631Snectar 3973120631Snectar *) Change AES_cbc_encrypt() so it outputs exact multiple of 3974120631Snectar blocks during encryption. 3975120631Snectar [Richard Levitte] 3976120631Snectar 3977120631Snectar *) Various fixes to base64 BIO and non blocking I/O. On write 3978120631Snectar flushes were not handled properly if the BIO retried. On read 3979120631Snectar data was not being buffered properly and had various logic bugs. 3980120631Snectar This also affects blocking I/O when the data being decoded is a 3981120631Snectar certain size. 3982120631Snectar [Steve Henson] 3983120631Snectar 3984120631Snectar *) Various S/MIME bugfixes and compatibility changes: 3985120631Snectar output correct application/pkcs7 MIME type if 3986120631Snectar PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures. 3987120631Snectar Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening 3988120631Snectar of files as .eml work). Correctly handle very long lines in MIME 3989120631Snectar parser. 3990120631Snectar [Steve Henson] 3991120631Snectar 3992120631Snectar Changes between 0.9.7a and 0.9.7b [10 Apr 2003] 3993120631Snectar 3994120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 3995120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 3996120631Snectar a protocol version number mismatch like a decryption error 3997120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 3998120631Snectar [Bodo Moeller] 3999120631Snectar 4000120631Snectar *) Turn on RSA blinding by default in the default implementation 4001120631Snectar to avoid a timing attack. Applications that don't want it can call 4002120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 4003120631Snectar They would be ill-advised to do so in most cases. 4004120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 4005120631Snectar 4006120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 4007120631Snectar seeded (in this case, the secret RSA exponent is abused as 4008120631Snectar an unpredictable seed -- if it is not unpredictable, there 4009120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 4010120631Snectar by remembering the creator's thread ID in rsa->blinding and 4011120631Snectar having all other threads use local one-time blinding factors 4012120631Snectar (this requires more computation than sharing rsa->blinding, but 4013120631Snectar avoids excessive locking; and if an RSA object is not shared 4014120631Snectar between threads, blinding will still be very fast). 4015120631Snectar [Bodo Moeller] 4016120631Snectar 4017120631Snectar *) Fixed a typo bug that would cause ENGINE_set_default() to set an 4018120631Snectar ENGINE as defaults for all supported algorithms irrespective of 4019120631Snectar the 'flags' parameter. 'flags' is now honoured, so applications 4020120631Snectar should make sure they are passing it correctly. 4021120631Snectar [Geoff Thorpe] 4022120631Snectar 4023120631Snectar *) Target "mingw" now allows native Windows code to be generated in 4024120631Snectar the Cygwin environment as well as with the MinGW compiler. 4025120631Snectar [Ulf Moeller] 4026120631Snectar 4027111147Snectar Changes between 0.9.7 and 0.9.7a [19 Feb 2003] 4028111147Snectar 4029111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 4030111147Snectar via timing by performing a MAC computation even if incorrrect 4031111147Snectar block cipher padding has been found. This is a countermeasure 4032111147Snectar against active attacks where the attacker has to distinguish 4033160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 4034111147Snectar 4035111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 4036111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 4037111147Snectar Martin Vuagnoux (EPFL, Ilion)] 4038111147Snectar 4039111147Snectar *) Make the no-err option work as intended. The intention with no-err 4040111147Snectar is not to have the whole error stack handling routines removed from 4041111147Snectar libcrypto, it's only intended to remove all the function name and 4042111147Snectar reason texts, thereby removing some of the footprint that may not 4043111147Snectar be interesting if those errors aren't displayed anyway. 4044111147Snectar 4045111147Snectar NOTE: it's still possible for any application or module to have it's 4046111147Snectar own set of error texts inserted. The routines are there, just not 4047111147Snectar used by default when no-err is given. 4048111147Snectar [Richard Levitte] 4049111147Snectar 4050111147Snectar *) Add support for FreeBSD on IA64. 4051111147Snectar [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454] 4052111147Snectar 4053111147Snectar *) Adjust DES_cbc_cksum() so it returns the same value as the MIT 4054111147Snectar Kerberos function mit_des_cbc_cksum(). Before this change, 4055111147Snectar the value returned by DES_cbc_cksum() was like the one from 4056111147Snectar mit_des_cbc_cksum(), except the bytes were swapped. 4057111147Snectar [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte] 4058111147Snectar 4059111147Snectar *) Allow an application to disable the automatic SSL chain building. 4060111147Snectar Before this a rather primitive chain build was always performed in 4061111147Snectar ssl3_output_cert_chain(): an application had no way to send the 4062111147Snectar correct chain if the automatic operation produced an incorrect result. 4063111147Snectar 4064111147Snectar Now the chain builder is disabled if either: 4065111147Snectar 4066111147Snectar 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert(). 4067111147Snectar 4068111147Snectar 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set. 4069111147Snectar 4070111147Snectar The reasoning behind this is that an application would not want the 4071111147Snectar auto chain building to take place if extra chain certificates are 4072111147Snectar present and it might also want a means of sending no additional 4073111147Snectar certificates (for example the chain has two certificates and the 4074111147Snectar root is omitted). 4075111147Snectar [Steve Henson] 4076111147Snectar 4077111147Snectar *) Add the possibility to build without the ENGINE framework. 4078111147Snectar [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 4079111147Snectar 4080111147Snectar *) Under Win32 gmtime() can return NULL: check return value in 4081111147Snectar OPENSSL_gmtime(). Add error code for case where gmtime() fails. 4082111147Snectar [Steve Henson] 4083111147Snectar 4084111147Snectar *) DSA routines: under certain error conditions uninitialized BN objects 4085111147Snectar could be freed. Solution: make sure initialization is performed early 4086111147Snectar enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>, 4087111147Snectar Nils Larsch <nla@trustcenter.de> via PR#459) 4088111147Snectar [Lutz Jaenicke] 4089111147Snectar 4090111147Snectar *) Another fix for SSLv2 session ID handling: the session ID was incorrectly 4091111147Snectar checked on reconnect on the client side, therefore session resumption 4092111147Snectar could still fail with a "ssl session id is different" error. This 4093111147Snectar behaviour is masked when SSL_OP_ALL is used due to 4094111147Snectar SSL_OP_MICROSOFT_SESS_ID_BUG being set. 4095111147Snectar Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 4096111147Snectar followup to PR #377. 4097111147Snectar [Lutz Jaenicke] 4098111147Snectar 4099111147Snectar *) IA-32 assembler support enhancements: unified ELF targets, support 4100111147Snectar for SCO/Caldera platforms, fix for Cygwin shared build. 4101111147Snectar [Andy Polyakov] 4102111147Snectar 4103111147Snectar *) Add support for FreeBSD on sparc64. As a consequence, support for 4104111147Snectar FreeBSD on non-x86 processors is separate from x86 processors on 4105111147Snectar the config script, much like the NetBSD support. 4106111147Snectar [Richard Levitte & Kris Kennaway <kris@obsecurity.org>] 4107111147Snectar 4108109998Smarkm Changes between 0.9.6h and 0.9.7 [31 Dec 2002] 4109109998Smarkm 4110120631Snectar [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after 4111120631Snectar OpenSSL 0.9.7.] 4112120631Snectar 4113109998Smarkm *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED 4114109998Smarkm code (06) was taken as the first octet of the session ID and the last 4115109998Smarkm octet was ignored consequently. As a result SSLv2 client side session 4116109998Smarkm caching could not have worked due to the session ID mismatch between 4117109998Smarkm client and server. 4118109998Smarkm Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 4119109998Smarkm PR #377. 4120109998Smarkm [Lutz Jaenicke] 4121109998Smarkm 4122109998Smarkm *) Change the declaration of needed Kerberos libraries to use EX_LIBS 4123109998Smarkm instead of the special (and badly supported) LIBKRB5. LIBKRB5 is 4124109998Smarkm removed entirely. 4125109998Smarkm [Richard Levitte] 4126109998Smarkm 4127109998Smarkm *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it 4128109998Smarkm seems that in spite of existing for more than a year, many application 4129109998Smarkm author have done nothing to provide the necessary callbacks, which 4130109998Smarkm means that this particular engine will not work properly anywhere. 4131109998Smarkm This is a very unfortunate situation which forces us, in the name 4132109998Smarkm of usability, to give the hw_ncipher.c a static lock, which is part 4133109998Smarkm of libcrypto. 4134109998Smarkm NOTE: This is for the 0.9.7 series ONLY. This hack will never 4135109998Smarkm appear in 0.9.8 or later. We EXPECT application authors to have 4136109998Smarkm dealt properly with this when 0.9.8 is released (unless we actually 4137109998Smarkm make such changes in the libcrypto locking code that changes will 4138109998Smarkm have to be made anyway). 4139109998Smarkm [Richard Levitte] 4140109998Smarkm 4141109998Smarkm *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content 4142109998Smarkm octets have been read, EOF or an error occurs. Without this change 4143109998Smarkm some truncated ASN1 structures will not produce an error. 4144109998Smarkm [Steve Henson] 4145109998Smarkm 4146109998Smarkm *) Disable Heimdal support, since it hasn't been fully implemented. 4147109998Smarkm Still give the possibility to force the use of Heimdal, but with 4148109998Smarkm warnings and a request that patches get sent to openssl-dev. 4149109998Smarkm [Richard Levitte] 4150109998Smarkm 4151109998Smarkm *) Add the VC-CE target, introduce the WINCE sysname, and add 4152109998Smarkm INSTALL.WCE and appropriate conditionals to make it build. 4153109998Smarkm [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 4154109998Smarkm 4155109998Smarkm *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and 4156109998Smarkm cygssl-x.y.z.dll, where x, y and z are the major, minor and 4157109998Smarkm edit numbers of the version. 4158109998Smarkm [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 4159109998Smarkm 4160109998Smarkm *) Introduce safe string copy and catenation functions 4161109998Smarkm (BUF_strlcpy() and BUF_strlcat()). 4162109998Smarkm [Ben Laurie (CHATS) and Richard Levitte] 4163109998Smarkm 4164109998Smarkm *) Avoid using fixed-size buffers for one-line DNs. 4165109998Smarkm [Ben Laurie (CHATS)] 4166109998Smarkm 4167109998Smarkm *) Add BUF_MEM_grow_clean() to avoid information leakage when 4168109998Smarkm resizing buffers containing secrets, and use where appropriate. 4169109998Smarkm [Ben Laurie (CHATS)] 4170109998Smarkm 4171109998Smarkm *) Avoid using fixed size buffers for configuration file location. 4172109998Smarkm [Ben Laurie (CHATS)] 4173109998Smarkm 4174109998Smarkm *) Avoid filename truncation for various CA files. 4175109998Smarkm [Ben Laurie (CHATS)] 4176109998Smarkm 4177109998Smarkm *) Use sizeof in preference to magic numbers. 4178109998Smarkm [Ben Laurie (CHATS)] 4179109998Smarkm 4180109998Smarkm *) Avoid filename truncation in cert requests. 4181109998Smarkm [Ben Laurie (CHATS)] 4182109998Smarkm 4183109998Smarkm *) Add assertions to check for (supposedly impossible) buffer 4184109998Smarkm overflows. 4185109998Smarkm [Ben Laurie (CHATS)] 4186109998Smarkm 4187109998Smarkm *) Don't cache truncated DNS entries in the local cache (this could 4188109998Smarkm potentially lead to a spoofing attack). 4189109998Smarkm [Ben Laurie (CHATS)] 4190109998Smarkm 4191109998Smarkm *) Fix various buffers to be large enough for hex/decimal 4192109998Smarkm representations in a platform independent manner. 4193109998Smarkm [Ben Laurie (CHATS)] 4194109998Smarkm 4195109998Smarkm *) Add CRYPTO_realloc_clean() to avoid information leakage when 4196109998Smarkm resizing buffers containing secrets, and use where appropriate. 4197109998Smarkm [Ben Laurie (CHATS)] 4198109998Smarkm 4199109998Smarkm *) Add BIO_indent() to avoid much slightly worrying code to do 4200109998Smarkm indents. 4201109998Smarkm [Ben Laurie (CHATS)] 4202109998Smarkm 4203109998Smarkm *) Convert sprintf()/BIO_puts() to BIO_printf(). 4204109998Smarkm [Ben Laurie (CHATS)] 4205109998Smarkm 4206109998Smarkm *) buffer_gets() could terminate with the buffer only half 4207109998Smarkm full. Fixed. 4208109998Smarkm [Ben Laurie (CHATS)] 4209109998Smarkm 4210109998Smarkm *) Add assertions to prevent user-supplied crypto functions from 4211109998Smarkm overflowing internal buffers by having large block sizes, etc. 4212109998Smarkm [Ben Laurie (CHATS)] 4213109998Smarkm 4214109998Smarkm *) New OPENSSL_assert() macro (similar to assert(), but enabled 4215109998Smarkm unconditionally). 4216109998Smarkm [Ben Laurie (CHATS)] 4217109998Smarkm 4218109998Smarkm *) Eliminate unused copy of key in RC4. 4219109998Smarkm [Ben Laurie (CHATS)] 4220109998Smarkm 4221109998Smarkm *) Eliminate unused and incorrectly sized buffers for IV in pem.h. 4222109998Smarkm [Ben Laurie (CHATS)] 4223109998Smarkm 4224109998Smarkm *) Fix off-by-one error in EGD path. 4225109998Smarkm [Ben Laurie (CHATS)] 4226109998Smarkm 4227109998Smarkm *) If RANDFILE path is too long, ignore instead of truncating. 4228109998Smarkm [Ben Laurie (CHATS)] 4229109998Smarkm 4230109998Smarkm *) Eliminate unused and incorrectly sized X.509 structure 4231109998Smarkm CBCParameter. 4232109998Smarkm [Ben Laurie (CHATS)] 4233109998Smarkm 4234109998Smarkm *) Eliminate unused and dangerous function knumber(). 4235109998Smarkm [Ben Laurie (CHATS)] 4236109998Smarkm 4237109998Smarkm *) Eliminate unused and dangerous structure, KSSL_ERR. 4238109998Smarkm [Ben Laurie (CHATS)] 4239109998Smarkm 4240109998Smarkm *) Protect against overlong session ID context length in an encoded 4241109998Smarkm session object. Since these are local, this does not appear to be 4242109998Smarkm exploitable. 4243109998Smarkm [Ben Laurie (CHATS)] 4244109998Smarkm 4245109998Smarkm *) Change from security patch (see 0.9.6e below) that did not affect 4246109998Smarkm the 0.9.6 release series: 4247109998Smarkm 4248109998Smarkm Remote buffer overflow in SSL3 protocol - an attacker could 4249109998Smarkm supply an oversized master key in Kerberos-enabled versions. 4250160814Ssimon (CVE-2002-0657) 4251109998Smarkm [Ben Laurie (CHATS)] 4252109998Smarkm 4253109998Smarkm *) Change the SSL kerb5 codes to match RFC 2712. 4254109998Smarkm [Richard Levitte] 4255109998Smarkm 4256109998Smarkm *) Make -nameopt work fully for req and add -reqopt switch. 4257109998Smarkm [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson] 4258109998Smarkm 4259109998Smarkm *) The "block size" for block ciphers in CFB and OFB mode should be 1. 4260109998Smarkm [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>] 4261109998Smarkm 4262109998Smarkm *) Make sure tests can be performed even if the corresponding algorithms 4263109998Smarkm have been removed entirely. This was also the last step to make 4264109998Smarkm OpenSSL compilable with DJGPP under all reasonable conditions. 4265109998Smarkm [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>] 4266109998Smarkm 4267109998Smarkm *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT 4268109998Smarkm to allow version independent disabling of normally unselected ciphers, 4269109998Smarkm which may be activated as a side-effect of selecting a single cipher. 4270109998Smarkm 4271109998Smarkm (E.g., cipher list string "RSA" enables ciphersuites that are left 4272109998Smarkm out of "ALL" because they do not provide symmetric encryption. 4273109998Smarkm "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) 4274109998Smarkm [Lutz Jaenicke, Bodo Moeller] 4275109998Smarkm 4276109998Smarkm *) Add appropriate support for separate platform-dependent build 4277109998Smarkm directories. The recommended way to make a platform-dependent 4278109998Smarkm build directory is the following (tested on Linux), maybe with 4279109998Smarkm some local tweaks: 4280109998Smarkm 4281109998Smarkm # Place yourself outside of the OpenSSL source tree. In 4282109998Smarkm # this example, the environment variable OPENSSL_SOURCE 4283109998Smarkm # is assumed to contain the absolute OpenSSL source directory. 4284109998Smarkm mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" 4285109998Smarkm cd objtree/"`uname -s`-`uname -r`-`uname -m`" 4286111147Snectar (cd $OPENSSL_SOURCE; find . -type f) | while read F; do 4287109998Smarkm mkdir -p `dirname $F` 4288109998Smarkm ln -s $OPENSSL_SOURCE/$F $F 4289109998Smarkm done 4290109998Smarkm 4291109998Smarkm To be absolutely sure not to disturb the source tree, a "make clean" 4292109998Smarkm is a good thing. If it isn't successfull, don't worry about it, 4293109998Smarkm it probably means the source directory is very clean. 4294109998Smarkm [Richard Levitte] 4295109998Smarkm 4296109998Smarkm *) Make sure any ENGINE control commands make local copies of string 4297109998Smarkm pointers passed to them whenever necessary. Otherwise it is possible 4298109998Smarkm the caller may have overwritten (or deallocated) the original string 4299109998Smarkm data when a later ENGINE operation tries to use the stored values. 4300109998Smarkm [G�tz Babin-Ebell <babinebell@trustcenter.de>] 4301109998Smarkm 4302109998Smarkm *) Improve diagnostics in file reading and command-line digests. 4303109998Smarkm [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>] 4304109998Smarkm 4305109998Smarkm *) Add AES modes CFB and OFB to the object database. Correct an 4306109998Smarkm error in AES-CFB decryption. 4307109998Smarkm [Richard Levitte] 4308109998Smarkm 4309109998Smarkm *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this 4310109998Smarkm allows existing EVP_CIPHER_CTX structures to be reused after 4311109998Smarkm calling EVP_*Final(). This behaviour is used by encryption 4312109998Smarkm BIOs and some applications. This has the side effect that 4313109998Smarkm applications must explicitly clean up cipher contexts with 4314109998Smarkm EVP_CIPHER_CTX_cleanup() or they will leak memory. 4315109998Smarkm [Steve Henson] 4316109998Smarkm 4317109998Smarkm *) Check the values of dna and dnb in bn_mul_recursive before calling 4318109998Smarkm bn_mul_comba (a non zero value means the a or b arrays do not contain 4319109998Smarkm n2 elements) and fallback to bn_mul_normal if either is not zero. 4320109998Smarkm [Steve Henson] 4321109998Smarkm 4322109998Smarkm *) Fix escaping of non-ASCII characters when using the -subj option 4323109998Smarkm of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) 4324109998Smarkm [Lutz Jaenicke] 4325109998Smarkm 4326109998Smarkm *) Make object definitions compliant to LDAP (RFC2256): SN is the short 4327109998Smarkm form for "surname", serialNumber has no short form. 4328109998Smarkm Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; 4329109998Smarkm therefore remove "mail" short name for "internet 7". 4330109998Smarkm The OID for unique identifiers in X509 certificates is 4331109998Smarkm x500UniqueIdentifier, not uniqueIdentifier. 4332109998Smarkm Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) 4333109998Smarkm [Lutz Jaenicke] 4334109998Smarkm 4335109998Smarkm *) Add an "init" command to the ENGINE config module and auto initialize 4336109998Smarkm ENGINEs. Without any "init" command the ENGINE will be initialized 4337109998Smarkm after all ctrl commands have been executed on it. If init=1 the 4338109998Smarkm ENGINE is initailized at that point (ctrls before that point are run 4339109998Smarkm on the uninitialized ENGINE and after on the initialized one). If 4340109998Smarkm init=0 then the ENGINE will not be iniatialized at all. 4341109998Smarkm [Steve Henson] 4342109998Smarkm 4343109998Smarkm *) Fix the 'app_verify_callback' interface so that the user-defined 4344109998Smarkm argument is actually passed to the callback: In the 4345109998Smarkm SSL_CTX_set_cert_verify_callback() prototype, the callback 4346109998Smarkm declaration has been changed from 4347109998Smarkm int (*cb)() 4348109998Smarkm into 4349109998Smarkm int (*cb)(X509_STORE_CTX *,void *); 4350109998Smarkm in ssl_verify_cert_chain (ssl/ssl_cert.c), the call 4351109998Smarkm i=s->ctx->app_verify_callback(&ctx) 4352109998Smarkm has been changed into 4353109998Smarkm i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). 4354109998Smarkm 4355109998Smarkm To update applications using SSL_CTX_set_cert_verify_callback(), 4356109998Smarkm a dummy argument can be added to their callback functions. 4357109998Smarkm [D. K. Smetters <smetters@parc.xerox.com>] 4358109998Smarkm 4359109998Smarkm *) Added the '4758cca' ENGINE to support IBM 4758 cards. 4360109998Smarkm [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe] 4361109998Smarkm 4362109998Smarkm *) Add and OPENSSL_LOAD_CONF define which will cause 4363109998Smarkm OpenSSL_add_all_algorithms() to load the openssl.cnf config file. 4364109998Smarkm This allows older applications to transparently support certain 4365109998Smarkm OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. 4366109998Smarkm Two new functions OPENSSL_add_all_algorithms_noconf() which will never 4367109998Smarkm load the config file and OPENSSL_add_all_algorithms_conf() which will 4368109998Smarkm always load it have also been added. 4369109998Smarkm [Steve Henson] 4370109998Smarkm 4371109998Smarkm *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. 4372109998Smarkm Adjust NIDs and EVP layer. 4373109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 4374109998Smarkm 4375109998Smarkm *) Config modules support in openssl utility. 4376109998Smarkm 4377109998Smarkm Most commands now load modules from the config file, 4378109998Smarkm though in a few (such as version) this isn't done 4379109998Smarkm because it couldn't be used for anything. 4380109998Smarkm 4381109998Smarkm In the case of ca and req the config file used is 4382109998Smarkm the same as the utility itself: that is the -config 4383109998Smarkm command line option can be used to specify an 4384109998Smarkm alternative file. 4385109998Smarkm [Steve Henson] 4386109998Smarkm 4387109998Smarkm *) Move default behaviour from OPENSSL_config(). If appname is NULL 4388109998Smarkm use "openssl_conf" if filename is NULL use default openssl config file. 4389109998Smarkm [Steve Henson] 4390109998Smarkm 4391109998Smarkm *) Add an argument to OPENSSL_config() to allow the use of an alternative 4392109998Smarkm config section name. Add a new flag to tolerate a missing config file 4393109998Smarkm and move code to CONF_modules_load_file(). 4394109998Smarkm [Steve Henson] 4395109998Smarkm 4396109998Smarkm *) Support for crypto accelerator cards from Accelerated Encryption 4397109998Smarkm Processing, www.aep.ie. (Use engine 'aep') 4398109998Smarkm The support was copied from 0.9.6c [engine] and adapted/corrected 4399109998Smarkm to work with the new engine framework. 4400109998Smarkm [AEP Inc. and Richard Levitte] 4401109998Smarkm 4402109998Smarkm *) Support for SureWare crypto accelerator cards from Baltimore 4403109998Smarkm Technologies. (Use engine 'sureware') 4404109998Smarkm The support was copied from 0.9.6c [engine] and adapted 4405109998Smarkm to work with the new engine framework. 4406109998Smarkm [Richard Levitte] 4407109998Smarkm 4408109998Smarkm *) Have the CHIL engine fork-safe (as defined by nCipher) and actually 4409109998Smarkm make the newer ENGINE framework commands for the CHIL engine work. 4410109998Smarkm [Toomas Kiisk <vix@cyber.ee> and Richard Levitte] 4411109998Smarkm 4412109998Smarkm *) Make it possible to produce shared libraries on ReliantUNIX. 4413109998Smarkm [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte] 4414109998Smarkm 4415109998Smarkm *) Add the configuration target debug-linux-ppro. 4416109998Smarkm Make 'openssl rsa' use the general key loading routines 4417109998Smarkm implemented in apps.c, and make those routines able to 4418109998Smarkm handle the key format FORMAT_NETSCAPE and the variant 4419109998Smarkm FORMAT_IISSGC. 4420109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 4421109998Smarkm 4422109998Smarkm *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 4423109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 4424109998Smarkm 4425109998Smarkm *) Add -keyform to rsautl, and document -engine. 4426109998Smarkm [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>] 4427109998Smarkm 4428109998Smarkm *) Change BIO_new_file (crypto/bio/bss_file.c) to use new 4429109998Smarkm BIO_R_NO_SUCH_FILE error code rather than the generic 4430109998Smarkm ERR_R_SYS_LIB error code if fopen() fails with ENOENT. 4431109998Smarkm [Ben Laurie] 4432109998Smarkm 4433109998Smarkm *) Add new functions 4434109998Smarkm ERR_peek_last_error 4435109998Smarkm ERR_peek_last_error_line 4436109998Smarkm ERR_peek_last_error_line_data. 4437109998Smarkm These are similar to 4438109998Smarkm ERR_peek_error 4439109998Smarkm ERR_peek_error_line 4440109998Smarkm ERR_peek_error_line_data, 4441109998Smarkm but report on the latest error recorded rather than the first one 4442109998Smarkm still in the error queue. 4443109998Smarkm [Ben Laurie, Bodo Moeller] 4444109998Smarkm 4445109998Smarkm *) default_algorithms option in ENGINE config module. This allows things 4446109998Smarkm like: 4447109998Smarkm default_algorithms = ALL 4448109998Smarkm default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS 4449109998Smarkm [Steve Henson] 4450109998Smarkm 4451109998Smarkm *) Prelminary ENGINE config module. 4452109998Smarkm [Steve Henson] 4453109998Smarkm 4454109998Smarkm *) New experimental application configuration code. 4455109998Smarkm [Steve Henson] 4456109998Smarkm 4457109998Smarkm *) Change the AES code to follow the same name structure as all other 4458109998Smarkm symmetric ciphers, and behave the same way. Move everything to 4459109998Smarkm the directory crypto/aes, thereby obsoleting crypto/rijndael. 4460109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 4461109998Smarkm 4462109998Smarkm *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. 4463109998Smarkm [Ben Laurie and Theo de Raadt] 4464109998Smarkm 4465109998Smarkm *) Add option to output public keys in req command. 4466109998Smarkm [Massimiliano Pala madwolf@openca.org] 4467109998Smarkm 4468109998Smarkm *) Use wNAFs in EC_POINTs_mul() for improved efficiency 4469109998Smarkm (up to about 10% better than before for P-192 and P-224). 4470109998Smarkm [Bodo Moeller] 4471109998Smarkm 4472109998Smarkm *) New functions/macros 4473109998Smarkm 4474109998Smarkm SSL_CTX_set_msg_callback(ctx, cb) 4475109998Smarkm SSL_CTX_set_msg_callback_arg(ctx, arg) 4476109998Smarkm SSL_set_msg_callback(ssl, cb) 4477109998Smarkm SSL_set_msg_callback_arg(ssl, arg) 4478109998Smarkm 4479109998Smarkm to request calling a callback function 4480109998Smarkm 4481109998Smarkm void cb(int write_p, int version, int content_type, 4482109998Smarkm const void *buf, size_t len, SSL *ssl, void *arg) 4483109998Smarkm 4484109998Smarkm whenever a protocol message has been completely received 4485109998Smarkm (write_p == 0) or sent (write_p == 1). Here 'version' is the 4486109998Smarkm protocol version according to which the SSL library interprets 4487109998Smarkm the current protocol message (SSL2_VERSION, SSL3_VERSION, or 4488109998Smarkm TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or 4489109998Smarkm the content type as defined in the SSL 3.0/TLS 1.0 protocol 4490109998Smarkm specification (change_cipher_spec(20), alert(21), handshake(22)). 4491109998Smarkm 'buf' and 'len' point to the actual message, 'ssl' to the 4492109998Smarkm SSL object, and 'arg' is the application-defined value set by 4493109998Smarkm SSL[_CTX]_set_msg_callback_arg(). 4494109998Smarkm 4495109998Smarkm 'openssl s_client' and 'openssl s_server' have new '-msg' options 4496109998Smarkm to enable a callback that displays all protocol messages. 4497109998Smarkm [Bodo Moeller] 4498109998Smarkm 4499109998Smarkm *) Change the shared library support so shared libraries are built as 4500109998Smarkm soon as the corresponding static library is finished, and thereby get 4501109998Smarkm openssl and the test programs linked against the shared library. 4502109998Smarkm This still only happens when the keyword "shard" has been given to 4503109998Smarkm the configuration scripts. 4504109998Smarkm 4505109998Smarkm NOTE: shared library support is still an experimental thing, and 4506109998Smarkm backward binary compatibility is still not guaranteed. 4507109998Smarkm ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte] 4508109998Smarkm 4509109998Smarkm *) Add support for Subject Information Access extension. 4510109998Smarkm [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 4511109998Smarkm 4512109998Smarkm *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero 4513109998Smarkm additional bytes when new memory had to be allocated, not just 4514109998Smarkm when reusing an existing buffer. 4515109998Smarkm [Bodo Moeller] 4516109998Smarkm 4517109998Smarkm *) New command line and configuration option 'utf8' for the req command. 4518109998Smarkm This allows field values to be specified as UTF8 strings. 4519109998Smarkm [Steve Henson] 4520109998Smarkm 4521109998Smarkm *) Add -multi and -mr options to "openssl speed" - giving multiple parallel 4522109998Smarkm runs for the former and machine-readable output for the latter. 4523109998Smarkm [Ben Laurie] 4524109998Smarkm 4525109998Smarkm *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion 4526109998Smarkm of the e-mail address in the DN (i.e., it will go into a certificate 4527109998Smarkm extension only). The new configuration file option 'email_in_dn = no' 4528109998Smarkm has the same effect. 4529109998Smarkm [Massimiliano Pala madwolf@openca.org] 4530109998Smarkm 4531109998Smarkm *) Change all functions with names starting with des_ to be starting 4532109998Smarkm with DES_ instead. Add wrappers that are compatible with libdes, 4533109998Smarkm but are named _ossl_old_des_*. Finally, add macros that map the 4534109998Smarkm des_* symbols to the corresponding _ossl_old_des_* if libdes 4535109998Smarkm compatibility is desired. If OpenSSL 0.9.6c compatibility is 4536109998Smarkm desired, the des_* symbols will be mapped to DES_*, with one 4537109998Smarkm exception. 4538109998Smarkm 4539109998Smarkm Since we provide two compatibility mappings, the user needs to 4540109998Smarkm define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes 4541109998Smarkm compatibility is desired. The default (i.e., when that macro 4542109998Smarkm isn't defined) is OpenSSL 0.9.6c compatibility. 4543109998Smarkm 4544109998Smarkm There are also macros that enable and disable the support of old 4545109998Smarkm des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT 4546109998Smarkm and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those 4547109998Smarkm are defined, the default will apply: to support the old des routines. 4548109998Smarkm 4549109998Smarkm In either case, one must include openssl/des.h to get the correct 4550109998Smarkm definitions. Do not try to just include openssl/des_old.h, that 4551109998Smarkm won't work. 4552109998Smarkm 4553109998Smarkm NOTE: This is a major break of an old API into a new one. Software 4554109998Smarkm authors are encouraged to switch to the DES_ style functions. Some 4555109998Smarkm time in the future, des_old.h and the libdes compatibility functions 4556109998Smarkm will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the 4557109998Smarkm default), and then completely removed. 4558109998Smarkm [Richard Levitte] 4559109998Smarkm 4560109998Smarkm *) Test for certificates which contain unsupported critical extensions. 4561109998Smarkm If such a certificate is found during a verify operation it is 4562109998Smarkm rejected by default: this behaviour can be overridden by either 4563109998Smarkm handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or 4564109998Smarkm by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function 4565109998Smarkm X509_supported_extension() has also been added which returns 1 if a 4566109998Smarkm particular extension is supported. 4567109998Smarkm [Steve Henson] 4568109998Smarkm 4569109998Smarkm *) Modify the behaviour of EVP cipher functions in similar way to digests 4570109998Smarkm to retain compatibility with existing code. 4571109998Smarkm [Steve Henson] 4572109998Smarkm 4573109998Smarkm *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain 4574109998Smarkm compatibility with existing code. In particular the 'ctx' parameter does 4575109998Smarkm not have to be to be initialized before the call to EVP_DigestInit() and 4576109998Smarkm it is tidied up after a call to EVP_DigestFinal(). New function 4577109998Smarkm EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function 4578109998Smarkm EVP_MD_CTX_copy() changed to not require the destination to be 4579109998Smarkm initialized valid and new function EVP_MD_CTX_copy_ex() added which 4580109998Smarkm requires the destination to be valid. 4581109998Smarkm 4582109998Smarkm Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), 4583109998Smarkm EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). 4584109998Smarkm [Steve Henson] 4585109998Smarkm 4586109998Smarkm *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it 4587109998Smarkm so that complete 'Handshake' protocol structures are kept in memory 4588109998Smarkm instead of overwriting 'msg_type' and 'length' with 'body' data. 4589109998Smarkm [Bodo Moeller] 4590109998Smarkm 4591109998Smarkm *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. 4592109998Smarkm [Massimo Santin via Richard Levitte] 4593109998Smarkm 4594109998Smarkm *) Major restructuring to the underlying ENGINE code. This includes 4595109998Smarkm reduction of linker bloat, separation of pure "ENGINE" manipulation 4596109998Smarkm (initialisation, etc) from functionality dealing with implementations 4597109998Smarkm of specific crypto iterfaces. This change also introduces integrated 4598109998Smarkm support for symmetric ciphers and digest implementations - so ENGINEs 4599109998Smarkm can now accelerate these by providing EVP_CIPHER and EVP_MD 4600109998Smarkm implementations of their own. This is detailed in crypto/engine/README 4601109998Smarkm as it couldn't be adequately described here. However, there are a few 4602109998Smarkm API changes worth noting - some RSA, DSA, DH, and RAND functions that 4603109998Smarkm were changed in the original introduction of ENGINE code have now 4604109998Smarkm reverted back - the hooking from this code to ENGINE is now a good 4605109998Smarkm deal more passive and at run-time, operations deal directly with 4606109998Smarkm RSA_METHODs, DSA_METHODs (etc) as they did before, rather than 4607109998Smarkm dereferencing through an ENGINE pointer any more. Also, the ENGINE 4608109998Smarkm functions dealing with BN_MOD_EXP[_CRT] handlers have been removed - 4609109998Smarkm they were not being used by the framework as there is no concept of a 4610109998Smarkm BIGNUM_METHOD and they could not be generalised to the new 4611109998Smarkm 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, 4612109998Smarkm ENGINE_cpy() has been removed as it cannot be consistently defined in 4613109998Smarkm the new code. 4614109998Smarkm [Geoff Thorpe] 4615109998Smarkm 4616109998Smarkm *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. 4617109998Smarkm [Steve Henson] 4618109998Smarkm 4619109998Smarkm *) Change mkdef.pl to sort symbols that get the same entry number, 4620109998Smarkm and make sure the automatically generated functions ERR_load_* 4621109998Smarkm become part of libeay.num as well. 4622109998Smarkm [Richard Levitte] 4623109998Smarkm 4624109998Smarkm *) New function SSL_renegotiate_pending(). This returns true once 4625109998Smarkm renegotiation has been requested (either SSL_renegotiate() call 4626109998Smarkm or HelloRequest/ClientHello receveived from the peer) and becomes 4627109998Smarkm false once a handshake has been completed. 4628109998Smarkm (For servers, SSL_renegotiate() followed by SSL_do_handshake() 4629109998Smarkm sends a HelloRequest, but does not ensure that a handshake takes 4630109998Smarkm place. SSL_renegotiate_pending() is useful for checking if the 4631109998Smarkm client has followed the request.) 4632109998Smarkm [Bodo Moeller] 4633109998Smarkm 4634109998Smarkm *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. 4635109998Smarkm By default, clients may request session resumption even during 4636109998Smarkm renegotiation (if session ID contexts permit); with this option, 4637109998Smarkm session resumption is possible only in the first handshake. 4638109998Smarkm 4639109998Smarkm SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes 4640109998Smarkm more bits available for options that should not be part of 4641109998Smarkm SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). 4642109998Smarkm [Bodo Moeller] 4643109998Smarkm 4644109998Smarkm *) Add some demos for certificate and certificate request creation. 4645109998Smarkm [Steve Henson] 4646109998Smarkm 4647109998Smarkm *) Make maximum certificate chain size accepted from the peer application 4648109998Smarkm settable (SSL*_get/set_max_cert_list()), as proposed by 4649109998Smarkm "Douglas E. Engert" <deengert@anl.gov>. 4650109998Smarkm [Lutz Jaenicke] 4651109998Smarkm 4652109998Smarkm *) Add support for shared libraries for Unixware-7 4653109998Smarkm (Boyd Lynn Gerber <gerberb@zenez.com>). 4654109998Smarkm [Lutz Jaenicke] 4655109998Smarkm 4656109998Smarkm *) Add a "destroy" handler to ENGINEs that allows structural cleanup to 4657109998Smarkm be done prior to destruction. Use this to unload error strings from 4658109998Smarkm ENGINEs that load their own error strings. NB: This adds two new API 4659109998Smarkm functions to "get" and "set" this destroy handler in an ENGINE. 4660109998Smarkm [Geoff Thorpe] 4661109998Smarkm 4662109998Smarkm *) Alter all existing ENGINE implementations (except "openssl" and 4663109998Smarkm "openbsd") to dynamically instantiate their own error strings. This 4664109998Smarkm makes them more flexible to be built both as statically-linked ENGINEs 4665109998Smarkm and self-contained shared-libraries loadable via the "dynamic" ENGINE. 4666109998Smarkm Also, add stub code to each that makes building them as self-contained 4667109998Smarkm shared-libraries easier (see README.ENGINE). 4668109998Smarkm [Geoff Thorpe] 4669109998Smarkm 4670109998Smarkm *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE 4671109998Smarkm implementations into applications that are completely implemented in 4672109998Smarkm self-contained shared-libraries. The "dynamic" ENGINE exposes control 4673109998Smarkm commands that can be used to configure what shared-library to load and 4674109998Smarkm to control aspects of the way it is handled. Also, made an update to 4675109998Smarkm the README.ENGINE file that brings its information up-to-date and 4676109998Smarkm provides some information and instructions on the "dynamic" ENGINE 4677109998Smarkm (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). 4678109998Smarkm [Geoff Thorpe] 4679109998Smarkm 4680109998Smarkm *) Make it possible to unload ranges of ERR strings with a new 4681109998Smarkm "ERR_unload_strings" function. 4682109998Smarkm [Geoff Thorpe] 4683109998Smarkm 4684109998Smarkm *) Add a copy() function to EVP_MD. 4685109998Smarkm [Ben Laurie] 4686109998Smarkm 4687109998Smarkm *) Make EVP_MD routines take a context pointer instead of just the 4688109998Smarkm md_data void pointer. 4689109998Smarkm [Ben Laurie] 4690109998Smarkm 4691109998Smarkm *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates 4692109998Smarkm that the digest can only process a single chunk of data 4693109998Smarkm (typically because it is provided by a piece of 4694109998Smarkm hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application 4695109998Smarkm is only going to provide a single chunk of data, and hence the 4696109998Smarkm framework needn't accumulate the data for oneshot drivers. 4697109998Smarkm [Ben Laurie] 4698109998Smarkm 4699109998Smarkm *) As with "ERR", make it possible to replace the underlying "ex_data" 4700109998Smarkm functions. This change also alters the storage and management of global 4701109998Smarkm ex_data state - it's now all inside ex_data.c and all "class" code (eg. 4702109998Smarkm RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class 4703109998Smarkm index counters. The API functions that use this state have been changed 4704109998Smarkm to take a "class_index" rather than pointers to the class's local STACK 4705109998Smarkm and counter, and there is now an API function to dynamically create new 4706109998Smarkm classes. This centralisation allows us to (a) plug a lot of the 4707109998Smarkm thread-safety problems that existed, and (b) makes it possible to clean 4708109998Smarkm up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) 4709109998Smarkm such data would previously have always leaked in application code and 4710109998Smarkm workarounds were in place to make the memory debugging turn a blind eye 4711109998Smarkm to it. Application code that doesn't use this new function will still 4712109998Smarkm leak as before, but their memory debugging output will announce it now 4713109998Smarkm rather than letting it slide. 4714109998Smarkm 4715109998Smarkm Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change 4716109998Smarkm induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now 4717109998Smarkm has a return value to indicate success or failure. 4718109998Smarkm [Geoff Thorpe] 4719109998Smarkm 4720109998Smarkm *) Make it possible to replace the underlying "ERR" functions such that the 4721109998Smarkm global state (2 LHASH tables and 2 locks) is only used by the "default" 4722109998Smarkm implementation. This change also adds two functions to "get" and "set" 4723109998Smarkm the implementation prior to it being automatically set the first time 4724109998Smarkm any other ERR function takes place. Ie. an application can call "get", 4725109998Smarkm pass the return value to a module it has just loaded, and that module 4726109998Smarkm can call its own "set" function using that value. This means the 4727109998Smarkm module's "ERR" operations will use (and modify) the error state in the 4728109998Smarkm application and not in its own statically linked copy of OpenSSL code. 4729109998Smarkm [Geoff Thorpe] 4730109998Smarkm 4731109998Smarkm *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment 4732109998Smarkm reference counts. This performs normal REF_PRINT/REF_CHECK macros on 4733109998Smarkm the operation, and provides a more encapsulated way for external code 4734109998Smarkm (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code 4735109998Smarkm to use these functions rather than manually incrementing the counts. 4736109998Smarkm 4737109998Smarkm Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". 4738109998Smarkm [Geoff Thorpe] 4739109998Smarkm 4740109998Smarkm *) Add EVP test program. 4741109998Smarkm [Ben Laurie] 4742109998Smarkm 4743109998Smarkm *) Add symmetric cipher support to ENGINE. Expect the API to change! 4744109998Smarkm [Ben Laurie] 4745109998Smarkm 4746109998Smarkm *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() 4747109998Smarkm X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), 4748109998Smarkm X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). 4749109998Smarkm These allow a CRL to be built without having to access X509_CRL fields 4750109998Smarkm directly. Modify 'ca' application to use new functions. 4751109998Smarkm [Steve Henson] 4752109998Smarkm 4753109998Smarkm *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended 4754109998Smarkm bug workarounds. Rollback attack detection is a security feature. 4755109998Smarkm The problem will only arise on OpenSSL servers when TLSv1 is not 4756109998Smarkm available (sslv3_server_method() or SSL_OP_NO_TLSv1). 4757109998Smarkm Software authors not wanting to support TLSv1 will have special reasons 4758109998Smarkm for their choice and can explicitly enable this option. 4759109998Smarkm [Bodo Moeller, Lutz Jaenicke] 4760109998Smarkm 4761109998Smarkm *) Rationalise EVP so it can be extended: don't include a union of 4762109998Smarkm cipher/digest structures, add init/cleanup functions for EVP_MD_CTX 4763109998Smarkm (similar to those existing for EVP_CIPHER_CTX). 4764109998Smarkm Usage example: 4765109998Smarkm 4766109998Smarkm EVP_MD_CTX md; 4767109998Smarkm 4768109998Smarkm EVP_MD_CTX_init(&md); /* new function call */ 4769109998Smarkm EVP_DigestInit(&md, EVP_sha1()); 4770109998Smarkm EVP_DigestUpdate(&md, in, len); 4771109998Smarkm EVP_DigestFinal(&md, out, NULL); 4772109998Smarkm EVP_MD_CTX_cleanup(&md); /* new function call */ 4773109998Smarkm 4774109998Smarkm [Ben Laurie] 4775109998Smarkm 4776109998Smarkm *) Make DES key schedule conform to the usual scheme, as well as 4777109998Smarkm correcting its structure. This means that calls to DES functions 4778109998Smarkm now have to pass a pointer to a des_key_schedule instead of a 4779109998Smarkm plain des_key_schedule (which was actually always a pointer 4780109998Smarkm anyway): E.g., 4781109998Smarkm 4782109998Smarkm des_key_schedule ks; 4783109998Smarkm 4784109998Smarkm des_set_key_checked(..., &ks); 4785109998Smarkm des_ncbc_encrypt(..., &ks, ...); 4786109998Smarkm 4787109998Smarkm (Note that a later change renames 'des_...' into 'DES_...'.) 4788109998Smarkm [Ben Laurie] 4789109998Smarkm 4790109998Smarkm *) Initial reduction of linker bloat: the use of some functions, such as 4791109998Smarkm PEM causes large amounts of unused functions to be linked in due to 4792109998Smarkm poor organisation. For example pem_all.c contains every PEM function 4793109998Smarkm which has a knock on effect of linking in large amounts of (unused) 4794109998Smarkm ASN1 code. Grouping together similar functions and splitting unrelated 4795109998Smarkm functions prevents this. 4796109998Smarkm [Steve Henson] 4797109998Smarkm 4798109998Smarkm *) Cleanup of EVP macros. 4799109998Smarkm [Ben Laurie] 4800109998Smarkm 4801109998Smarkm *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the 4802109998Smarkm correct _ecb suffix. 4803109998Smarkm [Ben Laurie] 4804109998Smarkm 4805109998Smarkm *) Add initial OCSP responder support to ocsp application. The 4806109998Smarkm revocation information is handled using the text based index 4807109998Smarkm use by the ca application. The responder can either handle 4808109998Smarkm requests generated internally, supplied in files (for example 4809109998Smarkm via a CGI script) or using an internal minimal server. 4810109998Smarkm [Steve Henson] 4811109998Smarkm 4812109998Smarkm *) Add configuration choices to get zlib compression for TLS. 4813109998Smarkm [Richard Levitte] 4814109998Smarkm 4815109998Smarkm *) Changes to Kerberos SSL for RFC 2712 compliance: 4816109998Smarkm 1. Implemented real KerberosWrapper, instead of just using 4817109998Smarkm KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] 4818109998Smarkm 2. Implemented optional authenticator field of KerberosWrapper. 4819109998Smarkm 4820109998Smarkm Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, 4821109998Smarkm and authenticator structs; see crypto/krb5/. 4822109998Smarkm 4823109998Smarkm Generalized Kerberos calls to support multiple Kerberos libraries. 4824109998Smarkm [Vern Staats <staatsvr@asc.hpc.mil>, 4825109998Smarkm Jeffrey Altman <jaltman@columbia.edu> 4826109998Smarkm via Richard Levitte] 4827109998Smarkm 4828109998Smarkm *) Cause 'openssl speed' to use fully hard-coded DSA keys as it 4829109998Smarkm already does with RSA. testdsa.h now has 'priv_key/pub_key' 4830109998Smarkm values for each of the key sizes rather than having just 4831109998Smarkm parameters (and 'speed' generating keys each time). 4832109998Smarkm [Geoff Thorpe] 4833109998Smarkm 4834109998Smarkm *) Speed up EVP routines. 4835109998Smarkm Before: 4836109998Smarkmencrypt 4837109998Smarkmtype 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 4838109998Smarkmdes-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k 4839109998Smarkmdes-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k 4840109998Smarkmdes-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k 4841109998Smarkmdecrypt 4842109998Smarkmdes-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k 4843109998Smarkmdes-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k 4844109998Smarkmdes-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k 4845109998Smarkm After: 4846109998Smarkmencrypt 4847109998Smarkmdes-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k 4848109998Smarkmdecrypt 4849109998Smarkmdes-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k 4850109998Smarkm [Ben Laurie] 4851109998Smarkm 4852109998Smarkm *) Added the OS2-EMX target. 4853109998Smarkm ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte] 4854109998Smarkm 4855109998Smarkm *) Rewrite apps to use NCONF routines instead of the old CONF. New functions 4856109998Smarkm to support NCONF routines in extension code. New function CONF_set_nconf() 4857109998Smarkm to allow functions which take an NCONF to also handle the old LHASH 4858109998Smarkm structure: this means that the old CONF compatible routines can be 4859109998Smarkm retained (in particular wrt extensions) without having to duplicate the 4860109998Smarkm code. New function X509V3_add_ext_nconf_sk to add extensions to a stack. 4861109998Smarkm [Steve Henson] 4862109998Smarkm 4863109998Smarkm *) Enhance the general user interface with mechanisms for inner control 4864109998Smarkm and with possibilities to have yes/no kind of prompts. 4865109998Smarkm [Richard Levitte] 4866109998Smarkm 4867109998Smarkm *) Change all calls to low level digest routines in the library and 4868109998Smarkm applications to use EVP. Add missing calls to HMAC_cleanup() and 4869109998Smarkm don't assume HMAC_CTX can be copied using memcpy(). 4870109998Smarkm [Verdon Walker <VWalker@novell.com>, Steve Henson] 4871109998Smarkm 4872109998Smarkm *) Add the possibility to control engines through control names but with 4873109998Smarkm arbitrary arguments instead of just a string. 4874109998Smarkm Change the key loaders to take a UI_METHOD instead of a callback 4875109998Smarkm function pointer. NOTE: this breaks binary compatibility with earlier 4876109998Smarkm versions of OpenSSL [engine]. 4877109998Smarkm Adapt the nCipher code for these new conditions and add a card insertion 4878109998Smarkm callback. 4879109998Smarkm [Richard Levitte] 4880109998Smarkm 4881109998Smarkm *) Enhance the general user interface with mechanisms to better support 4882109998Smarkm dialog box interfaces, application-defined prompts, the possibility 4883109998Smarkm to use defaults (for example default passwords from somewhere else) 4884109998Smarkm and interrupts/cancellations. 4885109998Smarkm [Richard Levitte] 4886109998Smarkm 4887109998Smarkm *) Tidy up PKCS#12 attribute handling. Add support for the CSP name 4888109998Smarkm attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. 4889109998Smarkm [Steve Henson] 4890109998Smarkm 4891109998Smarkm *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also 4892109998Smarkm tidy up some unnecessarily weird code in 'sk_new()'). 4893109998Smarkm [Geoff, reported by Diego Tartara <dtartara@novamens.com>] 4894109998Smarkm 4895109998Smarkm *) Change the key loading routines for ENGINEs to use the same kind 4896109998Smarkm callback (pem_password_cb) as all other routines that need this 4897109998Smarkm kind of callback. 4898109998Smarkm [Richard Levitte] 4899109998Smarkm 4900109998Smarkm *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with 4901109998Smarkm 256 bit (=32 byte) keys. Of course seeding with more entropy bytes 4902109998Smarkm than this minimum value is recommended. 4903109998Smarkm [Lutz Jaenicke] 4904109998Smarkm 4905109998Smarkm *) New random seeder for OpenVMS, using the system process statistics 4906109998Smarkm that are easily reachable. 4907109998Smarkm [Richard Levitte] 4908109998Smarkm 4909109998Smarkm *) Windows apparently can't transparently handle global 4910109998Smarkm variables defined in DLLs. Initialisations such as: 4911109998Smarkm 4912109998Smarkm const ASN1_ITEM *it = &ASN1_INTEGER_it; 4913109998Smarkm 4914109998Smarkm wont compile. This is used by the any applications that need to 4915109998Smarkm declare their own ASN1 modules. This was fixed by adding the option 4916109998Smarkm EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly 4917109998Smarkm needed for static libraries under Win32. 4918109998Smarkm [Steve Henson] 4919109998Smarkm 4920109998Smarkm *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle 4921109998Smarkm setting of purpose and trust fields. New X509_STORE trust and 4922109998Smarkm purpose functions and tidy up setting in other SSL functions. 4923109998Smarkm [Steve Henson] 4924109998Smarkm 4925109998Smarkm *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE 4926109998Smarkm structure. These are inherited by X509_STORE_CTX when it is 4927109998Smarkm initialised. This allows various defaults to be set in the 4928109998Smarkm X509_STORE structure (such as flags for CRL checking and custom 4929109998Smarkm purpose or trust settings) for functions which only use X509_STORE_CTX 4930109998Smarkm internally such as S/MIME. 4931109998Smarkm 4932109998Smarkm Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and 4933109998Smarkm trust settings if they are not set in X509_STORE. This allows X509_STORE 4934109998Smarkm purposes and trust (in S/MIME for example) to override any set by default. 4935109998Smarkm 4936109998Smarkm Add command line options for CRL checking to smime, s_client and s_server 4937109998Smarkm applications. 4938109998Smarkm [Steve Henson] 4939109998Smarkm 4940109998Smarkm *) Initial CRL based revocation checking. If the CRL checking flag(s) 4941109998Smarkm are set then the CRL is looked up in the X509_STORE structure and 4942109998Smarkm its validity and signature checked, then if the certificate is found 4943109998Smarkm in the CRL the verify fails with a revoked error. 4944109998Smarkm 4945109998Smarkm Various new CRL related callbacks added to X509_STORE_CTX structure. 4946109998Smarkm 4947109998Smarkm Command line options added to 'verify' application to support this. 4948109998Smarkm 4949109998Smarkm This needs some additional work, such as being able to handle multiple 4950109998Smarkm CRLs with different times, extension based lookup (rather than just 4951109998Smarkm by subject name) and ultimately more complete V2 CRL extension 4952109998Smarkm handling. 4953109998Smarkm [Steve Henson] 4954109998Smarkm 4955109998Smarkm *) Add a general user interface API (crypto/ui/). This is designed 4956109998Smarkm to replace things like des_read_password and friends (backward 4957109998Smarkm compatibility functions using this new API are provided). 4958109998Smarkm The purpose is to remove prompting functions from the DES code 4959109998Smarkm section as well as provide for prompting through dialog boxes in 4960109998Smarkm a window system and the like. 4961109998Smarkm [Richard Levitte] 4962109998Smarkm 4963109998Smarkm *) Add "ex_data" support to ENGINE so implementations can add state at a 4964109998Smarkm per-structure level rather than having to store it globally. 4965109998Smarkm [Geoff] 4966109998Smarkm 4967109998Smarkm *) Make it possible for ENGINE structures to be copied when retrieved by 4968109998Smarkm ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. 4969109998Smarkm This causes the "original" ENGINE structure to act like a template, 4970109998Smarkm analogous to the RSA vs. RSA_METHOD type of separation. Because of this 4971109998Smarkm operational state can be localised to each ENGINE structure, despite the 4972109998Smarkm fact they all share the same "methods". New ENGINE structures returned in 4973109998Smarkm this case have no functional references and the return value is the single 4974109998Smarkm structural reference. This matches the single structural reference returned 4975109998Smarkm by ENGINE_by_id() normally, when it is incremented on the pre-existing 4976109998Smarkm ENGINE structure. 4977109998Smarkm [Geoff] 4978109998Smarkm 4979109998Smarkm *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this 4980109998Smarkm needs to match any other type at all we need to manually clear the 4981109998Smarkm tag cache. 4982109998Smarkm [Steve Henson] 4983109998Smarkm 4984109998Smarkm *) Changes to the "openssl engine" utility to include; 4985109998Smarkm - verbosity levels ('-v', '-vv', and '-vvv') that provide information 4986109998Smarkm about an ENGINE's available control commands. 4987109998Smarkm - executing control commands from command line arguments using the 4988109998Smarkm '-pre' and '-post' switches. '-post' is only used if '-t' is 4989109998Smarkm specified and the ENGINE is successfully initialised. The syntax for 4990109998Smarkm the individual commands are colon-separated, for example; 4991109998Smarkm openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so 4992109998Smarkm [Geoff] 4993109998Smarkm 4994109998Smarkm *) New dynamic control command support for ENGINEs. ENGINEs can now 4995109998Smarkm declare their own commands (numbers), names (strings), descriptions, 4996109998Smarkm and input types for run-time discovery by calling applications. A 4997109998Smarkm subset of these commands are implicitly classed as "executable" 4998109998Smarkm depending on their input type, and only these can be invoked through 4999109998Smarkm the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this 5000109998Smarkm can be based on user input, config files, etc). The distinction is 5001109998Smarkm that "executable" commands cannot return anything other than a boolean 5002109998Smarkm result and can only support numeric or string input, whereas some 5003109998Smarkm discoverable commands may only be for direct use through 5004109998Smarkm ENGINE_ctrl(), eg. supporting the exchange of binary data, function 5005109998Smarkm pointers, or other custom uses. The "executable" commands are to 5006109998Smarkm support parameterisations of ENGINE behaviour that can be 5007109998Smarkm unambiguously defined by ENGINEs and used consistently across any 5008109998Smarkm OpenSSL-based application. Commands have been added to all the 5009109998Smarkm existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow 5010109998Smarkm control over shared-library paths without source code alterations. 5011109998Smarkm [Geoff] 5012109998Smarkm 5013109998Smarkm *) Changed all ENGINE implementations to dynamically allocate their 5014109998Smarkm ENGINEs rather than declaring them statically. Apart from this being 5015109998Smarkm necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, 5016109998Smarkm this also allows the implementations to compile without using the 5017109998Smarkm internal engine_int.h header. 5018109998Smarkm [Geoff] 5019109998Smarkm 5020109998Smarkm *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a 5021109998Smarkm 'const' value. Any code that should be able to modify a RAND_METHOD 5022109998Smarkm should already have non-const pointers to it (ie. they should only 5023109998Smarkm modify their own ones). 5024109998Smarkm [Geoff] 5025109998Smarkm 5026109998Smarkm *) Made a variety of little tweaks to the ENGINE code. 5027109998Smarkm - "atalla" and "ubsec" string definitions were moved from header files 5028109998Smarkm to C code. "nuron" string definitions were placed in variables 5029109998Smarkm rather than hard-coded - allowing parameterisation of these values 5030109998Smarkm later on via ctrl() commands. 5031109998Smarkm - Removed unused "#if 0"'d code. 5032109998Smarkm - Fixed engine list iteration code so it uses ENGINE_free() to release 5033109998Smarkm structural references. 5034109998Smarkm - Constified the RAND_METHOD element of ENGINE structures. 5035109998Smarkm - Constified various get/set functions as appropriate and added 5036109998Smarkm missing functions (including a catch-all ENGINE_cpy that duplicates 5037109998Smarkm all ENGINE values onto a new ENGINE except reference counts/state). 5038109998Smarkm - Removed NULL parameter checks in get/set functions. Setting a method 5039109998Smarkm or function to NULL is a way of cancelling out a previously set 5040109998Smarkm value. Passing a NULL ENGINE parameter is just plain stupid anyway 5041109998Smarkm and doesn't justify the extra error symbols and code. 5042109998Smarkm - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for 5043109998Smarkm flags from engine_int.h to engine.h. 5044109998Smarkm - Changed prototypes for ENGINE handler functions (init(), finish(), 5045109998Smarkm ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. 5046109998Smarkm [Geoff] 5047109998Smarkm 5048109998Smarkm *) Implement binary inversion algorithm for BN_mod_inverse in addition 5049109998Smarkm to the algorithm using long division. The binary algorithm can be 5050109998Smarkm used only if the modulus is odd. On 32-bit systems, it is faster 5051109998Smarkm only for relatively small moduli (roughly 20-30% for 128-bit moduli, 5052109998Smarkm roughly 5-15% for 256-bit moduli), so we use it only for moduli 5053109998Smarkm up to 450 bits. In 64-bit environments, the binary algorithm 5054109998Smarkm appears to be advantageous for much longer moduli; here we use it 5055109998Smarkm for moduli up to 2048 bits. 5056109998Smarkm [Bodo Moeller] 5057109998Smarkm 5058109998Smarkm *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code 5059109998Smarkm could not support the combine flag in choice fields. 5060109998Smarkm [Steve Henson] 5061109998Smarkm 5062109998Smarkm *) Add a 'copy_extensions' option to the 'ca' utility. This copies 5063109998Smarkm extensions from a certificate request to the certificate. 5064109998Smarkm [Steve Henson] 5065109998Smarkm 5066109998Smarkm *) Allow multiple 'certopt' and 'nameopt' options to be separated 5067109998Smarkm by commas. Add 'namopt' and 'certopt' options to the 'ca' config 5068109998Smarkm file: this allows the display of the certificate about to be 5069109998Smarkm signed to be customised, to allow certain fields to be included 5070109998Smarkm or excluded and extension details. The old system didn't display 5071109998Smarkm multicharacter strings properly, omitted fields not in the policy 5072109998Smarkm and couldn't display additional details such as extensions. 5073109998Smarkm [Steve Henson] 5074109998Smarkm 5075109998Smarkm *) Function EC_POINTs_mul for multiple scalar multiplication 5076109998Smarkm of an arbitrary number of elliptic curve points 5077109998Smarkm \sum scalars[i]*points[i], 5078109998Smarkm optionally including the generator defined for the EC_GROUP: 5079109998Smarkm scalar*generator + \sum scalars[i]*points[i]. 5080109998Smarkm 5081109998Smarkm EC_POINT_mul is a simple wrapper function for the typical case 5082109998Smarkm that the point list has just one item (besides the optional 5083109998Smarkm generator). 5084109998Smarkm [Bodo Moeller] 5085109998Smarkm 5086109998Smarkm *) First EC_METHODs for curves over GF(p): 5087109998Smarkm 5088109998Smarkm EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr 5089109998Smarkm operations and provides various method functions that can also 5090109998Smarkm operate with faster implementations of modular arithmetic. 5091109998Smarkm 5092109998Smarkm EC_GFp_mont_method() reuses most functions that are part of 5093109998Smarkm EC_GFp_simple_method, but uses Montgomery arithmetic. 5094109998Smarkm 5095109998Smarkm [Bodo Moeller; point addition and point doubling 5096109998Smarkm implementation directly derived from source code provided by 5097109998Smarkm Lenka Fibikova <fibikova@exp-math.uni-essen.de>] 5098109998Smarkm 5099109998Smarkm *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, 5100109998Smarkm crypto/ec/ec_lib.c): 5101109998Smarkm 5102109998Smarkm Curves are EC_GROUP objects (with an optional group generator) 5103109998Smarkm based on EC_METHODs that are built into the library. 5104109998Smarkm 5105109998Smarkm Points are EC_POINT objects based on EC_GROUP objects. 5106109998Smarkm 5107109998Smarkm Most of the framework would be able to handle curves over arbitrary 5108109998Smarkm finite fields, but as there are no obvious types for fields other 5109109998Smarkm than GF(p), some functions are limited to that for now. 5110109998Smarkm [Bodo Moeller] 5111109998Smarkm 5112109998Smarkm *) Add the -HTTP option to s_server. It is similar to -WWW, but requires 5113109998Smarkm that the file contains a complete HTTP response. 5114109998Smarkm [Richard Levitte] 5115109998Smarkm 5116109998Smarkm *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl 5117109998Smarkm change the def and num file printf format specifier from "%-40sXXX" 5118109998Smarkm to "%-39s XXX". The latter will always guarantee a space after the 5119109998Smarkm field while the former will cause them to run together if the field 5120109998Smarkm is 40 of more characters long. 5121109998Smarkm [Steve Henson] 5122109998Smarkm 5123109998Smarkm *) Constify the cipher and digest 'method' functions and structures 5124109998Smarkm and modify related functions to take constant EVP_MD and EVP_CIPHER 5125109998Smarkm pointers. 5126109998Smarkm [Steve Henson] 5127109998Smarkm 5128109998Smarkm *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them 5129109998Smarkm in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. 5130109998Smarkm [Bodo Moeller] 5131109998Smarkm 5132109998Smarkm *) Modify EVP_Digest*() routines so they now return values. Although the 5133109998Smarkm internal software routines can never fail additional hardware versions 5134109998Smarkm might. 5135109998Smarkm [Steve Henson] 5136109998Smarkm 5137109998Smarkm *) Clean up crypto/err/err.h and change some error codes to avoid conflicts: 5138109998Smarkm 5139109998Smarkm Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 5140109998Smarkm (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. 5141109998Smarkm 5142109998Smarkm ASN1 error codes 5143109998Smarkm ERR_R_NESTED_ASN1_ERROR 5144109998Smarkm ... 5145109998Smarkm ERR_R_MISSING_ASN1_EOS 5146109998Smarkm were 4 .. 9, conflicting with 5147109998Smarkm ERR_LIB_RSA (= ERR_R_RSA_LIB) 5148109998Smarkm ... 5149109998Smarkm ERR_LIB_PEM (= ERR_R_PEM_LIB). 5150109998Smarkm They are now 58 .. 63 (i.e., just below ERR_R_FATAL). 5151109998Smarkm 5152109998Smarkm Add new error code 'ERR_R_INTERNAL_ERROR'. 5153109998Smarkm [Bodo Moeller] 5154109998Smarkm 5155109998Smarkm *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock 5156109998Smarkm suffices. 5157109998Smarkm [Bodo Moeller] 5158109998Smarkm 5159109998Smarkm *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This 5160109998Smarkm sets the subject name for a new request or supersedes the 5161109998Smarkm subject name in a given request. Formats that can be parsed are 5162109998Smarkm 'CN=Some Name, OU=myOU, C=IT' 5163109998Smarkm and 5164109998Smarkm 'CN=Some Name/OU=myOU/C=IT'. 5165109998Smarkm 5166109998Smarkm Add options '-batch' and '-verbose' to 'openssl req'. 5167109998Smarkm [Massimiliano Pala <madwolf@hackmasters.net>] 5168109998Smarkm 5169109998Smarkm *) Introduce the possibility to access global variables through 5170109998Smarkm functions on platform were that's the best way to handle exporting 5171109998Smarkm global variables in shared libraries. To enable this functionality, 5172109998Smarkm one must configure with "EXPORT_VAR_AS_FN" or defined the C macro 5173109998Smarkm "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter 5174109998Smarkm is normally done by Configure or something similar). 5175109998Smarkm 5176109998Smarkm To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL 5177109998Smarkm in the source file (foo.c) like this: 5178109998Smarkm 5179109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; 5180109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(double,bar); 5181109998Smarkm 5182109998Smarkm To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL 5183109998Smarkm and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: 5184109998Smarkm 5185109998Smarkm OPENSSL_DECLARE_GLOBAL(int,foo); 5186109998Smarkm #define foo OPENSSL_GLOBAL_REF(foo) 5187109998Smarkm OPENSSL_DECLARE_GLOBAL(double,bar); 5188109998Smarkm #define bar OPENSSL_GLOBAL_REF(bar) 5189109998Smarkm 5190109998Smarkm The #defines are very important, and therefore so is including the 5191109998Smarkm header file everywhere where the defined globals are used. 5192109998Smarkm 5193109998Smarkm The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition 5194109998Smarkm of ASN.1 items, but that structure is a bit different. 5195109998Smarkm 5196109998Smarkm The largest change is in util/mkdef.pl which has been enhanced with 5197109998Smarkm better and easier to understand logic to choose which symbols should 5198109998Smarkm go into the Windows .def files as well as a number of fixes and code 5199109998Smarkm cleanup (among others, algorithm keywords are now sorted 5200109998Smarkm lexicographically to avoid constant rewrites). 5201109998Smarkm [Richard Levitte] 5202109998Smarkm 5203109998Smarkm *) In BN_div() keep a copy of the sign of 'num' before writing the 5204109998Smarkm result to 'rm' because if rm==num the value will be overwritten 5205109998Smarkm and produce the wrong result if 'num' is negative: this caused 5206109998Smarkm problems with BN_mod() and BN_nnmod(). 5207109998Smarkm [Steve Henson] 5208109998Smarkm 5209109998Smarkm *) Function OCSP_request_verify(). This checks the signature on an 5210109998Smarkm OCSP request and verifies the signer certificate. The signer 5211109998Smarkm certificate is just checked for a generic purpose and OCSP request 5212109998Smarkm trust settings. 5213109998Smarkm [Steve Henson] 5214109998Smarkm 5215109998Smarkm *) Add OCSP_check_validity() function to check the validity of OCSP 5216109998Smarkm responses. OCSP responses are prepared in real time and may only 5217109998Smarkm be a few seconds old. Simply checking that the current time lies 5218109998Smarkm between thisUpdate and nextUpdate max reject otherwise valid responses 5219109998Smarkm caused by either OCSP responder or client clock inaccuracy. Instead 5220109998Smarkm we allow thisUpdate and nextUpdate to fall within a certain period of 5221109998Smarkm the current time. The age of the response can also optionally be 5222109998Smarkm checked. Two new options -validity_period and -status_age added to 5223109998Smarkm ocsp utility. 5224109998Smarkm [Steve Henson] 5225109998Smarkm 5226109998Smarkm *) If signature or public key algorithm is unrecognized print out its 5227109998Smarkm OID rather that just UNKNOWN. 5228109998Smarkm [Steve Henson] 5229109998Smarkm 5230109998Smarkm *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and 5231109998Smarkm OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate 5232109998Smarkm ID to be generated from the issuer certificate alone which can then be 5233109998Smarkm passed to OCSP_id_issuer_cmp(). 5234109998Smarkm [Steve Henson] 5235109998Smarkm 5236109998Smarkm *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new 5237109998Smarkm ASN1 modules to export functions returning ASN1_ITEM pointers 5238109998Smarkm instead of the ASN1_ITEM structures themselves. This adds several 5239109998Smarkm new macros which allow the underlying ASN1 function/structure to 5240109998Smarkm be accessed transparently. As a result code should not use ASN1_ITEM 5241109998Smarkm references directly (such as &X509_it) but instead use the relevant 5242109998Smarkm macros (such as ASN1_ITEM_rptr(X509)). This option is to allow 5243109998Smarkm use of the new ASN1 code on platforms where exporting structures 5244109998Smarkm is problematical (for example in shared libraries) but exporting 5245109998Smarkm functions returning pointers to structures is not. 5246109998Smarkm [Steve Henson] 5247109998Smarkm 5248109998Smarkm *) Add support for overriding the generation of SSL/TLS session IDs. 5249109998Smarkm These callbacks can be registered either in an SSL_CTX or per SSL. 5250109998Smarkm The purpose of this is to allow applications to control, if they wish, 5251109998Smarkm the arbitrary values chosen for use as session IDs, particularly as it 5252109998Smarkm can be useful for session caching in multiple-server environments. A 5253109998Smarkm command-line switch for testing this (and any client code that wishes 5254109998Smarkm to use such a feature) has been added to "s_server". 5255109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 5256109998Smarkm 5257109998Smarkm *) Modify mkdef.pl to recognise and parse preprocessor conditionals 5258109998Smarkm of the form '#if defined(...) || defined(...) || ...' and 5259109998Smarkm '#if !defined(...) && !defined(...) && ...'. This also avoids 5260109998Smarkm the growing number of special cases it was previously handling. 5261109998Smarkm [Richard Levitte] 5262109998Smarkm 5263109998Smarkm *) Make all configuration macros available for application by making 5264109998Smarkm sure they are available in opensslconf.h, by giving them names starting 5265109998Smarkm with "OPENSSL_" to avoid conflicts with other packages and by making 5266109998Smarkm sure e_os2.h will cover all platform-specific cases together with 5267109998Smarkm opensslconf.h. 5268109998Smarkm Additionally, it is now possible to define configuration/platform- 5269109998Smarkm specific names (called "system identities"). In the C code, these 5270109998Smarkm are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another 5271109998Smarkm macro with the name beginning with "OPENSSL_SYS_", which is determined 5272109998Smarkm from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on 5273109998Smarkm what is available. 5274109998Smarkm [Richard Levitte] 5275109998Smarkm 5276109998Smarkm *) New option -set_serial to 'req' and 'x509' this allows the serial 5277109998Smarkm number to use to be specified on the command line. Previously self 5278109998Smarkm signed certificates were hard coded with serial number 0 and the 5279109998Smarkm CA options of 'x509' had to use a serial number in a file which was 5280109998Smarkm auto incremented. 5281109998Smarkm [Steve Henson] 5282109998Smarkm 5283109998Smarkm *) New options to 'ca' utility to support V2 CRL entry extensions. 5284109998Smarkm Currently CRL reason, invalidity date and hold instruction are 5285109998Smarkm supported. Add new CRL extensions to V3 code and some new objects. 5286109998Smarkm [Steve Henson] 5287109998Smarkm 5288109998Smarkm *) New function EVP_CIPHER_CTX_set_padding() this is used to 5289109998Smarkm disable standard block padding (aka PKCS#5 padding) in the EVP 5290109998Smarkm API, which was previously mandatory. This means that the data is 5291109998Smarkm not padded in any way and so the total length much be a multiple 5292109998Smarkm of the block size, otherwise an error occurs. 5293109998Smarkm [Steve Henson] 5294109998Smarkm 5295109998Smarkm *) Initial (incomplete) OCSP SSL support. 5296109998Smarkm [Steve Henson] 5297109998Smarkm 5298109998Smarkm *) New function OCSP_parse_url(). This splits up a URL into its host, 5299109998Smarkm port and path components: primarily to parse OCSP URLs. New -url 5300109998Smarkm option to ocsp utility. 5301109998Smarkm [Steve Henson] 5302109998Smarkm 5303109998Smarkm *) New nonce behavior. The return value of OCSP_check_nonce() now 5304109998Smarkm reflects the various checks performed. Applications can decide 5305109998Smarkm whether to tolerate certain situations such as an absent nonce 5306109998Smarkm in a response when one was present in a request: the ocsp application 5307109998Smarkm just prints out a warning. New function OCSP_add1_basic_nonce() 5308109998Smarkm this is to allow responders to include a nonce in a response even if 5309109998Smarkm the request is nonce-less. 5310109998Smarkm [Steve Henson] 5311109998Smarkm 5312109998Smarkm *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are 5313109998Smarkm skipped when using openssl x509 multiple times on a single input file, 5314109998Smarkm e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs". 5315109998Smarkm [Bodo Moeller] 5316109998Smarkm 5317109998Smarkm *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() 5318109998Smarkm set string type: to handle setting ASN1_TIME structures. Fix ca 5319109998Smarkm utility to correctly initialize revocation date of CRLs. 5320109998Smarkm [Steve Henson] 5321109998Smarkm 5322109998Smarkm *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override 5323109998Smarkm the clients preferred ciphersuites and rather use its own preferences. 5324109998Smarkm Should help to work around M$ SGC (Server Gated Cryptography) bug in 5325109998Smarkm Internet Explorer by ensuring unchanged hash method during stepup. 5326109998Smarkm (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) 5327109998Smarkm [Lutz Jaenicke] 5328109998Smarkm 5329109998Smarkm *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael 5330109998Smarkm to aes and add a new 'exist' option to print out symbols that don't 5331109998Smarkm appear to exist. 5332109998Smarkm [Steve Henson] 5333109998Smarkm 5334109998Smarkm *) Additional options to ocsp utility to allow flags to be set and 5335109998Smarkm additional certificates supplied. 5336109998Smarkm [Steve Henson] 5337109998Smarkm 5338109998Smarkm *) Add the option -VAfile to 'openssl ocsp', so the user can give the 5339109998Smarkm OCSP client a number of certificate to only verify the response 5340109998Smarkm signature against. 5341109998Smarkm [Richard Levitte] 5342109998Smarkm 5343109998Smarkm *) Update Rijndael code to version 3.0 and change EVP AES ciphers to 5344109998Smarkm handle the new API. Currently only ECB, CBC modes supported. Add new 5345109998Smarkm AES OIDs. 5346109998Smarkm 5347109998Smarkm Add TLS AES ciphersuites as described in RFC3268, "Advanced 5348109998Smarkm Encryption Standard (AES) Ciphersuites for Transport Layer 5349109998Smarkm Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were 5350109998Smarkm not enabled by default and were not part of the "ALL" ciphersuite 5351109998Smarkm alias because they were not yet official; they could be 5352109998Smarkm explicitly requested by specifying the "AESdraft" ciphersuite 5353109998Smarkm group alias. In the final release of OpenSSL 0.9.7, the group 5354109998Smarkm alias is called "AES" and is part of "ALL".) 5355109998Smarkm [Ben Laurie, Steve Henson, Bodo Moeller] 5356109998Smarkm 5357109998Smarkm *) New function OCSP_copy_nonce() to copy nonce value (if present) from 5358109998Smarkm request to response. 5359109998Smarkm [Steve Henson] 5360109998Smarkm 5361109998Smarkm *) Functions for OCSP responders. OCSP_request_onereq_count(), 5362109998Smarkm OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() 5363109998Smarkm extract information from a certificate request. OCSP_response_create() 5364109998Smarkm creates a response and optionally adds a basic response structure. 5365109998Smarkm OCSP_basic_add1_status() adds a complete single response to a basic 5366109998Smarkm response and returns the OCSP_SINGLERESP structure just added (to allow 5367109998Smarkm extensions to be included for example). OCSP_basic_add1_cert() adds a 5368109998Smarkm certificate to a basic response and OCSP_basic_sign() signs a basic 5369109998Smarkm response with various flags. New helper functions ASN1_TIME_check() 5370109998Smarkm (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() 5371109998Smarkm (converts ASN1_TIME to GeneralizedTime). 5372109998Smarkm [Steve Henson] 5373109998Smarkm 5374109998Smarkm *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() 5375109998Smarkm in a single operation. X509_get0_pubkey_bitstr() extracts the public_key 5376109998Smarkm structure from a certificate. X509_pubkey_digest() digests the public_key 5377109998Smarkm contents: this is used in various key identifiers. 5378109998Smarkm [Steve Henson] 5379109998Smarkm 5380109998Smarkm *) Make sk_sort() tolerate a NULL argument. 5381109998Smarkm [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>] 5382109998Smarkm 5383109998Smarkm *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates 5384109998Smarkm passed by the function are trusted implicitly. If any of them signed the 5385109998Smarkm response then it is assumed to be valid and is not verified. 5386109998Smarkm [Steve Henson] 5387109998Smarkm 5388109998Smarkm *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT 5389109998Smarkm to data. This was previously part of the PKCS7 ASN1 code. This 5390109998Smarkm was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. 5391109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 5392109998Smarkm <support@securenetterm.com>] 5393109998Smarkm 5394109998Smarkm *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 5395109998Smarkm routines: without these tracing memory leaks is very painful. 5396109998Smarkm Fix leaks in PKCS12 and PKCS7 routines. 5397109998Smarkm [Steve Henson] 5398109998Smarkm 5399109998Smarkm *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). 5400109998Smarkm Previously it initialised the 'type' argument to V_ASN1_UTCTIME which 5401109998Smarkm effectively meant GeneralizedTime would never be used. Now it 5402109998Smarkm is initialised to -1 but X509_time_adj() now has to check the value 5403109998Smarkm and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or 5404109998Smarkm V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. 5405109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 5406109998Smarkm <support@securenetterm.com>] 5407109998Smarkm 5408109998Smarkm *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously 5409109998Smarkm result in a zero length in the ASN1_INTEGER structure which was 5410109998Smarkm not consistent with the structure when d2i_ASN1_INTEGER() was used 5411109998Smarkm and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() 5412109998Smarkm to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() 5413109998Smarkm where it did not print out a minus for negative ASN1_INTEGER. 5414109998Smarkm [Steve Henson] 5415109998Smarkm 5416109998Smarkm *) Add summary printout to ocsp utility. The various functions which 5417109998Smarkm convert status values to strings have been renamed to: 5418109998Smarkm OCSP_response_status_str(), OCSP_cert_status_str() and 5419109998Smarkm OCSP_crl_reason_str() and are no longer static. New options 5420109998Smarkm to verify nonce values and to disable verification. OCSP response 5421109998Smarkm printout format cleaned up. 5422109998Smarkm [Steve Henson] 5423109998Smarkm 5424109998Smarkm *) Add additional OCSP certificate checks. These are those specified 5425109998Smarkm in RFC2560. This consists of two separate checks: the CA of the 5426109998Smarkm certificate being checked must either be the OCSP signer certificate 5427109998Smarkm or the issuer of the OCSP signer certificate. In the latter case the 5428109998Smarkm OCSP signer certificate must contain the OCSP signing extended key 5429109998Smarkm usage. This check is performed by attempting to match the OCSP 5430109998Smarkm signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash 5431109998Smarkm in the OCSP_CERTID structures of the response. 5432109998Smarkm [Steve Henson] 5433109998Smarkm 5434109998Smarkm *) Initial OCSP certificate verification added to OCSP_basic_verify() 5435109998Smarkm and related routines. This uses the standard OpenSSL certificate 5436109998Smarkm verify routines to perform initial checks (just CA validity) and 5437109998Smarkm to obtain the certificate chain. Then additional checks will be 5438109998Smarkm performed on the chain. Currently the root CA is checked to see 5439109998Smarkm if it is explicitly trusted for OCSP signing. This is used to set 5440109998Smarkm a root CA as a global signing root: that is any certificate that 5441109998Smarkm chains to that CA is an acceptable OCSP signing certificate. 5442109998Smarkm [Steve Henson] 5443109998Smarkm 5444109998Smarkm *) New '-extfile ...' option to 'openssl ca' for reading X.509v3 5445109998Smarkm extensions from a separate configuration file. 5446109998Smarkm As when reading extensions from the main configuration file, 5447109998Smarkm the '-extensions ...' option may be used for specifying the 5448109998Smarkm section to use. 5449109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 5450109998Smarkm 5451109998Smarkm *) New OCSP utility. Allows OCSP requests to be generated or 5452109998Smarkm read. The request can be sent to a responder and the output 5453109998Smarkm parsed, outputed or printed in text form. Not complete yet: 5454109998Smarkm still needs to check the OCSP response validity. 5455109998Smarkm [Steve Henson] 5456109998Smarkm 5457109998Smarkm *) New subcommands for 'openssl ca': 5458109998Smarkm 'openssl ca -status <serial>' prints the status of the cert with 5459109998Smarkm the given serial number (according to the index file). 5460109998Smarkm 'openssl ca -updatedb' updates the expiry status of certificates 5461109998Smarkm in the index file. 5462109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 5463109998Smarkm 5464109998Smarkm *) New '-newreq-nodes' command option to CA.pl. This is like 5465109998Smarkm '-newreq', but calls 'openssl req' with the '-nodes' option 5466109998Smarkm so that the resulting key is not encrypted. 5467109998Smarkm [Damien Miller <djm@mindrot.org>] 5468109998Smarkm 5469109998Smarkm *) New configuration for the GNU Hurd. 5470109998Smarkm [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte] 5471109998Smarkm 5472109998Smarkm *) Initial code to implement OCSP basic response verify. This 5473109998Smarkm is currently incomplete. Currently just finds the signer's 5474109998Smarkm certificate and verifies the signature on the response. 5475109998Smarkm [Steve Henson] 5476109998Smarkm 5477109998Smarkm *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in 5478109998Smarkm value of OPENSSLDIR. This is available via the new '-d' option 5479109998Smarkm to 'openssl version', and is also included in 'openssl version -a'. 5480109998Smarkm [Bodo Moeller] 5481109998Smarkm 5482109998Smarkm *) Allowing defining memory allocation callbacks that will be given 5483109998Smarkm file name and line number information in additional arguments 5484109998Smarkm (a const char* and an int). The basic functionality remains, as 5485109998Smarkm well as the original possibility to just replace malloc(), 5486109998Smarkm realloc() and free() by functions that do not know about these 5487109998Smarkm additional arguments. To register and find out the current 5488109998Smarkm settings for extended allocation functions, the following 5489109998Smarkm functions are provided: 5490109998Smarkm 5491109998Smarkm CRYPTO_set_mem_ex_functions 5492109998Smarkm CRYPTO_set_locked_mem_ex_functions 5493109998Smarkm CRYPTO_get_mem_ex_functions 5494109998Smarkm CRYPTO_get_locked_mem_ex_functions 5495109998Smarkm 5496109998Smarkm These work the same way as CRYPTO_set_mem_functions and friends. 5497109998Smarkm CRYPTO_get_[locked_]mem_functions now writes 0 where such an 5498109998Smarkm extended allocation function is enabled. 5499109998Smarkm Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where 5500109998Smarkm a conventional allocation function is enabled. 5501109998Smarkm [Richard Levitte, Bodo Moeller] 5502109998Smarkm 5503109998Smarkm *) Finish off removing the remaining LHASH function pointer casts. 5504109998Smarkm There should no longer be any prototype-casting required when using 5505109998Smarkm the LHASH abstraction, and any casts that remain are "bugs". See 5506109998Smarkm the callback types and macros at the head of lhash.h for details 5507109998Smarkm (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). 5508109998Smarkm [Geoff Thorpe] 5509109998Smarkm 5510109998Smarkm *) Add automatic query of EGD sockets in RAND_poll() for the unix variant. 5511109998Smarkm If /dev/[u]random devices are not available or do not return enough 5512109998Smarkm entropy, EGD style sockets (served by EGD or PRNGD) will automatically 5513109998Smarkm be queried. 5514109998Smarkm The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and 5515109998Smarkm /etc/entropy will be queried once each in this sequence, quering stops 5516109998Smarkm when enough entropy was collected without querying more sockets. 5517109998Smarkm [Lutz Jaenicke] 5518109998Smarkm 5519109998Smarkm *) Change the Unix RAND_poll() variant to be able to poll several 5520109998Smarkm random devices, as specified by DEVRANDOM, until a sufficient amount 5521109998Smarkm of data has been collected. We spend at most 10 ms on each file 5522109998Smarkm (select timeout) and read in non-blocking mode. DEVRANDOM now 5523109998Smarkm defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" 5524109998Smarkm (previously it was just the string "/dev/urandom"), so on typical 5525109998Smarkm platforms the 10 ms delay will never occur. 5526109998Smarkm Also separate out the Unix variant to its own file, rand_unix.c. 5527109998Smarkm For VMS, there's a currently-empty rand_vms.c. 5528109998Smarkm [Richard Levitte] 5529109998Smarkm 5530109998Smarkm *) Move OCSP client related routines to ocsp_cl.c. These 5531109998Smarkm provide utility functions which an application needing 5532109998Smarkm to issue a request to an OCSP responder and analyse the 5533109998Smarkm response will typically need: as opposed to those which an 5534109998Smarkm OCSP responder itself would need which will be added later. 5535109998Smarkm 5536109998Smarkm OCSP_request_sign() signs an OCSP request with an API similar 5537109998Smarkm to PKCS7_sign(). OCSP_response_status() returns status of OCSP 5538109998Smarkm response. OCSP_response_get1_basic() extracts basic response 5539109998Smarkm from response. OCSP_resp_find_status(): finds and extracts status 5540109998Smarkm information from an OCSP_CERTID structure (which will be created 5541109998Smarkm when the request structure is built). These are built from lower 5542109998Smarkm level functions which work on OCSP_SINGLERESP structures but 5543109998Smarkm wont normally be used unless the application wishes to examine 5544109998Smarkm extensions in the OCSP response for example. 5545109998Smarkm 5546109998Smarkm Replace nonce routines with a pair of functions. 5547109998Smarkm OCSP_request_add1_nonce() adds a nonce value and optionally 5548109998Smarkm generates a random value. OCSP_check_nonce() checks the 5549109998Smarkm validity of the nonce in an OCSP response. 5550109998Smarkm [Steve Henson] 5551109998Smarkm 5552109998Smarkm *) Change function OCSP_request_add() to OCSP_request_add0_id(). 5553109998Smarkm This doesn't copy the supplied OCSP_CERTID and avoids the 5554109998Smarkm need to free up the newly created id. Change return type 5555109998Smarkm to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. 5556109998Smarkm This can then be used to add extensions to the request. 5557109998Smarkm Deleted OCSP_request_new(), since most of its functionality 5558109998Smarkm is now in OCSP_REQUEST_new() (and the case insensitive name 5559109998Smarkm clash) apart from the ability to set the request name which 5560109998Smarkm will be added elsewhere. 5561109998Smarkm [Steve Henson] 5562109998Smarkm 5563109998Smarkm *) Update OCSP API. Remove obsolete extensions argument from 5564109998Smarkm various functions. Extensions are now handled using the new 5565109998Smarkm OCSP extension code. New simple OCSP HTTP function which 5566109998Smarkm can be used to send requests and parse the response. 5567109998Smarkm [Steve Henson] 5568109998Smarkm 5569109998Smarkm *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new 5570109998Smarkm ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN 5571109998Smarkm uses the special reorder version of SET OF to sort the attributes 5572109998Smarkm and reorder them to match the encoded order. This resolves a long 5573109998Smarkm standing problem: a verify on a PKCS7 structure just after signing 5574109998Smarkm it used to fail because the attribute order did not match the 5575109998Smarkm encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: 5576109998Smarkm it uses the received order. This is necessary to tolerate some broken 5577109998Smarkm software that does not order SET OF. This is handled by encoding 5578109998Smarkm as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) 5579109998Smarkm to produce the required SET OF. 5580109998Smarkm [Steve Henson] 5581109998Smarkm 5582109998Smarkm *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and 5583109998Smarkm OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header 5584109998Smarkm files to get correct declarations of the ASN.1 item variables. 5585109998Smarkm [Richard Levitte] 5586109998Smarkm 5587109998Smarkm *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many 5588109998Smarkm PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: 5589109998Smarkm asn1_check_tlen() would sometimes attempt to use 'ctx' when it was 5590109998Smarkm NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). 5591109998Smarkm New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant 5592109998Smarkm ASN1_ITEM and no wrapper functions. 5593109998Smarkm [Steve Henson] 5594109998Smarkm 5595109998Smarkm *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These 5596109998Smarkm replace the old function pointer based I/O routines. Change most of 5597109998Smarkm the *_d2i_bio() and *_d2i_fp() functions to use these. 5598109998Smarkm [Steve Henson] 5599109998Smarkm 5600109998Smarkm *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor 5601109998Smarkm lines, recognice more "algorithms" that can be deselected, and make 5602109998Smarkm it complain about algorithm deselection that isn't recognised. 5603109998Smarkm [Richard Levitte] 5604109998Smarkm 5605109998Smarkm *) New ASN1 functions to handle dup, sign, verify, digest, pack and 5606109998Smarkm unpack operations in terms of ASN1_ITEM. Modify existing wrappers 5607109998Smarkm to use new functions. Add NO_ASN1_OLD which can be set to remove 5608109998Smarkm some old style ASN1 functions: this can be used to determine if old 5609109998Smarkm code will still work when these eventually go away. 5610109998Smarkm [Steve Henson] 5611109998Smarkm 5612109998Smarkm *) New extension functions for OCSP structures, these follow the 5613109998Smarkm same conventions as certificates and CRLs. 5614109998Smarkm [Steve Henson] 5615109998Smarkm 5616109998Smarkm *) New function X509V3_add1_i2d(). This automatically encodes and 5617109998Smarkm adds an extension. Its behaviour can be customised with various 5618109998Smarkm flags to append, replace or delete. Various wrappers added for 5619109998Smarkm certifcates and CRLs. 5620109998Smarkm [Steve Henson] 5621109998Smarkm 5622109998Smarkm *) Fix to avoid calling the underlying ASN1 print routine when 5623109998Smarkm an extension cannot be parsed. Correct a typo in the 5624109998Smarkm OCSP_SERVICELOC extension. Tidy up print OCSP format. 5625109998Smarkm [Steve Henson] 5626109998Smarkm 5627109998Smarkm *) Make mkdef.pl parse some of the ASN1 macros and add apropriate 5628109998Smarkm entries for variables. 5629109998Smarkm [Steve Henson] 5630109998Smarkm 5631109998Smarkm *) Add functionality to apps/openssl.c for detecting locking 5632109998Smarkm problems: As the program is single-threaded, all we have 5633109998Smarkm to do is register a locking callback using an array for 5634109998Smarkm storing which locks are currently held by the program. 5635109998Smarkm [Bodo Moeller] 5636109998Smarkm 5637109998Smarkm *) Use a lock around the call to CRYPTO_get_ex_new_index() in 5638109998Smarkm SSL_get_ex_data_X509_STORE_idx(), which is used in 5639109998Smarkm ssl_verify_cert_chain() and thus can be called at any time 5640109998Smarkm during TLS/SSL handshakes so that thread-safety is essential. 5641109998Smarkm Unfortunately, the ex_data design is not at all suited 5642109998Smarkm for multi-threaded use, so it probably should be abolished. 5643109998Smarkm [Bodo Moeller] 5644109998Smarkm 5645109998Smarkm *) Added Broadcom "ubsec" ENGINE to OpenSSL. 5646109998Smarkm [Broadcom, tweaked and integrated by Geoff Thorpe] 5647109998Smarkm 5648109998Smarkm *) Move common extension printing code to new function 5649109998Smarkm X509V3_print_extensions(). Reorganise OCSP print routines and 5650109998Smarkm implement some needed OCSP ASN1 functions. Add OCSP extensions. 5651109998Smarkm [Steve Henson] 5652109998Smarkm 5653109998Smarkm *) New function X509_signature_print() to remove duplication in some 5654109998Smarkm print routines. 5655109998Smarkm [Steve Henson] 5656109998Smarkm 5657109998Smarkm *) Add a special meaning when SET OF and SEQUENCE OF flags are both 5658109998Smarkm set (this was treated exactly the same as SET OF previously). This 5659109998Smarkm is used to reorder the STACK representing the structure to match the 5660109998Smarkm encoding. This will be used to get round a problem where a PKCS7 5661109998Smarkm structure which was signed could not be verified because the STACK 5662109998Smarkm order did not reflect the encoded order. 5663109998Smarkm [Steve Henson] 5664109998Smarkm 5665109998Smarkm *) Reimplement the OCSP ASN1 module using the new code. 5666109998Smarkm [Steve Henson] 5667109998Smarkm 5668109998Smarkm *) Update the X509V3 code to permit the use of an ASN1_ITEM structure 5669109998Smarkm for its ASN1 operations. The old style function pointers still exist 5670109998Smarkm for now but they will eventually go away. 5671109998Smarkm [Steve Henson] 5672109998Smarkm 5673109998Smarkm *) Merge in replacement ASN1 code from the ASN1 branch. This almost 5674109998Smarkm completely replaces the old ASN1 functionality with a table driven 5675109998Smarkm encoder and decoder which interprets an ASN1_ITEM structure describing 5676109998Smarkm the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is 5677109998Smarkm largely maintained. Almost all of the old asn1_mac.h macro based ASN1 5678109998Smarkm has also been converted to the new form. 5679109998Smarkm [Steve Henson] 5680109998Smarkm 5681109998Smarkm *) Change BN_mod_exp_recp so that negative moduli are tolerated 5682109998Smarkm (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set 5683109998Smarkm so that BN_mod_exp_mont and BN_mod_exp_mont_word work 5684109998Smarkm for negative moduli. 5685109998Smarkm [Bodo Moeller] 5686109998Smarkm 5687109998Smarkm *) Fix BN_uadd and BN_usub: Always return non-negative results instead 5688109998Smarkm of not touching the result's sign bit. 5689109998Smarkm [Bodo Moeller] 5690109998Smarkm 5691109998Smarkm *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be 5692109998Smarkm set. 5693109998Smarkm [Bodo Moeller] 5694109998Smarkm 5695109998Smarkm *) Changed the LHASH code to use prototypes for callbacks, and created 5696109998Smarkm macros to declare and implement thin (optionally static) functions 5697109998Smarkm that provide type-safety and avoid function pointer casting for the 5698109998Smarkm type-specific callbacks. 5699109998Smarkm [Geoff Thorpe] 5700109998Smarkm 5701109998Smarkm *) Added Kerberos Cipher Suites to be used with TLS, as written in 5702109998Smarkm RFC 2712. 5703109998Smarkm [Veers Staats <staatsvr@asc.hpc.mil>, 5704109998Smarkm Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte] 5705109998Smarkm 5706109998Smarkm *) Reformat the FAQ so the different questions and answers can be divided 5707109998Smarkm in sections depending on the subject. 5708109998Smarkm [Richard Levitte] 5709109998Smarkm 5710109998Smarkm *) Have the zlib compression code load ZLIB.DLL dynamically under 5711109998Smarkm Windows. 5712109998Smarkm [Richard Levitte] 5713109998Smarkm 5714109998Smarkm *) New function BN_mod_sqrt for computing square roots modulo a prime 5715109998Smarkm (using the probabilistic Tonelli-Shanks algorithm unless 5716109998Smarkm p == 3 (mod 4) or p == 5 (mod 8), which are cases that can 5717109998Smarkm be handled deterministically). 5718109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 5719109998Smarkm 5720109998Smarkm *) Make BN_mod_inverse faster by explicitly handling small quotients 5721109998Smarkm in the Euclid loop. (Speed gain about 20% for small moduli [256 or 5722109998Smarkm 512 bits], about 30% for larger ones [1024 or 2048 bits].) 5723109998Smarkm [Bodo Moeller] 5724109998Smarkm 5725109998Smarkm *) New function BN_kronecker. 5726109998Smarkm [Bodo Moeller] 5727109998Smarkm 5728109998Smarkm *) Fix BN_gcd so that it works on negative inputs; the result is 5729109998Smarkm positive unless both parameters are zero. 5730109998Smarkm Previously something reasonably close to an infinite loop was 5731109998Smarkm possible because numbers could be growing instead of shrinking 5732109998Smarkm in the implementation of Euclid's algorithm. 5733109998Smarkm [Bodo Moeller] 5734109998Smarkm 5735109998Smarkm *) Fix BN_is_word() and BN_is_one() macros to take into account the 5736109998Smarkm sign of the number in question. 5737109998Smarkm 5738109998Smarkm Fix BN_is_word(a,w) to work correctly for w == 0. 5739109998Smarkm 5740109998Smarkm The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) 5741109998Smarkm because its test if the absolute value of 'a' equals 'w'. 5742109998Smarkm Note that BN_abs_is_word does *not* handle w == 0 reliably; 5743109998Smarkm it exists mostly for use in the implementations of BN_is_zero(), 5744109998Smarkm BN_is_one(), and BN_is_word(). 5745109998Smarkm [Bodo Moeller] 5746109998Smarkm 5747109998Smarkm *) New function BN_swap. 5748109998Smarkm [Bodo Moeller] 5749109998Smarkm 5750109998Smarkm *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that 5751109998Smarkm the exponentiation functions are more likely to produce reasonable 5752109998Smarkm results on negative inputs. 5753109998Smarkm [Bodo Moeller] 5754109998Smarkm 5755109998Smarkm *) Change BN_mod_mul so that the result is always non-negative. 5756109998Smarkm Previously, it could be negative if one of the factors was negative; 5757109998Smarkm I don't think anyone really wanted that behaviour. 5758109998Smarkm [Bodo Moeller] 5759109998Smarkm 5760109998Smarkm *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c 5761109998Smarkm (except for exponentiation, which stays in crypto/bn/bn_exp.c, 5762109998Smarkm and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c) 5763109998Smarkm and add new functions: 5764109998Smarkm 5765109998Smarkm BN_nnmod 5766109998Smarkm BN_mod_sqr 5767109998Smarkm BN_mod_add 5768109998Smarkm BN_mod_add_quick 5769109998Smarkm BN_mod_sub 5770109998Smarkm BN_mod_sub_quick 5771109998Smarkm BN_mod_lshift1 5772109998Smarkm BN_mod_lshift1_quick 5773109998Smarkm BN_mod_lshift 5774109998Smarkm BN_mod_lshift_quick 5775109998Smarkm 5776109998Smarkm These functions always generate non-negative results. 5777109998Smarkm 5778109998Smarkm BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r 5779109998Smarkm such that |m| < r < 0, BN_nnmod will output rem + |m| instead). 5780109998Smarkm 5781109998Smarkm BN_mod_XXX_quick(r, a, [b,] m) generates the same result as 5782109998Smarkm BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b] 5783109998Smarkm be reduced modulo m. 5784109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 5785109998Smarkm 5786111147Snectar#if 0 5787111147Snectar The following entry accidentily appeared in the CHANGES file 5788111147Snectar distributed with OpenSSL 0.9.7. The modifications described in 5789111147Snectar it do *not* apply to OpenSSL 0.9.7. 5790111147Snectar 5791109998Smarkm *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 5792109998Smarkm was actually never needed) and in BN_mul(). The removal in BN_mul() 5793109998Smarkm required a small change in bn_mul_part_recursive() and the addition 5794109998Smarkm of the functions bn_cmp_part_words(), bn_sub_part_words() and 5795109998Smarkm bn_add_part_words(), which do the same thing as bn_cmp_words(), 5796109998Smarkm bn_sub_words() and bn_add_words() except they take arrays with 5797109998Smarkm differing sizes. 5798109998Smarkm [Richard Levitte] 5799111147Snectar#endif 5800109998Smarkm 5801109998Smarkm *) In 'openssl passwd', verify passwords read from the terminal 5802109998Smarkm unless the '-salt' option is used (which usually means that 5803109998Smarkm verification would just waste user's time since the resulting 5804109998Smarkm hash is going to be compared with some given password hash) 5805109998Smarkm or the new '-noverify' option is used. 5806109998Smarkm 5807109998Smarkm This is an incompatible change, but it does not affect 5808109998Smarkm non-interactive use of 'openssl passwd' (passwords on the command 5809109998Smarkm line, '-stdin' option, '-in ...' option) and thus should not 5810109998Smarkm cause any problems. 5811109998Smarkm [Bodo Moeller] 5812109998Smarkm 5813109998Smarkm *) Remove all references to RSAref, since there's no more need for it. 5814109998Smarkm [Richard Levitte] 5815109998Smarkm 5816109998Smarkm *) Make DSO load along a path given through an environment variable 5817109998Smarkm (SHLIB_PATH) with shl_load(). 5818109998Smarkm [Richard Levitte] 5819109998Smarkm 5820109998Smarkm *) Constify the ENGINE code as a result of BIGNUM constification. 5821109998Smarkm Also constify the RSA code and most things related to it. In a 5822109998Smarkm few places, most notable in the depth of the ASN.1 code, ugly 5823109998Smarkm casts back to non-const were required (to be solved at a later 5824109998Smarkm time) 5825109998Smarkm [Richard Levitte] 5826109998Smarkm 5827109998Smarkm *) Make it so the openssl application has all engines loaded by default. 5828109998Smarkm [Richard Levitte] 5829109998Smarkm 5830109998Smarkm *) Constify the BIGNUM routines a little more. 5831109998Smarkm [Richard Levitte] 5832109998Smarkm 5833109998Smarkm *) Add the following functions: 5834109998Smarkm 5835109998Smarkm ENGINE_load_cswift() 5836109998Smarkm ENGINE_load_chil() 5837109998Smarkm ENGINE_load_atalla() 5838109998Smarkm ENGINE_load_nuron() 5839109998Smarkm ENGINE_load_builtin_engines() 5840109998Smarkm 5841109998Smarkm That way, an application can itself choose if external engines that 5842109998Smarkm are built-in in OpenSSL shall ever be used or not. The benefit is 5843109998Smarkm that applications won't have to be linked with libdl or other dso 5844109998Smarkm libraries unless it's really needed. 5845109998Smarkm 5846109998Smarkm Changed 'openssl engine' to load all engines on demand. 5847109998Smarkm Changed the engine header files to avoid the duplication of some 5848109998Smarkm declarations (they differed!). 5849109998Smarkm [Richard Levitte] 5850109998Smarkm 5851109998Smarkm *) 'openssl engine' can now list capabilities. 5852109998Smarkm [Richard Levitte] 5853109998Smarkm 5854109998Smarkm *) Better error reporting in 'openssl engine'. 5855109998Smarkm [Richard Levitte] 5856109998Smarkm 5857109998Smarkm *) Never call load_dh_param(NULL) in s_server. 5858109998Smarkm [Bodo Moeller] 5859109998Smarkm 5860109998Smarkm *) Add engine application. It can currently list engines by name and 5861109998Smarkm identity, and test if they are actually available. 5862109998Smarkm [Richard Levitte] 5863109998Smarkm 5864109998Smarkm *) Improve RPM specification file by forcing symbolic linking and making 5865109998Smarkm sure the installed documentation is also owned by root.root. 5866109998Smarkm [Damien Miller <djm@mindrot.org>] 5867109998Smarkm 5868109998Smarkm *) Give the OpenSSL applications more possibilities to make use of 5869109998Smarkm keys (public as well as private) handled by engines. 5870109998Smarkm [Richard Levitte] 5871109998Smarkm 5872109998Smarkm *) Add OCSP code that comes from CertCo. 5873109998Smarkm [Richard Levitte] 5874109998Smarkm 5875109998Smarkm *) Add VMS support for the Rijndael code. 5876109998Smarkm [Richard Levitte] 5877109998Smarkm 5878109998Smarkm *) Added untested support for Nuron crypto accelerator. 5879109998Smarkm [Ben Laurie] 5880109998Smarkm 5881109998Smarkm *) Add support for external cryptographic devices. This code was 5882109998Smarkm previously distributed separately as the "engine" branch. 5883109998Smarkm [Geoff Thorpe, Richard Levitte] 5884109998Smarkm 5885109998Smarkm *) Rework the filename-translation in the DSO code. It is now possible to 5886109998Smarkm have far greater control over how a "name" is turned into a filename 5887109998Smarkm depending on the operating environment and any oddities about the 5888109998Smarkm different shared library filenames on each system. 5889109998Smarkm [Geoff Thorpe] 5890109998Smarkm 5891109998Smarkm *) Support threads on FreeBSD-elf in Configure. 5892109998Smarkm [Richard Levitte] 5893109998Smarkm 5894109998Smarkm *) Fix for SHA1 assembly problem with MASM: it produces 5895109998Smarkm warnings about corrupt line number information when assembling 5896109998Smarkm with debugging information. This is caused by the overlapping 5897109998Smarkm of two sections. 5898109998Smarkm [Bernd Matthes <mainbug@celocom.de>, Steve Henson] 5899109998Smarkm 5900109998Smarkm *) NCONF changes. 5901109998Smarkm NCONF_get_number() has no error checking at all. As a replacement, 5902109998Smarkm NCONF_get_number_e() is defined (_e for "error checking") and is 5903109998Smarkm promoted strongly. The old NCONF_get_number is kept around for 5904109998Smarkm binary backward compatibility. 5905109998Smarkm Make it possible for methods to load from something other than a BIO, 5906109998Smarkm by providing a function pointer that is given a name instead of a BIO. 5907109998Smarkm For example, this could be used to load configuration data from an 5908109998Smarkm LDAP server. 5909109998Smarkm [Richard Levitte] 5910109998Smarkm 5911109998Smarkm *) Fix for non blocking accept BIOs. Added new I/O special reason 5912109998Smarkm BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs 5913109998Smarkm with non blocking I/O was not possible because no retry code was 5914109998Smarkm implemented. Also added new SSL code SSL_WANT_ACCEPT to cover 5915109998Smarkm this case. 5916109998Smarkm [Steve Henson] 5917109998Smarkm 5918109998Smarkm *) Added the beginnings of Rijndael support. 5919109998Smarkm [Ben Laurie] 5920109998Smarkm 5921109998Smarkm *) Fix for bug in DirectoryString mask setting. Add support for 5922109998Smarkm X509_NAME_print_ex() in 'req' and X509_print_ex() function 5923109998Smarkm to allow certificate printing to more controllable, additional 5924109998Smarkm 'certopt' option to 'x509' to allow new printing options to be 5925109998Smarkm set. 5926109998Smarkm [Steve Henson] 5927109998Smarkm 5928109998Smarkm *) Clean old EAY MD5 hack from e_os.h. 5929109998Smarkm [Richard Levitte] 5930109998Smarkm 5931142425Snectar Changes between 0.9.6l and 0.9.6m [17 Mar 2004] 5932142425Snectar 5933142425Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 5934160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 5935142425Snectar [Joe Orton, Steve Henson] 5936142425Snectar 5937142425Snectar Changes between 0.9.6k and 0.9.6l [04 Nov 2003] 5938142425Snectar 5939142425Snectar *) Fix additional bug revealed by the NISCC test suite: 5940142425Snectar 5941142425Snectar Stop bug triggering large recursion when presented with 5942160814Ssimon certain ASN.1 tags (CVE-2003-0851) 5943142425Snectar [Steve Henson] 5944142425Snectar 5945120631Snectar Changes between 0.9.6j and 0.9.6k [30 Sep 2003] 5946120631Snectar 5947120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 5948120631Snectar 5949120631Snectar Stop out of bounds reads in the ASN1 code when presented with 5950160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 5951120631Snectar 5952120631Snectar If verify callback ignores invalid public key errors don't try to check 5953120631Snectar certificate signature with the NULL public key. 5954120631Snectar 5955120631Snectar [Steve Henson] 5956120631Snectar 5957120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 5958120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 5959120631Snectar specifications. 5960120631Snectar [Steve Henson] 5961120631Snectar 5962120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 5963120631Snectar extra data after the compression methods not only for TLS 1.0 5964120631Snectar but also for SSL 3.0 (as required by the specification). 5965120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 5966120631Snectar 5967120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 5968120631Snectar when it's 512 *bits* long, not 512 bytes. 5969120631Snectar [Richard Levitte] 5970120631Snectar 5971120631Snectar Changes between 0.9.6i and 0.9.6j [10 Apr 2003] 5972120631Snectar 5973120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 5974120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 5975120631Snectar a protocol version number mismatch like a decryption error 5976120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 5977120631Snectar [Bodo Moeller] 5978120631Snectar 5979120631Snectar *) Turn on RSA blinding by default in the default implementation 5980120631Snectar to avoid a timing attack. Applications that don't want it can call 5981120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 5982120631Snectar They would be ill-advised to do so in most cases. 5983120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 5984120631Snectar 5985120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 5986120631Snectar seeded (in this case, the secret RSA exponent is abused as 5987120631Snectar an unpredictable seed -- if it is not unpredictable, there 5988120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 5989120631Snectar by remembering the creator's thread ID in rsa->blinding and 5990120631Snectar having all other threads use local one-time blinding factors 5991120631Snectar (this requires more computation than sharing rsa->blinding, but 5992120631Snectar avoids excessive locking; and if an RSA object is not shared 5993120631Snectar between threads, blinding will still be very fast). 5994120631Snectar [Bodo Moeller] 5995120631Snectar 5996111147Snectar Changes between 0.9.6h and 0.9.6i [19 Feb 2003] 5997111147Snectar 5998111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 5999111147Snectar via timing by performing a MAC computation even if incorrrect 6000111147Snectar block cipher padding has been found. This is a countermeasure 6001111147Snectar against active attacks where the attacker has to distinguish 6002160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 6003111147Snectar 6004111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 6005111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 6006111147Snectar Martin Vuagnoux (EPFL, Ilion)] 6007111147Snectar 6008109998Smarkm Changes between 0.9.6g and 0.9.6h [5 Dec 2002] 6009109998Smarkm 6010109998Smarkm *) New function OPENSSL_cleanse(), which is used to cleanse a section of 6011109998Smarkm memory from it's contents. This is done with a counter that will 6012109998Smarkm place alternating values in each byte. This can be used to solve 6013109998Smarkm two issues: 1) the removal of calls to memset() by highly optimizing 6014109998Smarkm compilers, and 2) cleansing with other values than 0, since those can 6015109998Smarkm be read through on certain media, for example a swap space on disk. 6016109998Smarkm [Geoff Thorpe] 6017109998Smarkm 6018109998Smarkm *) Bugfix: client side session caching did not work with external caching, 6019109998Smarkm because the session->cipher setting was not restored when reloading 6020109998Smarkm from the external cache. This problem was masked, when 6021109998Smarkm SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. 6022109998Smarkm (Found by Steve Haslam <steve@araqnid.ddts.net>.) 6023109998Smarkm [Lutz Jaenicke] 6024109998Smarkm 6025109998Smarkm *) Fix client_certificate (ssl/s2_clnt.c): The permissible total 6026109998Smarkm length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. 6027109998Smarkm [Zeev Lieber <zeev-l@yahoo.com>] 6028109998Smarkm 6029109998Smarkm *) Undo an undocumented change introduced in 0.9.6e which caused 6030109998Smarkm repeated calls to OpenSSL_add_all_ciphers() and 6031109998Smarkm OpenSSL_add_all_digests() to be ignored, even after calling 6032109998Smarkm EVP_cleanup(). 6033109998Smarkm [Richard Levitte] 6034109998Smarkm 6035109998Smarkm *) Change the default configuration reader to deal with last line not 6036109998Smarkm being properly terminated. 6037109998Smarkm [Richard Levitte] 6038109998Smarkm 6039109998Smarkm *) Change X509_NAME_cmp() so it applies the special rules on handling 6040109998Smarkm DN values that are of type PrintableString, as well as RDNs of type 6041109998Smarkm emailAddress where the value has the type ia5String. 6042109998Smarkm [stefank@valicert.com via Richard Levitte] 6043109998Smarkm 6044109998Smarkm *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half 6045109998Smarkm the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently 6046109998Smarkm doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be 6047109998Smarkm the bitwise-OR of the two for use by the majority of applications 6048109998Smarkm wanting this behaviour, and update the docs. The documented 6049109998Smarkm behaviour and actual behaviour were inconsistent and had been 6050109998Smarkm changing anyway, so this is more a bug-fix than a behavioural 6051109998Smarkm change. 6052109998Smarkm [Geoff Thorpe, diagnosed by Nadav Har'El] 6053109998Smarkm 6054109998Smarkm *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c 6055109998Smarkm (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). 6056109998Smarkm [Bodo Moeller] 6057109998Smarkm 6058109998Smarkm *) Fix initialization code race conditions in 6059109998Smarkm SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), 6060109998Smarkm SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), 6061109998Smarkm SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), 6062109998Smarkm TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), 6063109998Smarkm ssl2_get_cipher_by_char(), 6064109998Smarkm ssl3_get_cipher_by_char(). 6065109998Smarkm [Patrick McCormick <patrick@tellme.com>, Bodo Moeller] 6066109998Smarkm 6067109998Smarkm *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after 6068109998Smarkm the cached sessions are flushed, as the remove_cb() might use ex_data 6069109998Smarkm contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> 6070109998Smarkm (see [openssl.org #212]). 6071109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 6072109998Smarkm 6073109998Smarkm *) Fix typo in OBJ_txt2obj which incorrectly passed the content 6074109998Smarkm length, instead of the encoding length to d2i_ASN1_OBJECT. 6075109998Smarkm [Steve Henson] 6076109998Smarkm 6077101618Snectar Changes between 0.9.6f and 0.9.6g [9 Aug 2002] 6078101618Snectar 6079101618Snectar *) [In 0.9.6g-engine release:] 6080101618Snectar Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall'). 6081101618Snectar [Lynn Gazis <lgazis@rainbow.com>] 6082101618Snectar 6083101613Snectar Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 6084101613Snectar 6085101613Snectar *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 6086101613Snectar and get fix the header length calculation. 6087101613Snectar [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 6088101613Snectar Alon Kantor <alonk@checkpoint.com> (and others), 6089101613Snectar Steve Henson] 6090101613Snectar 6091101613Snectar *) Use proper error handling instead of 'assertions' in buffer 6092101613Snectar overflow checks added in 0.9.6e. This prevents DoS (the 6093101613Snectar assertions could call abort()). 6094101613Snectar [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] 6095101613Snectar 6096100936Snectar Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 6097100936Snectar 6098109998Smarkm *) Add various sanity checks to asn1_get_length() to reject 6099109998Smarkm the ASN1 length bytes if they exceed sizeof(long), will appear 6100109998Smarkm negative or the content length exceeds the length of the 6101109998Smarkm supplied buffer. 6102109998Smarkm [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 6103109998Smarkm 6104100936Snectar *) Fix cipher selection routines: ciphers without encryption had no flags 6105100936Snectar for the cipher strength set and where therefore not handled correctly 6106100936Snectar by the selection routines (PR #130). 6107100936Snectar [Lutz Jaenicke] 6108100936Snectar 6109100936Snectar *) Fix EVP_dsa_sha macro. 6110100936Snectar [Nils Larsch] 6111100936Snectar 6112100936Snectar *) New option 6113100936Snectar SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 6114100936Snectar for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 6115100936Snectar that was added in OpenSSL 0.9.6d. 6116100936Snectar 6117100936Snectar As the countermeasure turned out to be incompatible with some 6118100936Snectar broken SSL implementations, the new option is part of SSL_OP_ALL. 6119100936Snectar SSL_OP_ALL is usually employed when compatibility with weird SSL 6120100936Snectar implementations is desired (e.g. '-bugs' option to 's_client' and 6121100936Snectar 's_server'), so the new option is automatically set in many 6122100936Snectar applications. 6123100936Snectar [Bodo Moeller] 6124100936Snectar 6125100936Snectar *) Changes in security patch: 6126100936Snectar 6127100936Snectar Changes marked "(CHATS)" were sponsored by the Defense Advanced 6128100936Snectar Research Projects Agency (DARPA) and Air Force Research Laboratory, 6129100936Snectar Air Force Materiel Command, USAF, under agreement number 6130100936Snectar F30602-01-2-0537. 6131100936Snectar 6132100936Snectar *) Add various sanity checks to asn1_get_length() to reject 6133100936Snectar the ASN1 length bytes if they exceed sizeof(long), will appear 6134100936Snectar negative or the content length exceeds the length of the 6135160814Ssimon supplied buffer. (CVE-2002-0659) 6136100936Snectar [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 6137100936Snectar 6138100936Snectar *) Assertions for various potential buffer overflows, not known to 6139100936Snectar happen in practice. 6140100936Snectar [Ben Laurie (CHATS)] 6141100936Snectar 6142100936Snectar *) Various temporary buffers to hold ASCII versions of integers were 6143160814Ssimon too small for 64 bit platforms. (CVE-2002-0655) 6144100936Snectar [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 6145100936Snectar 6146100936Snectar *) Remote buffer overflow in SSL3 protocol - an attacker could 6147160814Ssimon supply an oversized session ID to a client. (CVE-2002-0656) 6148100936Snectar [Ben Laurie (CHATS)] 6149100936Snectar 6150100936Snectar *) Remote buffer overflow in SSL2 protocol - an attacker could 6151160814Ssimon supply an oversized client master key. (CVE-2002-0656) 6152100936Snectar [Ben Laurie (CHATS)] 6153100936Snectar 6154100928Snectar Changes between 0.9.6c and 0.9.6d [9 May 2002] 6155100928Snectar 6156100928Snectar *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 6157100928Snectar encoded as NULL) with id-dsa-with-sha1. 6158100928Snectar [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 6159100928Snectar 6160100928Snectar *) Check various X509_...() return values in apps/req.c. 6161100928Snectar [Nils Larsch <nla@trustcenter.de>] 6162100928Snectar 6163100928Snectar *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 6164100928Snectar an end-of-file condition would erronously be flagged, when the CRLF 6165100928Snectar was just at the end of a processed block. The bug was discovered when 6166100928Snectar processing data through a buffering memory BIO handing the data to a 6167100928Snectar BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 6168100928Snectar <ptsekov@syntrex.com> and Nedelcho Stanev. 6169100928Snectar [Lutz Jaenicke] 6170100928Snectar 6171100928Snectar *) Implement a countermeasure against a vulnerability recently found 6172100928Snectar in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 6173100928Snectar before application data chunks to avoid the use of known IVs 6174100928Snectar with data potentially chosen by the attacker. 6175100928Snectar [Bodo Moeller] 6176100928Snectar 6177100928Snectar *) Fix length checks in ssl3_get_client_hello(). 6178100928Snectar [Bodo Moeller] 6179100928Snectar 6180100928Snectar *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 6181100928Snectar to prevent ssl3_read_internal() from incorrectly assuming that 6182100928Snectar ssl3_read_bytes() found application data while handshake 6183100928Snectar processing was enabled when in fact s->s3->in_read_app_data was 6184100928Snectar merely automatically cleared during the initial handshake. 6185100928Snectar [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 6186100928Snectar 6187100928Snectar *) Fix object definitions for Private and Enterprise: they were not 6188100928Snectar recognized in their shortname (=lowercase) representation. Extend 6189100928Snectar obj_dat.pl to issue an error when using undefined keywords instead 6190100928Snectar of silently ignoring the problem (Svenning Sorensen 6191100928Snectar <sss@sss.dnsalias.net>). 6192100928Snectar [Lutz Jaenicke] 6193100928Snectar 6194100928Snectar *) Fix DH_generate_parameters() so that it works for 'non-standard' 6195100928Snectar generators, i.e. generators other than 2 and 5. (Previously, the 6196100928Snectar code did not properly initialise the 'add' and 'rem' values to 6197100928Snectar BN_generate_prime().) 6198100928Snectar 6199100928Snectar In the new general case, we do not insist that 'generator' is 6200100928Snectar actually a primitive root: This requirement is rather pointless; 6201100928Snectar a generator of the order-q subgroup is just as good, if not 6202100928Snectar better. 6203100928Snectar [Bodo Moeller] 6204100928Snectar 6205100928Snectar *) Map new X509 verification errors to alerts. Discovered and submitted by 6206100928Snectar Tom Wu <tom@arcot.com>. 6207100928Snectar [Lutz Jaenicke] 6208100928Snectar 6209100928Snectar *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 6210100928Snectar returning non-zero before the data has been completely received 6211100928Snectar when using non-blocking I/O. 6212100928Snectar [Bodo Moeller; problem pointed out by John Hughes] 6213100928Snectar 6214100928Snectar *) Some of the ciphers missed the strength entry (SSL_LOW etc). 6215100928Snectar [Ben Laurie, Lutz Jaenicke] 6216100928Snectar 6217100928Snectar *) Fix bug in SSL_clear(): bad sessions were not removed (found by 6218100928Snectar Yoram Zahavi <YoramZ@gilian.com>). 6219100928Snectar [Lutz Jaenicke] 6220100928Snectar 6221100928Snectar *) Add information about CygWin 1.3 and on, and preserve proper 6222100928Snectar configuration for the versions before that. 6223100928Snectar [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 6224100928Snectar 6225100928Snectar *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 6226100928Snectar check whether we deal with a copy of a session and do not delete from 6227100928Snectar the cache in this case. Problem reported by "Izhar Shoshani Levi" 6228100928Snectar <izhar@checkpoint.com>. 6229100928Snectar [Lutz Jaenicke] 6230100928Snectar 6231100928Snectar *) Do not store session data into the internal session cache, if it 6232100928Snectar is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 6233100928Snectar flag is set). Proposed by Aslam <aslam@funk.com>. 6234100928Snectar [Lutz Jaenicke] 6235100928Snectar 6236100928Snectar *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 6237100928Snectar value is 0. 6238100928Snectar [Richard Levitte] 6239100928Snectar 6240109998Smarkm *) [In 0.9.6d-engine release:] 6241109998Smarkm Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 6242100928Snectar [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 6243100928Snectar 6244100928Snectar *) Add the configuration target linux-s390x. 6245100928Snectar [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 6246100928Snectar 6247100928Snectar *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 6248100928Snectar ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 6249100928Snectar variable as an indication that a ClientHello message has been 6250100928Snectar received. As the flag value will be lost between multiple 6251100928Snectar invocations of ssl3_accept when using non-blocking I/O, the 6252100928Snectar function may not be aware that a handshake has actually taken 6253100928Snectar place, thus preventing a new session from being added to the 6254100928Snectar session cache. 6255100928Snectar 6256100928Snectar To avoid this problem, we now set s->new_session to 2 instead of 6257100928Snectar using a local variable. 6258100928Snectar [Lutz Jaenicke, Bodo Moeller] 6259100928Snectar 6260100928Snectar *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 6261100928Snectar if the SSL_R_LENGTH_MISMATCH error is detected. 6262100928Snectar [Geoff Thorpe, Bodo Moeller] 6263100928Snectar 6264100928Snectar *) New 'shared_ldflag' column in Configure platform table. 6265100928Snectar [Richard Levitte] 6266100928Snectar 6267100928Snectar *) Fix EVP_CIPHER_mode macro. 6268100928Snectar ["Dan S. Camper" <dan@bti.net>] 6269100928Snectar 6270100928Snectar *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 6271100928Snectar type, we must throw them away by setting rr->length to 0. 6272100928Snectar [D P Chang <dpc@qualys.com>] 6273100928Snectar 627489837Skris Changes between 0.9.6b and 0.9.6c [21 dec 2001] 627589837Skris 627689837Skris *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 627789837Skris <Dominikus.Scherkl@biodata.com>. (The previous implementation 627889837Skris worked incorrectly for those cases where range = 10..._2 and 627989837Skris 3*range is two bits longer than range.) 628089837Skris [Bodo Moeller] 628189837Skris 628289837Skris *) Only add signing time to PKCS7 structures if it is not already 628389837Skris present. 628489837Skris [Steve Henson] 628589837Skris 628689837Skris *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 628789837Skris OBJ_ld_ce should be OBJ_id_ce. 628889837Skris Also some ip-pda OIDs in crypto/objects/objects.txt were 628989837Skris incorrect (cf. RFC 3039). 629089837Skris [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 629189837Skris 629289837Skris *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 629389837Skris returns early because it has nothing to do. 629489837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 629589837Skris 629689837Skris *) [In 0.9.6c-engine release:] 629789837Skris Fix mutex callback return values in crypto/engine/hw_ncipher.c. 629889837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 629989837Skris 630089837Skris *) [In 0.9.6c-engine release:] 630189837Skris Add support for Cryptographic Appliance's keyserver technology. 630289837Skris (Use engine 'keyclient') 630389837Skris [Cryptographic Appliances and Geoff Thorpe] 630489837Skris 630589837Skris *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 630689837Skris is called via tools/c89.sh because arguments have to be 630789837Skris rearranged (all '-L' options must appear before the first object 630889837Skris modules). 630989837Skris [Richard Shapiro <rshapiro@abinitio.com>] 631089837Skris 631189837Skris *) [In 0.9.6c-engine release:] 631289837Skris Add support for Broadcom crypto accelerator cards, backported 631389837Skris from 0.9.7. 631489837Skris [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 631589837Skris 631689837Skris *) [In 0.9.6c-engine release:] 631789837Skris Add support for SureWare crypto accelerator cards from 631889837Skris Baltimore Technologies. (Use engine 'sureware') 631989837Skris [Baltimore Technologies and Mark Cox] 632089837Skris 632189837Skris *) [In 0.9.6c-engine release:] 632289837Skris Add support for crypto accelerator cards from Accelerated 632389837Skris Encryption Processing, www.aep.ie. (Use engine 'aep') 632489837Skris [AEP Inc. and Mark Cox] 632589837Skris 632689837Skris *) Add a configuration entry for gcc on UnixWare. 632789837Skris [Gary Benson <gbenson@redhat.com>] 632889837Skris 632989837Skris *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 633089837Skris messages are stored in a single piece (fixed-length part and 633189837Skris variable-length part combined) and fix various bugs found on the way. 633289837Skris [Bodo Moeller] 633389837Skris 633489837Skris *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 633589837Skris instead. BIO_gethostbyname() does not know what timeouts are 633689837Skris appropriate, so entries would stay in cache even when they have 633789837Skris become invalid. 633889837Skris [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 633989837Skris 634089837Skris *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 634189837Skris faced with a pathologically small ClientHello fragment that does 634289837Skris not contain client_version: Instead of aborting with an error, 634389837Skris simply choose the highest available protocol version (i.e., 634489837Skris TLS 1.0 unless it is disabled). In practice, ClientHello 634589837Skris messages are never sent like this, but this change gives us 634689837Skris strictly correct behaviour at least for TLS. 634789837Skris [Bodo Moeller] 634889837Skris 634989837Skris *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 635089837Skris never resets s->method to s->ctx->method when called from within 635189837Skris one of the SSL handshake functions. 635289837Skris [Bodo Moeller; problem pointed out by Niko Baric] 635389837Skris 635489837Skris *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 635589837Skris (sent using the client's version number) if client_version is 635689837Skris smaller than the protocol version in use. Also change 635789837Skris ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 635889837Skris the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 635989837Skris the client will at least see that alert. 636089837Skris [Bodo Moeller] 636189837Skris 636289837Skris *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 636389837Skris correctly. 636489837Skris [Bodo Moeller] 636589837Skris 636689837Skris *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 636789837Skris client receives HelloRequest while in a handshake. 636889837Skris [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 636989837Skris 637089837Skris *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 637189837Skris should end in 'break', not 'goto end' which circuments various 637289837Skris cleanups done in state SSL_ST_OK. But session related stuff 637389837Skris must be disabled for SSL_ST_OK in the case that we just sent a 637489837Skris HelloRequest. 637589837Skris 637689837Skris Also avoid some overhead by not calling ssl_init_wbio_buffer() 637789837Skris before just sending a HelloRequest. 637889837Skris [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 637989837Skris 638089837Skris *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 638189837Skris reveal whether illegal block cipher padding was found or a MAC 638289837Skris verification error occured. (Neither SSLerr() codes nor alerts 638389837Skris are directly visible to potential attackers, but the information 638489837Skris may leak via logfiles.) 638589837Skris 638689837Skris Similar changes are not required for the SSL 2.0 implementation 638789837Skris because the number of padding bytes is sent in clear for SSL 2.0, 638889837Skris and the extra bytes are just ignored. However ssl/s2_pkt.c 638989837Skris failed to verify that the purported number of padding bytes is in 639089837Skris the legal range. 639189837Skris [Bodo Moeller] 639289837Skris 639389837Skris *) Add OpenUNIX-8 support including shared libraries 639489837Skris (Boyd Lynn Gerber <gerberb@zenez.com>). 639589837Skris [Lutz Jaenicke] 639689837Skris 639789837Skris *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 639889837Skris 'wristwatch attack' using huge encoding parameters (cf. 639989837Skris James H. Manger's CRYPTO 2001 paper). Note that the 640089837Skris RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 640189837Skris encoding parameters and hence was not vulnerable. 640289837Skris [Bodo Moeller] 640389837Skris 640489837Skris *) BN_sqr() bug fix. 640589837Skris [Ulf M�ller, reported by Jim Ellis <jim.ellis@cavium.com>] 640689837Skris 640789837Skris *) Rabin-Miller test analyses assume uniformly distributed witnesses, 640889837Skris so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 640989837Skris followed by modular reduction. 641089837Skris [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 641189837Skris 641289837Skris *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 641389837Skris equivalent based on BN_pseudo_rand() instead of BN_rand(). 641489837Skris [Bodo Moeller] 641589837Skris 641689837Skris *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 641789837Skris This function was broken, as the check for a new client hello message 641889837Skris to handle SGC did not allow these large messages. 641989837Skris (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 642089837Skris [Lutz Jaenicke] 642189837Skris 642289837Skris *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 642389837Skris [Lutz Jaenicke] 642489837Skris 642589837Skris *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 642689837Skris for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 642789837Skris [Lutz Jaenicke] 642889837Skris 642989837Skris *) Rework the configuration and shared library support for Tru64 Unix. 643089837Skris The configuration part makes use of modern compiler features and 643189837Skris still retains old compiler behavior for those that run older versions 643289837Skris of the OS. The shared library support part includes a variant that 643389837Skris uses the RPATH feature, and is available through the special 643489837Skris configuration target "alpha-cc-rpath", which will never be selected 643589837Skris automatically. 643689837Skris [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 643789837Skris 643889837Skris *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 643989837Skris with the same message size as in ssl3_get_certificate_request(). 644089837Skris Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 644189837Skris messages might inadvertently be reject as too long. 644289837Skris [Petr Lampa <lampa@fee.vutbr.cz>] 644389837Skris 644489837Skris *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 644589837Skris [Andy Polyakov] 644689837Skris 644789837Skris *) Modified SSL library such that the verify_callback that has been set 644889837Skris specificly for an SSL object with SSL_set_verify() is actually being 644989837Skris used. Before the change, a verify_callback set with this function was 645089837Skris ignored and the verify_callback() set in the SSL_CTX at the time of 645189837Skris the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 645289837Skris to allow the necessary settings. 645389837Skris [Lutz Jaenicke] 645489837Skris 645589837Skris *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 645689837Skris explicitly to NULL, as at least on Solaris 8 this seems not always to be 645789837Skris done automatically (in contradiction to the requirements of the C 645889837Skris standard). This made problems when used from OpenSSH. 645989837Skris [Lutz Jaenicke] 646089837Skris 646189837Skris *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 646289837Skris dh->length and always used 646389837Skris 646489837Skris BN_rand_range(priv_key, dh->p). 646589837Skris 646689837Skris BN_rand_range() is not necessary for Diffie-Hellman, and this 646789837Skris specific range makes Diffie-Hellman unnecessarily inefficient if 646889837Skris dh->length (recommended exponent length) is much smaller than the 646989837Skris length of dh->p. We could use BN_rand_range() if the order of 647089837Skris the subgroup was stored in the DH structure, but we only have 647189837Skris dh->length. 647289837Skris 647389837Skris So switch back to 647489837Skris 647589837Skris BN_rand(priv_key, l, ...) 647689837Skris 647789837Skris where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 647889837Skris otherwise. 647989837Skris [Bodo Moeller] 648089837Skris 648189837Skris *) In 648289837Skris 648389837Skris RSA_eay_public_encrypt 648489837Skris RSA_eay_private_decrypt 648589837Skris RSA_eay_private_encrypt (signing) 648689837Skris RSA_eay_public_decrypt (signature verification) 648789837Skris 648889837Skris (default implementations for RSA_public_encrypt, 648989837Skris RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 649089837Skris always reject numbers >= n. 649189837Skris [Bodo Moeller] 649289837Skris 649389837Skris *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 649489837Skris to synchronize access to 'locking_thread'. This is necessary on 649589837Skris systems where access to 'locking_thread' (an 'unsigned long' 649689837Skris variable) is not atomic. 649789837Skris [Bodo Moeller] 649889837Skris 649989837Skris *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 650089837Skris *before* setting the 'crypto_lock_rand' flag. The previous code had 650189837Skris a race condition if 0 is a valid thread ID. 650289837Skris [Travis Vitek <vitek@roguewave.com>] 650389837Skris 650489837Skris *) Add support for shared libraries under Irix. 650589837Skris [Albert Chin-A-Young <china@thewrittenword.com>] 650689837Skris 650789837Skris *) Add configuration option to build on Linux on both big-endian and 650889837Skris little-endian MIPS. 650989837Skris [Ralf Baechle <ralf@uni-koblenz.de>] 651089837Skris 651189837Skris *) Add the possibility to create shared libraries on HP-UX. 651289837Skris [Richard Levitte] 651389837Skris 651479998Skris Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 651579998Skris 651679998Skris *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 651779998Skris to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 651879998Skris Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 651979998Skris PRNG state recovery was possible based on the output of 652079998Skris one PRNG request appropriately sized to gain knowledge on 652179998Skris 'md' followed by enough consecutive 1-byte PRNG requests 652279998Skris to traverse all of 'state'. 652379998Skris 652479998Skris 1. When updating 'md_local' (the current thread's copy of 'md') 652579998Skris during PRNG output generation, hash all of the previous 652679998Skris 'md_local' value, not just the half used for PRNG output. 652779998Skris 652879998Skris 2. Make the number of bytes from 'state' included into the hash 652979998Skris independent from the number of PRNG bytes requested. 653079998Skris 653179998Skris The first measure alone would be sufficient to avoid 653279998Skris Markku-Juhani's attack. (Actually it had never occurred 653379998Skris to me that the half of 'md_local' used for chaining was the 653479998Skris half from which PRNG output bytes were taken -- I had always 653579998Skris assumed that the secret half would be used.) The second 653679998Skris measure makes sure that additional data from 'state' is never 653779998Skris mixed into 'md_local' in small portions; this heuristically 653879998Skris further strengthens the PRNG. 653979998Skris [Bodo Moeller] 654079998Skris 654179998Skris *) Fix crypto/bn/asm/mips3.s. 654279998Skris [Andy Polyakov] 654379998Skris 654479998Skris *) When only the key is given to "enc", the IV is undefined. Print out 654579998Skris an error message in this case. 654679998Skris [Lutz Jaenicke] 654779998Skris 654879998Skris *) Handle special case when X509_NAME is empty in X509 printing routines. 654979998Skris [Steve Henson] 655079998Skris 655179998Skris *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 655279998Skris positive and less than q. 655379998Skris [Bodo Moeller] 655479998Skris 655579998Skris *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 655679998Skris used: it isn't thread safe and the add_lock_callback should handle 655779998Skris that itself. 655879998Skris [Paul Rose <Paul.Rose@bridge.com>] 655979998Skris 656079998Skris *) Verify that incoming data obeys the block size in 656179998Skris ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 656279998Skris [Bodo Moeller] 656379998Skris 656479998Skris *) Fix OAEP check. 656579998Skris [Ulf M�ller, Bodo M�ller] 656679998Skris 656779998Skris *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 656889837Skris RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 656979998Skris when fixing the server behaviour for backwards-compatible 'client 657079998Skris hello' messages. (Note that the attack is impractical against 657179998Skris SSL 3.0 and TLS 1.0 anyway because length and version checking 657279998Skris means that the probability of guessing a valid ciphertext is 657379998Skris around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 657479998Skris paper.) 657579998Skris 657679998Skris Before 0.9.5, the countermeasure (hide the error by generating a 657779998Skris random 'decryption result') did not work properly because 657879998Skris ERR_clear_error() was missing, meaning that SSL_get_error() would 657979998Skris detect the supposedly ignored error. 658079998Skris 658179998Skris Both problems are now fixed. 658279998Skris [Bodo Moeller] 658379998Skris 658479998Skris *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 658579998Skris (previously it was 1024). 658679998Skris [Bodo Moeller] 658779998Skris 658879998Skris *) Fix for compatibility mode trust settings: ignore trust settings 658979998Skris unless some valid trust or reject settings are present. 659079998Skris [Steve Henson] 659179998Skris 659279998Skris *) Fix for blowfish EVP: its a variable length cipher. 659379998Skris [Steve Henson] 659479998Skris 659579998Skris *) Fix various bugs related to DSA S/MIME verification. Handle missing 659679998Skris parameters in DSA public key structures and return an error in the 659779998Skris DSA routines if parameters are absent. 659879998Skris [Steve Henson] 659979998Skris 660079998Skris *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 660179998Skris in the current directory if neither $RANDFILE nor $HOME was set. 660279998Skris RAND_file_name() in 0.9.6a returned NULL in this case. This has 660379998Skris caused some confusion to Windows users who haven't defined $HOME. 660479998Skris Thus RAND_file_name() is changed again: e_os.h can define a 660579998Skris DEFAULT_HOME, which will be used if $HOME is not set. 660679998Skris For Windows, we use "C:"; on other platforms, we still require 660779998Skris environment variables. 660879998Skris 660979998Skris *) Move 'if (!initialized) RAND_poll()' into regions protected by 661079998Skris CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 661179998Skris having multiple threads call RAND_poll() concurrently. 661279998Skris [Bodo Moeller] 661379998Skris 661479998Skris *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 661579998Skris combination of a flag and a thread ID variable. 661679998Skris Otherwise while one thread is in ssleay_rand_bytes (which sets the 661779998Skris flag), *other* threads can enter ssleay_add_bytes without obeying 661889837Skris the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 661979998Skris that they do not hold after the first thread unsets add_do_not_lock). 662079998Skris [Bodo Moeller] 662179998Skris 662279998Skris *) Change bctest again: '-x' expressions are not available in all 662379998Skris versions of 'test'. 662479998Skris [Bodo Moeller] 662579998Skris 662676866Skris Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 662772613Skris 662876866Skris *) Fix a couple of memory leaks in PKCS7_dataDecode() 662976866Skris [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 663076866Skris 663176866Skris *) Change Configure and Makefiles to provide EXE_EXT, which will contain 663276866Skris the default extension for executables, if any. Also, make the perl 663376866Skris scripts that use symlink() to test if it really exists and use "cp" 663476866Skris if it doesn't. All this made OpenSSL compilable and installable in 663576866Skris CygWin. 663676866Skris [Richard Levitte] 663776866Skris 663876866Skris *) Fix for asn1_GetSequence() for indefinite length constructed data. 663976866Skris If SEQUENCE is length is indefinite just set c->slen to the total 664076866Skris amount of data available. 664176866Skris [Steve Henson, reported by shige@FreeBSD.org] 664276866Skris [This change does not apply to 0.9.7.] 664376866Skris 664476866Skris *) Change bctest to avoid here-documents inside command substitution 664576866Skris (workaround for FreeBSD /bin/sh bug). 664676866Skris For compatibility with Ultrix, avoid shell functions (introduced 664776866Skris in the bctest version that searches along $PATH). 664876866Skris [Bodo Moeller] 664976866Skris 665076866Skris *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 665176866Skris with des_encrypt() defined on some operating systems, like Solaris 665276866Skris and UnixWare. 665376866Skris [Richard Levitte] 665476866Skris 665576866Skris *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 665676866Skris On the Importance of Eliminating Errors in Cryptographic 665776866Skris Computations, J. Cryptology 14 (2001) 2, 101-119, 665876866Skris http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 665976866Skris [Ulf Moeller] 666076866Skris 666176866Skris *) MIPS assembler BIGNUM division bug fix. 666276866Skris [Andy Polyakov] 666376866Skris 666476866Skris *) Disabled incorrect Alpha assembler code. 666576866Skris [Richard Levitte] 666676866Skris 666776866Skris *) Fix PKCS#7 decode routines so they correctly update the length 666876866Skris after reading an EOC for the EXPLICIT tag. 666976866Skris [Steve Henson] 667076866Skris [This change does not apply to 0.9.7.] 667176866Skris 667276866Skris *) Fix bug in PKCS#12 key generation routines. This was triggered 667376866Skris if a 3DES key was generated with a 0 initial byte. Include 667476866Skris PKCS12_BROKEN_KEYGEN compilation option to retain the old 667576866Skris (but broken) behaviour. 667676866Skris [Steve Henson] 667776866Skris 667876866Skris *) Enhance bctest to search for a working bc along $PATH and print 667976866Skris it when found. 668076866Skris [Tim Rice <tim@multitalents.net> via Richard Levitte] 668176866Skris 668276866Skris *) Fix memory leaks in err.c: free err_data string if necessary; 668376866Skris don't write to the wrong index in ERR_set_error_data. 668476866Skris [Bodo Moeller] 668576866Skris 668676866Skris *) Implement ssl23_peek (analogous to ssl23_read), which previously 668776866Skris did not exist. 668876866Skris [Bodo Moeller] 668976866Skris 669076866Skris *) Replace rdtsc with _emit statements for VC++ version 5. 669176866Skris [Jeremy Cooper <jeremy@baymoo.org>] 669276866Skris 669376866Skris *) Make it possible to reuse SSLv2 sessions. 669476866Skris [Richard Levitte] 669576866Skris 669676866Skris *) In copy_email() check for >= 0 as a return value for 669776866Skris X509_NAME_get_index_by_NID() since 0 is a valid index. 669876866Skris [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 669976866Skris 670076866Skris *) Avoid coredump with unsupported or invalid public keys by checking if 670176866Skris X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 670276866Skris PKCS7_verify() fails with non detached data. 670376866Skris [Steve Henson] 670476866Skris 670576866Skris *) Don't use getenv in library functions when run as setuid/setgid. 670676866Skris New function OPENSSL_issetugid(). 670776866Skris [Ulf Moeller] 670876866Skris 670976866Skris *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 671076866Skris due to incorrect handling of multi-threading: 671176866Skris 671276866Skris 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 671376866Skris 671476866Skris 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 671576866Skris 671676866Skris 3. Count how many times MemCheck_off() has been called so that 671776866Skris nested use can be treated correctly. This also avoids 671876866Skris inband-signalling in the previous code (which relied on the 671976866Skris assumption that thread ID 0 is impossible). 672076866Skris [Bodo Moeller] 672176866Skris 672276866Skris *) Add "-rand" option also to s_client and s_server. 672376866Skris [Lutz Jaenicke] 672476866Skris 672576866Skris *) Fix CPU detection on Irix 6.x. 672676866Skris [Kurt Hockenbury <khockenb@stevens-tech.edu> and 672776866Skris "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 672876866Skris 672976866Skris *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 673076866Skris was empty. 673176866Skris [Steve Henson] 673276866Skris [This change does not apply to 0.9.7.] 673376866Skris 673476866Skris *) Use the cached encoding of an X509_NAME structure rather than 673576866Skris copying it. This is apparently the reason for the libsafe "errors" 673676866Skris but the code is actually correct. 673776866Skris [Steve Henson] 673876866Skris 673972613Skris *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 674072613Skris Bleichenbacher's DSA attack. 674176866Skris Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 674276866Skris to be set and top=0 forces the highest bit to be set; top=-1 is new 674376866Skris and leaves the highest bit random. 674476866Skris [Ulf Moeller, Bodo Moeller] 674572613Skris 674672613Skris *) In the NCONF_...-based implementations for CONF_... queries 674772613Skris (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 674872613Skris a temporary CONF structure with the data component set to NULL 674972613Skris (which gives segmentation faults in lh_retrieve). 675072613Skris Instead, use NULL for the CONF pointer in CONF_get_string and 675172613Skris CONF_get_number (which may use environment variables) and directly 675272613Skris return NULL from CONF_get_section. 675372613Skris [Bodo Moeller] 675472613Skris 675572613Skris *) Fix potential buffer overrun for EBCDIC. 675672613Skris [Ulf Moeller] 675772613Skris 675872613Skris *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 675972613Skris keyUsage if basicConstraints absent for a CA. 676072613Skris [Steve Henson] 676172613Skris 676272613Skris *) Make SMIME_write_PKCS7() write mail header values with a format that 676372613Skris is more generally accepted (no spaces before the semicolon), since 676472613Skris some programs can't parse those values properly otherwise. Also make 676572613Skris sure BIO's that break lines after each write do not create invalid 676672613Skris headers. 676772613Skris [Richard Levitte] 676872613Skris 676972613Skris *) Make the CRL encoding routines work with empty SEQUENCE OF. The 677072613Skris macros previously used would not encode an empty SEQUENCE OF 677172613Skris and break the signature. 677272613Skris [Steve Henson] 677376866Skris [This change does not apply to 0.9.7.] 677472613Skris 677572613Skris *) Zero the premaster secret after deriving the master secret in 677672613Skris DH ciphersuites. 677772613Skris [Steve Henson] 677872613Skris 677972613Skris *) Add some EVP_add_digest_alias registrations (as found in 678072613Skris OpenSSL_add_all_digests()) to SSL_library_init() 678172613Skris aka OpenSSL_add_ssl_algorithms(). This provides improved 678272613Skris compatibility with peers using X.509 certificates 678372613Skris with unconventional AlgorithmIdentifier OIDs. 678472613Skris [Bodo Moeller] 678572613Skris 678672613Skris *) Fix for Irix with NO_ASM. 678772613Skris ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 678872613Skris 678972613Skris *) ./config script fixes. 679072613Skris [Ulf Moeller, Richard Levitte] 679172613Skris 679272613Skris *) Fix 'openssl passwd -1'. 679372613Skris [Bodo Moeller] 679472613Skris 679572613Skris *) Change PKCS12_key_gen_asc() so it can cope with non null 679672613Skris terminated strings whose length is passed in the passlen 679772613Skris parameter, for example from PEM callbacks. This was done 679872613Skris by adding an extra length parameter to asc2uni(). 679972613Skris [Steve Henson, reported by <oddissey@samsung.co.kr>] 680072613Skris 680172613Skris *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 680272613Skris call failed, free the DSA structure. 680372613Skris [Bodo Moeller] 680472613Skris 680572613Skris *) Fix to uni2asc() to cope with zero length Unicode strings. 680672613Skris These are present in some PKCS#12 files. 680772613Skris [Steve Henson] 680872613Skris 680972613Skris *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 681072613Skris Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 681172613Skris when writing a 32767 byte record. 681272613Skris [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 681372613Skris 681472613Skris *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 681572613Skris obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 681672613Skris 681772613Skris (RSA objects have a reference count access to which is protected 681872613Skris by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 681972613Skris so they are meant to be shared between threads.) 682072613Skris [Bodo Moeller, Geoff Thorpe; original patch submitted by 682172613Skris "Reddie, Steven" <Steven.Reddie@ca.com>] 682272613Skris 682372613Skris *) Fix a deadlock in CRYPTO_mem_leaks(). 682472613Skris [Bodo Moeller] 682572613Skris 682676866Skris *) Use better test patterns in bntest. 682776866Skris [Ulf M�ller] 682876866Skris 682972613Skris *) rand_win.c fix for Borland C. 683072613Skris [Ulf M�ller] 683172613Skris 683272613Skris *) BN_rshift bugfix for n == 0. 683372613Skris [Bodo Moeller] 683472613Skris 683576866Skris *) Add a 'bctest' script that checks for some known 'bc' bugs 683676866Skris so that 'make test' does not abort just because 'bc' is broken. 683776866Skris [Bodo Moeller] 683876866Skris 683972613Skris *) Store verify_result within SSL_SESSION also for client side to 684072613Skris avoid potential security hole. (Re-used sessions on the client side 684172613Skris always resulted in verify_result==X509_V_OK, not using the original 684272613Skris result of the server certificate verification.) 684372613Skris [Lutz Jaenicke] 684472613Skris 684572613Skris *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 684672613Skris SSL3_RT_APPLICATION_DATA, return 0. 684772613Skris Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 684872613Skris [Bodo Moeller] 684972613Skris 685072613Skris *) Fix SSL_peek: 685172613Skris Both ssl2_peek and ssl3_peek, which were totally broken in earlier 685272613Skris releases, have been re-implemented by renaming the previous 685372613Skris implementations of ssl2_read and ssl3_read to ssl2_read_internal 685472613Skris and ssl3_read_internal, respectively, and adding 'peek' parameters 685572613Skris to them. The new ssl[23]_{read,peek} functions are calls to 685672613Skris ssl[23]_read_internal with the 'peek' flag set appropriately. 685772613Skris A 'peek' parameter has also been added to ssl3_read_bytes, which 685872613Skris does the actual work for ssl3_read_internal. 685972613Skris [Bodo Moeller] 686072613Skris 686176866Skris *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 686276866Skris the method-specific "init()" handler. Also clean up ex_data after 686376866Skris calling the method-specific "finish()" handler. Previously, this was 686476866Skris happening the other way round. 686576866Skris [Geoff Thorpe] 686676866Skris 686772613Skris *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 686872613Skris The previous value, 12, was not always sufficient for BN_mod_exp(). 686972613Skris [Bodo Moeller] 687072613Skris 687176866Skris *) Make sure that shared libraries get the internal name engine with 687276866Skris the full version number and not just 0. This should mark the 687376866Skris shared libraries as not backward compatible. Of course, this should 687476866Skris be changed again when we can guarantee backward binary compatibility. 687576866Skris [Richard Levitte] 687676866Skris 687772613Skris *) Fix typo in get_cert_by_subject() in by_dir.c 687872613Skris [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 687972613Skris 688076866Skris *) Rework the system to generate shared libraries: 688176866Skris 688276866Skris - Make note of the expected extension for the shared libraries and 688376866Skris if there is a need for symbolic links from for example libcrypto.so.0 688476866Skris to libcrypto.so.0.9.7. There is extended info in Configure for 688576866Skris that. 688676866Skris 688776866Skris - Make as few rebuilds of the shared libraries as possible. 688876866Skris 688976866Skris - Still avoid linking the OpenSSL programs with the shared libraries. 689076866Skris 689176866Skris - When installing, install the shared libraries separately from the 689276866Skris static ones. 689376866Skris [Richard Levitte] 689476866Skris 689572613Skris *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 689672613Skris 689772613Skris Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 689872613Skris and not in SSL_clear because the latter is also used by the 689972613Skris accept/connect functions; previously, the settings made by 690072613Skris SSL_set_read_ahead would be lost during the handshake. 690172613Skris [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 690272613Skris 690372613Skris *) Correct util/mkdef.pl to be selective about disabled algorithms. 690472613Skris Previously, it would create entries for disableed algorithms no 690572613Skris matter what. 690672613Skris [Richard Levitte] 690772613Skris 690872613Skris *) Added several new manual pages for SSL_* function. 690972613Skris [Lutz Jaenicke] 691072613Skris 691168651Skris Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 691268651Skris 691368651Skris *) In ssl23_get_client_hello, generate an error message when faced 691468651Skris with an initial SSL 3.0/TLS record that is too small to contain the 691568651Skris first two bytes of the ClientHello message, i.e. client_version. 691668651Skris (Note that this is a pathologic case that probably has never happened 691768651Skris in real life.) The previous approach was to use the version number 691868651Skris from the record header as a substitute; but our protocol choice 691968651Skris should not depend on that one because it is not authenticated 692068651Skris by the Finished messages. 692168651Skris [Bodo Moeller] 692268651Skris 692368651Skris *) More robust randomness gathering functions for Windows. 692468651Skris [Jeffrey Altman <jaltman@columbia.edu>] 692568651Skris 692668651Skris *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 692768651Skris not set then we don't setup the error code for issuer check errors 692868651Skris to avoid possibly overwriting other errors which the callback does 692968651Skris handle. If an application does set the flag then we assume it knows 693068651Skris what it is doing and can handle the new informational codes 693168651Skris appropriately. 693268651Skris [Steve Henson] 693368651Skris 693468651Skris *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 693568651Skris a general "ANY" type, as such it should be able to decode anything 693668651Skris including tagged types. However it didn't check the class so it would 693768651Skris wrongly interpret tagged types in the same way as their universal 693868651Skris counterpart and unknown types were just rejected. Changed so that the 693968651Skris tagged and unknown types are handled in the same way as a SEQUENCE: 694068651Skris that is the encoding is stored intact. There is also a new type 694168651Skris "V_ASN1_OTHER" which is used when the class is not universal, in this 694268651Skris case we have no idea what the actual type is so we just lump them all 694368651Skris together. 694468651Skris [Steve Henson] 694568651Skris 694668651Skris *) On VMS, stdout may very well lead to a file that is written to 694768651Skris in a record-oriented fashion. That means that every write() will 694868651Skris write a separate record, which will be read separately by the 694968651Skris programs trying to read from it. This can be very confusing. 695068651Skris 695168651Skris The solution is to put a BIO filter in the way that will buffer 695268651Skris text until a linefeed is reached, and then write everything a 695368651Skris line at a time, so every record written will be an actual line, 695468651Skris not chunks of lines and not (usually doesn't happen, but I've 695568651Skris seen it once) several lines in one record. BIO_f_linebuffer() is 695668651Skris the answer. 695768651Skris 695868651Skris Currently, it's a VMS-only method, because that's where it has 695968651Skris been tested well enough. 696068651Skris [Richard Levitte] 696168651Skris 696268651Skris *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 696368651Skris it can return incorrect results. 696468651Skris (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 696568651Skris but it was in 0.9.6-beta[12].) 696668651Skris [Bodo Moeller] 696768651Skris 696868651Skris *) Disable the check for content being present when verifying detached 696968651Skris signatures in pk7_smime.c. Some versions of Netscape (wrongly) 697068651Skris include zero length content when signing messages. 697168651Skris [Steve Henson] 697268651Skris 697368651Skris *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 697468651Skris BIO_ctrl (for BIO pairs). 697568651Skris [Bodo M�ller] 697668651Skris 697768651Skris *) Add DSO method for VMS. 697868651Skris [Richard Levitte] 697968651Skris 698068651Skris *) Bug fix: Montgomery multiplication could produce results with the 698168651Skris wrong sign. 698268651Skris [Ulf M�ller] 698368651Skris 698468651Skris *) Add RPM specification openssl.spec and modify it to build three 698568651Skris packages. The default package contains applications, application 698668651Skris documentation and run-time libraries. The devel package contains 698768651Skris include files, static libraries and function documentation. The 698868651Skris doc package contains the contents of the doc directory. The original 698968651Skris openssl.spec was provided by Damien Miller <djm@mindrot.org>. 699068651Skris [Richard Levitte] 699168651Skris 699268651Skris *) Add a large number of documentation files for many SSL routines. 699368651Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 699468651Skris 699568651Skris *) Add a configuration entry for Sony News 4. 699668651Skris [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 699768651Skris 699868651Skris *) Don't set the two most significant bits to one when generating a 699968651Skris random number < q in the DSA library. 700068651Skris [Ulf M�ller] 700168651Skris 700268651Skris *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 700368651Skris behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 700468651Skris the underlying transport is blocking) if a handshake took place. 700568651Skris (The default behaviour is needed by applications such as s_client 700668651Skris and s_server that use select() to determine when to use SSL_read; 700768651Skris but for applications that know in advance when to expect data, it 700868651Skris just makes things more complicated.) 700968651Skris [Bodo Moeller] 701068651Skris 701168651Skris *) Add RAND_egd_bytes(), which gives control over the number of bytes read 701268651Skris from EGD. 701368651Skris [Ben Laurie] 701468651Skris 701568651Skris *) Add a few more EBCDIC conditionals that make `req' and `x509' 701668651Skris work better on such systems. 701768651Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 701868651Skris 701968651Skris *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 702068651Skris Update PKCS12_parse() so it copies the friendlyName and the 702168651Skris keyid to the certificates aux info. 702268651Skris [Steve Henson] 702368651Skris 702468651Skris *) Fix bug in PKCS7_verify() which caused an infinite loop 702568651Skris if there was more than one signature. 702668651Skris [Sven Uszpelkat <su@celocom.de>] 702768651Skris 702868651Skris *) Major change in util/mkdef.pl to include extra information 702968651Skris about each symbol, as well as presentig variables as well 703068651Skris as functions. This change means that there's n more need 703168651Skris to rebuild the .num files when some algorithms are excluded. 703268651Skris [Richard Levitte] 703368651Skris 703468651Skris *) Allow the verify time to be set by an application, 703568651Skris rather than always using the current time. 703668651Skris [Steve Henson] 703768651Skris 703868651Skris *) Phase 2 verify code reorganisation. The certificate 703968651Skris verify code now looks up an issuer certificate by a 704068651Skris number of criteria: subject name, authority key id 704168651Skris and key usage. It also verifies self signed certificates 704268651Skris by the same criteria. The main comparison function is 704368651Skris X509_check_issued() which performs these checks. 704468651Skris 704568651Skris Lot of changes were necessary in order to support this 704668651Skris without completely rewriting the lookup code. 704768651Skris 704868651Skris Authority and subject key identifier are now cached. 704968651Skris 705068651Skris The LHASH 'certs' is X509_STORE has now been replaced 705168651Skris by a STACK_OF(X509_OBJECT). This is mainly because an 705268651Skris LHASH can't store or retrieve multiple objects with 705368651Skris the same hash value. 705468651Skris 705568651Skris As a result various functions (which were all internal 705668651Skris use only) have changed to handle the new X509_STORE 705768651Skris structure. This will break anything that messed round 705868651Skris with X509_STORE internally. 705968651Skris 706068651Skris The functions X509_STORE_add_cert() now checks for an 706168651Skris exact match, rather than just subject name. 706268651Skris 706368651Skris The X509_STORE API doesn't directly support the retrieval 706468651Skris of multiple certificates matching a given criteria, however 706568651Skris this can be worked round by performing a lookup first 706668651Skris (which will fill the cache with candidate certificates) 706768651Skris and then examining the cache for matches. This is probably 706868651Skris the best we can do without throwing out X509_LOOKUP 706968651Skris entirely (maybe later...). 707068651Skris 707168651Skris The X509_VERIFY_CTX structure has been enhanced considerably. 707268651Skris 707368651Skris All certificate lookup operations now go via a get_issuer() 707468651Skris callback. Although this currently uses an X509_STORE it 707568651Skris can be replaced by custom lookups. This is a simple way 707668651Skris to bypass the X509_STORE hackery necessary to make this 707768651Skris work and makes it possible to use more efficient techniques 707868651Skris in future. A very simple version which uses a simple 707968651Skris STACK for its trusted certificate store is also provided 708068651Skris using X509_STORE_CTX_trusted_stack(). 708168651Skris 708268651Skris The verify_cb() and verify() callbacks now have equivalents 708368651Skris in the X509_STORE_CTX structure. 708468651Skris 708568651Skris X509_STORE_CTX also has a 'flags' field which can be used 708668651Skris to customise the verify behaviour. 708768651Skris [Steve Henson] 708868651Skris 708968651Skris *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 709068651Skris excludes S/MIME capabilities. 709168651Skris [Steve Henson] 709268651Skris 709368651Skris *) When a certificate request is read in keep a copy of the 709468651Skris original encoding of the signed data and use it when outputing 709568651Skris again. Signatures then use the original encoding rather than 709668651Skris a decoded, encoded version which may cause problems if the 709768651Skris request is improperly encoded. 709868651Skris [Steve Henson] 709968651Skris 710068651Skris *) For consistency with other BIO_puts implementations, call 710168651Skris buffer_write(b, ...) directly in buffer_puts instead of calling 710268651Skris BIO_write(b, ...). 710368651Skris 710468651Skris In BIO_puts, increment b->num_write as in BIO_write. 710568651Skris [Peter.Sylvester@EdelWeb.fr] 710668651Skris 710768651Skris *) Fix BN_mul_word for the case where the word is 0. (We have to use 710868651Skris BN_zero, we may not return a BIGNUM with an array consisting of 710968651Skris words set to zero.) 711068651Skris [Bodo Moeller] 711168651Skris 711268651Skris *) Avoid calling abort() from within the library when problems are 711368651Skris detected, except if preprocessor symbols have been defined 711468651Skris (such as REF_CHECK, BN_DEBUG etc.). 711568651Skris [Bodo Moeller] 711668651Skris 711768651Skris *) New openssl application 'rsautl'. This utility can be 711868651Skris used for low level RSA operations. DER public key 711968651Skris BIO/fp routines also added. 712068651Skris [Steve Henson] 712168651Skris 712268651Skris *) New Configure entry and patches for compiling on QNX 4. 712368651Skris [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 712468651Skris 712568651Skris *) A demo state-machine implementation was sponsored by 712668651Skris Nuron (http://www.nuron.com/) and is now available in 712768651Skris demos/state_machine. 712868651Skris [Ben Laurie] 712968651Skris 713068651Skris *) New options added to the 'dgst' utility for signature 713168651Skris generation and verification. 713268651Skris [Steve Henson] 713368651Skris 713468651Skris *) Unrecognized PKCS#7 content types are now handled via a 713568651Skris catch all ASN1_TYPE structure. This allows unsupported 713668651Skris types to be stored as a "blob" and an application can 713768651Skris encode and decode it manually. 713868651Skris [Steve Henson] 713968651Skris 714068651Skris *) Fix various signed/unsigned issues to make a_strex.c 714168651Skris compile under VC++. 714268651Skris [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 714368651Skris 714468651Skris *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 714568651Skris length if passed a buffer. ASN1_INTEGER_to_BN failed 714668651Skris if passed a NULL BN and its argument was negative. 714768651Skris [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 714868651Skris 714968651Skris *) Modification to PKCS#7 encoding routines to output definite 715068651Skris length encoding. Since currently the whole structures are in 715168651Skris memory there's not real point in using indefinite length 715268651Skris constructed encoding. However if OpenSSL is compiled with 715368651Skris the flag PKCS7_INDEFINITE_ENCODING the old form is used. 715468651Skris [Steve Henson] 715568651Skris 715668651Skris *) Added BIO_vprintf() and BIO_vsnprintf(). 715768651Skris [Richard Levitte] 715868651Skris 715968651Skris *) Added more prefixes to parse for in the the strings written 716068651Skris through a logging bio, to cover all the levels that are available 716168651Skris through syslog. The prefixes are now: 716268651Skris 716368651Skris PANIC, EMERG, EMR => LOG_EMERG 716468651Skris ALERT, ALR => LOG_ALERT 716568651Skris CRIT, CRI => LOG_CRIT 716668651Skris ERROR, ERR => LOG_ERR 716768651Skris WARNING, WARN, WAR => LOG_WARNING 716868651Skris NOTICE, NOTE, NOT => LOG_NOTICE 716968651Skris INFO, INF => LOG_INFO 717068651Skris DEBUG, DBG => LOG_DEBUG 717168651Skris 717268651Skris and as before, if none of those prefixes are present at the 717368651Skris beginning of the string, LOG_ERR is chosen. 717468651Skris 717568651Skris On Win32, the LOG_* levels are mapped according to this: 717668651Skris 717768651Skris LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 717868651Skris LOG_WARNING => EVENTLOG_WARNING_TYPE 717968651Skris LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 718068651Skris 718168651Skris [Richard Levitte] 718268651Skris 718368651Skris *) Made it possible to reconfigure with just the configuration 718468651Skris argument "reconf" or "reconfigure". The command line arguments 718568651Skris are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 718668651Skris and are retrieved from there when reconfiguring. 718768651Skris [Richard Levitte] 718868651Skris 718968651Skris *) MD4 implemented. 719068651Skris [Assar Westerlund <assar@sics.se>, Richard Levitte] 719168651Skris 719268651Skris *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 719368651Skris [Richard Levitte] 719468651Skris 719568651Skris *) The obj_dat.pl script was messing up the sorting of object 719668651Skris names. The reason was that it compared the quoted version 719768651Skris of strings as a result "OCSP" > "OCSP Signing" because 719868651Skris " > SPACE. Changed script to store unquoted versions of 719968651Skris names and add quotes on output. It was also omitting some 720068651Skris names from the lookup table if they were given a default 720168651Skris value (that is if SN is missing it is given the same 720268651Skris value as LN and vice versa), these are now added on the 720368651Skris grounds that if an object has a name we should be able to 720468651Skris look it up. Finally added warning output when duplicate 720568651Skris short or long names are found. 720668651Skris [Steve Henson] 720768651Skris 720868651Skris *) Changes needed for Tandem NSK. 720968651Skris [Scott Uroff <scott@xypro.com>] 721068651Skris 721168651Skris *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 721268651Skris RSA_padding_check_SSLv23(), special padding was never detected 721368651Skris and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 721468651Skris version rollback attacks was not effective. 721568651Skris 721668651Skris In s23_clnt.c, don't use special rollback-attack detection padding 721768651Skris (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 721868651Skris client; similarly, in s23_srvr.c, don't do the rollback check if 721968651Skris SSL 2.0 is the only protocol enabled in the server. 722068651Skris [Bodo Moeller] 722168651Skris 722268651Skris *) Make it possible to get hexdumps of unprintable data with 'openssl 722368651Skris asn1parse'. By implication, the functions ASN1_parse_dump() and 722468651Skris BIO_dump_indent() are added. 722568651Skris [Richard Levitte] 722668651Skris 722768651Skris *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 722868651Skris these print out strings and name structures based on various 722968651Skris flags including RFC2253 support and proper handling of 723068651Skris multibyte characters. Added options to the 'x509' utility 723168651Skris to allow the various flags to be set. 723268651Skris [Steve Henson] 723368651Skris 723468651Skris *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 723568651Skris Also change the functions X509_cmp_current_time() and 723668651Skris X509_gmtime_adj() work with an ASN1_TIME structure, 723768651Skris this will enable certificates using GeneralizedTime in validity 723868651Skris dates to be checked. 723968651Skris [Steve Henson] 724068651Skris 724168651Skris *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 724268651Skris negative public key encodings) on by default, 724368651Skris NO_NEG_PUBKEY_BUG can be set to disable it. 724468651Skris [Steve Henson] 724568651Skris 724668651Skris *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 724768651Skris content octets. An i2c_ASN1_OBJECT is unnecessary because 724868651Skris the encoding can be trivially obtained from the structure. 724968651Skris [Steve Henson] 725068651Skris 725168651Skris *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 725268651Skris not read locks (CRYPTO_r_[un]lock). 725368651Skris [Bodo Moeller] 725468651Skris 725568651Skris *) A first attempt at creating official support for shared 725668651Skris libraries through configuration. I've kept it so the 725768651Skris default is static libraries only, and the OpenSSL programs 725868651Skris are always statically linked for now, but there are 725968651Skris preparations for dynamic linking in place. 726089837Skris This has been tested on Linux and Tru64. 726168651Skris [Richard Levitte] 726268651Skris 726368651Skris *) Randomness polling function for Win9x, as described in: 726468651Skris Peter Gutmann, Software Generation of Practically Strong 726568651Skris Random Numbers. 726668651Skris [Ulf M�ller] 726768651Skris 726868651Skris *) Fix so PRNG is seeded in req if using an already existing 726968651Skris DSA key. 727068651Skris [Steve Henson] 727168651Skris 727268651Skris *) New options to smime application. -inform and -outform 727368651Skris allow alternative formats for the S/MIME message including 727468651Skris PEM and DER. The -content option allows the content to be 727568651Skris specified separately. This should allow things like Netscape 727668651Skris form signing output easier to verify. 727768651Skris [Steve Henson] 727868651Skris 727968651Skris *) Fix the ASN1 encoding of tags using the 'long form'. 728068651Skris [Steve Henson] 728168651Skris 728268651Skris *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 728368651Skris STRING types. These convert content octets to and from the 728468651Skris underlying type. The actual tag and length octets are 728568651Skris already assumed to have been read in and checked. These 728668651Skris are needed because all other string types have virtually 728768651Skris identical handling apart from the tag. By having versions 728868651Skris of the ASN1 functions that just operate on content octets 728968651Skris IMPLICIT tagging can be handled properly. It also allows 729068651Skris the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 729168651Skris and ASN1_INTEGER are identical apart from the tag. 729268651Skris [Steve Henson] 729368651Skris 729468651Skris *) Change the handling of OID objects as follows: 729568651Skris 729668651Skris - New object identifiers are inserted in objects.txt, following 729768651Skris the syntax given in objects.README. 729868651Skris - objects.pl is used to process obj_mac.num and create a new 729968651Skris obj_mac.h. 730068651Skris - obj_dat.pl is used to create a new obj_dat.h, using the data in 730168651Skris obj_mac.h. 730268651Skris 730368651Skris This is currently kind of a hack, and the perl code in objects.pl 730468651Skris isn't very elegant, but it works as I intended. The simplest way 730568651Skris to check that it worked correctly is to look in obj_dat.h and 730668651Skris check the array nid_objs and make sure the objects haven't moved 730768651Skris around (this is important!). Additions are OK, as well as 730868651Skris consistent name changes. 730968651Skris [Richard Levitte] 731068651Skris 731168651Skris *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 731268651Skris [Bodo Moeller] 731368651Skris 731468651Skris *) Addition of the command line parameter '-rand file' to 'openssl req'. 731568651Skris The given file adds to whatever has already been seeded into the 731668651Skris random pool through the RANDFILE configuration file option or 731768651Skris environment variable, or the default random state file. 731868651Skris [Richard Levitte] 731968651Skris 732068651Skris *) mkstack.pl now sorts each macro group into lexical order. 732168651Skris Previously the output order depended on the order the files 732268651Skris appeared in the directory, resulting in needless rewriting 732368651Skris of safestack.h . 732468651Skris [Steve Henson] 732568651Skris 732668651Skris *) Patches to make OpenSSL compile under Win32 again. Mostly 732768651Skris work arounds for the VC++ problem that it treats func() as 732868651Skris func(void). Also stripped out the parts of mkdef.pl that 732968651Skris added extra typesafe functions: these no longer exist. 733068651Skris [Steve Henson] 733168651Skris 733268651Skris *) Reorganisation of the stack code. The macros are now all 733368651Skris collected in safestack.h . Each macro is defined in terms of 733468651Skris a "stack macro" of the form SKM_<name>(type, a, b). The 733568651Skris DEBUG_SAFESTACK is now handled in terms of function casts, 733668651Skris this has the advantage of retaining type safety without the 733768651Skris use of additional functions. If DEBUG_SAFESTACK is not defined 733868651Skris then the non typesafe macros are used instead. Also modified the 733968651Skris mkstack.pl script to handle the new form. Needs testing to see 734068651Skris if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 734168651Skris the default if no major problems. Similar behaviour for ASN1_SET_OF 734268651Skris and PKCS12_STACK_OF. 734368651Skris [Steve Henson] 734468651Skris 734568651Skris *) When some versions of IIS use the 'NET' form of private key the 734668651Skris key derivation algorithm is different. Normally MD5(password) is 734768651Skris used as a 128 bit RC4 key. In the modified case 734868651Skris MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 734968651Skris new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 735068651Skris as the old Netscape_RSA functions except they have an additional 735168651Skris 'sgckey' parameter which uses the modified algorithm. Also added 735268651Skris an -sgckey command line option to the rsa utility. Thanks to 735368651Skris Adrian Peck <bertie@ncipher.com> for posting details of the modified 735468651Skris algorithm to openssl-dev. 735568651Skris [Steve Henson] 735668651Skris 735768651Skris *) The evp_local.h macros were using 'c.##kname' which resulted in 735868651Skris invalid expansion on some systems (SCO 5.0.5 for example). 735968651Skris Corrected to 'c.kname'. 736068651Skris [Phillip Porch <root@theporch.com>] 736168651Skris 736268651Skris *) New X509_get1_email() and X509_REQ_get1_email() functions that return 736368651Skris a STACK of email addresses from a certificate or request, these look 736468651Skris in the subject name and the subject alternative name extensions and 736568651Skris omit any duplicate addresses. 736668651Skris [Steve Henson] 736768651Skris 736868651Skris *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 736968651Skris This makes DSA verification about 2 % faster. 737068651Skris [Bodo Moeller] 737168651Skris 737268651Skris *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 737368651Skris (meaning that now 2^5 values will be precomputed, which is only 4 KB 737468651Skris plus overhead for 1024 bit moduli). 737568651Skris This makes exponentiations about 0.5 % faster for 1024 bit 737668651Skris exponents (as measured by "openssl speed rsa2048"). 737768651Skris [Bodo Moeller] 737868651Skris 737968651Skris *) Rename memory handling macros to avoid conflicts with other 738068651Skris software: 738168651Skris Malloc => OPENSSL_malloc 738268651Skris Malloc_locked => OPENSSL_malloc_locked 738368651Skris Realloc => OPENSSL_realloc 738468651Skris Free => OPENSSL_free 738568651Skris [Richard Levitte] 738668651Skris 738768651Skris *) New function BN_mod_exp_mont_word for small bases (roughly 15% 738868651Skris faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 738968651Skris [Bodo Moeller] 739068651Skris 739168651Skris *) CygWin32 support. 739268651Skris [John Jarvie <jjarvie@newsguy.com>] 739368651Skris 739468651Skris *) The type-safe stack code has been rejigged. It is now only compiled 739568651Skris in when OpenSSL is configured with the DEBUG_SAFESTACK option and 739668651Skris by default all type-specific stack functions are "#define"d back to 739768651Skris standard stack functions. This results in more streamlined output 739868651Skris but retains the type-safety checking possibilities of the original 739968651Skris approach. 740068651Skris [Geoff Thorpe] 740168651Skris 740268651Skris *) The STACK code has been cleaned up, and certain type declarations 740368651Skris that didn't make a lot of sense have been brought in line. This has 740468651Skris also involved a cleanup of sorts in safestack.h to more correctly 740568651Skris map type-safe stack functions onto their plain stack counterparts. 740668651Skris This work has also resulted in a variety of "const"ifications of 740768651Skris lots of the code, especially "_cmp" operations which should normally 740868651Skris be prototyped with "const" parameters anyway. 740968651Skris [Geoff Thorpe] 741068651Skris 741168651Skris *) When generating bytes for the first time in md_rand.c, 'stir the pool' 741268651Skris by seeding with STATE_SIZE dummy bytes (with zero entropy count). 741368651Skris (The PRNG state consists of two parts, the large pool 'state' and 'md', 741468651Skris where all of 'md' is used each time the PRNG is used, but 'state' 741568651Skris is used only indexed by a cyclic counter. As entropy may not be 741668651Skris well distributed from the beginning, 'md' is important as a 741768651Skris chaining variable. However, the output function chains only half 741868651Skris of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 741968651Skris all of 'md', and seeding with STATE_SIZE dummy bytes will result 742068651Skris in all of 'state' being rewritten, with the new values depending 742168651Skris on virtually all of 'md'. This overcomes the 80 bit limitation.) 742268651Skris [Bodo Moeller] 742368651Skris 742468651Skris *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 742568651Skris the handshake is continued after ssl_verify_cert_chain(); 742668651Skris otherwise, if SSL_VERIFY_NONE is set, remaining error codes 742768651Skris can lead to 'unexplainable' connection aborts later. 742868651Skris [Bodo Moeller; problem tracked down by Lutz Jaenicke] 742968651Skris 743068651Skris *) Major EVP API cipher revision. 743168651Skris Add hooks for extra EVP features. This allows various cipher 743268651Skris parameters to be set in the EVP interface. Support added for variable 743368651Skris key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 743468651Skris setting of RC2 and RC5 parameters. 743568651Skris 743668651Skris Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 743768651Skris ciphers. 743868651Skris 743968651Skris Remove lots of duplicated code from the EVP library. For example *every* 744068651Skris cipher init() function handles the 'iv' in the same way according to the 744168651Skris cipher mode. They also all do nothing if the 'key' parameter is NULL and 744268651Skris for CFB and OFB modes they zero ctx->num. 744368651Skris 744468651Skris New functionality allows removal of S/MIME code RC2 hack. 744568651Skris 744668651Skris Most of the routines have the same form and so can be declared in terms 744768651Skris of macros. 744868651Skris 744968651Skris By shifting this to the top level EVP_CipherInit() it can be removed from 745068651Skris all individual ciphers. If the cipher wants to handle IVs or keys 745168651Skris differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 745268651Skris flags. 745368651Skris 745468651Skris Change lots of functions like EVP_EncryptUpdate() to now return a 745568651Skris value: although software versions of the algorithms cannot fail 745668651Skris any installed hardware versions can. 745768651Skris [Steve Henson] 745868651Skris 745968651Skris *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 746068651Skris this option is set, tolerate broken clients that send the negotiated 746168651Skris protocol version number instead of the requested protocol version 746268651Skris number. 746368651Skris [Bodo Moeller] 746468651Skris 746568651Skris *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 746668651Skris i.e. non-zero for export ciphersuites, zero otherwise. 746768651Skris Previous versions had this flag inverted, inconsistent with 746868651Skris rsa_tmp_cb (..._TMP_RSA_CB). 746968651Skris [Bodo Moeller; problem reported by Amit Chopra] 747068651Skris 747168651Skris *) Add missing DSA library text string. Work around for some IIS 747268651Skris key files with invalid SEQUENCE encoding. 747368651Skris [Steve Henson] 747468651Skris 747568651Skris *) Add a document (doc/standards.txt) that list all kinds of standards 747668651Skris and so on that are implemented in OpenSSL. 747768651Skris [Richard Levitte] 747868651Skris 747968651Skris *) Enhance c_rehash script. Old version would mishandle certificates 748068651Skris with the same subject name hash and wouldn't handle CRLs at all. 748168651Skris Added -fingerprint option to crl utility, to support new c_rehash 748268651Skris features. 748368651Skris [Steve Henson] 748468651Skris 748568651Skris *) Eliminate non-ANSI declarations in crypto.h and stack.h. 748668651Skris [Ulf M�ller] 748768651Skris 748868651Skris *) Fix for SSL server purpose checking. Server checking was 748968651Skris rejecting certificates which had extended key usage present 749068651Skris but no ssl client purpose. 749168651Skris [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 749268651Skris 749368651Skris *) Make PKCS#12 code work with no password. The PKCS#12 spec 749468651Skris is a little unclear about how a blank password is handled. 749568651Skris Since the password in encoded as a BMPString with terminating 749668651Skris double NULL a zero length password would end up as just the 749768651Skris double NULL. However no password at all is different and is 749868651Skris handled differently in the PKCS#12 key generation code. NS 749968651Skris treats a blank password as zero length. MSIE treats it as no 750068651Skris password on export: but it will try both on import. We now do 750168651Skris the same: PKCS12_parse() tries zero length and no password if 750268651Skris the password is set to "" or NULL (NULL is now a valid password: 750368651Skris it wasn't before) as does the pkcs12 application. 750468651Skris [Steve Henson] 750568651Skris 750668651Skris *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 750768651Skris perror when PEM_read_bio_X509_REQ fails, the error message must 750868651Skris be obtained from the error queue. 750968651Skris [Bodo Moeller] 751068651Skris 751168651Skris *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 751268651Skris it in ERR_remove_state if appropriate, and change ERR_get_state 751368651Skris accordingly to avoid race conditions (this is necessary because 751468651Skris thread_hash is no longer constant once set). 751568651Skris [Bodo Moeller] 751668651Skris 751768651Skris *) Bugfix for linux-elf makefile.one. 751868651Skris [Ulf M�ller] 751968651Skris 752068651Skris *) RSA_get_default_method() will now cause a default 752168651Skris RSA_METHOD to be chosen if one doesn't exist already. 752268651Skris Previously this was only set during a call to RSA_new() 752368651Skris or RSA_new_method(NULL) meaning it was possible for 752468651Skris RSA_get_default_method() to return NULL. 752568651Skris [Geoff Thorpe] 752668651Skris 752768651Skris *) Added native name translation to the existing DSO code 752868651Skris that will convert (if the flag to do so is set) filenames 752968651Skris that are sufficiently small and have no path information 753068651Skris into a canonical native form. Eg. "blah" converted to 753168651Skris "libblah.so" or "blah.dll" etc. 753268651Skris [Geoff Thorpe] 753368651Skris 753468651Skris *) New function ERR_error_string_n(e, buf, len) which is like 753568651Skris ERR_error_string(e, buf), but writes at most 'len' bytes 753668651Skris including the 0 terminator. For ERR_error_string_n, 'buf' 753768651Skris may not be NULL. 753868651Skris [Damien Miller <djm@mindrot.org>, Bodo Moeller] 753968651Skris 754068651Skris *) CONF library reworked to become more general. A new CONF 754168651Skris configuration file reader "class" is implemented as well as a 754268651Skris new functions (NCONF_*, for "New CONF") to handle it. The now 754368651Skris old CONF_* functions are still there, but are reimplemented to 754468651Skris work in terms of the new functions. Also, a set of functions 754568651Skris to handle the internal storage of the configuration data is 754668651Skris provided to make it easier to write new configuration file 754768651Skris reader "classes" (I can definitely see something reading a 754868651Skris configuration file in XML format, for example), called _CONF_*, 754968651Skris or "the configuration storage API"... 755068651Skris 755168651Skris The new configuration file reading functions are: 755268651Skris 755368651Skris NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 755468651Skris NCONF_get_section, NCONF_get_string, NCONF_get_numbre 755568651Skris 755668651Skris NCONF_default, NCONF_WIN32 755768651Skris 755868651Skris NCONF_dump_fp, NCONF_dump_bio 755968651Skris 756068651Skris NCONF_default and NCONF_WIN32 are method (or "class") choosers, 756168651Skris NCONF_new creates a new CONF object. This works in the same way 756268651Skris as other interfaces in OpenSSL, like the BIO interface. 756368651Skris NCONF_dump_* dump the internal storage of the configuration file, 756468651Skris which is useful for debugging. All other functions take the same 756568651Skris arguments as the old CONF_* functions wth the exception of the 756668651Skris first that must be a `CONF *' instead of a `LHASH *'. 756768651Skris 756868651Skris To make it easer to use the new classes with the old CONF_* functions, 756968651Skris the function CONF_set_default_method is provided. 757068651Skris [Richard Levitte] 757168651Skris 757268651Skris *) Add '-tls1' option to 'openssl ciphers', which was already 757368651Skris mentioned in the documentation but had not been implemented. 757468651Skris (This option is not yet really useful because even the additional 757568651Skris experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 757668651Skris [Bodo Moeller] 757768651Skris 757868651Skris *) Initial DSO code added into libcrypto for letting OpenSSL (and 757968651Skris OpenSSL-based applications) load shared libraries and bind to 758068651Skris them in a portable way. 758168651Skris [Geoff Thorpe, with contributions from Richard Levitte] 758268651Skris 758359191Skris Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 758459191Skris 758559191Skris *) Make sure _lrotl and _lrotr are only used with MSVC. 758659191Skris 758759191Skris *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 758859191Skris (the default implementation of RAND_status). 758959191Skris 759059191Skris *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 759159191Skris to '-clrext' (= clear extensions), as intended and documented. 759259191Skris [Bodo Moeller; inconsistency pointed out by Michael Attili 759359191Skris <attili@amaxo.com>] 759459191Skris 759559191Skris *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 759659191Skris was larger than the MD block size. 759759191Skris [Steve Henson, pointed out by Yost William <YostW@tce.com>] 759859191Skris 759959191Skris *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 760059191Skris fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 760159191Skris using the passed key: if the passed key was a private key the result 760259191Skris of X509_print(), for example, would be to print out all the private key 760359191Skris components. 760459191Skris [Steve Henson] 760559191Skris 760659191Skris *) des_quad_cksum() byte order bug fix. 760759191Skris [Ulf M�ller, using the problem description in krb4-0.9.7, where 760859191Skris the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 760959191Skris 761059191Skris *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 761159191Skris discouraged. 761259191Skris [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 761359191Skris 761459191Skris *) For easily testing in shell scripts whether some command 761559191Skris 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 761659191Skris returns with exit code 0 iff no command of the given name is available. 761759191Skris 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 761859191Skris the output goes to stdout and nothing is printed to stderr. 761959191Skris Additional arguments are always ignored. 762059191Skris 762159191Skris Since for each cipher there is a command of the same name, 762259191Skris the 'no-cipher' compilation switches can be tested this way. 762359191Skris 762459191Skris ('openssl no-XXX' is not able to detect pseudo-commands such 762559191Skris as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 762659191Skris [Bodo Moeller] 762759191Skris 762859191Skris *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 762959191Skris [Bodo Moeller] 763059191Skris 763159191Skris *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 763259191Skris is set; it will be thrown away anyway because each handshake creates 763359191Skris its own key. 763459191Skris ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 763559191Skris to parameters -- in previous versions (since OpenSSL 0.9.3) the 763659191Skris 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 763759191Skris you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 763859191Skris [Bodo Moeller] 763959191Skris 764059191Skris *) New s_client option -ign_eof: EOF at stdin is ignored, and 764159191Skris 'Q' and 'R' lose their special meanings (quit/renegotiate). 764259191Skris This is part of what -quiet does; unlike -quiet, -ign_eof 764359191Skris does not suppress any output. 764459191Skris [Richard Levitte] 764559191Skris 764659191Skris *) Add compatibility options to the purpose and trust code. The 764759191Skris purpose X509_PURPOSE_ANY is "any purpose" which automatically 764859191Skris accepts a certificate or CA, this was the previous behaviour, 764959191Skris with all the associated security issues. 765059191Skris 765159191Skris X509_TRUST_COMPAT is the old trust behaviour: only and 765259191Skris automatically trust self signed roots in certificate store. A 765359191Skris new trust setting X509_TRUST_DEFAULT is used to specify that 765459191Skris a purpose has no associated trust setting and it should instead 765559191Skris use the value in the default purpose. 765659191Skris [Steve Henson] 765759191Skris 765859191Skris *) Fix the PKCS#8 DSA private key code so it decodes keys again 765959191Skris and fix a memory leak. 766059191Skris [Steve Henson] 766159191Skris 766259191Skris *) In util/mkerr.pl (which implements 'make errors'), preserve 766359191Skris reason strings from the previous version of the .c file, as 766459191Skris the default to have only downcase letters (and digits) in 766559191Skris automatically generated reasons codes is not always appropriate. 766659191Skris [Bodo Moeller] 766759191Skris 766859191Skris *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 766959191Skris using strerror. Previously, ERR_reason_error_string() returned 767059191Skris library names as reason strings for SYSerr; but SYSerr is a special 767159191Skris case where small numbers are errno values, not library numbers. 767259191Skris [Bodo Moeller] 767359191Skris 767459191Skris *) Add '-dsaparam' option to 'openssl dhparam' application. This 767559191Skris converts DSA parameters into DH parameters. (When creating parameters, 767659191Skris DSA_generate_parameters is used.) 767759191Skris [Bodo Moeller] 767859191Skris 767959191Skris *) Include 'length' (recommended exponent length) in C code generated 768059191Skris by 'openssl dhparam -C'. 768159191Skris [Bodo Moeller] 768259191Skris 768359191Skris *) The second argument to set_label in perlasm was already being used 768459191Skris so couldn't be used as a "file scope" flag. Moved to third argument 768559191Skris which was free. 768659191Skris [Steve Henson] 768759191Skris 768859191Skris *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 768959191Skris instead of RAND_bytes for encryption IVs and salts. 769059191Skris [Bodo Moeller] 769159191Skris 769259191Skris *) Include RAND_status() into RAND_METHOD instead of implementing 769359191Skris it only for md_rand.c Otherwise replacing the PRNG by calling 769459191Skris RAND_set_rand_method would be impossible. 769559191Skris [Bodo Moeller] 769659191Skris 769759191Skris *) Don't let DSA_generate_key() enter an infinite loop if the random 769859191Skris number generation fails. 769959191Skris [Bodo Moeller] 770059191Skris 770159191Skris *) New 'rand' application for creating pseudo-random output. 770259191Skris [Bodo Moeller] 770359191Skris 770459191Skris *) Added configuration support for Linux/IA64 770559191Skris [Rolf Haberrecker <rolf@suse.de>] 770659191Skris 770759191Skris *) Assembler module support for Mingw32. 770859191Skris [Ulf M�ller] 770959191Skris 771059191Skris *) Shared library support for HPUX (in shlib/). 771159191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 771259191Skris 771359191Skris *) Shared library support for Solaris gcc. 771459191Skris [Lutz Behnke <behnke@trustcenter.de>] 771559191Skris 771659191Skris Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 771759191Skris 771859191Skris *) PKCS7_encrypt() was adding text MIME headers twice because they 771959191Skris were added manually and by SMIME_crlf_copy(). 772059191Skris [Steve Henson] 772159191Skris 772259191Skris *) In bntest.c don't call BN_rand with zero bits argument. 772359191Skris [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 772459191Skris 772559191Skris *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 772659191Skris case was implemented. This caused BN_div_recp() to fail occasionally. 772759191Skris [Ulf M�ller] 772859191Skris 772959191Skris *) Add an optional second argument to the set_label() in the perl 773059191Skris assembly language builder. If this argument exists and is set 773159191Skris to 1 it signals that the assembler should use a symbol whose 773259191Skris scope is the entire file, not just the current function. This 773359191Skris is needed with MASM which uses the format label:: for this scope. 773459191Skris [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 773559191Skris 773659191Skris *) Change the ASN1 types so they are typedefs by default. Before 773759191Skris almost all types were #define'd to ASN1_STRING which was causing 773859191Skris STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 773959191Skris for example. 774059191Skris [Steve Henson] 774159191Skris 774259191Skris *) Change names of new functions to the new get1/get0 naming 774359191Skris convention: After 'get1', the caller owns a reference count 774459191Skris and has to call ..._free; 'get0' returns a pointer to some 774559191Skris data structure without incrementing reference counters. 774659191Skris (Some of the existing 'get' functions increment a reference 774759191Skris counter, some don't.) 774859191Skris Similarly, 'set1' and 'add1' functions increase reference 774959191Skris counters or duplicate objects. 775059191Skris [Steve Henson] 775159191Skris 775259191Skris *) Allow for the possibility of temp RSA key generation failure: 775359191Skris the code used to assume it always worked and crashed on failure. 775459191Skris [Steve Henson] 775559191Skris 775659191Skris *) Fix potential buffer overrun problem in BIO_printf(). 775759191Skris [Ulf M�ller, using public domain code by Patrick Powell; problem 775859191Skris pointed out by David Sacerdote <das33@cornell.edu>] 775959191Skris 776059191Skris *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 776159191Skris RAND_egd() and RAND_status(). In the command line application, 776259191Skris the EGD socket can be specified like a seed file using RANDFILE 776359191Skris or -rand. 776459191Skris [Ulf M�ller] 776559191Skris 776659191Skris *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 776759191Skris Some CAs (e.g. Verisign) distribute certificates in this form. 776859191Skris [Steve Henson] 776959191Skris 777059191Skris *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 777159191Skris list to exclude them. This means that no special compilation option 777259191Skris is needed to use anonymous DH: it just needs to be included in the 777359191Skris cipher list. 777459191Skris [Steve Henson] 777559191Skris 777659191Skris *) Change the EVP_MD_CTX_type macro so its meaning consistent with 777759191Skris EVP_MD_type. The old functionality is available in a new macro called 777859191Skris EVP_MD_md(). Change code that uses it and update docs. 777959191Skris [Steve Henson] 778059191Skris 778159191Skris *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 778259191Skris where the 'void *' argument is replaced by a function pointer argument. 778359191Skris Previously 'void *' was abused to point to functions, which works on 778459191Skris many platforms, but is not correct. As these functions are usually 778559191Skris called by macros defined in OpenSSL header files, most source code 778659191Skris should work without changes. 778759191Skris [Richard Levitte] 778859191Skris 778959191Skris *) <openssl/opensslconf.h> (which is created by Configure) now contains 779059191Skris sections with information on -D... compiler switches used for 779159191Skris compiling the library so that applications can see them. To enable 779259191Skris one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 779359191Skris must be defined. E.g., 779459191Skris #define OPENSSL_ALGORITHM_DEFINES 779559191Skris #include <openssl/opensslconf.h> 779659191Skris defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 779759191Skris [Richard Levitte, Ulf and Bodo M�ller] 779859191Skris 779959191Skris *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 780059191Skris record layer. 780159191Skris [Bodo Moeller] 780259191Skris 780359191Skris *) Change the 'other' type in certificate aux info to a STACK_OF 780459191Skris X509_ALGOR. Although not an AlgorithmIdentifier as such it has 780559191Skris the required ASN1 format: arbitrary types determined by an OID. 780659191Skris [Steve Henson] 780759191Skris 780859191Skris *) Add some PEM_write_X509_REQ_NEW() functions and a command line 780959191Skris argument to 'req'. This is not because the function is newer or 781059191Skris better than others it just uses the work 'NEW' in the certificate 781159191Skris request header lines. Some software needs this. 781259191Skris [Steve Henson] 781359191Skris 781459191Skris *) Reorganise password command line arguments: now passwords can be 781559191Skris obtained from various sources. Delete the PEM_cb function and make 781659191Skris it the default behaviour: i.e. if the callback is NULL and the 781759191Skris usrdata argument is not NULL interpret it as a null terminated pass 781859191Skris phrase. If usrdata and the callback are NULL then the pass phrase 781959191Skris is prompted for as usual. 782059191Skris [Steve Henson] 782159191Skris 782259191Skris *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 782359191Skris the support is automatically enabled. The resulting binaries will 782459191Skris autodetect the card and use it if present. 782559191Skris [Ben Laurie and Compaq Inc.] 782659191Skris 782759191Skris *) Work around for Netscape hang bug. This sends certificate request 782859191Skris and server done in one record. Since this is perfectly legal in the 782959191Skris SSL/TLS protocol it isn't a "bug" option and is on by default. See 783059191Skris the bugs/SSLv3 entry for more info. 783159191Skris [Steve Henson] 783259191Skris 783359191Skris *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 783459191Skris [Andy Polyakov] 783559191Skris 783659191Skris *) Add -rand argument to smime and pkcs12 applications and read/write 783759191Skris of seed file. 783859191Skris [Steve Henson] 783959191Skris 784059191Skris *) New 'passwd' tool for crypt(3) and apr1 password hashes. 784159191Skris [Bodo Moeller] 784259191Skris 784359191Skris *) Add command line password options to the remaining applications. 784459191Skris [Steve Henson] 784559191Skris 784659191Skris *) Bug fix for BN_div_recp() for numerators with an even number of 784759191Skris bits. 784859191Skris [Ulf M�ller] 784959191Skris 785059191Skris *) More tests in bntest.c, and changed test_bn output. 785159191Skris [Ulf M�ller] 785259191Skris 785359191Skris *) ./config recognizes MacOS X now. 785459191Skris [Andy Polyakov] 785559191Skris 785659191Skris *) Bug fix for BN_div() when the first words of num and divsor are 785759191Skris equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 785859191Skris [Ulf M�ller] 785959191Skris 786059191Skris *) Add support for various broken PKCS#8 formats, and command line 786159191Skris options to produce them. 786259191Skris [Steve Henson] 786359191Skris 786459191Skris *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 786559191Skris get temporary BIGNUMs from a BN_CTX. 786659191Skris [Ulf M�ller] 786759191Skris 786859191Skris *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 786959191Skris for p == 0. 787059191Skris [Ulf M�ller] 787159191Skris 787259191Skris *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 787359191Skris include a #define from the old name to the new. The original intent 787459191Skris was that statically linked binaries could for example just call 787559191Skris SSLeay_add_all_ciphers() to just add ciphers to the table and not 787659191Skris link with digests. This never worked becayse SSLeay_add_all_digests() 787759191Skris and SSLeay_add_all_ciphers() were in the same source file so calling 787859191Skris one would link with the other. They are now in separate source files. 787959191Skris [Steve Henson] 788059191Skris 788159191Skris *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 788259191Skris [Steve Henson] 788359191Skris 788459191Skris *) Use a less unusual form of the Miller-Rabin primality test (it used 788559191Skris a binary algorithm for exponentiation integrated into the Miller-Rabin 788659191Skris loop, our standard modexp algorithms are faster). 788759191Skris [Bodo Moeller] 788859191Skris 788959191Skris *) Support for the EBCDIC character set completed. 789059191Skris [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 789159191Skris 789259191Skris *) Source code cleanups: use const where appropriate, eliminate casts, 789359191Skris use void * instead of char * in lhash. 789459191Skris [Ulf M�ller] 789559191Skris 789659191Skris *) Bugfix: ssl3_send_server_key_exchange was not restartable 789759191Skris (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 789859191Skris this the server could overwrite ephemeral keys that the client 789959191Skris has already seen). 790059191Skris [Bodo Moeller] 790159191Skris 790259191Skris *) Turn DSA_is_prime into a macro that calls BN_is_prime, 790359191Skris using 50 iterations of the Rabin-Miller test. 790459191Skris 790559191Skris DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 790659191Skris iterations of the Rabin-Miller test as required by the appendix 790759191Skris to FIPS PUB 186[-1]) instead of DSA_is_prime. 790859191Skris As BN_is_prime_fasttest includes trial division, DSA parameter 790959191Skris generation becomes much faster. 791059191Skris 791159191Skris This implies a change for the callback functions in DSA_is_prime 791259191Skris and DSA_generate_parameters: The callback function is called once 791359191Skris for each positive witness in the Rabin-Miller test, not just 791459191Skris occasionally in the inner loop; and the parameters to the 791559191Skris callback function now provide an iteration count for the outer 791659191Skris loop rather than for the current invocation of the inner loop. 791759191Skris DSA_generate_parameters additionally can call the callback 791859191Skris function with an 'iteration count' of -1, meaning that a 791959191Skris candidate has passed the trial division test (when q is generated 792059191Skris from an application-provided seed, trial division is skipped). 792159191Skris [Bodo Moeller] 792259191Skris 792359191Skris *) New function BN_is_prime_fasttest that optionally does trial 792459191Skris division before starting the Rabin-Miller test and has 792559191Skris an additional BN_CTX * argument (whereas BN_is_prime always 792659191Skris has to allocate at least one BN_CTX). 792759191Skris 'callback(1, -1, cb_arg)' is called when a number has passed the 792859191Skris trial division stage. 792959191Skris [Bodo Moeller] 793059191Skris 793159191Skris *) Fix for bug in CRL encoding. The validity dates weren't being handled 793259191Skris as ASN1_TIME. 793359191Skris [Steve Henson] 793459191Skris 793559191Skris *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 793659191Skris [Steve Henson] 793759191Skris 793859191Skris *) New function BN_pseudo_rand(). 793959191Skris [Ulf M�ller] 794059191Skris 794159191Skris *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 794259191Skris bignum version of BN_from_montgomery() with the working code from 794359191Skris SSLeay 0.9.0 (the word based version is faster anyway), and clean up 794459191Skris the comments. 794559191Skris [Ulf M�ller] 794659191Skris 794759191Skris *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 794859191Skris made it impossible to use the same SSL_SESSION data structure in 794959191Skris SSL2 clients in multiple threads. 795059191Skris [Bodo Moeller] 795159191Skris 795259191Skris *) The return value of RAND_load_file() no longer counts bytes obtained 795359191Skris by stat(). RAND_load_file(..., -1) is new and uses the complete file 795459191Skris to seed the PRNG (previously an explicit byte count was required). 795559191Skris [Ulf M�ller, Bodo M�ller] 795659191Skris 795759191Skris *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 795859191Skris used (char *) instead of (void *) and had casts all over the place. 795959191Skris [Steve Henson] 796059191Skris 796159191Skris *) Make BN_generate_prime() return NULL on error if ret!=NULL. 796259191Skris [Ulf M�ller] 796359191Skris 796459191Skris *) Retain source code compatibility for BN_prime_checks macro: 796559191Skris BN_is_prime(..., BN_prime_checks, ...) now uses 796659191Skris BN_prime_checks_for_size to determine the appropriate number of 796759191Skris Rabin-Miller iterations. 796859191Skris [Ulf M�ller] 796959191Skris 797059191Skris *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 797159191Skris DH_CHECK_P_NOT_SAFE_PRIME. 797259191Skris (Check if this is true? OpenPGP calls them "strong".) 797359191Skris [Ulf M�ller] 797459191Skris 797559191Skris *) Merge the functionality of "dh" and "gendh" programs into a new program 797659191Skris "dhparam". The old programs are retained for now but will handle DH keys 797759191Skris (instead of parameters) in future. 797859191Skris [Steve Henson] 797959191Skris 798059191Skris *) Make the ciphers, s_server and s_client programs check the return values 798159191Skris when a new cipher list is set. 798259191Skris [Steve Henson] 798359191Skris 798459191Skris *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 798559191Skris ciphers. Before when the 56bit ciphers were enabled the sorting was 798659191Skris wrong. 798759191Skris 798859191Skris The syntax for the cipher sorting has been extended to support sorting by 798959191Skris cipher-strength (using the strength_bits hard coded in the tables). 799059191Skris The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 799159191Skris 799259191Skris Fix a bug in the cipher-command parser: when supplying a cipher command 799359191Skris string with an "undefined" symbol (neither command nor alphanumeric 799459191Skris [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 799559191Skris an error is flagged. 799659191Skris 799759191Skris Due to the strength-sorting extension, the code of the 799859191Skris ssl_create_cipher_list() function was completely rearranged. I hope that 799959191Skris the readability was also increased :-) 800059191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 800159191Skris 800259191Skris *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 800359191Skris for the first serial number and places 2 in the serial number file. This 800459191Skris avoids problems when the root CA is created with serial number zero and 800559191Skris the first user certificate has the same issuer name and serial number 800659191Skris as the root CA. 800759191Skris [Steve Henson] 800859191Skris 800959191Skris *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 801059191Skris the new code. Add documentation for this stuff. 801159191Skris [Steve Henson] 801259191Skris 801359191Skris *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 801459191Skris X509_*() to X509at_*() on the grounds that they don't handle X509 801559191Skris structures and behave in an analagous way to the X509v3 functions: 801659191Skris they shouldn't be called directly but wrapper functions should be used 801759191Skris instead. 801859191Skris 801959191Skris So we also now have some wrapper functions that call the X509at functions 802059191Skris when passed certificate requests. (TO DO: similar things can be done with 802159191Skris PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 802259191Skris things. Some of these need some d2i or i2d and print functionality 802359191Skris because they handle more complex structures.) 802459191Skris [Steve Henson] 802559191Skris 802659191Skris *) Add missing #ifndefs that caused missing symbols when building libssl 802759191Skris as a shared library without RSA. Use #ifndef NO_SSL2 instead of 802859191Skris NO_RSA in ssl/s2*.c. 802959191Skris [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 803059191Skris 803159191Skris *) Precautions against using the PRNG uninitialized: RAND_bytes() now 803259191Skris has a return value which indicates the quality of the random data 803359191Skris (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 803459191Skris error queue. New function RAND_pseudo_bytes() generates output that is 803559191Skris guaranteed to be unique but not unpredictable. RAND_add is like 803659191Skris RAND_seed, but takes an extra argument for an entropy estimate 803759191Skris (RAND_seed always assumes full entropy). 803859191Skris [Ulf M�ller] 803959191Skris 804059191Skris *) Do more iterations of Rabin-Miller probable prime test (specifically, 804159191Skris 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 804259191Skris instead of only 2 for all lengths; see BN_prime_checks_for_size definition 804359191Skris in crypto/bn/bn_prime.c for the complete table). This guarantees a 804459191Skris false-positive rate of at most 2^-80 for random input. 804559191Skris [Bodo Moeller] 804659191Skris 804759191Skris *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 804859191Skris [Bodo Moeller] 804959191Skris 805059191Skris *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 805159191Skris in the 0.9.5 release), this returns the chain 805259191Skris from an X509_CTX structure with a dup of the stack and all 805359191Skris the X509 reference counts upped: so the stack will exist 805459191Skris after X509_CTX_cleanup() has been called. Modify pkcs12.c 805559191Skris to use this. 805659191Skris 805759191Skris Also make SSL_SESSION_print() print out the verify return 805859191Skris code. 805959191Skris [Steve Henson] 806059191Skris 806159191Skris *) Add manpage for the pkcs12 command. Also change the default 806259191Skris behaviour so MAC iteration counts are used unless the new 806359191Skris -nomaciter option is used. This improves file security and 806459191Skris only older versions of MSIE (4.0 for example) need it. 806559191Skris [Steve Henson] 806659191Skris 806759191Skris *) Honor the no-xxx Configure options when creating .DEF files. 806859191Skris [Ulf M�ller] 806959191Skris 807059191Skris *) Add PKCS#10 attributes to field table: challengePassword, 807159191Skris unstructuredName and unstructuredAddress. These are taken from 807259191Skris draft PKCS#9 v2.0 but are compatible with v1.2 provided no 807359191Skris international characters are used. 807459191Skris 807559191Skris More changes to X509_ATTRIBUTE code: allow the setting of types 807659191Skris based on strings. Remove the 'loc' parameter when adding 807759191Skris attributes because these will be a SET OF encoding which is sorted 807859191Skris in ASN1 order. 807959191Skris [Steve Henson] 808059191Skris 808159191Skris *) Initial changes to the 'req' utility to allow request generation 808259191Skris automation. This will allow an application to just generate a template 808359191Skris file containing all the field values and have req construct the 808459191Skris request. 808559191Skris 808659191Skris Initial support for X509_ATTRIBUTE handling. Stacks of these are 808759191Skris used all over the place including certificate requests and PKCS#7 808859191Skris structures. They are currently handled manually where necessary with 808959191Skris some primitive wrappers for PKCS#7. The new functions behave in a 809059191Skris manner analogous to the X509 extension functions: they allow 809159191Skris attributes to be looked up by NID and added. 809259191Skris 809359191Skris Later something similar to the X509V3 code would be desirable to 809459191Skris automatically handle the encoding, decoding and printing of the 809559191Skris more complex types. The string types like challengePassword can 809659191Skris be handled by the string table functions. 809759191Skris 809859191Skris Also modified the multi byte string table handling. Now there is 809959191Skris a 'global mask' which masks out certain types. The table itself 810059191Skris can use the flag STABLE_NO_MASK to ignore the mask setting: this 810159191Skris is useful when for example there is only one permissible type 810259191Skris (as in countryName) and using the mask might result in no valid 810359191Skris types at all. 810459191Skris [Steve Henson] 810559191Skris 810659191Skris *) Clean up 'Finished' handling, and add functions SSL_get_finished and 810759191Skris SSL_get_peer_finished to allow applications to obtain the latest 810859191Skris Finished messages sent to the peer or expected from the peer, 810959191Skris respectively. (SSL_get_peer_finished is usually the Finished message 811059191Skris actually received from the peer, otherwise the protocol will be aborted.) 811159191Skris 811259191Skris As the Finished message are message digests of the complete handshake 811359191Skris (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 811459191Skris be used for external authentication procedures when the authentication 811559191Skris provided by SSL/TLS is not desired or is not enough. 811659191Skris [Bodo Moeller] 811759191Skris 811859191Skris *) Enhanced support for Alpha Linux is added. Now ./config checks if 811959191Skris the host supports BWX extension and if Compaq C is present on the 812059191Skris $PATH. Just exploiting of the BWX extension results in 20-30% 812159191Skris performance kick for some algorithms, e.g. DES and RC4 to mention 812259191Skris a couple. Compaq C in turn generates ~20% faster code for MD5 and 812359191Skris SHA1. 812459191Skris [Andy Polyakov] 812559191Skris 812659191Skris *) Add support for MS "fast SGC". This is arguably a violation of the 812759191Skris SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 812859191Skris weak crypto and after checking the certificate is SGC a second one 812959191Skris with strong crypto. MS SGC stops the first handshake after receiving 813059191Skris the server certificate message and sends a second client hello. Since 813159191Skris a server will typically do all the time consuming operations before 813259191Skris expecting any further messages from the client (server key exchange 813359191Skris is the most expensive) there is little difference between the two. 813459191Skris 813559191Skris To get OpenSSL to support MS SGC we have to permit a second client 813659191Skris hello message after we have sent server done. In addition we have to 813759191Skris reset the MAC if we do get this second client hello. 813859191Skris [Steve Henson] 813959191Skris 814059191Skris *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 814159191Skris if a DER encoded private key is RSA or DSA traditional format. Changed 814259191Skris d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 814359191Skris format DER encoded private key. Newer code should use PKCS#8 format which 814459191Skris has the key type encoded in the ASN1 structure. Added DER private key 814559191Skris support to pkcs8 application. 814659191Skris [Steve Henson] 814759191Skris 814859191Skris *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 814959191Skris ciphersuites has been selected (as required by the SSL 3/TLS 1 815059191Skris specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 815159191Skris is set, we interpret this as a request to violate the specification 815259191Skris (the worst that can happen is a handshake failure, and 'correct' 815359191Skris behaviour would result in a handshake failure anyway). 815459191Skris [Bodo Moeller] 815559191Skris 815659191Skris *) In SSL_CTX_add_session, take into account that there might be multiple 815759191Skris SSL_SESSION structures with the same session ID (e.g. when two threads 815859191Skris concurrently obtain them from an external cache). 815959191Skris The internal cache can handle only one SSL_SESSION with a given ID, 816059191Skris so if there's a conflict, we now throw out the old one to achieve 816159191Skris consistency. 816259191Skris [Bodo Moeller] 816359191Skris 816459191Skris *) Add OIDs for idea and blowfish in CBC mode. This will allow both 816559191Skris to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 816659191Skris some routines that use cipher OIDs: some ciphers do not have OIDs 816759191Skris defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 816859191Skris example. 816959191Skris [Steve Henson] 817059191Skris 817159191Skris *) Simplify the trust setting structure and code. Now we just have 817259191Skris two sequences of OIDs for trusted and rejected settings. These will 817359191Skris typically have values the same as the extended key usage extension 817459191Skris and any application specific purposes. 817559191Skris 817659191Skris The trust checking code now has a default behaviour: it will just 817759191Skris check for an object with the same NID as the passed id. Functions can 817859191Skris be provided to override either the default behaviour or the behaviour 817959191Skris for a given id. SSL client, server and email already have functions 818059191Skris in place for compatibility: they check the NID and also return "trusted" 818159191Skris if the certificate is self signed. 818259191Skris [Steve Henson] 818359191Skris 818459191Skris *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 818559191Skris traditional format into an EVP_PKEY structure. 818659191Skris [Steve Henson] 818759191Skris 818859191Skris *) Add a password callback function PEM_cb() which either prompts for 818959191Skris a password if usr_data is NULL or otherwise assumes it is a null 819059191Skris terminated password. Allow passwords to be passed on command line 819159191Skris environment or config files in a few more utilities. 819259191Skris [Steve Henson] 819359191Skris 819459191Skris *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 819559191Skris keys. Add some short names for PKCS#8 PBE algorithms and allow them 819659191Skris to be specified on the command line for the pkcs8 and pkcs12 utilities. 819759191Skris Update documentation. 819859191Skris [Steve Henson] 819959191Skris 820059191Skris *) Support for ASN1 "NULL" type. This could be handled before by using 820159191Skris ASN1_TYPE but there wasn't any function that would try to read a NULL 820259191Skris and produce an error if it couldn't. For compatibility we also have 820359191Skris ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 820459191Skris don't allocate anything because they don't need to. 820559191Skris [Steve Henson] 820659191Skris 820759191Skris *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 820859191Skris for details. 820959191Skris [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 821059191Skris 821159191Skris *) Rebuild of the memory allocation routines used by OpenSSL code and 821259191Skris possibly others as well. The purpose is to make an interface that 821359191Skris provide hooks so anyone can build a separate set of allocation and 821459191Skris deallocation routines to be used by OpenSSL, for example memory 821559191Skris pool implementations, or something else, which was previously hard 821659191Skris since Malloc(), Realloc() and Free() were defined as macros having 821759191Skris the values malloc, realloc and free, respectively (except for Win32 821859191Skris compilations). The same is provided for memory debugging code. 821959191Skris OpenSSL already comes with functionality to find memory leaks, but 822059191Skris this gives people a chance to debug other memory problems. 822159191Skris 822259191Skris With these changes, a new set of functions and macros have appeared: 822359191Skris 822468651Skris CRYPTO_set_mem_debug_functions() [F] 822568651Skris CRYPTO_get_mem_debug_functions() [F] 822668651Skris CRYPTO_dbg_set_options() [F] 822768651Skris CRYPTO_dbg_get_options() [F] 822868651Skris CRYPTO_malloc_debug_init() [M] 822959191Skris 823059191Skris The memory debug functions are NULL by default, unless the library 823159191Skris is compiled with CRYPTO_MDEBUG or friends is defined. If someone 823259191Skris wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 823359191Skris gives the standard debugging functions that come with OpenSSL) or 823459191Skris CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 823559191Skris provided by the library user) must be used. When the standard 823659191Skris debugging functions are used, CRYPTO_dbg_set_options can be used to 823759191Skris request additional information: 823859191Skris CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 823959191Skris the CRYPTO_MDEBUG_xxx macro when compiling the library. 824059191Skris 824159191Skris Also, things like CRYPTO_set_mem_functions will always give the 824259191Skris expected result (the new set of functions is used for allocation 824359191Skris and deallocation) at all times, regardless of platform and compiler 824459191Skris options. 824559191Skris 824659191Skris To finish it up, some functions that were never use in any other 824759191Skris way than through macros have a new API and new semantic: 824859191Skris 824959191Skris CRYPTO_dbg_malloc() 825059191Skris CRYPTO_dbg_realloc() 825159191Skris CRYPTO_dbg_free() 825259191Skris 825359191Skris All macros of value have retained their old syntax. 825459191Skris [Richard Levitte and Bodo Moeller] 825559191Skris 825659191Skris *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 825759191Skris ordering of SMIMECapabilities wasn't in "strength order" and there 825859191Skris was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 825959191Skris algorithm. 826059191Skris [Steve Henson] 826159191Skris 826259191Skris *) Some ASN1 types with illegal zero length encoding (INTEGER, 826359191Skris ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 826459191Skris [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 826559191Skris 826659191Skris *) Merge in my S/MIME library for OpenSSL. This provides a simple 826759191Skris S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 826859191Skris functionality to handle multipart/signed properly) and a utility 826959191Skris called 'smime' to call all this stuff. This is based on code I 827059191Skris originally wrote for Celo who have kindly allowed it to be 827159191Skris included in OpenSSL. 827259191Skris [Steve Henson] 827359191Skris 827459191Skris *) Add variants des_set_key_checked and des_set_key_unchecked of 827559191Skris des_set_key (aka des_key_sched). Global variable des_check_key 827659191Skris decides which of these is called by des_set_key; this way 827759191Skris des_check_key behaves as it always did, but applications and 827859191Skris the library itself, which was buggy for des_check_key == 1, 827959191Skris have a cleaner way to pick the version they need. 828059191Skris [Bodo Moeller] 828159191Skris 828259191Skris *) New function PKCS12_newpass() which changes the password of a 828359191Skris PKCS12 structure. 828459191Skris [Steve Henson] 828559191Skris 828659191Skris *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 828759191Skris dynamic mix. In both cases the ids can be used as an index into the 828859191Skris table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 828959191Skris functions so they accept a list of the field values and the 829059191Skris application doesn't need to directly manipulate the X509_TRUST 829159191Skris structure. 829259191Skris [Steve Henson] 829359191Skris 829459191Skris *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 829559191Skris need initialising. 829659191Skris [Steve Henson] 829759191Skris 829859191Skris *) Modify the way the V3 extension code looks up extensions. This now 829959191Skris works in a similar way to the object code: we have some "standard" 830059191Skris extensions in a static table which is searched with OBJ_bsearch() 830159191Skris and the application can add dynamic ones if needed. The file 830259191Skris crypto/x509v3/ext_dat.h now has the info: this file needs to be 830359191Skris updated whenever a new extension is added to the core code and kept 830459191Skris in ext_nid order. There is a simple program 'tabtest.c' which checks 830559191Skris this. New extensions are not added too often so this file can readily 830659191Skris be maintained manually. 830759191Skris 830859191Skris There are two big advantages in doing things this way. The extensions 830959191Skris can be looked up immediately and no longer need to be "added" using 831059191Skris X509V3_add_standard_extensions(): this function now does nothing. 831159191Skris [Side note: I get *lots* of email saying the extension code doesn't 831259191Skris work because people forget to call this function] 831359191Skris Also no dynamic allocation is done unless new extensions are added: 831459191Skris so if we don't add custom extensions there is no need to call 831559191Skris X509V3_EXT_cleanup(). 831659191Skris [Steve Henson] 831759191Skris 831859191Skris *) Modify enc utility's salting as follows: make salting the default. Add a 831959191Skris magic header, so unsalted files fail gracefully instead of just decrypting 832059191Skris to garbage. This is because not salting is a big security hole, so people 832159191Skris should be discouraged from doing it. 832259191Skris [Ben Laurie] 832359191Skris 832459191Skris *) Fixes and enhancements to the 'x509' utility. It allowed a message 832559191Skris digest to be passed on the command line but it only used this 832659191Skris parameter when signing a certificate. Modified so all relevant 832759191Skris operations are affected by the digest parameter including the 832859191Skris -fingerprint and -x509toreq options. Also -x509toreq choked if a 832959191Skris DSA key was used because it didn't fix the digest. 833059191Skris [Steve Henson] 833159191Skris 833259191Skris *) Initial certificate chain verify code. Currently tests the untrusted 833359191Skris certificates for consistency with the verify purpose (which is set 833459191Skris when the X509_STORE_CTX structure is set up) and checks the pathlength. 833559191Skris 833659191Skris There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 833759191Skris this is because it will reject chains with invalid extensions whereas 833859191Skris every previous version of OpenSSL and SSLeay made no checks at all. 833959191Skris 834059191Skris Trust code: checks the root CA for the relevant trust settings. Trust 834159191Skris settings have an initial value consistent with the verify purpose: e.g. 834259191Skris if the verify purpose is for SSL client use it expects the CA to be 834359191Skris trusted for SSL client use. However the default value can be changed to 834459191Skris permit custom trust settings: one example of this would be to only trust 834559191Skris certificates from a specific "secure" set of CAs. 834659191Skris 834759191Skris Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 834859191Skris which should be used for version portability: especially since the 834959191Skris verify structure is likely to change more often now. 835059191Skris 835159191Skris SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 835259191Skris to set them. If not set then assume SSL clients will verify SSL servers 835359191Skris and vice versa. 835459191Skris 835559191Skris Two new options to the verify program: -untrusted allows a set of 835659191Skris untrusted certificates to be passed in and -purpose which sets the 835759191Skris intended purpose of the certificate. If a purpose is set then the 835859191Skris new chain verify code is used to check extension consistency. 835959191Skris [Steve Henson] 836059191Skris 836159191Skris *) Support for the authority information access extension. 836259191Skris [Steve Henson] 836359191Skris 836459191Skris *) Modify RSA and DSA PEM read routines to transparently handle 836559191Skris PKCS#8 format private keys. New *_PUBKEY_* functions that handle 836659191Skris public keys in a format compatible with certificate 836759191Skris SubjectPublicKeyInfo structures. Unfortunately there were already 836859191Skris functions called *_PublicKey_* which used various odd formats so 836959191Skris these are retained for compatibility: however the DSA variants were 837059191Skris never in a public release so they have been deleted. Changed dsa/rsa 837159191Skris utilities to handle the new format: note no releases ever handled public 837259191Skris keys so we should be OK. 837359191Skris 837459191Skris The primary motivation for this change is to avoid the same fiasco 837559191Skris that dogs private keys: there are several incompatible private key 837659191Skris formats some of which are standard and some OpenSSL specific and 837759191Skris require various evil hacks to allow partial transparent handling and 837859191Skris even then it doesn't work with DER formats. Given the option anything 837959191Skris other than PKCS#8 should be dumped: but the other formats have to 838059191Skris stay in the name of compatibility. 838159191Skris 838259191Skris With public keys and the benefit of hindsight one standard format 838359191Skris is used which works with EVP_PKEY, RSA or DSA structures: though 838459191Skris it clearly returns an error if you try to read the wrong kind of key. 838559191Skris 838659191Skris Added a -pubkey option to the 'x509' utility to output the public key. 838759191Skris Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 838859191Skris (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 838959191Skris EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 839059191Skris that do the same as the EVP_PKEY_assign_*() except they up the 839159191Skris reference count of the added key (they don't "swallow" the 839259191Skris supplied key). 839359191Skris [Steve Henson] 839459191Skris 839559191Skris *) Fixes to crypto/x509/by_file.c the code to read in certificates and 839659191Skris CRLs would fail if the file contained no certificates or no CRLs: 839759191Skris added a new function to read in both types and return the number 839859191Skris read: this means that if none are read it will be an error. The 839959191Skris DER versions of the certificate and CRL reader would always fail 840059191Skris because it isn't possible to mix certificates and CRLs in DER format 840159191Skris without choking one or the other routine. Changed this to just read 840259191Skris a certificate: this is the best we can do. Also modified the code 840359191Skris in apps/verify.c to take notice of return codes: it was previously 840459191Skris attempting to read in certificates from NULL pointers and ignoring 840559191Skris any errors: this is one reason why the cert and CRL reader seemed 840659191Skris to work. It doesn't check return codes from the default certificate 840759191Skris routines: these may well fail if the certificates aren't installed. 840859191Skris [Steve Henson] 840959191Skris 841059191Skris *) Code to support otherName option in GeneralName. 841159191Skris [Steve Henson] 841259191Skris 841359191Skris *) First update to verify code. Change the verify utility 841459191Skris so it warns if it is passed a self signed certificate: 841559191Skris for consistency with the normal behaviour. X509_verify 841659191Skris has been modified to it will now verify a self signed 841759191Skris certificate if *exactly* the same certificate appears 841859191Skris in the store: it was previously impossible to trust a 841959191Skris single self signed certificate. This means that: 842059191Skris openssl verify ss.pem 842159191Skris now gives a warning about a self signed certificate but 842259191Skris openssl verify -CAfile ss.pem ss.pem 842359191Skris is OK. 842459191Skris [Steve Henson] 842559191Skris 842659191Skris *) For servers, store verify_result in SSL_SESSION data structure 842759191Skris (and add it to external session representation). 842859191Skris This is needed when client certificate verifications fails, 842959191Skris but an application-provided verification callback (set by 843059191Skris SSL_CTX_set_cert_verify_callback) allows accepting the session 843159191Skris anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 843259191Skris but returns 1): When the session is reused, we have to set 843359191Skris ssl->verify_result to the appropriate error code to avoid 843459191Skris security holes. 843559191Skris [Bodo Moeller, problem pointed out by Lutz Jaenicke] 843659191Skris 843759191Skris *) Fix a bug in the new PKCS#7 code: it didn't consider the 843859191Skris case in PKCS7_dataInit() where the signed PKCS7 structure 843959191Skris didn't contain any existing data because it was being created. 844059191Skris [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 844159191Skris 844259191Skris *) Add a salt to the key derivation routines in enc.c. This 844359191Skris forms the first 8 bytes of the encrypted file. Also add a 844459191Skris -S option to allow a salt to be input on the command line. 844559191Skris [Steve Henson] 844659191Skris 844759191Skris *) New function X509_cmp(). Oddly enough there wasn't a function 844859191Skris to compare two certificates. We do this by working out the SHA1 844959191Skris hash and comparing that. X509_cmp() will be needed by the trust 845059191Skris code. 845159191Skris [Steve Henson] 845259191Skris 845359191Skris *) SSL_get1_session() is like SSL_get_session(), but increments 845459191Skris the reference count in the SSL_SESSION returned. 845559191Skris [Geoff Thorpe <geoff@eu.c2.net>] 845659191Skris 845759191Skris *) Fix for 'req': it was adding a null to request attributes. 845859191Skris Also change the X509_LOOKUP and X509_INFO code to handle 845959191Skris certificate auxiliary information. 846059191Skris [Steve Henson] 846159191Skris 846259191Skris *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 846359191Skris the 'enc' command. 846459191Skris [Steve Henson] 846559191Skris 846659191Skris *) Add the possibility to add extra information to the memory leak 846759191Skris detecting output, to form tracebacks, showing from where each 846859191Skris allocation was originated: CRYPTO_push_info("constant string") adds 846959191Skris the string plus current file name and line number to a per-thread 847059191Skris stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 847159191Skris is like calling CYRPTO_pop_info() until the stack is empty. 847259191Skris Also updated memory leak detection code to be multi-thread-safe. 847359191Skris [Richard Levitte] 847459191Skris 847559191Skris *) Add options -text and -noout to pkcs7 utility and delete the 847659191Skris encryption options which never did anything. Update docs. 847759191Skris [Steve Henson] 847859191Skris 847959191Skris *) Add options to some of the utilities to allow the pass phrase 848059191Skris to be included on either the command line (not recommended on 848159191Skris OSes like Unix) or read from the environment. Update the 848259191Skris manpages and fix a few bugs. 848359191Skris [Steve Henson] 848459191Skris 848559191Skris *) Add a few manpages for some of the openssl commands. 848659191Skris [Steve Henson] 848759191Skris 848859191Skris *) Fix the -revoke option in ca. It was freeing up memory twice, 848959191Skris leaking and not finding already revoked certificates. 849059191Skris [Steve Henson] 849159191Skris 849259191Skris *) Extensive changes to support certificate auxiliary information. 849359191Skris This involves the use of X509_CERT_AUX structure and X509_AUX 849459191Skris functions. An X509_AUX function such as PEM_read_X509_AUX() 849559191Skris can still read in a certificate file in the usual way but it 849659191Skris will also read in any additional "auxiliary information". By 849759191Skris doing things this way a fair degree of compatibility can be 849859191Skris retained: existing certificates can have this information added 849959191Skris using the new 'x509' options. 850059191Skris 850159191Skris Current auxiliary information includes an "alias" and some trust 850259191Skris settings. The trust settings will ultimately be used in enhanced 850359191Skris certificate chain verification routines: currently a certificate 850459191Skris can only be trusted if it is self signed and then it is trusted 850559191Skris for all purposes. 850659191Skris [Steve Henson] 850759191Skris 850859191Skris *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 850959191Skris The problem was that one of the replacement routines had not been working 851059191Skris since SSLeay releases. For now the offending routine has been replaced 851159191Skris with non-optimised assembler. Even so, this now gives around 95% 851259191Skris performance improvement for 1024 bit RSA signs. 851359191Skris [Mark Cox] 851459191Skris 851559191Skris *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 851659191Skris handling. Most clients have the effective key size in bits equal to 851759191Skris the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 851859191Skris A few however don't do this and instead use the size of the decrypted key 851959191Skris to determine the RC2 key length and the AlgorithmIdentifier to determine 852059191Skris the effective key length. In this case the effective key length can still 852159191Skris be 40 bits but the key length can be 168 bits for example. This is fixed 852259191Skris by manually forcing an RC2 key into the EVP_PKEY structure because the 852359191Skris EVP code can't currently handle unusual RC2 key sizes: it always assumes 852459191Skris the key length and effective key length are equal. 852559191Skris [Steve Henson] 852659191Skris 852759191Skris *) Add a bunch of functions that should simplify the creation of 852859191Skris X509_NAME structures. Now you should be able to do: 852959191Skris X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 853059191Skris and have it automatically work out the correct field type and fill in 853159191Skris the structures. The more adventurous can try: 853259191Skris X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 853359191Skris and it will (hopefully) work out the correct multibyte encoding. 853459191Skris [Steve Henson] 853559191Skris 853659191Skris *) Change the 'req' utility to use the new field handling and multibyte 853759191Skris copy routines. Before the DN field creation was handled in an ad hoc 853859191Skris way in req, ca, and x509 which was rather broken and didn't support 853959191Skris BMPStrings or UTF8Strings. Since some software doesn't implement 854059191Skris BMPStrings or UTF8Strings yet, they can be enabled using the config file 854159191Skris using the dirstring_type option. See the new comment in the default 854259191Skris openssl.cnf for more info. 854359191Skris [Steve Henson] 854459191Skris 854559191Skris *) Make crypto/rand/md_rand.c more robust: 854659191Skris - Assure unique random numbers after fork(). 854759191Skris - Make sure that concurrent threads access the global counter and 854859191Skris md serializably so that we never lose entropy in them 854959191Skris or use exactly the same state in multiple threads. 855059191Skris Access to the large state is not always serializable because 855159191Skris the additional locking could be a performance killer, and 855259191Skris md should be large enough anyway. 855359191Skris [Bodo Moeller] 855459191Skris 855559191Skris *) New file apps/app_rand.c with commonly needed functionality 855659191Skris for handling the random seed file. 855759191Skris 855859191Skris Use the random seed file in some applications that previously did not: 855959191Skris ca, 856059191Skris dsaparam -genkey (which also ignored its '-rand' option), 856159191Skris s_client, 856259191Skris s_server, 856359191Skris x509 (when signing). 856459191Skris Except on systems with /dev/urandom, it is crucial to have a random 856559191Skris seed file at least for key creation, DSA signing, and for DH exchanges; 856659191Skris for RSA signatures we could do without one. 856759191Skris 856859191Skris gendh and gendsa (unlike genrsa) used to read only the first byte 856959191Skris of each file listed in the '-rand' option. The function as previously 857059191Skris found in genrsa is now in app_rand.c and is used by all programs 857159191Skris that support '-rand'. 857259191Skris [Bodo Moeller] 857359191Skris 857459191Skris *) In RAND_write_file, use mode 0600 for creating files; 857559191Skris don't just chmod when it may be too late. 857659191Skris [Bodo Moeller] 857759191Skris 857859191Skris *) Report an error from X509_STORE_load_locations 857959191Skris when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 858059191Skris [Bill Perry] 858159191Skris 858259191Skris *) New function ASN1_mbstring_copy() this copies a string in either 858359191Skris ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 858459191Skris into an ASN1_STRING type. A mask of permissible types is passed 858559191Skris and it chooses the "minimal" type to use or an error if not type 858659191Skris is suitable. 858759191Skris [Steve Henson] 858859191Skris 858959191Skris *) Add function equivalents to the various macros in asn1.h. The old 859059191Skris macros are retained with an M_ prefix. Code inside the library can 859159191Skris use the M_ macros. External code (including the openssl utility) 859259191Skris should *NOT* in order to be "shared library friendly". 859359191Skris [Steve Henson] 859459191Skris 859559191Skris *) Add various functions that can check a certificate's extensions 859659191Skris to see if it usable for various purposes such as SSL client, 859759191Skris server or S/MIME and CAs of these types. This is currently 859859191Skris VERY EXPERIMENTAL but will ultimately be used for certificate chain 859959191Skris verification. Also added a -purpose flag to x509 utility to 860059191Skris print out all the purposes. 860159191Skris [Steve Henson] 860259191Skris 860359191Skris *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 860459191Skris functions. 860559191Skris [Steve Henson] 860659191Skris 860759191Skris *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 860859191Skris for, obtain and decode and extension and obtain its critical flag. 860959191Skris This allows all the necessary extension code to be handled in a 861059191Skris single function call. 861159191Skris [Steve Henson] 861259191Skris 861359191Skris *) RC4 tune-up featuring 30-40% performance improvement on most RISC 861459191Skris platforms. See crypto/rc4/rc4_enc.c for further details. 861559191Skris [Andy Polyakov] 861659191Skris 861759191Skris *) New -noout option to asn1parse. This causes no output to be produced 861859191Skris its main use is when combined with -strparse and -out to extract data 861959191Skris from a file (which may not be in ASN.1 format). 862059191Skris [Steve Henson] 862159191Skris 862259191Skris *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 862359191Skris when producing the local key id. 862459191Skris [Richard Levitte <levitte@stacken.kth.se>] 862559191Skris 862659191Skris *) New option -dhparam in s_server. This allows a DH parameter file to be 862759191Skris stated explicitly. If it is not stated then it tries the first server 862859191Skris certificate file. The previous behaviour hard coded the filename 862959191Skris "server.pem". 863059191Skris [Steve Henson] 863159191Skris 863259191Skris *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 863359191Skris a public key to be input or output. For example: 863459191Skris openssl rsa -in key.pem -pubout -out pubkey.pem 863559191Skris Also added necessary DSA public key functions to handle this. 863659191Skris [Steve Henson] 863759191Skris 863859191Skris *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 863959191Skris in the message. This was handled by allowing 864059191Skris X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 864159191Skris [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 864259191Skris 864359191Skris *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 864459191Skris to the end of the strings whereas this didn't. This would cause problems 864559191Skris if strings read with d2i_ASN1_bytes() were later modified. 864659191Skris [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 864759191Skris 864859191Skris *) Fix for base64 decode bug. When a base64 bio reads only one line of 864959191Skris data and it contains EOF it will end up returning an error. This is 865059191Skris caused by input 46 bytes long. The cause is due to the way base64 865159191Skris BIOs find the start of base64 encoded data. They do this by trying a 865259191Skris trial decode on each line until they find one that works. When they 865359191Skris do a flag is set and it starts again knowing it can pass all the 865459191Skris data directly through the decoder. Unfortunately it doesn't reset 865559191Skris the context it uses. This means that if EOF is reached an attempt 865659191Skris is made to pass two EOFs through the context and this causes the 865759191Skris resulting error. This can also cause other problems as well. As is 865859191Skris usual with these problems it takes *ages* to find and the fix is 865959191Skris trivial: move one line. 866059191Skris [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 866159191Skris 866259191Skris *) Ugly workaround to get s_client and s_server working under Windows. The 866359191Skris old code wouldn't work because it needed to select() on sockets and the 866459191Skris tty (for keypresses and to see if data could be written). Win32 only 866559191Skris supports select() on sockets so we select() with a 1s timeout on the 866659191Skris sockets and then see if any characters are waiting to be read, if none 866759191Skris are present then we retry, we also assume we can always write data to 866859191Skris the tty. This isn't nice because the code then blocks until we've 866959191Skris received a complete line of data and it is effectively polling the 867059191Skris keyboard at 1s intervals: however it's quite a bit better than not 867159191Skris working at all :-) A dedicated Windows application might handle this 867259191Skris with an event loop for example. 867359191Skris [Steve Henson] 867459191Skris 867559191Skris *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 867659191Skris and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 867759191Skris will be called when RSA_sign() and RSA_verify() are used. This is useful 867859191Skris if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 867959191Skris For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 868059191Skris should *not* be used: RSA_sign() and RSA_verify() must be used instead. 868159191Skris This necessitated the support of an extra signature type NID_md5_sha1 868259191Skris for SSL signatures and modifications to the SSL library to use it instead 868359191Skris of calling RSA_public_decrypt() and RSA_private_encrypt(). 868459191Skris [Steve Henson] 868559191Skris 868659191Skris *) Add new -verify -CAfile and -CApath options to the crl program, these 868759191Skris will lookup a CRL issuers certificate and verify the signature in a 868859191Skris similar way to the verify program. Tidy up the crl program so it 868959191Skris no longer accesses structures directly. Make the ASN1 CRL parsing a bit 869059191Skris less strict. It will now permit CRL extensions even if it is not 869159191Skris a V2 CRL: this will allow it to tolerate some broken CRLs. 869259191Skris [Steve Henson] 869359191Skris 869459191Skris *) Initialize all non-automatic variables each time one of the openssl 869559191Skris sub-programs is started (this is necessary as they may be started 869659191Skris multiple times from the "OpenSSL>" prompt). 869759191Skris [Lennart Bang, Bodo Moeller] 869859191Skris 869959191Skris *) Preliminary compilation option RSA_NULL which disables RSA crypto without 870059191Skris removing all other RSA functionality (this is what NO_RSA does). This 870159191Skris is so (for example) those in the US can disable those operations covered 870259191Skris by the RSA patent while allowing storage and parsing of RSA keys and RSA 870359191Skris key generation. 870459191Skris [Steve Henson] 870559191Skris 870659191Skris *) Non-copying interface to BIO pairs. 870759191Skris (still largely untested) 870859191Skris [Bodo Moeller] 870959191Skris 871059191Skris *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 871159191Skris ASCII string. This was handled independently in various places before. 871259191Skris [Steve Henson] 871359191Skris 871459191Skris *) New functions UTF8_getc() and UTF8_putc() that parse and generate 871559191Skris UTF8 strings a character at a time. 871659191Skris [Steve Henson] 871759191Skris 871859191Skris *) Use client_version from client hello to select the protocol 871959191Skris (s23_srvr.c) and for RSA client key exchange verification 872059191Skris (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 872159191Skris [Bodo Moeller] 872259191Skris 872359191Skris *) Add various utility functions to handle SPKACs, these were previously 872459191Skris handled by poking round in the structure internals. Added new function 872559191Skris NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 872659191Skris print, verify and generate SPKACs. Based on an original idea from 872759191Skris Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 872859191Skris [Steve Henson] 872959191Skris 873059191Skris *) RIPEMD160 is operational on all platforms and is back in 'make test'. 873159191Skris [Andy Polyakov] 873259191Skris 873359191Skris *) Allow the config file extension section to be overwritten on the 873459191Skris command line. Based on an original idea from Massimiliano Pala 873559191Skris <madwolf@comune.modena.it>. The new option is called -extensions 873659191Skris and can be applied to ca, req and x509. Also -reqexts to override 873759191Skris the request extensions in req and -crlexts to override the crl extensions 873859191Skris in ca. 873959191Skris [Steve Henson] 874059191Skris 874159191Skris *) Add new feature to the SPKAC handling in ca. Now you can include 874259191Skris the same field multiple times by preceding it by "XXXX." for example: 874359191Skris 1.OU="Unit name 1" 874459191Skris 2.OU="Unit name 2" 874559191Skris this is the same syntax as used in the req config file. 874659191Skris [Steve Henson] 874759191Skris 874859191Skris *) Allow certificate extensions to be added to certificate requests. These 874959191Skris are specified in a 'req_extensions' option of the req section of the 875059191Skris config file. They can be printed out with the -text option to req but 875159191Skris are otherwise ignored at present. 875259191Skris [Steve Henson] 875359191Skris 875459191Skris *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 875559191Skris data read consists of only the final block it would not decrypted because 875659191Skris EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 875759191Skris A misplaced 'break' also meant the decrypted final block might not be 875859191Skris copied until the next read. 875959191Skris [Steve Henson] 876059191Skris 876159191Skris *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 876259191Skris a few extra parameters to the DH structure: these will be useful if 876359191Skris for example we want the value of 'q' or implement X9.42 DH. 876459191Skris [Steve Henson] 876559191Skris 876659191Skris *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 876759191Skris provides hooks that allow the default DSA functions or functions on a 876859191Skris "per key" basis to be replaced. This allows hardware acceleration and 876959191Skris hardware key storage to be handled without major modification to the 877059191Skris library. Also added low level modexp hooks and CRYPTO_EX structure and 877159191Skris associated functions. 877259191Skris [Steve Henson] 877359191Skris 877459191Skris *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 877559191Skris as "read only": it can't be written to and the buffer it points to will 877659191Skris not be freed. Reading from a read only BIO is much more efficient than 877759191Skris a normal memory BIO. This was added because there are several times when 877859191Skris an area of memory needs to be read from a BIO. The previous method was 877959191Skris to create a memory BIO and write the data to it, this results in two 878059191Skris copies of the data and an O(n^2) reading algorithm. There is a new 878159191Skris function BIO_new_mem_buf() which creates a read only memory BIO from 878259191Skris an area of memory. Also modified the PKCS#7 routines to use read only 878359191Skris memory BIOs. 878459191Skris [Steve Henson] 878559191Skris 878659191Skris *) Bugfix: ssl23_get_client_hello did not work properly when called in 878759191Skris state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 878859191Skris a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 878959191Skris but a retry condition occured while trying to read the rest. 879059191Skris [Bodo Moeller] 879159191Skris 879259191Skris *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 879359191Skris NID_pkcs7_encrypted by default: this was wrong since this should almost 879459191Skris always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 879559191Skris the encrypted data type: this is a more sensible place to put it and it 879659191Skris allows the PKCS#12 code to be tidied up that duplicated this 879759191Skris functionality. 879859191Skris [Steve Henson] 879959191Skris 880059191Skris *) Changed obj_dat.pl script so it takes its input and output files on 880159191Skris the command line. This should avoid shell escape redirection problems 880259191Skris under Win32. 880359191Skris [Steve Henson] 880459191Skris 880559191Skris *) Initial support for certificate extension requests, these are included 880659191Skris in things like Xenroll certificate requests. Included functions to allow 880759191Skris extensions to be obtained and added. 880859191Skris [Steve Henson] 880959191Skris 881059191Skris *) -crlf option to s_client and s_server for sending newlines as 881159191Skris CRLF (as required by many protocols). 881259191Skris [Bodo Moeller] 881359191Skris 881455714Skris Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 881555714Skris 881655714Skris *) Install libRSAglue.a when OpenSSL is built with RSAref. 881755714Skris [Ralf S. Engelschall] 881855714Skris 881955714Skris *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 882055714Skris [Andrija Antonijevic <TheAntony2@bigfoot.com>] 882155714Skris 882255714Skris *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 882355714Skris program. 882455714Skris [Steve Henson] 882555714Skris 882655714Skris *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 882755714Skris DH parameters/keys (q is lost during that conversion, but the resulting 882855714Skris DH parameters contain its length). 882955714Skris 883055714Skris For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 883155714Skris much faster than DH_generate_parameters (which creates parameters 883255714Skris where p = 2*q + 1), and also the smaller q makes DH computations 883355714Skris much more efficient (160-bit exponentiation instead of 1024-bit 883455714Skris exponentiation); so this provides a convenient way to support DHE 883555714Skris ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 883655714Skris utter importance to use 883755714Skris SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 883855714Skris or 883955714Skris SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 884055714Skris when such DH parameters are used, because otherwise small subgroup 884155714Skris attacks may become possible! 884255714Skris [Bodo Moeller] 884355714Skris 884455714Skris *) Avoid memory leak in i2d_DHparams. 884555714Skris [Bodo Moeller] 884655714Skris 884755714Skris *) Allow the -k option to be used more than once in the enc program: 884855714Skris this allows the same encrypted message to be read by multiple recipients. 884955714Skris [Steve Henson] 885055714Skris 885155714Skris *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 885255714Skris an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 885355714Skris it will always use the numerical form of the OID, even if it has a short 885455714Skris or long name. 885555714Skris [Steve Henson] 885655714Skris 885755714Skris *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 885855714Skris method only got called if p,q,dmp1,dmq1,iqmp components were present, 885955714Skris otherwise bn_mod_exp was called. In the case of hardware keys for example 886055714Skris no private key components need be present and it might store extra data 886159191Skris in the RSA structure, which cannot be accessed from bn_mod_exp. 886259191Skris By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 886359191Skris private key operations. 886455714Skris [Steve Henson] 886555714Skris 886655714Skris *) Added support for SPARC Linux. 886755714Skris [Andy Polyakov] 886855714Skris 886955714Skris *) pem_password_cb function type incompatibly changed from 887055714Skris typedef int pem_password_cb(char *buf, int size, int rwflag); 887155714Skris to 887255714Skris ....(char *buf, int size, int rwflag, void *userdata); 887355714Skris so that applications can pass data to their callbacks: 887455714Skris The PEM[_ASN1]_{read,write}... functions and macros now take an 887555714Skris additional void * argument, which is just handed through whenever 887655714Skris the password callback is called. 887759191Skris [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 887855714Skris 887955714Skris New function SSL_CTX_set_default_passwd_cb_userdata. 888055714Skris 888155714Skris Compatibility note: As many C implementations push function arguments 888255714Skris onto the stack in reverse order, the new library version is likely to 888355714Skris interoperate with programs that have been compiled with the old 888455714Skris pem_password_cb definition (PEM_whatever takes some data that 888555714Skris happens to be on the stack as its last argument, and the callback 888655714Skris just ignores this garbage); but there is no guarantee whatsoever that 888755714Skris this will work. 888855714Skris 888955714Skris *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 889055714Skris (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 889155714Skris problems not only on Windows, but also on some Unix platforms. 889255714Skris To avoid problematic command lines, these definitions are now in an 889355714Skris auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 889455714Skris for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 889555714Skris [Bodo Moeller] 889655714Skris 889755714Skris *) MIPS III/IV assembler module is reimplemented. 889855714Skris [Andy Polyakov] 889955714Skris 890055714Skris *) More DES library cleanups: remove references to srand/rand and 890155714Skris delete an unused file. 890255714Skris [Ulf M�ller] 890355714Skris 890455714Skris *) Add support for the the free Netwide assembler (NASM) under Win32, 890555714Skris since not many people have MASM (ml) and it can be hard to obtain. 890655714Skris This is currently experimental but it seems to work OK and pass all 890755714Skris the tests. Check out INSTALL.W32 for info. 890855714Skris [Steve Henson] 890955714Skris 891055714Skris *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 891155714Skris without temporary keys kept an extra copy of the server key, 891255714Skris and connections with temporary keys did not free everything in case 891355714Skris of an error. 891455714Skris [Bodo Moeller] 891555714Skris 891655714Skris *) New function RSA_check_key and new openssl rsa option -check 891755714Skris for verifying the consistency of RSA keys. 891855714Skris [Ulf Moeller, Bodo Moeller] 891955714Skris 892055714Skris *) Various changes to make Win32 compile work: 892155714Skris 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 892255714Skris 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 892355714Skris comparison" warnings. 892455714Skris 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 892555714Skris [Steve Henson] 892655714Skris 892755714Skris *) Add a debugging option to PKCS#5 v2 key generation function: when 892855714Skris you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 892955714Skris derived keys are printed to stderr. 893055714Skris [Steve Henson] 893155714Skris 893255714Skris *) Copy the flags in ASN1_STRING_dup(). 893355714Skris [Roman E. Pavlov <pre@mo.msk.ru>] 893455714Skris 893555714Skris *) The x509 application mishandled signing requests containing DSA 893655714Skris keys when the signing key was also DSA and the parameters didn't match. 893755714Skris 893855714Skris It was supposed to omit the parameters when they matched the signing key: 893955714Skris the verifying software was then supposed to automatically use the CA's 894055714Skris parameters if they were absent from the end user certificate. 894155714Skris 894255714Skris Omitting parameters is no longer recommended. The test was also 894355714Skris the wrong way round! This was probably due to unusual behaviour in 894455714Skris EVP_cmp_parameters() which returns 1 if the parameters match. 894555714Skris This meant that parameters were omitted when they *didn't* match and 894655714Skris the certificate was useless. Certificates signed with 'ca' didn't have 894755714Skris this bug. 894855714Skris [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 894955714Skris 895055714Skris *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 895155714Skris The interface is as follows: 895255714Skris Applications can use 895355714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 895455714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 895555714Skris "off" is now the default. 895655714Skris The library internally uses 895755714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 895855714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 895955714Skris to disable memory-checking temporarily. 896055714Skris 896155714Skris Some inconsistent states that previously were possible (and were 896255714Skris even the default) are now avoided. 896355714Skris 896455714Skris -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 896555714Skris with each memory chunk allocated; this is occasionally more helpful 896655714Skris than just having a counter. 896755714Skris 896855714Skris -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 896955714Skris 897055714Skris -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 897155714Skris extensions. 897255714Skris [Bodo Moeller] 897355714Skris 897455714Skris *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 897555714Skris which largely parallels "options", but is for changing API behaviour, 897655714Skris whereas "options" are about protocol behaviour. 897755714Skris Initial "mode" flags are: 897855714Skris 897955714Skris SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 898055714Skris a single record has been written. 898155714Skris SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 898255714Skris retries use the same buffer location. 898355714Skris (But all of the contents must be 898455714Skris copied!) 898555714Skris [Bodo Moeller] 898655714Skris 898779998Skris *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 898855714Skris worked. 898955714Skris 899055714Skris *) Fix problems with no-hmac etc. 899155714Skris [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 899255714Skris 899355714Skris *) New functions RSA_get_default_method(), RSA_set_method() and 899455714Skris RSA_get_method(). These allows replacement of RSA_METHODs without having 899555714Skris to mess around with the internals of an RSA structure. 899655714Skris [Steve Henson] 899755714Skris 899855714Skris *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 899955714Skris Also really enable memory leak checks in openssl.c and in some 900055714Skris test programs. 900155714Skris [Chad C. Mulligan, Bodo Moeller] 900255714Skris 900355714Skris *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 900455714Skris up the length of negative integers. This has now been simplified to just 900555714Skris store the length when it is first determined and use it later, rather 900655714Skris than trying to keep track of where data is copied and updating it to 900755714Skris point to the end. 900855714Skris [Steve Henson, reported by Brien Wheeler 900955714Skris <bwheeler@authentica-security.com>] 901055714Skris 901155714Skris *) Add a new function PKCS7_signatureVerify. This allows the verification 901255714Skris of a PKCS#7 signature but with the signing certificate passed to the 901355714Skris function itself. This contrasts with PKCS7_dataVerify which assumes the 901455714Skris certificate is present in the PKCS#7 structure. This isn't always the 901555714Skris case: certificates can be omitted from a PKCS#7 structure and be 901655714Skris distributed by "out of band" means (such as a certificate database). 901755714Skris [Steve Henson] 901855714Skris 901955714Skris *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 902055714Skris function prototypes in pem.h, also change util/mkdef.pl to add the 902155714Skris necessary function names. 902255714Skris [Steve Henson] 902355714Skris 902455714Skris *) mk1mf.pl (used by Windows builds) did not properly read the 902555714Skris options set by Configure in the top level Makefile, and Configure 902655714Skris was not even able to write more than one option correctly. 902755714Skris Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 902855714Skris [Bodo Moeller] 902955714Skris 903055714Skris *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 903155714Skris file to be loaded from a BIO or FILE pointer. The BIO version will 903255714Skris for example allow memory BIOs to contain config info. 903355714Skris [Steve Henson] 903455714Skris 903555714Skris *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 903655714Skris Whoever hopes to achieve shared-library compatibility across versions 903755714Skris must use this, not the compile-time macro. 903855714Skris (Exercise 0.9.4: Which is the minimum library version required by 903955714Skris such programs?) 904055714Skris Note: All this applies only to multi-threaded programs, others don't 904155714Skris need locks. 904255714Skris [Bodo Moeller] 904355714Skris 904455714Skris *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 904555714Skris through a BIO pair triggered the default case, i.e. 904655714Skris SSLerr(...,SSL_R_UNKNOWN_STATE). 904755714Skris [Bodo Moeller] 904855714Skris 904955714Skris *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 905055714Skris can use the SSL library even if none of the specific BIOs is 905155714Skris appropriate. 905255714Skris [Bodo Moeller] 905355714Skris 905455714Skris *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 905555714Skris for the encoded length. 905655714Skris [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 905755714Skris 905855714Skris *) Add initial documentation of the X509V3 functions. 905955714Skris [Steve Henson] 906055714Skris 906155714Skris *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 906255714Skris PEM_write_bio_PKCS8PrivateKey() that are equivalent to 906355714Skris PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 906455714Skris secure PKCS#8 private key format with a high iteration count. 906555714Skris [Steve Henson] 906655714Skris 906755714Skris *) Fix determination of Perl interpreter: A perl or perl5 906855714Skris _directory_ in $PATH was also accepted as the interpreter. 906955714Skris [Ralf S. Engelschall] 907055714Skris 907155714Skris *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 907255714Skris wrong with it but it was very old and did things like calling 907355714Skris PEM_ASN1_read() directly and used MD5 for the hash not to mention some 907455714Skris unusual formatting. 907555714Skris [Steve Henson] 907655714Skris 907755714Skris *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 907855714Skris to use the new extension code. 907955714Skris [Steve Henson] 908055714Skris 908155714Skris *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 908255714Skris with macros. This should make it easier to change their form, add extra 908355714Skris arguments etc. Fix a few PEM prototypes which didn't have cipher as a 908455714Skris constant. 908555714Skris [Steve Henson] 908655714Skris 908755714Skris *) Add to configuration table a new entry that can specify an alternative 908855714Skris name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 908955714Skris according to Mark Crispin <MRC@Panda.COM>. 909055714Skris [Bodo Moeller] 909155714Skris 909255714Skris#if 0 909355714Skris *) DES CBC did not update the IV. Weird. 909455714Skris [Ben Laurie] 909555714Skris#else 909655714Skris des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 909755714Skris Changing the behaviour of the former might break existing programs -- 909855714Skris where IV updating is needed, des_ncbc_encrypt can be used. 909955714Skris#endif 910055714Skris 910155714Skris *) When bntest is run from "make test" it drives bc to check its 910255714Skris calculations, as well as internally checking them. If an internal check 910355714Skris fails, it needs to cause bc to give a non-zero result or make test carries 910455714Skris on without noticing the failure. Fixed. 910555714Skris [Ben Laurie] 910655714Skris 910755714Skris *) DES library cleanups. 910855714Skris [Ulf M�ller] 910955714Skris 911055714Skris *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 911155714Skris used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 911255714Skris ciphers. NOTE: although the key derivation function has been verified 911355714Skris against some published test vectors it has not been extensively tested 911455714Skris yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 911555714Skris of v2.0. 911655714Skris [Steve Henson] 911755714Skris 911855714Skris *) Instead of "mkdir -p", which is not fully portable, use new 911955714Skris Perl script "util/mkdir-p.pl". 912055714Skris [Bodo Moeller] 912155714Skris 912255714Skris *) Rewrite the way password based encryption (PBE) is handled. It used to 912355714Skris assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 912455714Skris structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 912555714Skris but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 912655714Skris the 'parameter' field of the AlgorithmIdentifier is passed to the 912755714Skris underlying key generation function so it must do its own ASN1 parsing. 912855714Skris This has also changed the EVP_PBE_CipherInit() function which now has a 912955714Skris 'parameter' argument instead of literal salt and iteration count values 913055714Skris and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 913155714Skris [Steve Henson] 913255714Skris 913355714Skris *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 913455714Skris and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 913555714Skris Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 913655714Skris KEY" because this clashed with PKCS#8 unencrypted string. Since this 913755714Skris value was just used as a "magic string" and not used directly its 913855714Skris value doesn't matter. 913955714Skris [Steve Henson] 914055714Skris 914155714Skris *) Introduce some semblance of const correctness to BN. Shame C doesn't 914255714Skris support mutable. 914355714Skris [Ben Laurie] 914455714Skris 914555714Skris *) "linux-sparc64" configuration (ultrapenguin). 914655714Skris [Ray Miller <ray.miller@oucs.ox.ac.uk>] 914755714Skris "linux-sparc" configuration. 914855714Skris [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 914955714Skris 915055714Skris *) config now generates no-xxx options for missing ciphers. 915155714Skris [Ulf M�ller] 915255714Skris 915355714Skris *) Support the EBCDIC character set (work in progress). 915455714Skris File ebcdic.c not yet included because it has a different license. 915555714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 915655714Skris 915755714Skris *) Support BS2000/OSD-POSIX. 915855714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 915955714Skris 916055714Skris *) Make callbacks for key generation use void * instead of char *. 916155714Skris [Ben Laurie] 916255714Skris 916355714Skris *) Make S/MIME samples compile (not yet tested). 916455714Skris [Ben Laurie] 916555714Skris 916655714Skris *) Additional typesafe stacks. 916755714Skris [Ben Laurie] 916855714Skris 916955714Skris *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 917055714Skris [Bodo Moeller] 917155714Skris 917255714Skris 917355714Skris Changes between 0.9.3 and 0.9.3a [29 May 1999] 917455714Skris 917555714Skris *) New configuration variant "sco5-gcc". 917655714Skris 917755714Skris *) Updated some demos. 917855714Skris [Sean O Riordain, Wade Scholine] 917955714Skris 918055714Skris *) Add missing BIO_free at exit of pkcs12 application. 918155714Skris [Wu Zhigang] 918255714Skris 918355714Skris *) Fix memory leak in conf.c. 918455714Skris [Steve Henson] 918555714Skris 918655714Skris *) Updates for Win32 to assembler version of MD5. 918755714Skris [Steve Henson] 918855714Skris 918955714Skris *) Set #! path to perl in apps/der_chop to where we found it 919055714Skris instead of using a fixed path. 919155714Skris [Bodo Moeller] 919255714Skris 919355714Skris *) SHA library changes for irix64-mips4-cc. 919455714Skris [Andy Polyakov] 919555714Skris 919655714Skris *) Improvements for VMS support. 919755714Skris [Richard Levitte] 919855714Skris 919955714Skris 920055714Skris Changes between 0.9.2b and 0.9.3 [24 May 1999] 920155714Skris 920255714Skris *) Bignum library bug fix. IRIX 6 passes "make test" now! 920355714Skris This also avoids the problems with SC4.2 and unpatched SC5. 920455714Skris [Andy Polyakov <appro@fy.chalmers.se>] 920555714Skris 920655714Skris *) New functions sk_num, sk_value and sk_set to replace the previous macros. 920755714Skris These are required because of the typesafe stack would otherwise break 920855714Skris existing code. If old code used a structure member which used to be STACK 920955714Skris and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 921055714Skris sk_num or sk_value it would produce an error because the num, data members 921155714Skris are not present in STACK_OF. Now it just produces a warning. sk_set 921255714Skris replaces the old method of assigning a value to sk_value 921355714Skris (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 921455714Skris that does this will no longer work (and should use sk_set instead) but 921555714Skris this could be regarded as a "questionable" behaviour anyway. 921655714Skris [Steve Henson] 921755714Skris 921855714Skris *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 921955714Skris correctly handle encrypted S/MIME data. 922055714Skris [Steve Henson] 922155714Skris 922255714Skris *) Change type of various DES function arguments from des_cblock 922355714Skris (which means, in function argument declarations, pointer to char) 922455714Skris to des_cblock * (meaning pointer to array with 8 char elements), 922555714Skris which allows the compiler to do more typechecking; it was like 922655714Skris that back in SSLeay, but with lots of ugly casts. 922755714Skris 922855714Skris Introduce new type const_des_cblock. 922955714Skris [Bodo Moeller] 923055714Skris 923155714Skris *) Reorganise the PKCS#7 library and get rid of some of the more obvious 923255714Skris problems: find RecipientInfo structure that matches recipient certificate 923355714Skris and initialise the ASN1 structures properly based on passed cipher. 923455714Skris [Steve Henson] 923555714Skris 923655714Skris *) Belatedly make the BN tests actually check the results. 923755714Skris [Ben Laurie] 923855714Skris 923955714Skris *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 924055714Skris to and from BNs: it was completely broken. New compilation option 924155714Skris NEG_PUBKEY_BUG to allow for some broken certificates that encode public 924255714Skris key elements as negative integers. 924355714Skris [Steve Henson] 924455714Skris 924555714Skris *) Reorganize and speed up MD5. 924655714Skris [Andy Polyakov <appro@fy.chalmers.se>] 924755714Skris 924855714Skris *) VMS support. 924955714Skris [Richard Levitte <richard@levitte.org>] 925055714Skris 925155714Skris *) New option -out to asn1parse to allow the parsed structure to be 925255714Skris output to a file. This is most useful when combined with the -strparse 925355714Skris option to examine the output of things like OCTET STRINGS. 925455714Skris [Steve Henson] 925555714Skris 925655714Skris *) Make SSL library a little more fool-proof by not requiring any longer 925755714Skris that SSL_set_{accept,connect}_state be called before 925855714Skris SSL_{accept,connect} may be used (SSL_set_..._state is omitted 925955714Skris in many applications because usually everything *appeared* to work as 926055714Skris intended anyway -- now it really works as intended). 926155714Skris [Bodo Moeller] 926255714Skris 926355714Skris *) Move openssl.cnf out of lib/. 926455714Skris [Ulf M�ller] 926555714Skris 926655714Skris *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 926755714Skris -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 926855714Skris -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 926955714Skris [Ralf S. Engelschall] 927055714Skris 927155714Skris *) Various fixes to the EVP and PKCS#7 code. It may now be able to 927255714Skris handle PKCS#7 enveloped data properly. 927355714Skris [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 927455714Skris 927555714Skris *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 927655714Skris copying pointers. The cert_st handling is changed by this in 927755714Skris various ways (and thus what used to be known as ctx->default_cert 927855714Skris is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 927955714Skris any longer when s->cert does not give us what we need). 928055714Skris ssl_cert_instantiate becomes obsolete by this change. 928155714Skris As soon as we've got the new code right (possibly it already is?), 928255714Skris we have solved a couple of bugs of the earlier code where s->cert 928355714Skris was used as if it could not have been shared with other SSL structures. 928455714Skris 928555714Skris Note that using the SSL API in certain dirty ways now will result 928655714Skris in different behaviour than observed with earlier library versions: 928755714Skris Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 928855714Skris does not influence s as it used to. 928955714Skris 929055714Skris In order to clean up things more thoroughly, inside SSL_SESSION 929155714Skris we don't use CERT any longer, but a new structure SESS_CERT 929255714Skris that holds per-session data (if available); currently, this is 929355714Skris the peer's certificate chain and, for clients, the server's certificate 929455714Skris and temporary key. CERT holds only those values that can have 929555714Skris meaningful defaults in an SSL_CTX. 929655714Skris [Bodo Moeller] 929755714Skris 929855714Skris *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 929955714Skris from the internal representation. Various PKCS#7 fixes: remove some 930055714Skris evil casts and set the enc_dig_alg field properly based on the signing 930155714Skris key type. 930255714Skris [Steve Henson] 930355714Skris 930455714Skris *) Allow PKCS#12 password to be set from the command line or the 930555714Skris environment. Let 'ca' get its config file name from the environment 930655714Skris variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 930755714Skris and 'x509'). 930855714Skris [Steve Henson] 930955714Skris 931055714Skris *) Allow certificate policies extension to use an IA5STRING for the 931155714Skris organization field. This is contrary to the PKIX definition but 931255714Skris VeriSign uses it and IE5 only recognises this form. Document 'x509' 931355714Skris extension option. 931455714Skris [Steve Henson] 931555714Skris 931655714Skris *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 931755714Skris without disallowing inline assembler and the like for non-pedantic builds. 931855714Skris [Ben Laurie] 931955714Skris 932055714Skris *) Support Borland C++ builder. 932155714Skris [Janez Jere <jj@void.si>, modified by Ulf M�ller] 932255714Skris 932355714Skris *) Support Mingw32. 932455714Skris [Ulf M�ller] 932555714Skris 932655714Skris *) SHA-1 cleanups and performance enhancements. 932755714Skris [Andy Polyakov <appro@fy.chalmers.se>] 932855714Skris 932955714Skris *) Sparc v8plus assembler for the bignum library. 933055714Skris [Andy Polyakov <appro@fy.chalmers.se>] 933155714Skris 933255714Skris *) Accept any -xxx and +xxx compiler options in Configure. 933355714Skris [Ulf M�ller] 933455714Skris 933555714Skris *) Update HPUX configuration. 933655714Skris [Anonymous] 933755714Skris 933855714Skris *) Add missing sk_<type>_unshift() function to safestack.h 933955714Skris [Ralf S. Engelschall] 934055714Skris 934155714Skris *) New function SSL_CTX_use_certificate_chain_file that sets the 934255714Skris "extra_cert"s in addition to the certificate. (This makes sense 934355714Skris only for "PEM" format files, as chains as a whole are not 934455714Skris DER-encoded.) 934555714Skris [Bodo Moeller] 934655714Skris 934755714Skris *) Support verify_depth from the SSL API. 934855714Skris x509_vfy.c had what can be considered an off-by-one-error: 934955714Skris Its depth (which was not part of the external interface) 935055714Skris was actually counting the number of certificates in a chain; 935155714Skris now it really counts the depth. 935255714Skris [Bodo Moeller] 935355714Skris 935455714Skris *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 935555714Skris instead of X509err, which often resulted in confusing error 935655714Skris messages since the error codes are not globally unique 935755714Skris (e.g. an alleged error in ssl3_accept when a certificate 935855714Skris didn't match the private key). 935955714Skris 936055714Skris *) New function SSL_CTX_set_session_id_context that allows to set a default 936155714Skris value (so that you don't need SSL_set_session_id_context for each 936255714Skris connection using the SSL_CTX). 936355714Skris [Bodo Moeller] 936455714Skris 936555714Skris *) OAEP decoding bug fix. 936655714Skris [Ulf M�ller] 936755714Skris 936855714Skris *) Support INSTALL_PREFIX for package builders, as proposed by 936955714Skris David Harris. 937055714Skris [Bodo Moeller] 937155714Skris 937255714Skris *) New Configure options "threads" and "no-threads". For systems 937355714Skris where the proper compiler options are known (currently Solaris 937455714Skris and Linux), "threads" is the default. 937555714Skris [Bodo Moeller] 937655714Skris 937755714Skris *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 937855714Skris [Bodo Moeller] 937955714Skris 938055714Skris *) Install various scripts to $(OPENSSLDIR)/misc, not to 938155714Skris $(INSTALLTOP)/bin -- they shouldn't clutter directories 938255714Skris such as /usr/local/bin. 938355714Skris [Bodo Moeller] 938455714Skris 938555714Skris *) "make linux-shared" to build shared libraries. 938655714Skris [Niels Poppe <niels@netbox.org>] 938755714Skris 938855714Skris *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 938955714Skris [Ulf M�ller] 939055714Skris 939155714Skris *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 939255714Skris extension adding in x509 utility. 939355714Skris [Steve Henson] 939455714Skris 939555714Skris *) Remove NOPROTO sections and error code comments. 939655714Skris [Ulf M�ller] 939755714Skris 939855714Skris *) Partial rewrite of the DEF file generator to now parse the ANSI 939955714Skris prototypes. 940055714Skris [Steve Henson] 940155714Skris 940255714Skris *) New Configure options --prefix=DIR and --openssldir=DIR. 940355714Skris [Ulf M�ller] 940455714Skris 940555714Skris *) Complete rewrite of the error code script(s). It is all now handled 940655714Skris by one script at the top level which handles error code gathering, 940755714Skris header rewriting and C source file generation. It should be much better 940855714Skris than the old method: it now uses a modified version of Ulf's parser to 940955714Skris read the ANSI prototypes in all header files (thus the old K&R definitions 941055714Skris aren't needed for error creation any more) and do a better job of 941155714Skris translating function codes into names. The old 'ASN1 error code imbedded 941255714Skris in a comment' is no longer necessary and it doesn't use .err files which 941355714Skris have now been deleted. Also the error code call doesn't have to appear all 941455714Skris on one line (which resulted in some large lines...). 941555714Skris [Steve Henson] 941655714Skris 941755714Skris *) Change #include filenames from <foo.h> to <openssl/foo.h>. 941855714Skris [Bodo Moeller] 941955714Skris 942055714Skris *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 942155714Skris 0 (which usually indicates a closed connection), but continue reading. 942255714Skris [Bodo Moeller] 942355714Skris 942455714Skris *) Fix some race conditions. 942555714Skris [Bodo Moeller] 942655714Skris 942755714Skris *) Add support for CRL distribution points extension. Add Certificate 942855714Skris Policies and CRL distribution points documentation. 942955714Skris [Steve Henson] 943055714Skris 943155714Skris *) Move the autogenerated header file parts to crypto/opensslconf.h. 943255714Skris [Ulf M�ller] 943355714Skris 943455714Skris *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 943555714Skris 8 of keying material. Merlin has also confirmed interop with this fix 943655714Skris between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 943755714Skris [Merlin Hughes <merlin@baltimore.ie>] 943855714Skris 943955714Skris *) Fix lots of warnings. 944055714Skris [Richard Levitte <levitte@stacken.kth.se>] 944155714Skris 944255714Skris *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 944355714Skris the directory spec didn't end with a LIST_SEPARATOR_CHAR. 944455714Skris [Richard Levitte <levitte@stacken.kth.se>] 944555714Skris 944655714Skris *) Fix problems with sizeof(long) == 8. 944755714Skris [Andy Polyakov <appro@fy.chalmers.se>] 944855714Skris 944955714Skris *) Change functions to ANSI C. 945055714Skris [Ulf M�ller] 945155714Skris 945255714Skris *) Fix typos in error codes. 945355714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 945455714Skris 945555714Skris *) Remove defunct assembler files from Configure. 945655714Skris [Ulf M�ller] 945755714Skris 945855714Skris *) SPARC v8 assembler BIGNUM implementation. 945955714Skris [Andy Polyakov <appro@fy.chalmers.se>] 946055714Skris 946155714Skris *) Support for Certificate Policies extension: both print and set. 946255714Skris Various additions to support the r2i method this uses. 946355714Skris [Steve Henson] 946455714Skris 946555714Skris *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 946655714Skris return a const string when you are expecting an allocated buffer. 946755714Skris [Ben Laurie] 946855714Skris 946955714Skris *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 947055714Skris types DirectoryString and DisplayText. 947155714Skris [Steve Henson] 947255714Skris 947355714Skris *) Add code to allow r2i extensions to access the configuration database, 947455714Skris add an LHASH database driver and add several ctx helper functions. 947555714Skris [Steve Henson] 947655714Skris 947755714Skris *) Fix an evil bug in bn_expand2() which caused various BN functions to 947855714Skris fail when they extended the size of a BIGNUM. 947955714Skris [Steve Henson] 948055714Skris 948155714Skris *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 948255714Skris support typesafe stack. 948355714Skris [Steve Henson] 948455714Skris 948555714Skris *) Fix typo in SSL_[gs]et_options(). 948655714Skris [Nils Frostberg <nils@medcom.se>] 948755714Skris 948855714Skris *) Delete various functions and files that belonged to the (now obsolete) 948955714Skris old X509V3 handling code. 949055714Skris [Steve Henson] 949155714Skris 949255714Skris *) New Configure option "rsaref". 949355714Skris [Ulf M�ller] 949455714Skris 949555714Skris *) Don't auto-generate pem.h. 949655714Skris [Bodo Moeller] 949755714Skris 949855714Skris *) Introduce type-safe ASN.1 SETs. 949955714Skris [Ben Laurie] 950055714Skris 950155714Skris *) Convert various additional casted stacks to type-safe STACK_OF() variants. 950255714Skris [Ben Laurie, Ralf S. Engelschall, Steve Henson] 950355714Skris 950455714Skris *) Introduce type-safe STACKs. This will almost certainly break lots of code 950555714Skris that links with OpenSSL (well at least cause lots of warnings), but fear 950655714Skris not: the conversion is trivial, and it eliminates loads of evil casts. A 950755714Skris few STACKed things have been converted already. Feel free to convert more. 950855714Skris In the fullness of time, I'll do away with the STACK type altogether. 950955714Skris [Ben Laurie] 951055714Skris 951155714Skris *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 951255714Skris specified in <certfile> by updating the entry in the index.txt file. 951355714Skris This way one no longer has to edit the index.txt file manually for 951455714Skris revoking a certificate. The -revoke option does the gory details now. 951555714Skris [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 951655714Skris 951755714Skris *) Fix `openssl crl -noout -text' combination where `-noout' killed the 951855714Skris `-text' option at all and this way the `-noout -text' combination was 951955714Skris inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 952055714Skris [Ralf S. Engelschall] 952155714Skris 952255714Skris *) Make sure a corresponding plain text error message exists for the 952355714Skris X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 952455714Skris verify callback function determined that a certificate was revoked. 952555714Skris [Ralf S. Engelschall] 952655714Skris 952755714Skris *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 952855714Skris ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 952955714Skris all available cipers including rc5, which was forgotten until now. 953055714Skris In order to let the testing shell script know which algorithms 953155714Skris are available, a new (up to now undocumented) command 953255714Skris "openssl list-cipher-commands" is used. 953355714Skris [Bodo Moeller] 953455714Skris 953555714Skris *) Bugfix: s_client occasionally would sleep in select() when 953655714Skris it should have checked SSL_pending() first. 953755714Skris [Bodo Moeller] 953855714Skris 953955714Skris *) New functions DSA_do_sign and DSA_do_verify to provide access to 954055714Skris the raw DSA values prior to ASN.1 encoding. 954155714Skris [Ulf M�ller] 954255714Skris 954355714Skris *) Tweaks to Configure 954455714Skris [Niels Poppe <niels@netbox.org>] 954555714Skris 954655714Skris *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 954755714Skris yet... 954855714Skris [Steve Henson] 954955714Skris 955055714Skris *) New variables $(RANLIB) and $(PERL) in the Makefiles. 955155714Skris [Ulf M�ller] 955255714Skris 955355714Skris *) New config option to avoid instructions that are illegal on the 80386. 955455714Skris The default code is faster, but requires at least a 486. 955555714Skris [Ulf M�ller] 955655714Skris 955755714Skris *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 955855714Skris SSL2_SERVER_VERSION (not used at all) macros, which are now the 955955714Skris same as SSL2_VERSION anyway. 956055714Skris [Bodo Moeller] 956155714Skris 956255714Skris *) New "-showcerts" option for s_client. 956355714Skris [Bodo Moeller] 956455714Skris 956555714Skris *) Still more PKCS#12 integration. Add pkcs12 application to openssl 956655714Skris application. Various cleanups and fixes. 956755714Skris [Steve Henson] 956855714Skris 956955714Skris *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 957055714Skris modify error routines to work internally. Add error codes and PBE init 957155714Skris to library startup routines. 957255714Skris [Steve Henson] 957355714Skris 957455714Skris *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 957555714Skris packing functions to asn1 and evp. Changed function names and error 957655714Skris codes along the way. 957755714Skris [Steve Henson] 957855714Skris 957955714Skris *) PKCS12 integration: and so it begins... First of several patches to 958055714Skris slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 958155714Skris objects to objects.h 958255714Skris [Steve Henson] 958355714Skris 958455714Skris *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 958555714Skris and display support for Thawte strong extranet extension. 958655714Skris [Steve Henson] 958755714Skris 958855714Skris *) Add LinuxPPC support. 958955714Skris [Jeff Dubrule <igor@pobox.org>] 959055714Skris 959155714Skris *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 959255714Skris bn_div_words in alpha.s. 959355714Skris [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 959455714Skris 959555714Skris *) Make sure the RSA OAEP test is skipped under -DRSAref because 959655714Skris OAEP isn't supported when OpenSSL is built with RSAref. 959755714Skris [Ulf Moeller <ulf@fitug.de>] 959855714Skris 959955714Skris *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 960055714Skris so they no longer are missing under -DNOPROTO. 960155714Skris [Soren S. Jorvang <soren@t.dk>] 960255714Skris 960355714Skris 960455714Skris Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 960555714Skris 960655714Skris *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 960755714Skris doesn't work when the session is reused. Coming soon! 960855714Skris [Ben Laurie] 960955714Skris 961055714Skris *) Fix a security hole, that allows sessions to be reused in the wrong 961155714Skris context thus bypassing client cert protection! All software that uses 961255714Skris client certs and session caches in multiple contexts NEEDS PATCHING to 961355714Skris allow session reuse! A fuller solution is in the works. 961455714Skris [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 961555714Skris 961655714Skris *) Some more source tree cleanups (removed obsolete files 961755714Skris crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 961855714Skris permission on "config" script to be executable) and a fix for the INSTALL 961955714Skris document. 962055714Skris [Ulf Moeller <ulf@fitug.de>] 962155714Skris 962255714Skris *) Remove some legacy and erroneous uses of malloc, free instead of 962355714Skris Malloc, Free. 962455714Skris [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 962555714Skris 962655714Skris *) Make rsa_oaep_test return non-zero on error. 962755714Skris [Ulf Moeller <ulf@fitug.de>] 962855714Skris 962955714Skris *) Add support for native Solaris shared libraries. Configure 963055714Skris solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 963155714Skris if someone would make that last step automatic. 963255714Skris [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 963355714Skris 963455714Skris *) ctx_size was not built with the right compiler during "make links". Fixed. 963555714Skris [Ben Laurie] 963655714Skris 963755714Skris *) Change the meaning of 'ALL' in the cipher list. It now means "everything 963855714Skris except NULL ciphers". This means the default cipher list will no longer 963955714Skris enable NULL ciphers. They need to be specifically enabled e.g. with 964055714Skris the string "DEFAULT:eNULL". 964155714Skris [Steve Henson] 964255714Skris 964355714Skris *) Fix to RSA private encryption routines: if p < q then it would 964455714Skris occasionally produce an invalid result. This will only happen with 964555714Skris externally generated keys because OpenSSL (and SSLeay) ensure p > q. 964655714Skris [Steve Henson] 964755714Skris 964855714Skris *) Be less restrictive and allow also `perl util/perlpath.pl 964955714Skris /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 965055714Skris because this way one can also use an interpreter named `perl5' (which is 965155714Skris usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 965255714Skris installed as `perl'). 965355714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 965455714Skris 965555714Skris *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 965655714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 965755714Skris 965855714Skris *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 965955714Skris advapi32.lib to Win32 build and change the pem test comparision 966055714Skris to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 966155714Skris suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 966255714Skris and crypto/des/ede_cbcm_enc.c. 966355714Skris [Steve Henson] 966455714Skris 966555714Skris *) DES quad checksum was broken on big-endian architectures. Fixed. 966655714Skris [Ben Laurie] 966755714Skris 966855714Skris *) Comment out two functions in bio.h that aren't implemented. Fix up the 966955714Skris Win32 test batch file so it (might) work again. The Win32 test batch file 967055714Skris is horrible: I feel ill.... 967155714Skris [Steve Henson] 967255714Skris 967355714Skris *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 967455714Skris in e_os.h. Audit of header files to check ANSI and non ANSI 967555714Skris sections: 10 functions were absent from non ANSI section and not exported 967655714Skris from Windows DLLs. Fixed up libeay.num for new functions. 967755714Skris [Steve Henson] 967855714Skris 967955714Skris *) Make `openssl version' output lines consistent. 968055714Skris [Ralf S. Engelschall] 968155714Skris 968255714Skris *) Fix Win32 symbol export lists for BIO functions: Added 968355714Skris BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 968455714Skris to ms/libeay{16,32}.def. 968555714Skris [Ralf S. Engelschall] 968655714Skris 968755714Skris *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 968855714Skris fine under Unix and passes some trivial tests I've now added. But the 968955714Skris whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 969055714Skris added to make sure no one expects that this stuff really works in the 969155714Skris OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 969255714Skris up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 969355714Skris openssl_bio.xs. 969455714Skris [Ralf S. Engelschall] 969555714Skris 969655714Skris *) Fix the generation of two part addresses in perl. 969755714Skris [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 969855714Skris 969955714Skris *) Add config entry for Linux on MIPS. 970055714Skris [John Tobey <jtobey@channel1.com>] 970155714Skris 970255714Skris *) Make links whenever Configure is run, unless we are on Windoze. 970355714Skris [Ben Laurie] 970455714Skris 970555714Skris *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 970655714Skris Currently only issuerAltName and AuthorityKeyIdentifier make any sense 970755714Skris in CRLs. 970855714Skris [Steve Henson] 970955714Skris 971055714Skris *) Add a useful kludge to allow package maintainers to specify compiler and 971155714Skris other platforms details on the command line without having to patch the 971255714Skris Configure script everytime: One now can use ``perl Configure 971355714Skris <id>:<details>'', i.e. platform ids are allowed to have details appended 971455714Skris to them (seperated by colons). This is treated as there would be a static 971555714Skris pre-configured entry in Configure's %table under key <id> with value 971655714Skris <details> and ``perl Configure <id>'' is called. So, when you want to 971755714Skris perform a quick test-compile under FreeBSD 3.1 with pgcc and without 971855714Skris assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 971955714Skris now, which overrides the FreeBSD-elf entry on-the-fly. 972055714Skris [Ralf S. Engelschall] 972155714Skris 972255714Skris *) Disable new TLS1 ciphersuites by default: they aren't official yet. 972355714Skris [Ben Laurie] 972455714Skris 972555714Skris *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 972655714Skris on the `perl Configure ...' command line. This way one can compile 972755714Skris OpenSSL libraries with Position Independent Code (PIC) which is needed 972855714Skris for linking it into DSOs. 972955714Skris [Ralf S. Engelschall] 973055714Skris 973155714Skris *) Remarkably, export ciphers were totally broken and no-one had noticed! 973255714Skris Fixed. 973355714Skris [Ben Laurie] 973455714Skris 973555714Skris *) Cleaned up the LICENSE document: The official contact for any license 973655714Skris questions now is the OpenSSL core team under openssl-core@openssl.org. 973755714Skris And add a paragraph about the dual-license situation to make sure people 973855714Skris recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 973955714Skris to the OpenSSL toolkit. 974055714Skris [Ralf S. Engelschall] 974155714Skris 974255714Skris *) General source tree makefile cleanups: Made `making xxx in yyy...' 974355714Skris display consistent in the source tree and replaced `/bin/rm' by `rm'. 974455714Skris Additonally cleaned up the `make links' target: Remove unnecessary 974555714Skris semicolons, subsequent redundant removes, inline point.sh into mklink.sh 974655714Skris to speed processing and no longer clutter the display with confusing 974755714Skris stuff. Instead only the actually done links are displayed. 974855714Skris [Ralf S. Engelschall] 974955714Skris 975055714Skris *) Permit null encryption ciphersuites, used for authentication only. It used 975155714Skris to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 975255714Skris It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 975355714Skris encryption. 975455714Skris [Ben Laurie] 975555714Skris 975655714Skris *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 975755714Skris signed attributes when verifying signatures (this would break them), 975855714Skris the detached data encoding was wrong and public keys obtained using 975955714Skris X509_get_pubkey() weren't freed. 976055714Skris [Steve Henson] 976155714Skris 976255714Skris *) Add text documentation for the BUFFER functions. Also added a work around 976355714Skris to a Win95 console bug. This was triggered by the password read stuff: the 976455714Skris last character typed gets carried over to the next fread(). If you were 976555714Skris generating a new cert request using 'req' for example then the last 976655714Skris character of the passphrase would be CR which would then enter the first 976755714Skris field as blank. 976855714Skris [Steve Henson] 976955714Skris 977055714Skris *) Added the new `Includes OpenSSL Cryptography Software' button as 977155714Skris doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 977255714Skris button and can be used by applications based on OpenSSL to show the 977355714Skris relationship to the OpenSSL project. 977455714Skris [Ralf S. Engelschall] 977555714Skris 977655714Skris *) Remove confusing variables in function signatures in files 977755714Skris ssl/ssl_lib.c and ssl/ssl.h. 977855714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 977955714Skris 978055714Skris *) Don't install bss_file.c under PREFIX/include/ 978155714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 978255714Skris 978355714Skris *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 978455714Skris functions that return function pointers and has support for NT specific 978555714Skris stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 978655714Skris #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 978755714Skris unsigned to signed types: this was killing the Win32 compile. 978855714Skris [Steve Henson] 978955714Skris 979055714Skris *) Add new certificate file to stack functions, 979155714Skris SSL_add_dir_cert_subjects_to_stack() and 979255714Skris SSL_add_file_cert_subjects_to_stack(). These largely supplant 979355714Skris SSL_load_client_CA_file(), and can be used to add multiple certs easily 979455714Skris to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 979555714Skris This means that Apache-SSL and similar packages don't have to mess around 979655714Skris to add as many CAs as they want to the preferred list. 979755714Skris [Ben Laurie] 979855714Skris 979955714Skris *) Experiment with doxygen documentation. Currently only partially applied to 980055714Skris ssl/ssl_lib.c. 980155714Skris See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 980255714Skris openssl.doxy as the configuration file. 980355714Skris [Ben Laurie] 980455714Skris 980555714Skris *) Get rid of remaining C++-style comments which strict C compilers hate. 980655714Skris [Ralf S. Engelschall, pointed out by Carlos Amengual] 980755714Skris 980855714Skris *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 980955714Skris compiled in by default: it has problems with large keys. 981055714Skris [Steve Henson] 981155714Skris 981255714Skris *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 981355714Skris DH private keys and/or callback functions which directly correspond to 981455714Skris their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 981555714Skris is needed for applications which have to configure certificates on a 981655714Skris per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 981755714Skris (e.g. s_server). 981855714Skris For the RSA certificate situation is makes no difference, but 981955714Skris for the DSA certificate situation this fixes the "no shared cipher" 982055714Skris problem where the OpenSSL cipher selection procedure failed because the 982155714Skris temporary keys were not overtaken from the context and the API provided 982255714Skris no way to reconfigure them. 982355714Skris The new functions now let applications reconfigure the stuff and they 982455714Skris are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 982555714Skris SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 982655714Skris non-public-API function ssl_cert_instantiate() is used as a helper 982755714Skris function and also to reduce code redundancy inside ssl_rsa.c. 982855714Skris [Ralf S. Engelschall] 982955714Skris 983055714Skris *) Move s_server -dcert and -dkey options out of the undocumented feature 983155714Skris area because they are useful for the DSA situation and should be 983255714Skris recognized by the users. 983355714Skris [Ralf S. Engelschall] 983455714Skris 983555714Skris *) Fix the cipher decision scheme for export ciphers: the export bits are 983655714Skris *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 983755714Skris SSL_EXP_MASK. So, the original variable has to be used instead of the 983855714Skris already masked variable. 983955714Skris [Richard Levitte <levitte@stacken.kth.se>] 984055714Skris 984155714Skris *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 984255714Skris [Richard Levitte <levitte@stacken.kth.se>] 984355714Skris 984455714Skris *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 984555714Skris from `int' to `unsigned int' because it's a length and initialized by 984655714Skris EVP_DigestFinal() which expects an `unsigned int *'. 984755714Skris [Richard Levitte <levitte@stacken.kth.se>] 984855714Skris 984955714Skris *) Don't hard-code path to Perl interpreter on shebang line of Configure 985055714Skris script. Instead use the usual Shell->Perl transition trick. 985155714Skris [Ralf S. Engelschall] 985255714Skris 985355714Skris *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 985455714Skris (in addition to RSA certificates) to match the behaviour of `openssl dsa 985555714Skris -noout -modulus' as it's already the case for `openssl rsa -noout 985655714Skris -modulus'. For RSA the -modulus is the real "modulus" while for DSA 985755714Skris currently the public key is printed (a decision which was already done by 985855714Skris `openssl dsa -modulus' in the past) which serves a similar purpose. 985955714Skris Additionally the NO_RSA no longer completely removes the whole -modulus 986055714Skris option; it now only avoids using the RSA stuff. Same applies to NO_DSA 986155714Skris now, too. 986255714Skris [Ralf S. Engelschall] 986355714Skris 986455714Skris *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 986555714Skris BIO. See the source (crypto/evp/bio_ok.c) for more info. 986655714Skris [Arne Ansper <arne@ats.cyber.ee>] 986755714Skris 986855714Skris *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 986955714Skris to be added. Now both 'req' and 'ca' can use new objects defined in the 987055714Skris config file. 987155714Skris [Steve Henson] 987255714Skris 987355714Skris *) Add cool BIO that does syslog (or event log on NT). 987455714Skris [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 987555714Skris 987655714Skris *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 987755714Skris TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 987855714Skris TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 987955714Skris Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 988055714Skris [Ben Laurie] 988155714Skris 988255714Skris *) Add preliminary config info for new extension code. 988355714Skris [Steve Henson] 988455714Skris 988555714Skris *) Make RSA_NO_PADDING really use no padding. 988655714Skris [Ulf Moeller <ulf@fitug.de>] 988755714Skris 988855714Skris *) Generate errors when private/public key check is done. 988955714Skris [Ben Laurie] 989055714Skris 989155714Skris *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 989255714Skris for some CRL extensions and new objects added. 989355714Skris [Steve Henson] 989455714Skris 989555714Skris *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 989655714Skris key usage extension and fuller support for authority key id. 989755714Skris [Steve Henson] 989855714Skris 989955714Skris *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 990055714Skris padding method for RSA, which is recommended for new applications in PKCS 990155714Skris #1 v2.0 (RFC 2437, October 1998). 990255714Skris OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 990355714Skris foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 990455714Skris against Bleichbacher's attack on RSA. 990555714Skris [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 990655714Skris Ben Laurie] 990755714Skris 990855714Skris *) Updates to the new SSL compression code 990955714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 991055714Skris 991155714Skris *) Fix so that the version number in the master secret, when passed 991255714Skris via RSA, checks that if TLS was proposed, but we roll back to SSLv3 991355714Skris (because the server will not accept higher), that the version number 991455714Skris is 0x03,0x01, not 0x03,0x00 991555714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 991655714Skris 991755714Skris *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 991855714Skris leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 991955714Skris in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 992055714Skris [Steve Henson] 992155714Skris 992255714Skris *) Support for RAW extensions where an arbitrary extension can be 992355714Skris created by including its DER encoding. See apps/openssl.cnf for 992455714Skris an example. 992555714Skris [Steve Henson] 992655714Skris 992755714Skris *) Make sure latest Perl versions don't interpret some generated C array 992855714Skris code as Perl array code in the crypto/err/err_genc.pl script. 992955714Skris [Lars Weber <3weber@informatik.uni-hamburg.de>] 993055714Skris 993155714Skris *) Modify ms/do_ms.bat to not generate assembly language makefiles since 993255714Skris not many people have the assembler. Various Win32 compilation fixes and 993355714Skris update to the INSTALL.W32 file with (hopefully) more accurate Win32 993455714Skris build instructions. 993555714Skris [Steve Henson] 993655714Skris 993755714Skris *) Modify configure script 'Configure' to automatically create crypto/date.h 993855714Skris file under Win32 and also build pem.h from pem.org. New script 993955714Skris util/mkfiles.pl to create the MINFO file on environments that can't do a 994055714Skris 'make files': perl util/mkfiles.pl >MINFO should work. 994155714Skris [Steve Henson] 994255714Skris 994355714Skris *) Major rework of DES function declarations, in the pursuit of correctness 994455714Skris and purity. As a result, many evil casts evaporated, and some weirdness, 994555714Skris too. You may find this causes warnings in your code. Zapping your evil 994655714Skris casts will probably fix them. Mostly. 994755714Skris [Ben Laurie] 994855714Skris 994955714Skris *) Fix for a typo in asn1.h. Bug fix to object creation script 995055714Skris obj_dat.pl. It considered a zero in an object definition to mean 995155714Skris "end of object": none of the objects in objects.h have any zeros 995255714Skris so it wasn't spotted. 995355714Skris [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 995455714Skris 995555714Skris *) Add support for Triple DES Cipher Block Chaining with Output Feedback 995655714Skris Masking (CBCM). In the absence of test vectors, the best I have been able 995755714Skris to do is check that the decrypt undoes the encrypt, so far. Send me test 995855714Skris vectors if you have them. 995955714Skris [Ben Laurie] 996055714Skris 996155714Skris *) Correct calculation of key length for export ciphers (too much space was 996255714Skris allocated for null ciphers). This has not been tested! 996355714Skris [Ben Laurie] 996455714Skris 996555714Skris *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 996655714Skris message is now correct (it understands "crypto" and "ssl" on its 996755714Skris command line). There is also now an "update" option. This will update 996855714Skris the util/ssleay.num and util/libeay.num files with any new functions. 996955714Skris If you do a: 997055714Skris perl util/mkdef.pl crypto ssl update 997155714Skris it will update them. 997255714Skris [Steve Henson] 997355714Skris 997455714Skris *) Overhauled the Perl interface (perl/*): 997555714Skris - ported BN stuff to OpenSSL's different BN library 997655714Skris - made the perl/ source tree CVS-aware 997755714Skris - renamed the package from SSLeay to OpenSSL (the files still contain 997855714Skris their history because I've copied them in the repository) 997955714Skris - removed obsolete files (the test scripts will be replaced 998055714Skris by better Test::Harness variants in the future) 998155714Skris [Ralf S. Engelschall] 998255714Skris 998355714Skris *) First cut for a very conservative source tree cleanup: 998455714Skris 1. merge various obsolete readme texts into doc/ssleay.txt 998555714Skris where we collect the old documents and readme texts. 998655714Skris 2. remove the first part of files where I'm already sure that we no 998755714Skris longer need them because of three reasons: either they are just temporary 998855714Skris files which were left by Eric or they are preserved original files where 998955714Skris I've verified that the diff is also available in the CVS via "cvs diff 999055714Skris -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 999155714Skris the crypto/md/ stuff). 999255714Skris [Ralf S. Engelschall] 999355714Skris 999455714Skris *) More extension code. Incomplete support for subject and issuer alt 999555714Skris name, issuer and authority key id. Change the i2v function parameters 999655714Skris and add an extra 'crl' parameter in the X509V3_CTX structure: guess 999755714Skris what that's for :-) Fix to ASN1 macro which messed up 999855714Skris IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 999955714Skris [Steve Henson] 1000055714Skris 1000155714Skris *) Preliminary support for ENUMERATED type. This is largely copied from the 1000255714Skris INTEGER code. 1000355714Skris [Steve Henson] 1000455714Skris 1000555714Skris *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 1000655714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 1000755714Skris 1000855714Skris *) Make sure `make rehash' target really finds the `openssl' program. 1000955714Skris [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 1001055714Skris 1001155714Skris *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 1001255714Skris like to hear about it if this slows down other processors. 1001355714Skris [Ben Laurie] 1001455714Skris 1001555714Skris *) Add CygWin32 platform information to Configure script. 1001655714Skris [Alan Batie <batie@aahz.jf.intel.com>] 1001755714Skris 1001855714Skris *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 1001955714Skris [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 1002055714Skris 1002155714Skris *) New program nseq to manipulate netscape certificate sequences 1002255714Skris [Steve Henson] 1002355714Skris 1002455714Skris *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 1002555714Skris few typos. 1002655714Skris [Steve Henson] 1002755714Skris 1002855714Skris *) Fixes to BN code. Previously the default was to define BN_RECURSION 1002955714Skris but the BN code had some problems that would cause failures when 1003055714Skris doing certificate verification and some other functions. 1003155714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 1003255714Skris 1003355714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 1003455714Skris [Steve Henson] 1003555714Skris 1003655714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 1003755714Skris [Steve Henson] 1003855714Skris 1003955714Skris *) Add several PKIX and private extended key usage OIDs. 1004055714Skris [Steve Henson] 1004155714Skris 1004255714Skris *) Modify the 'ca' program to handle the new extension code. Modify 1004355714Skris openssl.cnf for new extension format, add comments. 1004455714Skris [Steve Henson] 1004555714Skris 1004655714Skris *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 1004755714Skris and add a sample to openssl.cnf so req -x509 now adds appropriate 1004855714Skris CA extensions. 1004955714Skris [Steve Henson] 1005055714Skris 1005155714Skris *) Continued X509 V3 changes. Add to other makefiles, integrate with the 1005255714Skris error code, add initial support to X509_print() and x509 application. 1005355714Skris [Steve Henson] 1005455714Skris 1005555714Skris *) Takes a deep breath and start addding X509 V3 extension support code. Add 1005655714Skris files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 1005755714Skris stuff is currently isolated and isn't even compiled yet. 1005855714Skris [Steve Henson] 1005955714Skris 1006055714Skris *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 1006155714Skris ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 1006255714Skris Removed the versions check from X509 routines when loading extensions: 1006355714Skris this allows certain broken certificates that don't set the version 1006455714Skris properly to be processed. 1006555714Skris [Steve Henson] 1006655714Skris 1006755714Skris *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 1006855714Skris Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 1006955714Skris can still be regenerated with "make depend". 1007055714Skris [Ben Laurie] 1007155714Skris 1007255714Skris *) Spelling mistake in C version of CAST-128. 1007355714Skris [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 1007455714Skris 1007555714Skris *) Changes to the error generation code. The perl script err-code.pl 1007655714Skris now reads in the old error codes and retains the old numbers, only 1007755714Skris adding new ones if necessary. It also only changes the .err files if new 1007855714Skris codes are added. The makefiles have been modified to only insert errors 1007955714Skris when needed (to avoid needlessly modifying header files). This is done 1008055714Skris by only inserting errors if the .err file is newer than the auto generated 1008155714Skris C file. To rebuild all the error codes from scratch (the old behaviour) 1008255714Skris either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 1008355714Skris or delete all the .err files. 1008455714Skris [Steve Henson] 1008555714Skris 1008655714Skris *) CAST-128 was incorrectly implemented for short keys. The C version has 1008755714Skris been fixed, but is untested. The assembler versions are also fixed, but 1008855714Skris new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 1008955714Skris to regenerate it if needed. 1009055714Skris [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 1009155714Skris Hagino <itojun@kame.net>] 1009255714Skris 1009355714Skris *) File was opened incorrectly in randfile.c. 1009455714Skris [Ulf M�ller <ulf@fitug.de>] 1009555714Skris 1009655714Skris *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 1009755714Skris functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 1009855714Skris GeneralizedTime. ASN1_TIME is the proper type used in certificates et 1009955714Skris al: it's just almost always a UTCTime. Note this patch adds new error 1010055714Skris codes so do a "make errors" if there are problems. 1010155714Skris [Steve Henson] 1010255714Skris 1010355714Skris *) Correct Linux 1 recognition in config. 1010455714Skris [Ulf M�ller <ulf@fitug.de>] 1010555714Skris 1010655714Skris *) Remove pointless MD5 hash when using DSA keys in ca. 1010755714Skris [Anonymous <nobody@replay.com>] 1010855714Skris 1010955714Skris *) Generate an error if given an empty string as a cert directory. Also 1011055714Skris generate an error if handed NULL (previously returned 0 to indicate an 1011155714Skris error, but didn't set one). 1011255714Skris [Ben Laurie, reported by Anonymous <nobody@replay.com>] 1011355714Skris 1011455714Skris *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 1011555714Skris [Ben Laurie] 1011655714Skris 1011755714Skris *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 1011855714Skris parameters. This was causing a warning which killed off the Win32 compile. 1011955714Skris [Steve Henson] 1012055714Skris 1012155714Skris *) Remove C++ style comments from crypto/bn/bn_local.h. 1012255714Skris [Neil Costigan <neil.costigan@celocom.com>] 1012355714Skris 1012455714Skris *) The function OBJ_txt2nid was broken. It was supposed to return a nid 1012555714Skris based on a text string, looking up short and long names and finally 1012655714Skris "dot" format. The "dot" format stuff didn't work. Added new function 1012755714Skris OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 1012855714Skris OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 1012955714Skris OID is not part of the table. 1013055714Skris [Steve Henson] 1013155714Skris 1013255714Skris *) Add prototypes to X509 lookup/verify methods, fixing a bug in 1013355714Skris X509_LOOKUP_by_alias(). 1013455714Skris [Ben Laurie] 1013555714Skris 1013655714Skris *) Sort openssl functions by name. 1013755714Skris [Ben Laurie] 1013855714Skris 1013955714Skris *) Get the gendsa program working (hopefully) and add it to app list. Remove 1014055714Skris encryption from sample DSA keys (in case anyone is interested the password 1014155714Skris was "1234"). 1014255714Skris [Steve Henson] 1014355714Skris 1014455714Skris *) Make _all_ *_free functions accept a NULL pointer. 1014555714Skris [Frans Heymans <fheymans@isaserver.be>] 1014655714Skris 1014755714Skris *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 1014855714Skris NULL pointers. 1014955714Skris [Anonymous <nobody@replay.com>] 1015055714Skris 1015155714Skris *) s_server should send the CAfile as acceptable CAs, not its own cert. 1015255714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 1015355714Skris 1015455714Skris *) Don't blow it for numeric -newkey arguments to apps/req. 1015555714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 1015655714Skris 1015755714Skris *) Temp key "for export" tests were wrong in s3_srvr.c. 1015855714Skris [Anonymous <nobody@replay.com>] 1015955714Skris 1016055714Skris *) Add prototype for temp key callback functions 1016155714Skris SSL_CTX_set_tmp_{rsa,dh}_callback(). 1016255714Skris [Ben Laurie] 1016355714Skris 1016455714Skris *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 1016555714Skris DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 1016655714Skris [Steve Henson] 1016755714Skris 1016855714Skris *) X509_name_add_entry() freed the wrong thing after an error. 1016955714Skris [Arne Ansper <arne@ats.cyber.ee>] 1017055714Skris 1017155714Skris *) rsa_eay.c would attempt to free a NULL context. 1017255714Skris [Arne Ansper <arne@ats.cyber.ee>] 1017355714Skris 1017455714Skris *) BIO_s_socket() had a broken should_retry() on Windoze. 1017555714Skris [Arne Ansper <arne@ats.cyber.ee>] 1017655714Skris 1017755714Skris *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 1017855714Skris [Arne Ansper <arne@ats.cyber.ee>] 1017955714Skris 1018055714Skris *) Make sure the already existing X509_STORE->depth variable is initialized 1018155714Skris in X509_STORE_new(), but document the fact that this variable is still 1018255714Skris unused in the certificate verification process. 1018355714Skris [Ralf S. Engelschall] 1018455714Skris 1018555714Skris *) Fix the various library and apps files to free up pkeys obtained from 1018655714Skris X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 1018755714Skris [Steve Henson] 1018855714Skris 1018955714Skris *) Fix reference counting in X509_PUBKEY_get(). This makes 1019055714Skris demos/maurice/example2.c work, amongst others, probably. 1019155714Skris [Steve Henson and Ben Laurie] 1019255714Skris 1019355714Skris *) First cut of a cleanup for apps/. First the `ssleay' program is now named 1019455714Skris `openssl' and second, the shortcut symlinks for the `openssl <command>' 1019555714Skris are no longer created. This way we have a single and consistent command 1019655714Skris line interface `openssl <command>', similar to `cvs <command>'. 1019755714Skris [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 1019855714Skris 1019955714Skris *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 1020055714Skris BIT STRING wrapper always have zero unused bits. 1020155714Skris [Steve Henson] 1020255714Skris 1020355714Skris *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 1020455714Skris [Steve Henson] 1020555714Skris 1020655714Skris *) Make the top-level INSTALL documentation easier to understand. 1020755714Skris [Paul Sutton] 1020855714Skris 1020955714Skris *) Makefiles updated to exit if an error occurs in a sub-directory 1021055714Skris make (including if user presses ^C) [Paul Sutton] 1021155714Skris 1021255714Skris *) Make Montgomery context stuff explicit in RSA data structure. 1021355714Skris [Ben Laurie] 1021455714Skris 1021555714Skris *) Fix build order of pem and err to allow for generated pem.h. 1021655714Skris [Ben Laurie] 1021755714Skris 1021855714Skris *) Fix renumbering bug in X509_NAME_delete_entry(). 1021955714Skris [Ben Laurie] 1022055714Skris 1022155714Skris *) Enhanced the err-ins.pl script so it makes the error library number 1022255714Skris global and can add a library name. This is needed for external ASN1 and 1022355714Skris other error libraries. 1022455714Skris [Steve Henson] 1022555714Skris 1022655714Skris *) Fixed sk_insert which never worked properly. 1022755714Skris [Steve Henson] 1022855714Skris 1022955714Skris *) Fix ASN1 macros so they can handle indefinite length construted 1023055714Skris EXPLICIT tags. Some non standard certificates use these: they can now 1023155714Skris be read in. 1023255714Skris [Steve Henson] 1023355714Skris 1023455714Skris *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 1023555714Skris into a single doc/ssleay.txt bundle. This way the information is still 1023655714Skris preserved but no longer messes up this directory. Now it's new room for 1023755714Skris the new set of documenation files. 1023855714Skris [Ralf S. Engelschall] 1023955714Skris 1024055714Skris *) SETs were incorrectly DER encoded. This was a major pain, because they 1024155714Skris shared code with SEQUENCEs, which aren't coded the same. This means that 1024255714Skris almost everything to do with SETs or SEQUENCEs has either changed name or 1024355714Skris number of arguments. 1024455714Skris [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 1024555714Skris 1024655714Skris *) Fix test data to work with the above. 1024755714Skris [Ben Laurie] 1024855714Skris 1024955714Skris *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 1025055714Skris was already fixed by Eric for 0.9.1 it seems. 1025155714Skris [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 1025255714Skris 1025355714Skris *) Autodetect FreeBSD3. 1025455714Skris [Ben Laurie] 1025555714Skris 1025655714Skris *) Fix various bugs in Configure. This affects the following platforms: 1025755714Skris nextstep 1025855714Skris ncr-scde 1025955714Skris unixware-2.0 1026055714Skris unixware-2.0-pentium 1026155714Skris sco5-cc. 1026255714Skris [Ben Laurie] 1026355714Skris 1026455714Skris *) Eliminate generated files from CVS. Reorder tests to regenerate files 1026555714Skris before they are needed. 1026655714Skris [Ben Laurie] 1026755714Skris 1026855714Skris *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 1026955714Skris [Ben Laurie] 1027055714Skris 1027155714Skris 1027255714Skris Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 1027355714Skris 1027455714Skris *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 1027555714Skris changed SSLeay to OpenSSL in version strings. 1027655714Skris [Ralf S. Engelschall] 1027755714Skris 1027855714Skris *) Some fixups to the top-level documents. 1027955714Skris [Paul Sutton] 1028055714Skris 1028155714Skris *) Fixed the nasty bug where rsaref.h was not found under compile-time 1028255714Skris because the symlink to include/ was missing. 1028355714Skris [Ralf S. Engelschall] 1028455714Skris 1028555714Skris *) Incorporated the popular no-RSA/DSA-only patches 1028655714Skris which allow to compile a RSA-free SSLeay. 1028755714Skris [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 1028855714Skris 1028955714Skris *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 1029055714Skris when "ssleay" is still not found. 1029155714Skris [Ralf S. Engelschall] 1029255714Skris 1029355714Skris *) Added more platforms to Configure: Cray T3E, HPUX 11, 1029455714Skris [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 1029555714Skris 1029655714Skris *) Updated the README file. 1029755714Skris [Ralf S. Engelschall] 1029855714Skris 1029955714Skris *) Added various .cvsignore files in the CVS repository subdirs 1030055714Skris to make a "cvs update" really silent. 1030155714Skris [Ralf S. Engelschall] 1030255714Skris 1030355714Skris *) Recompiled the error-definition header files and added 1030455714Skris missing symbols to the Win32 linker tables. 1030555714Skris [Ralf S. Engelschall] 1030655714Skris 1030755714Skris *) Cleaned up the top-level documents; 1030855714Skris o new files: CHANGES and LICENSE 1030955714Skris o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 1031055714Skris o merged COPYRIGHT into LICENSE 1031155714Skris o removed obsolete TODO file 1031255714Skris o renamed MICROSOFT to INSTALL.W32 1031355714Skris [Ralf S. Engelschall] 1031455714Skris 1031555714Skris *) Removed dummy files from the 0.9.1b source tree: 1031655714Skris crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 1031755714Skris crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 1031855714Skris crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 1031955714Skris crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 1032055714Skris util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 1032155714Skris [Ralf S. Engelschall] 1032255714Skris 1032355714Skris *) Added various platform portability fixes. 1032455714Skris [Mark J. Cox] 1032555714Skris 1032655714Skris *) The Genesis of the OpenSSL rpject: 1032755714Skris We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 1032855714Skris Young and Tim J. Hudson created while they were working for C2Net until 1032955714Skris summer 1998. 1033055714Skris [The OpenSSL Project] 1033155714Skris 1033255714Skris 1033355714Skris Changes between 0.9.0b and 0.9.1b [not released] 1033455714Skris 1033555714Skris *) Updated a few CA certificates under certs/ 1033655714Skris [Eric A. Young] 1033755714Skris 1033855714Skris *) Changed some BIGNUM api stuff. 1033955714Skris [Eric A. Young] 1034055714Skris 1034155714Skris *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 1034255714Skris DGUX x86, Linux Alpha, etc. 1034355714Skris [Eric A. Young] 1034455714Skris 1034555714Skris *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 1034655714Skris RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 1034755714Skris available). 1034855714Skris [Eric A. Young] 1034955714Skris 1035055714Skris *) Add -strparse option to asn1pars program which parses nested 1035155714Skris binary structures 1035255714Skris [Dr Stephen Henson <shenson@bigfoot.com>] 1035355714Skris 1035455714Skris *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 1035555714Skris [Eric A. Young] 1035655714Skris 1035755714Skris *) DSA fix for "ca" program. 1035855714Skris [Eric A. Young] 1035955714Skris 1036055714Skris *) Added "-genkey" option to "dsaparam" program. 1036155714Skris [Eric A. Young] 1036255714Skris 1036355714Skris *) Added RIPE MD160 (rmd160) message digest. 1036455714Skris [Eric A. Young] 1036555714Skris 1036655714Skris *) Added -a (all) option to "ssleay version" command. 1036755714Skris [Eric A. Young] 1036855714Skris 1036955714Skris *) Added PLATFORM define which is the id given to Configure. 1037055714Skris [Eric A. Young] 1037155714Skris 1037255714Skris *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 1037355714Skris [Eric A. Young] 1037455714Skris 1037555714Skris *) Extended the ASN.1 parser routines. 1037655714Skris [Eric A. Young] 1037755714Skris 1037855714Skris *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 1037955714Skris [Eric A. Young] 1038055714Skris 1038155714Skris *) Added a BN_CTX to the BN library. 1038255714Skris [Eric A. Young] 1038355714Skris 1038455714Skris *) Fixed the weak key values in DES library 1038555714Skris [Eric A. Young] 1038655714Skris 1038755714Skris *) Changed API in EVP library for cipher aliases. 1038855714Skris [Eric A. Young] 1038955714Skris 1039055714Skris *) Added support for RC2/64bit cipher. 1039155714Skris [Eric A. Young] 1039255714Skris 1039355714Skris *) Converted the lhash library to the crypto/mem.c functions. 1039455714Skris [Eric A. Young] 1039555714Skris 1039655714Skris *) Added more recognized ASN.1 object ids. 1039755714Skris [Eric A. Young] 1039855714Skris 1039955714Skris *) Added more RSA padding checks for SSL/TLS. 1040055714Skris [Eric A. Young] 1040155714Skris 1040255714Skris *) Added BIO proxy/filter functionality. 1040355714Skris [Eric A. Young] 1040455714Skris 1040555714Skris *) Added extra_certs to SSL_CTX which can be used 1040655714Skris send extra CA certificates to the client in the CA cert chain sending 1040755714Skris process. It can be configured with SSL_CTX_add_extra_chain_cert(). 1040855714Skris [Eric A. Young] 1040955714Skris 1041055714Skris *) Now Fortezza is denied in the authentication phase because 1041155714Skris this is key exchange mechanism is not supported by SSLeay at all. 1041255714Skris [Eric A. Young] 1041355714Skris 1041455714Skris *) Additional PKCS1 checks. 1041555714Skris [Eric A. Young] 1041655714Skris 1041755714Skris *) Support the string "TLSv1" for all TLS v1 ciphers. 1041855714Skris [Eric A. Young] 1041955714Skris 1042055714Skris *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 1042155714Skris ex_data index of the SSL context in the X509_STORE_CTX ex_data. 1042255714Skris [Eric A. Young] 1042355714Skris 1042455714Skris *) Fixed a few memory leaks. 1042555714Skris [Eric A. Young] 1042655714Skris 1042755714Skris *) Fixed various code and comment typos. 1042855714Skris [Eric A. Young] 1042955714Skris 1043055714Skris *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 1043155714Skris bytes sent in the client random. 1043255714Skris [Edward Bishop <ebishop@spyglass.com>] 1043355714Skris 10434