CHANGES revision 284330
155714Skris 255714Skris OpenSSL CHANGES 355714Skris _______________ 455714Skris 5284330Sjkim Changes between 1.0.1n and 1.0.1o [12 Jun 2015] 6284330Sjkim 7284330Sjkim *) Fix HMAC ABI incompatibility. The previous version introduced an ABI 8284330Sjkim incompatibility in the handling of HMAC. The previous ABI has now been 9284330Sjkim restored. 10284330Sjkim 11284285Sjkim Changes between 1.0.1m and 1.0.1n [11 Jun 2015] 12284285Sjkim 13284285Sjkim *) Malformed ECParameters causes infinite loop 14284285Sjkim 15284285Sjkim When processing an ECParameters structure OpenSSL enters an infinite loop 16284285Sjkim if the curve specified is over a specially malformed binary polynomial 17284285Sjkim field. 18284285Sjkim 19284285Sjkim This can be used to perform denial of service against any 20284285Sjkim system which processes public keys, certificate requests or 21284285Sjkim certificates. This includes TLS clients and TLS servers with 22284285Sjkim client authentication enabled. 23284285Sjkim 24284285Sjkim This issue was reported to OpenSSL by Joseph Barr-Pixton. 25284285Sjkim (CVE-2015-1788) 26284285Sjkim [Andy Polyakov] 27284285Sjkim 28284285Sjkim *) Exploitable out-of-bounds read in X509_cmp_time 29284285Sjkim 30284285Sjkim X509_cmp_time does not properly check the length of the ASN1_TIME 31284285Sjkim string and can read a few bytes out of bounds. In addition, 32284285Sjkim X509_cmp_time accepts an arbitrary number of fractional seconds in the 33284285Sjkim time string. 34284285Sjkim 35284285Sjkim An attacker can use this to craft malformed certificates and CRLs of 36284285Sjkim various sizes and potentially cause a segmentation fault, resulting in 37284285Sjkim a DoS on applications that verify certificates or CRLs. TLS clients 38284285Sjkim that verify CRLs are affected. TLS clients and servers with client 39284285Sjkim authentication enabled may be affected if they use custom verification 40284285Sjkim callbacks. 41284285Sjkim 42284285Sjkim This issue was reported to OpenSSL by Robert Swiecki (Google), and 43284285Sjkim independently by Hanno B�ck. 44284285Sjkim (CVE-2015-1789) 45284285Sjkim [Emilia K�sper] 46284285Sjkim 47284285Sjkim *) PKCS7 crash with missing EnvelopedContent 48284285Sjkim 49284285Sjkim The PKCS#7 parsing code does not handle missing inner EncryptedContent 50284285Sjkim correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs 51284285Sjkim with missing content and trigger a NULL pointer dereference on parsing. 52284285Sjkim 53284285Sjkim Applications that decrypt PKCS#7 data or otherwise parse PKCS#7 54284285Sjkim structures from untrusted sources are affected. OpenSSL clients and 55284285Sjkim servers are not affected. 56284285Sjkim 57284285Sjkim This issue was reported to OpenSSL by Michal Zalewski (Google). 58284285Sjkim (CVE-2015-1790) 59284285Sjkim [Emilia K�sper] 60284285Sjkim 61284285Sjkim *) CMS verify infinite loop with unknown hash function 62284285Sjkim 63284285Sjkim When verifying a signedData message the CMS code can enter an infinite loop 64284285Sjkim if presented with an unknown hash function OID. This can be used to perform 65284285Sjkim denial of service against any system which verifies signedData messages using 66284285Sjkim the CMS code. 67284285Sjkim This issue was reported to OpenSSL by Johannes Bauer. 68284285Sjkim (CVE-2015-1792) 69284285Sjkim [Stephen Henson] 70284285Sjkim 71284285Sjkim *) Race condition handling NewSessionTicket 72284285Sjkim 73284285Sjkim If a NewSessionTicket is received by a multi-threaded client when attempting to 74284285Sjkim reuse a previous ticket then a race condition can occur potentially leading to 75284285Sjkim a double free of the ticket data. 76284285Sjkim (CVE-2015-1791) 77284285Sjkim [Matt Caswell] 78284285Sjkim 79284285Sjkim *) Reject DH handshakes with parameters shorter than 768 bits. 80284285Sjkim [Kurt Roeckx and Emilia Kasper] 81284285Sjkim 82280304Sjkim Changes between 1.0.1l and 1.0.1m [19 Mar 2015] 83280304Sjkim 84280304Sjkim *) Segmentation fault in ASN1_TYPE_cmp fix 85280304Sjkim 86280304Sjkim The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is 87280304Sjkim made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check 88280304Sjkim certificate signature algorithm consistency this can be used to crash any 89280304Sjkim certificate verification operation and exploited in a DoS attack. Any 90280304Sjkim application which performs certificate verification is vulnerable including 91280304Sjkim OpenSSL clients and servers which enable client authentication. 92280304Sjkim (CVE-2015-0286) 93280304Sjkim [Stephen Henson] 94280304Sjkim 95280304Sjkim *) ASN.1 structure reuse memory corruption fix 96280304Sjkim 97280304Sjkim Reusing a structure in ASN.1 parsing may allow an attacker to cause 98280304Sjkim memory corruption via an invalid write. Such reuse is and has been 99280304Sjkim strongly discouraged and is believed to be rare. 100280304Sjkim 101280304Sjkim Applications that parse structures containing CHOICE or ANY DEFINED BY 102280304Sjkim components may be affected. Certificate parsing (d2i_X509 and related 103280304Sjkim functions) are however not affected. OpenSSL clients and servers are 104280304Sjkim not affected. 105280304Sjkim (CVE-2015-0287) 106280304Sjkim [Stephen Henson] 107280304Sjkim 108280304Sjkim *) PKCS7 NULL pointer dereferences fix 109280304Sjkim 110280304Sjkim The PKCS#7 parsing code does not handle missing outer ContentInfo 111280304Sjkim correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with 112280304Sjkim missing content and trigger a NULL pointer dereference on parsing. 113280304Sjkim 114280304Sjkim Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or 115280304Sjkim otherwise parse PKCS#7 structures from untrusted sources are 116280304Sjkim affected. OpenSSL clients and servers are not affected. 117280304Sjkim 118280304Sjkim This issue was reported to OpenSSL by Michal Zalewski (Google). 119280304Sjkim (CVE-2015-0289) 120280304Sjkim [Emilia K�sper] 121280304Sjkim 122280304Sjkim *) DoS via reachable assert in SSLv2 servers fix 123280304Sjkim 124280304Sjkim A malicious client can trigger an OPENSSL_assert (i.e., an abort) in 125280304Sjkim servers that both support SSLv2 and enable export cipher suites by sending 126280304Sjkim a specially crafted SSLv2 CLIENT-MASTER-KEY message. 127280304Sjkim 128280304Sjkim This issue was discovered by Sean Burford (Google) and Emilia K�sper 129280304Sjkim (OpenSSL development team). 130280304Sjkim (CVE-2015-0293) 131280304Sjkim [Emilia K�sper] 132280304Sjkim 133280304Sjkim *) Use After Free following d2i_ECPrivatekey error fix 134280304Sjkim 135280304Sjkim A malformed EC private key file consumed via the d2i_ECPrivateKey function 136280304Sjkim could cause a use after free condition. This, in turn, could cause a double 137280304Sjkim free in several private key parsing functions (such as d2i_PrivateKey 138280304Sjkim or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption 139280304Sjkim for applications that receive EC private keys from untrusted 140280304Sjkim sources. This scenario is considered rare. 141280304Sjkim 142280304Sjkim This issue was discovered by the BoringSSL project and fixed in their 143280304Sjkim commit 517073cd4b. 144280304Sjkim (CVE-2015-0209) 145280304Sjkim [Matt Caswell] 146280304Sjkim 147280304Sjkim *) X509_to_X509_REQ NULL pointer deref fix 148280304Sjkim 149280304Sjkim The function X509_to_X509_REQ will crash with a NULL pointer dereference if 150280304Sjkim the certificate key is invalid. This function is rarely used in practice. 151280304Sjkim 152280304Sjkim This issue was discovered by Brian Carpenter. 153280304Sjkim (CVE-2015-0288) 154280304Sjkim [Stephen Henson] 155280304Sjkim 156280304Sjkim *) Removed the export ciphers from the DEFAULT ciphers 157280304Sjkim [Kurt Roeckx] 158280304Sjkim 159277597Sjkim Changes between 1.0.1k and 1.0.1l [15 Jan 2015] 160277597Sjkim 161277597Sjkim *) Build fixes for the Windows and OpenVMS platforms 162277597Sjkim [Matt Caswell and Richard Levitte] 163277597Sjkim 164276864Sjkim Changes between 1.0.1j and 1.0.1k [8 Jan 2015] 165276864Sjkim 166276864Sjkim *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS 167276864Sjkim message can cause a segmentation fault in OpenSSL due to a NULL pointer 168276864Sjkim dereference. This could lead to a Denial Of Service attack. Thanks to 169276864Sjkim Markus Stenberg of Cisco Systems, Inc. for reporting this issue. 170276864Sjkim (CVE-2014-3571) 171276864Sjkim [Steve Henson] 172276864Sjkim 173276864Sjkim *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the 174276864Sjkim dtls1_buffer_record function under certain conditions. In particular this 175276864Sjkim could occur if an attacker sent repeated DTLS records with the same 176276864Sjkim sequence number but for the next epoch. The memory leak could be exploited 177276864Sjkim by an attacker in a Denial of Service attack through memory exhaustion. 178276864Sjkim Thanks to Chris Mueller for reporting this issue. 179276864Sjkim (CVE-2015-0206) 180276864Sjkim [Matt Caswell] 181276864Sjkim 182276864Sjkim *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is 183276864Sjkim built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl 184276864Sjkim method would be set to NULL which could later result in a NULL pointer 185276864Sjkim dereference. Thanks to Frank Schmirler for reporting this issue. 186276864Sjkim (CVE-2014-3569) 187276864Sjkim [Kurt Roeckx] 188276864Sjkim 189276864Sjkim *) Abort handshake if server key exchange message is omitted for ephemeral 190276864Sjkim ECDH ciphersuites. 191276864Sjkim 192276864Sjkim Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for 193276864Sjkim reporting this issue. 194276864Sjkim (CVE-2014-3572) 195276864Sjkim [Steve Henson] 196276864Sjkim 197276864Sjkim *) Remove non-export ephemeral RSA code on client and server. This code 198276864Sjkim violated the TLS standard by allowing the use of temporary RSA keys in 199276864Sjkim non-export ciphersuites and could be used by a server to effectively 200276864Sjkim downgrade the RSA key length used to a value smaller than the server 201276864Sjkim certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at 202276864Sjkim INRIA or reporting this issue. 203276864Sjkim (CVE-2015-0204) 204276864Sjkim [Steve Henson] 205276864Sjkim 206276864Sjkim *) Fixed issue where DH client certificates are accepted without verification. 207276864Sjkim An OpenSSL server will accept a DH certificate for client authentication 208276864Sjkim without the certificate verify message. This effectively allows a client to 209276864Sjkim authenticate without the use of a private key. This only affects servers 210276864Sjkim which trust a client certificate authority which issues certificates 211276864Sjkim containing DH keys: these are extremely rare and hardly ever encountered. 212276864Sjkim Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting 213276864Sjkim this issue. 214276864Sjkim (CVE-2015-0205) 215276864Sjkim [Steve Henson] 216276864Sjkim 217276864Sjkim *) Ensure that the session ID context of an SSL is updated when its 218276864Sjkim SSL_CTX is updated via SSL_set_SSL_CTX. 219276864Sjkim 220276864Sjkim The session ID context is typically set from the parent SSL_CTX, 221276864Sjkim and can vary with the CTX. 222276864Sjkim [Adam Langley] 223276864Sjkim 224276864Sjkim *) Fix various certificate fingerprint issues. 225276864Sjkim 226276864Sjkim By using non-DER or invalid encodings outside the signed portion of a 227276864Sjkim certificate the fingerprint can be changed without breaking the signature. 228276864Sjkim Although no details of the signed portion of the certificate can be changed 229276864Sjkim this can cause problems with some applications: e.g. those using the 230276864Sjkim certificate fingerprint for blacklists. 231276864Sjkim 232276864Sjkim 1. Reject signatures with non zero unused bits. 233276864Sjkim 234276864Sjkim If the BIT STRING containing the signature has non zero unused bits reject 235276864Sjkim the signature. All current signature algorithms require zero unused bits. 236276864Sjkim 237276864Sjkim 2. Check certificate algorithm consistency. 238276864Sjkim 239276864Sjkim Check the AlgorithmIdentifier inside TBS matches the one in the 240276864Sjkim certificate signature. NB: this will result in signature failure 241276864Sjkim errors for some broken certificates. 242276864Sjkim 243276864Sjkim Thanks to Konrad Kraszewski from Google for reporting this issue. 244276864Sjkim 245276864Sjkim 3. Check DSA/ECDSA signatures use DER. 246276864Sjkim 247276864Sjkim Reencode DSA/ECDSA signatures and compare with the original received 248276864Sjkim signature. Return an error if there is a mismatch. 249276864Sjkim 250276864Sjkim This will reject various cases including garbage after signature 251276864Sjkim (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS 252276864Sjkim program for discovering this case) and use of BER or invalid ASN.1 INTEGERs 253276864Sjkim (negative or with leading zeroes). 254276864Sjkim 255276864Sjkim Further analysis was conducted and fixes were developed by Stephen Henson 256276864Sjkim of the OpenSSL core team. 257276864Sjkim 258276864Sjkim (CVE-2014-8275) 259276864Sjkim [Steve Henson] 260276864Sjkim 261276864Sjkim *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect 262276864Sjkim results on some platforms, including x86_64. This bug occurs at random 263276864Sjkim with a very low probability, and is not known to be exploitable in any 264276864Sjkim way, though its exact impact is difficult to determine. Thanks to Pieter 265276864Sjkim Wuille (Blockstream) who reported this issue and also suggested an initial 266276864Sjkim fix. Further analysis was conducted by the OpenSSL development team and 267276864Sjkim Adam Langley of Google. The final fix was developed by Andy Polyakov of 268276864Sjkim the OpenSSL core team. 269276864Sjkim (CVE-2014-3570) 270276864Sjkim [Andy Polyakov] 271276864Sjkim 272276864Sjkim *) Do not resume sessions on the server if the negotiated protocol 273276864Sjkim version does not match the session's version. Resuming with a different 274276864Sjkim version, while not strictly forbidden by the RFC, is of questionable 275276864Sjkim sanity and breaks all known clients. 276276864Sjkim [David Benjamin, Emilia K�sper] 277276864Sjkim 278276864Sjkim *) Tighten handling of the ChangeCipherSpec (CCS) message: reject 279276864Sjkim early CCS messages during renegotiation. (Note that because 280276864Sjkim renegotiation is encrypted, this early CCS was not exploitable.) 281276864Sjkim [Emilia K�sper] 282276864Sjkim 283276864Sjkim *) Tighten client-side session ticket handling during renegotiation: 284276864Sjkim ensure that the client only accepts a session ticket if the server sends 285276864Sjkim the extension anew in the ServerHello. Previously, a TLS client would 286276864Sjkim reuse the old extension state and thus accept a session ticket if one was 287276864Sjkim announced in the initial ServerHello. 288276864Sjkim 289276864Sjkim Similarly, ensure that the client requires a session ticket if one 290276864Sjkim was advertised in the ServerHello. Previously, a TLS client would 291276864Sjkim ignore a missing NewSessionTicket message. 292276864Sjkim [Emilia K�sper] 293276864Sjkim 294273149Sjkim Changes between 1.0.1i and 1.0.1j [15 Oct 2014] 295273149Sjkim 296273149Sjkim *) SRTP Memory Leak. 297273149Sjkim 298273149Sjkim A flaw in the DTLS SRTP extension parsing code allows an attacker, who 299273149Sjkim sends a carefully crafted handshake message, to cause OpenSSL to fail 300273149Sjkim to free up to 64k of memory causing a memory leak. This could be 301273149Sjkim exploited in a Denial Of Service attack. This issue affects OpenSSL 302273149Sjkim 1.0.1 server implementations for both SSL/TLS and DTLS regardless of 303273149Sjkim whether SRTP is used or configured. Implementations of OpenSSL that 304273149Sjkim have been compiled with OPENSSL_NO_SRTP defined are not affected. 305273149Sjkim 306273149Sjkim The fix was developed by the OpenSSL team. 307273149Sjkim (CVE-2014-3513) 308273149Sjkim [OpenSSL team] 309273149Sjkim 310273149Sjkim *) Session Ticket Memory Leak. 311273149Sjkim 312273149Sjkim When an OpenSSL SSL/TLS/DTLS server receives a session ticket the 313273149Sjkim integrity of that ticket is first verified. In the event of a session 314273149Sjkim ticket integrity check failing, OpenSSL will fail to free memory 315273149Sjkim causing a memory leak. By sending a large number of invalid session 316273149Sjkim tickets an attacker could exploit this issue in a Denial Of Service 317273149Sjkim attack. 318273149Sjkim (CVE-2014-3567) 319273149Sjkim [Steve Henson] 320273149Sjkim 321273149Sjkim *) Build option no-ssl3 is incomplete. 322273149Sjkim 323273149Sjkim When OpenSSL is configured with "no-ssl3" as a build option, servers 324273149Sjkim could accept and complete a SSL 3.0 handshake, and clients could be 325273149Sjkim configured to send them. 326273149Sjkim (CVE-2014-3568) 327273149Sjkim [Akamai and the OpenSSL team] 328273149Sjkim 329273149Sjkim *) Add support for TLS_FALLBACK_SCSV. 330273149Sjkim Client applications doing fallback retries should call 331273149Sjkim SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV). 332273149Sjkim (CVE-2014-3566) 333273149Sjkim [Adam Langley, Bodo Moeller] 334273149Sjkim 335273149Sjkim *) Add additional DigestInfo checks. 336273149Sjkim 337273149Sjkim Reencode DigestInto in DER and check against the original when 338273149Sjkim verifying RSA signature: this will reject any improperly encoded 339273149Sjkim DigestInfo structures. 340273149Sjkim 341273149Sjkim Note: this is a precautionary measure and no attacks are currently known. 342273149Sjkim 343273149Sjkim [Steve Henson] 344273149Sjkim 345269686Sjkim Changes between 1.0.1h and 1.0.1i [6 Aug 2014] 346269686Sjkim 347269686Sjkim *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the 348269686Sjkim SRP code can be overrun an internal buffer. Add sanity check that 349269686Sjkim g, A, B < N to SRP code. 350269686Sjkim 351269686Sjkim Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC 352269686Sjkim Group for discovering this issue. 353269686Sjkim (CVE-2014-3512) 354269686Sjkim [Steve Henson] 355269686Sjkim 356269686Sjkim *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate 357269686Sjkim TLS 1.0 instead of higher protocol versions when the ClientHello message 358269686Sjkim is badly fragmented. This allows a man-in-the-middle attacker to force a 359269686Sjkim downgrade to TLS 1.0 even if both the server and the client support a 360269686Sjkim higher protocol version, by modifying the client's TLS records. 361269686Sjkim 362269686Sjkim Thanks to David Benjamin and Adam Langley (Google) for discovering and 363269686Sjkim researching this issue. 364269686Sjkim (CVE-2014-3511) 365269686Sjkim [David Benjamin] 366269686Sjkim 367269686Sjkim *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject 368269686Sjkim to a denial of service attack. A malicious server can crash the client 369269686Sjkim with a null pointer dereference (read) by specifying an anonymous (EC)DH 370269686Sjkim ciphersuite and sending carefully crafted handshake messages. 371269686Sjkim 372269686Sjkim Thanks to Felix Gr�bert (Google) for discovering and researching this 373269686Sjkim issue. 374269686Sjkim (CVE-2014-3510) 375269686Sjkim [Emilia K�sper] 376269686Sjkim 377269686Sjkim *) By sending carefully crafted DTLS packets an attacker could cause openssl 378269686Sjkim to leak memory. This can be exploited through a Denial of Service attack. 379269686Sjkim Thanks to Adam Langley for discovering and researching this issue. 380269686Sjkim (CVE-2014-3507) 381269686Sjkim [Adam Langley] 382269686Sjkim 383269686Sjkim *) An attacker can force openssl to consume large amounts of memory whilst 384269686Sjkim processing DTLS handshake messages. This can be exploited through a 385269686Sjkim Denial of Service attack. 386269686Sjkim Thanks to Adam Langley for discovering and researching this issue. 387269686Sjkim (CVE-2014-3506) 388269686Sjkim [Adam Langley] 389269686Sjkim 390269686Sjkim *) An attacker can force an error condition which causes openssl to crash 391269686Sjkim whilst processing DTLS packets due to memory being freed twice. This 392269686Sjkim can be exploited through a Denial of Service attack. 393269686Sjkim Thanks to Adam Langley and Wan-Teh Chang for discovering and researching 394269686Sjkim this issue. 395269686Sjkim (CVE-2014-3505) 396269686Sjkim [Adam Langley] 397269686Sjkim 398269686Sjkim *) If a multithreaded client connects to a malicious server using a resumed 399269686Sjkim session and the server sends an ec point format extension it could write 400269686Sjkim up to 255 bytes to freed memory. 401269686Sjkim 402269686Sjkim Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this 403269686Sjkim issue. 404269686Sjkim (CVE-2014-3509) 405269686Sjkim [Gabor Tyukasz] 406269686Sjkim 407269686Sjkim *) A malicious server can crash an OpenSSL client with a null pointer 408269686Sjkim dereference (read) by specifying an SRP ciphersuite even though it was not 409269686Sjkim properly negotiated with the client. This can be exploited through a 410269686Sjkim Denial of Service attack. 411269686Sjkim 412269686Sjkim Thanks to Joonas Kuorilehto and Riku Hietam�ki (Codenomicon) for 413269686Sjkim discovering and researching this issue. 414269686Sjkim (CVE-2014-5139) 415269686Sjkim [Steve Henson] 416269686Sjkim 417269686Sjkim *) A flaw in OBJ_obj2txt may cause pretty printing functions such as 418269686Sjkim X509_name_oneline, X509_name_print_ex et al. to leak some information 419269686Sjkim from the stack. Applications may be affected if they echo pretty printing 420269686Sjkim output to the attacker. 421269686Sjkim 422269686Sjkim Thanks to Ivan Fratric (Google) for discovering this issue. 423269686Sjkim (CVE-2014-3508) 424269686Sjkim [Emilia K�sper, and Steve Henson] 425269686Sjkim 426269686Sjkim *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.) 427269686Sjkim for corner cases. (Certain input points at infinity could lead to 428269686Sjkim bogus results, with non-infinity inputs mapped to infinity too.) 429269686Sjkim [Bodo Moeller] 430269686Sjkim 431267258Sjkim Changes between 1.0.1g and 1.0.1h [5 Jun 2014] 432267258Sjkim 433267258Sjkim *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted 434267258Sjkim handshake can force the use of weak keying material in OpenSSL 435267258Sjkim SSL/TLS clients and servers. 436267258Sjkim 437267258Sjkim Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and 438267258Sjkim researching this issue. (CVE-2014-0224) 439267258Sjkim [KIKUCHI Masashi, Steve Henson] 440267258Sjkim 441267258Sjkim *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an 442267258Sjkim OpenSSL DTLS client the code can be made to recurse eventually crashing 443267258Sjkim in a DoS attack. 444267258Sjkim 445267258Sjkim Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue. 446267258Sjkim (CVE-2014-0221) 447267258Sjkim [Imre Rad, Steve Henson] 448267258Sjkim 449267258Sjkim *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can 450267258Sjkim be triggered by sending invalid DTLS fragments to an OpenSSL DTLS 451267258Sjkim client or server. This is potentially exploitable to run arbitrary 452267258Sjkim code on a vulnerable client or server. 453267258Sjkim 454267258Sjkim Thanks to J�ri Aedla for reporting this issue. (CVE-2014-0195) 455267258Sjkim [J�ri Aedla, Steve Henson] 456267258Sjkim 457267258Sjkim *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites 458267258Sjkim are subject to a denial of service attack. 459267258Sjkim 460267258Sjkim Thanks to Felix Gr�bert and Ivan Fratric at Google for discovering 461267258Sjkim this issue. (CVE-2014-3470) 462267258Sjkim [Felix Gr�bert, Ivan Fratric, Steve Henson] 463267258Sjkim 464267258Sjkim *) Harmonize version and its documentation. -f flag is used to display 465267258Sjkim compilation flags. 466267258Sjkim [mancha <mancha1@zoho.com>] 467267258Sjkim 468267258Sjkim *) Fix eckey_priv_encode so it immediately returns an error upon a failure 469267258Sjkim in i2d_ECPrivateKey. 470267258Sjkim [mancha <mancha1@zoho.com>] 471267258Sjkim 472267258Sjkim *) Fix some double frees. These are not thought to be exploitable. 473267258Sjkim [mancha <mancha1@zoho.com>] 474267258Sjkim 475264331Sjkim Changes between 1.0.1f and 1.0.1g [7 Apr 2014] 476264331Sjkim 477264331Sjkim *) A missing bounds check in the handling of the TLS heartbeat extension 478264331Sjkim can be used to reveal up to 64k of memory to a connected client or 479264331Sjkim server. 480264331Sjkim 481264331Sjkim Thanks for Neel Mehta of Google Security for discovering this bug and to 482264331Sjkim Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for 483264331Sjkim preparing the fix (CVE-2014-0160) 484264331Sjkim [Adam Langley, Bodo Moeller] 485264331Sjkim 486264331Sjkim *) Fix for the attack described in the paper "Recovering OpenSSL 487264331Sjkim ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack" 488264331Sjkim by Yuval Yarom and Naomi Benger. Details can be obtained from: 489264331Sjkim http://eprint.iacr.org/2014/140 490264331Sjkim 491264331Sjkim Thanks to Yuval Yarom and Naomi Benger for discovering this 492264331Sjkim flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076) 493264331Sjkim [Yuval Yarom and Naomi Benger] 494264331Sjkim 495264331Sjkim *) TLS pad extension: draft-agl-tls-padding-03 496264331Sjkim 497264331Sjkim Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the 498264331Sjkim TLS client Hello record length value would otherwise be > 255 and 499264331Sjkim less that 512 pad with a dummy extension containing zeroes so it 500264331Sjkim is at least 512 bytes long. 501264331Sjkim 502264331Sjkim [Adam Langley, Steve Henson] 503264331Sjkim 504264331Sjkim Changes between 1.0.1e and 1.0.1f [6 Jan 2014] 505264331Sjkim 506264331Sjkim *) Fix for TLS record tampering bug. A carefully crafted invalid 507264331Sjkim handshake could crash OpenSSL with a NULL pointer exception. 508264331Sjkim Thanks to Anton Johansson for reporting this issues. 509264331Sjkim (CVE-2013-4353) 510264331Sjkim 511264331Sjkim *) Keep original DTLS digest and encryption contexts in retransmission 512264331Sjkim structures so we can use the previous session parameters if they need 513264331Sjkim to be resent. (CVE-2013-6450) 514264331Sjkim [Steve Henson] 515264331Sjkim 516264331Sjkim *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which 517264331Sjkim avoids preferring ECDHE-ECDSA ciphers when the client appears to be 518264331Sjkim Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for 519264331Sjkim several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug 520264331Sjkim is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing 521264331Sjkim 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer. 522264331Sjkim [Rob Stradling, Adam Langley] 523264331Sjkim 524246772Sjkim Changes between 1.0.1d and 1.0.1e [11 Feb 2013] 525246772Sjkim 526264331Sjkim *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI 527264331Sjkim supporting platforms or when small records were transferred. 528264331Sjkim [Andy Polyakov, Steve Henson] 529246772Sjkim 530246772Sjkim Changes between 1.0.1c and 1.0.1d [5 Feb 2013] 531246772Sjkim 532246772Sjkim *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time. 533246772Sjkim 534246772Sjkim This addresses the flaw in CBC record processing discovered by 535246772Sjkim Nadhem Alfardan and Kenny Paterson. Details of this attack can be found 536246772Sjkim at: http://www.isg.rhul.ac.uk/tls/ 537246772Sjkim 538246772Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 539246772Sjkim Security Group at Royal Holloway, University of London 540246772Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and 541246772Sjkim Emilia K�sper for the initial patch. 542246772Sjkim (CVE-2013-0169) 543246772Sjkim [Emilia K�sper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson] 544246772Sjkim 545246772Sjkim *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode 546246772Sjkim ciphersuites which can be exploited in a denial of service attack. 547246772Sjkim Thanks go to and to Adam Langley <agl@chromium.org> for discovering 548246772Sjkim and detecting this bug and to Wolfgang Ettlinger 549246772Sjkim <wolfgang.ettlinger@gmail.com> for independently discovering this issue. 550246772Sjkim (CVE-2012-2686) 551246772Sjkim [Adam Langley] 552246772Sjkim 553246772Sjkim *) Return an error when checking OCSP signatures when key is NULL. 554246772Sjkim This fixes a DoS attack. (CVE-2013-0166) 555246772Sjkim [Steve Henson] 556246772Sjkim 557246772Sjkim *) Make openssl verify return errors. 558246772Sjkim [Chris Palmer <palmer@google.com> and Ben Laurie] 559246772Sjkim 560246772Sjkim *) Call OCSP Stapling callback after ciphersuite has been chosen, so 561246772Sjkim the right response is stapled. Also change SSL_get_certificate() 562246772Sjkim so it returns the certificate actually sent. 563246772Sjkim See http://rt.openssl.org/Ticket/Display.html?id=2836. 564246772Sjkim [Rob Stradling <rob.stradling@comodo.com>] 565246772Sjkim 566246772Sjkim *) Fix possible deadlock when decoding public keys. 567246772Sjkim [Steve Henson] 568246772Sjkim 569246772Sjkim *) Don't use TLS 1.0 record version number in initial client hello 570246772Sjkim if renegotiating. 571246772Sjkim [Steve Henson] 572246772Sjkim 573238405Sjkim Changes between 1.0.1b and 1.0.1c [10 May 2012] 574237657Sjkim 575238405Sjkim *) Sanity check record length before skipping explicit IV in TLS 576238405Sjkim 1.2, 1.1 and DTLS to fix DoS attack. 577237657Sjkim 578237657Sjkim Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic 579237657Sjkim fuzzing as a service testing platform. 580237657Sjkim (CVE-2012-2333) 581237657Sjkim [Steve Henson] 582237657Sjkim 583237657Sjkim *) Initialise tkeylen properly when encrypting CMS messages. 584237657Sjkim Thanks to Solar Designer of Openwall for reporting this issue. 585237657Sjkim [Steve Henson] 586237657Sjkim 587238405Sjkim *) In FIPS mode don't try to use composite ciphers as they are not 588238405Sjkim approved. 589238405Sjkim [Steve Henson] 590237657Sjkim 591238405Sjkim Changes between 1.0.1a and 1.0.1b [26 Apr 2012] 592237657Sjkim 593238405Sjkim *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and 594238405Sjkim 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately 595238405Sjkim mean any application compiled against OpenSSL 1.0.0 headers setting 596238405Sjkim SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng 597238405Sjkim TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to 598238405Sjkim 0x10000000L Any application which was previously compiled against 599238405Sjkim OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1 600238405Sjkim will need to be recompiled as a result. Letting be results in 601238405Sjkim inability to disable specifically TLS 1.1 and in client context, 602238405Sjkim in unlike event, limit maximum offered version to TLS 1.0 [see below]. 603238405Sjkim [Steve Henson] 604237657Sjkim 605238405Sjkim *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not 606238405Sjkim disable just protocol X, but all protocols above X *if* there are 607238405Sjkim protocols *below* X still enabled. In more practical terms it means 608238405Sjkim that if application wants to disable TLS1.0 in favor of TLS1.1 and 609238405Sjkim above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass 610238405Sjkim SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to 611238405Sjkim client side. 612238405Sjkim [Andy Polyakov] 613238405Sjkim 614238405Sjkim Changes between 1.0.1 and 1.0.1a [19 Apr 2012] 615238405Sjkim 616237657Sjkim *) Check for potentially exploitable overflows in asn1_d2i_read_bio 617237657Sjkim BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer 618237657Sjkim in CRYPTO_realloc_clean. 619237657Sjkim 620237657Sjkim Thanks to Tavis Ormandy, Google Security Team, for discovering this 621237657Sjkim issue and to Adam Langley <agl@chromium.org> for fixing it. 622237657Sjkim (CVE-2012-2110) 623237657Sjkim [Adam Langley (Google), Tavis Ormandy, Google Security Team] 624237657Sjkim 625238405Sjkim *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections. 626238405Sjkim [Adam Langley] 627237657Sjkim 628238405Sjkim *) Workarounds for some broken servers that "hang" if a client hello 629238405Sjkim record length exceeds 255 bytes. 630238405Sjkim 631238405Sjkim 1. Do not use record version number > TLS 1.0 in initial client 632238405Sjkim hello: some (but not all) hanging servers will now work. 633238405Sjkim 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate 634238405Sjkim the number of ciphers sent in the client hello. This should be 635238405Sjkim set to an even number, such as 50, for example by passing: 636238405Sjkim -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure. 637238405Sjkim Most broken servers should now work. 638238405Sjkim 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable 639238405Sjkim TLS 1.2 client support entirely. 640238405Sjkim [Steve Henson] 641238405Sjkim 642238405Sjkim *) Fix SEGV in Vector Permutation AES module observed in OpenSSH. 643238405Sjkim [Andy Polyakov] 644238405Sjkim 645238405Sjkim Changes between 1.0.0h and 1.0.1 [14 Mar 2012] 646238405Sjkim 647238405Sjkim *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET 648238405Sjkim STRING form instead of a DigestInfo. 649238405Sjkim [Steve Henson] 650238405Sjkim 651238405Sjkim *) The format used for MDC2 RSA signatures is inconsistent between EVP 652238405Sjkim and the RSA_sign/RSA_verify functions. This was made more apparent when 653238405Sjkim OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular 654238405Sjkim those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect 655238405Sjkim the correct format in RSA_verify so both forms transparently work. 656238405Sjkim [Steve Henson] 657238405Sjkim 658238405Sjkim *) Some servers which support TLS 1.0 can choke if we initially indicate 659238405Sjkim support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA 660238405Sjkim encrypted premaster secret. As a workaround use the maximum pemitted 661238405Sjkim client version in client hello, this should keep such servers happy 662238405Sjkim and still work with previous versions of OpenSSL. 663238405Sjkim [Steve Henson] 664238405Sjkim 665238405Sjkim *) Add support for TLS/DTLS heartbeats. 666238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>] 667238405Sjkim 668238405Sjkim *) Add support for SCTP. 669238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>] 670238405Sjkim 671238405Sjkim *) Improved PRNG seeding for VOS. 672238405Sjkim [Paul Green <Paul.Green@stratus.com>] 673238405Sjkim 674238405Sjkim *) Extensive assembler packs updates, most notably: 675238405Sjkim 676238405Sjkim - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support; 677238405Sjkim - x86[_64]: SSSE3 support (SHA1, vector-permutation AES); 678238405Sjkim - x86_64: bit-sliced AES implementation; 679238405Sjkim - ARM: NEON support, contemporary platforms optimizations; 680238405Sjkim - s390x: z196 support; 681238405Sjkim - *: GHASH and GF(2^m) multiplication implementations; 682238405Sjkim 683238405Sjkim [Andy Polyakov] 684238405Sjkim 685238405Sjkim *) Make TLS-SRP code conformant with RFC 5054 API cleanup 686238405Sjkim (removal of unnecessary code) 687238405Sjkim [Peter Sylvester <peter.sylvester@edelweb.fr>] 688238405Sjkim 689238405Sjkim *) Add TLS key material exporter from RFC 5705. 690238405Sjkim [Eric Rescorla] 691238405Sjkim 692238405Sjkim *) Add DTLS-SRTP negotiation from RFC 5764. 693238405Sjkim [Eric Rescorla] 694238405Sjkim 695238405Sjkim *) Add Next Protocol Negotiation, 696238405Sjkim http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be 697238405Sjkim disabled with a no-npn flag to config or Configure. Code donated 698238405Sjkim by Google. 699238405Sjkim [Adam Langley <agl@google.com> and Ben Laurie] 700238405Sjkim 701238405Sjkim *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224, 702238405Sjkim NIST-P256, NIST-P521, with constant-time single point multiplication on 703238405Sjkim typical inputs. Compiler support for the nonstandard type __uint128_t is 704238405Sjkim required to use this (present in gcc 4.4 and later, for 64-bit builds). 705238405Sjkim Code made available under Apache License version 2.0. 706238405Sjkim 707238405Sjkim Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command 708238405Sjkim line to include this in your build of OpenSSL, and run "make depend" (or 709238405Sjkim "make update"). This enables the following EC_METHODs: 710238405Sjkim 711238405Sjkim EC_GFp_nistp224_method() 712238405Sjkim EC_GFp_nistp256_method() 713238405Sjkim EC_GFp_nistp521_method() 714238405Sjkim 715238405Sjkim EC_GROUP_new_by_curve_name() will automatically use these (while 716238405Sjkim EC_GROUP_new_curve_GFp() currently prefers the more flexible 717238405Sjkim implementations). 718238405Sjkim [Emilia K�sper, Adam Langley, Bodo Moeller (Google)] 719238405Sjkim 720238405Sjkim *) Use type ossl_ssize_t instad of ssize_t which isn't available on 721238405Sjkim all platforms. Move ssize_t definition from e_os.h to the public 722238405Sjkim header file e_os2.h as it now appears in public header file cms.h 723238405Sjkim [Steve Henson] 724238405Sjkim 725238405Sjkim *) New -sigopt option to the ca, req and x509 utilities. Additional 726238405Sjkim signature parameters can be passed using this option and in 727238405Sjkim particular PSS. 728238405Sjkim [Steve Henson] 729238405Sjkim 730238405Sjkim *) Add RSA PSS signing function. This will generate and set the 731238405Sjkim appropriate AlgorithmIdentifiers for PSS based on those in the 732238405Sjkim corresponding EVP_MD_CTX structure. No application support yet. 733238405Sjkim [Steve Henson] 734238405Sjkim 735238405Sjkim *) Support for companion algorithm specific ASN1 signing routines. 736238405Sjkim New function ASN1_item_sign_ctx() signs a pre-initialised 737238405Sjkim EVP_MD_CTX structure and sets AlgorithmIdentifiers based on 738238405Sjkim the appropriate parameters. 739238405Sjkim [Steve Henson] 740238405Sjkim 741238405Sjkim *) Add new algorithm specific ASN1 verification initialisation function 742238405Sjkim to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1 743238405Sjkim handling will be the same no matter what EVP_PKEY_METHOD is used. 744238405Sjkim Add a PSS handler to support verification of PSS signatures: checked 745238405Sjkim against a number of sample certificates. 746238405Sjkim [Steve Henson] 747238405Sjkim 748238405Sjkim *) Add signature printing for PSS. Add PSS OIDs. 749238405Sjkim [Steve Henson, Martin Kaiser <lists@kaiser.cx>] 750238405Sjkim 751238405Sjkim *) Add algorithm specific signature printing. An individual ASN1 method 752238405Sjkim can now print out signatures instead of the standard hex dump. 753238405Sjkim 754238405Sjkim More complex signatures (e.g. PSS) can print out more meaningful 755238405Sjkim information. Include DSA version that prints out the signature 756238405Sjkim parameters r, s. 757238405Sjkim [Steve Henson] 758238405Sjkim 759238405Sjkim *) Password based recipient info support for CMS library: implementing 760238405Sjkim RFC3211. 761238405Sjkim [Steve Henson] 762238405Sjkim 763238405Sjkim *) Split password based encryption into PBES2 and PBKDF2 functions. This 764238405Sjkim neatly separates the code into cipher and PBE sections and is required 765238405Sjkim for some algorithms that split PBES2 into separate pieces (such as 766238405Sjkim password based CMS). 767238405Sjkim [Steve Henson] 768238405Sjkim 769238405Sjkim *) Session-handling fixes: 770238405Sjkim - Fix handling of connections that are resuming with a session ID, 771238405Sjkim but also support Session Tickets. 772238405Sjkim - Fix a bug that suppressed issuing of a new ticket if the client 773238405Sjkim presented a ticket with an expired session. 774238405Sjkim - Try to set the ticket lifetime hint to something reasonable. 775238405Sjkim - Make tickets shorter by excluding irrelevant information. 776238405Sjkim - On the client side, don't ignore renewed tickets. 777238405Sjkim [Adam Langley, Bodo Moeller (Google)] 778238405Sjkim 779238405Sjkim *) Fix PSK session representation. 780238405Sjkim [Bodo Moeller] 781238405Sjkim 782238405Sjkim *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations. 783238405Sjkim 784238405Sjkim This work was sponsored by Intel. 785238405Sjkim [Andy Polyakov] 786238405Sjkim 787238405Sjkim *) Add GCM support to TLS library. Some custom code is needed to split 788238405Sjkim the IV between the fixed (from PRF) and explicit (from TLS record) 789238405Sjkim portions. This adds all GCM ciphersuites supported by RFC5288 and 790238405Sjkim RFC5289. Generalise some AES* cipherstrings to inlclude GCM and 791238405Sjkim add a special AESGCM string for GCM only. 792238405Sjkim [Steve Henson] 793238405Sjkim 794238405Sjkim *) Expand range of ctrls for AES GCM. Permit setting invocation 795238405Sjkim field on decrypt and retrieval of invocation field only on encrypt. 796238405Sjkim [Steve Henson] 797238405Sjkim 798238405Sjkim *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support. 799238405Sjkim As required by RFC5289 these ciphersuites cannot be used if for 800238405Sjkim versions of TLS earlier than 1.2. 801238405Sjkim [Steve Henson] 802238405Sjkim 803238405Sjkim *) For FIPS capable OpenSSL interpret a NULL default public key method 804238405Sjkim as unset and return the appopriate default but do *not* set the default. 805238405Sjkim This means we can return the appopriate method in applications that 806238405Sjkim swicth between FIPS and non-FIPS modes. 807238405Sjkim [Steve Henson] 808238405Sjkim 809238405Sjkim *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an 810238405Sjkim ENGINE is used then we cannot handle that in the FIPS module so we 811238405Sjkim keep original code iff non-FIPS operations are allowed. 812238405Sjkim [Steve Henson] 813238405Sjkim 814238405Sjkim *) Add -attime option to openssl utilities. 815238405Sjkim [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson] 816238405Sjkim 817238405Sjkim *) Redirect DSA and DH operations to FIPS module in FIPS mode. 818238405Sjkim [Steve Henson] 819238405Sjkim 820238405Sjkim *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use 821238405Sjkim FIPS EC methods unconditionally for now. 822238405Sjkim [Steve Henson] 823238405Sjkim 824238405Sjkim *) New build option no-ec2m to disable characteristic 2 code. 825238405Sjkim [Steve Henson] 826238405Sjkim 827238405Sjkim *) Backport libcrypto audit of return value checking from 1.1.0-dev; not 828238405Sjkim all cases can be covered as some introduce binary incompatibilities. 829238405Sjkim [Steve Henson] 830238405Sjkim 831238405Sjkim *) Redirect RSA operations to FIPS module including keygen, 832238405Sjkim encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods. 833238405Sjkim [Steve Henson] 834238405Sjkim 835238405Sjkim *) Add similar low level API blocking to ciphers. 836238405Sjkim [Steve Henson] 837238405Sjkim 838238405Sjkim *) Low level digest APIs are not approved in FIPS mode: any attempt 839238405Sjkim to use these will cause a fatal error. Applications that *really* want 840238405Sjkim to use them can use the private_* version instead. 841238405Sjkim [Steve Henson] 842238405Sjkim 843238405Sjkim *) Redirect cipher operations to FIPS module for FIPS builds. 844238405Sjkim [Steve Henson] 845238405Sjkim 846238405Sjkim *) Redirect digest operations to FIPS module for FIPS builds. 847238405Sjkim [Steve Henson] 848238405Sjkim 849238405Sjkim *) Update build system to add "fips" flag which will link in fipscanister.o 850238405Sjkim for static and shared library builds embedding a signature if needed. 851238405Sjkim [Steve Henson] 852238405Sjkim 853238405Sjkim *) Output TLS supported curves in preference order instead of numerical 854238405Sjkim order. This is currently hardcoded for the highest order curves first. 855238405Sjkim This should be configurable so applications can judge speed vs strength. 856238405Sjkim [Steve Henson] 857238405Sjkim 858238405Sjkim *) Add TLS v1.2 server support for client authentication. 859238405Sjkim [Steve Henson] 860238405Sjkim 861238405Sjkim *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers 862238405Sjkim and enable MD5. 863238405Sjkim [Steve Henson] 864238405Sjkim 865238405Sjkim *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying 866238405Sjkim FIPS modules versions. 867238405Sjkim [Steve Henson] 868238405Sjkim 869238405Sjkim *) Add TLS v1.2 client side support for client authentication. Keep cache 870238405Sjkim of handshake records longer as we don't know the hash algorithm to use 871238405Sjkim until after the certificate request message is received. 872238405Sjkim [Steve Henson] 873238405Sjkim 874238405Sjkim *) Initial TLS v1.2 client support. Add a default signature algorithms 875238405Sjkim extension including all the algorithms we support. Parse new signature 876238405Sjkim format in client key exchange. Relax some ECC signing restrictions for 877238405Sjkim TLS v1.2 as indicated in RFC5246. 878238405Sjkim [Steve Henson] 879238405Sjkim 880238405Sjkim *) Add server support for TLS v1.2 signature algorithms extension. Switch 881238405Sjkim to new signature format when needed using client digest preference. 882238405Sjkim All server ciphersuites should now work correctly in TLS v1.2. No client 883238405Sjkim support yet and no support for client certificates. 884238405Sjkim [Steve Henson] 885238405Sjkim 886238405Sjkim *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch 887238405Sjkim to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based 888238405Sjkim ciphersuites. At present only RSA key exchange ciphersuites work with 889238405Sjkim TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete 890238405Sjkim SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods 891238405Sjkim and version checking. 892238405Sjkim [Steve Henson] 893238405Sjkim 894238405Sjkim *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled 895238405Sjkim with this defined it will not be affected by any changes to ssl internal 896238405Sjkim structures. Add several utility functions to allow openssl application 897238405Sjkim to work with OPENSSL_NO_SSL_INTERN defined. 898238405Sjkim [Steve Henson] 899238405Sjkim 900238405Sjkim *) Add SRP support. 901238405Sjkim [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie] 902238405Sjkim 903238405Sjkim *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id. 904238405Sjkim [Steve Henson] 905238405Sjkim 906238405Sjkim *) Permit abbreviated handshakes when renegotiating using the function 907238405Sjkim SSL_renegotiate_abbreviated(). 908238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>] 909238405Sjkim 910238405Sjkim *) Add call to ENGINE_register_all_complete() to 911238405Sjkim ENGINE_load_builtin_engines(), so some implementations get used 912238405Sjkim automatically instead of needing explicit application support. 913238405Sjkim [Steve Henson] 914238405Sjkim 915238405Sjkim *) Add support for TLS key exporter as described in RFC5705. 916238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson] 917238405Sjkim 918238405Sjkim *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only 919238405Sjkim a few changes are required: 920238405Sjkim 921238405Sjkim Add SSL_OP_NO_TLSv1_1 flag. 922238405Sjkim Add TLSv1_1 methods. 923238405Sjkim Update version checking logic to handle version 1.1. 924238405Sjkim Add explicit IV handling (ported from DTLS code). 925238405Sjkim Add command line options to s_client/s_server. 926238405Sjkim [Steve Henson] 927238405Sjkim 928238405Sjkim Changes between 1.0.0g and 1.0.0h [12 Mar 2012] 929238405Sjkim 930237657Sjkim *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness 931237657Sjkim in CMS and PKCS7 code. When RSA decryption fails use a random key for 932237657Sjkim content decryption and always return the same error. Note: this attack 933237657Sjkim needs on average 2^20 messages so it only affects automated senders. The 934237657Sjkim old behaviour can be reenabled in the CMS code by setting the 935237657Sjkim CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where 936237657Sjkim an MMA defence is not necessary. 937237657Sjkim Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering 938237657Sjkim this issue. (CVE-2012-0884) 939237657Sjkim [Steve Henson] 940237657Sjkim 941237657Sjkim *) Fix CVE-2011-4619: make sure we really are receiving a 942237657Sjkim client hello before rejecting multiple SGC restarts. Thanks to 943237657Sjkim Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug. 944237657Sjkim [Steve Henson] 945237657Sjkim 946238405Sjkim Changes between 1.0.0f and 1.0.0g [18 Jan 2012] 947238405Sjkim 948238405Sjkim *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109. 949238405Sjkim Thanks to Antonio Martin, Enterprise Secure Access Research and 950238405Sjkim Development, Cisco Systems, Inc. for discovering this bug and 951238405Sjkim preparing a fix. (CVE-2012-0050) 952238405Sjkim [Antonio Martin] 953238405Sjkim 954238405Sjkim Changes between 1.0.0e and 1.0.0f [4 Jan 2012] 955238405Sjkim 956238405Sjkim *) Nadhem Alfardan and Kenny Paterson have discovered an extension 957238405Sjkim of the Vaudenay padding oracle attack on CBC mode encryption 958238405Sjkim which enables an efficient plaintext recovery attack against 959238405Sjkim the OpenSSL implementation of DTLS. Their attack exploits timing 960238405Sjkim differences arising during decryption processing. A research 961238405Sjkim paper describing this attack can be found at: 962238405Sjkim http://www.isg.rhul.ac.uk/~kp/dtls.pdf 963238405Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 964238405Sjkim Security Group at Royal Holloway, University of London 965238405Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann 966238405Sjkim <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de> 967238405Sjkim for preparing the fix. (CVE-2011-4108) 968238405Sjkim [Robin Seggelmann, Michael Tuexen] 969238405Sjkim 970238405Sjkim *) Clear bytes used for block padding of SSL 3.0 records. 971238405Sjkim (CVE-2011-4576) 972238405Sjkim [Adam Langley (Google)] 973238405Sjkim 974238405Sjkim *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George 975238405Sjkim Kadianakis <desnacked@gmail.com> for discovering this issue and 976238405Sjkim Adam Langley for preparing the fix. (CVE-2011-4619) 977238405Sjkim [Adam Langley (Google)] 978238405Sjkim 979238405Sjkim *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027) 980238405Sjkim [Andrey Kulikov <amdeich@gmail.com>] 981238405Sjkim 982238405Sjkim *) Prevent malformed RFC3779 data triggering an assertion failure. 983238405Sjkim Thanks to Andrew Chi, BBN Technologies, for discovering the flaw 984238405Sjkim and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577) 985238405Sjkim [Rob Austein <sra@hactrn.net>] 986238405Sjkim 987238405Sjkim *) Improved PRNG seeding for VOS. 988238405Sjkim [Paul Green <Paul.Green@stratus.com>] 989238405Sjkim 990238405Sjkim *) Fix ssl_ciph.c set-up race. 991238405Sjkim [Adam Langley (Google)] 992238405Sjkim 993238405Sjkim *) Fix spurious failures in ecdsatest.c. 994238405Sjkim [Emilia K�sper (Google)] 995238405Sjkim 996238405Sjkim *) Fix the BIO_f_buffer() implementation (which was mixing different 997238405Sjkim interpretations of the '..._len' fields). 998238405Sjkim [Adam Langley (Google)] 999238405Sjkim 1000238405Sjkim *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than 1001238405Sjkim BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent 1002238405Sjkim threads won't reuse the same blinding coefficients. 1003238405Sjkim 1004238405Sjkim This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING 1005238405Sjkim lock to call BN_BLINDING_invert_ex, and avoids one use of 1006238405Sjkim BN_BLINDING_update for each BN_BLINDING structure (previously, 1007238405Sjkim the last update always remained unused). 1008238405Sjkim [Emilia K�sper (Google)] 1009238405Sjkim 1010238405Sjkim *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf. 1011238405Sjkim [Bob Buckholz (Google)] 1012238405Sjkim 1013238405Sjkim Changes between 1.0.0d and 1.0.0e [6 Sep 2011] 1014238405Sjkim 1015238405Sjkim *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted 1016238405Sjkim by initialising X509_STORE_CTX properly. (CVE-2011-3207) 1017238405Sjkim [Kaspar Brand <ossl@velox.ch>] 1018238405Sjkim 1019238405Sjkim *) Fix SSL memory handling for (EC)DH ciphersuites, in particular 1020238405Sjkim for multi-threaded use of ECDH. (CVE-2011-3210) 1021238405Sjkim [Adam Langley (Google)] 1022238405Sjkim 1023238405Sjkim *) Fix x509_name_ex_d2i memory leak on bad inputs. 1024238405Sjkim [Bodo Moeller] 1025238405Sjkim 1026238405Sjkim *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check 1027238405Sjkim signature public key algorithm by using OID xref utilities instead. 1028238405Sjkim Before this you could only use some ECC ciphersuites with SHA1 only. 1029238405Sjkim [Steve Henson] 1030238405Sjkim 1031238405Sjkim *) Add protection against ECDSA timing attacks as mentioned in the paper 1032238405Sjkim by Billy Bob Brumley and Nicola Tuveri, see: 1033238405Sjkim 1034238405Sjkim http://eprint.iacr.org/2011/232.pdf 1035238405Sjkim 1036238405Sjkim [Billy Bob Brumley and Nicola Tuveri] 1037238405Sjkim 1038238405Sjkim Changes between 1.0.0c and 1.0.0d [8 Feb 2011] 1039238405Sjkim 1040238405Sjkim *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014 1041238405Sjkim [Neel Mehta, Adam Langley, Bodo Moeller (Google)] 1042238405Sjkim 1043238405Sjkim *) Fix bug in string printing code: if *any* escaping is enabled we must 1044238405Sjkim escape the escape character (backslash) or the resulting string is 1045238405Sjkim ambiguous. 1046238405Sjkim [Steve Henson] 1047238405Sjkim 1048238405Sjkim Changes between 1.0.0b and 1.0.0c [2 Dec 2010] 1049238405Sjkim 1050238405Sjkim *) Disable code workaround for ancient and obsolete Netscape browsers 1051238405Sjkim and servers: an attacker can use it in a ciphersuite downgrade attack. 1052238405Sjkim Thanks to Martin Rex for discovering this bug. CVE-2010-4180 1053238405Sjkim [Steve Henson] 1054238405Sjkim 1055238405Sjkim *) Fixed J-PAKE implementation error, originally discovered by 1056238405Sjkim Sebastien Martini, further info and confirmation from Stefan 1057238405Sjkim Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252 1058238405Sjkim [Ben Laurie] 1059238405Sjkim 1060238405Sjkim Changes between 1.0.0a and 1.0.0b [16 Nov 2010] 1061238405Sjkim 1062238405Sjkim *) Fix extension code to avoid race conditions which can result in a buffer 1063238405Sjkim overrun vulnerability: resumed sessions must not be modified as they can 1064238405Sjkim be shared by multiple threads. CVE-2010-3864 1065238405Sjkim [Steve Henson] 1066238405Sjkim 1067238405Sjkim *) Fix WIN32 build system to correctly link an ENGINE directory into 1068238405Sjkim a DLL. 1069238405Sjkim [Steve Henson] 1070238405Sjkim 1071238405Sjkim Changes between 1.0.0 and 1.0.0a [01 Jun 2010] 1072238405Sjkim 1073238405Sjkim *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover 1074238405Sjkim (CVE-2010-1633) 1075238405Sjkim [Steve Henson, Peter-Michael Hager <hager@dortmund.net>] 1076238405Sjkim 1077238405Sjkim Changes between 0.9.8n and 1.0.0 [29 Mar 2010] 1078238405Sjkim 1079238405Sjkim *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher 1080238405Sjkim context. The operation can be customised via the ctrl mechanism in 1081238405Sjkim case ENGINEs want to include additional functionality. 1082238405Sjkim [Steve Henson] 1083238405Sjkim 1084238405Sjkim *) Tolerate yet another broken PKCS#8 key format: private key value negative. 1085238405Sjkim [Steve Henson] 1086238405Sjkim 1087238405Sjkim *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to 1088238405Sjkim output hashes compatible with older versions of OpenSSL. 1089238405Sjkim [Willy Weisz <weisz@vcpc.univie.ac.at>] 1090238405Sjkim 1091238405Sjkim *) Fix compression algorithm handling: if resuming a session use the 1092238405Sjkim compression algorithm of the resumed session instead of determining 1093238405Sjkim it from client hello again. Don't allow server to change algorithm. 1094238405Sjkim [Steve Henson] 1095238405Sjkim 1096238405Sjkim *) Add load_crls() function to apps tidying load_certs() too. Add option 1097238405Sjkim to verify utility to allow additional CRLs to be included. 1098238405Sjkim [Steve Henson] 1099238405Sjkim 1100238405Sjkim *) Update OCSP request code to permit adding custom headers to the request: 1101238405Sjkim some responders need this. 1102238405Sjkim [Steve Henson] 1103238405Sjkim 1104238405Sjkim *) The function EVP_PKEY_sign() returns <=0 on error: check return code 1105238405Sjkim correctly. 1106238405Sjkim [Julia Lawall <julia@diku.dk>] 1107238405Sjkim 1108238405Sjkim *) Update verify callback code in apps/s_cb.c and apps/verify.c, it 1109238405Sjkim needlessly dereferenced structures, used obsolete functions and 1110238405Sjkim didn't handle all updated verify codes correctly. 1111238405Sjkim [Steve Henson] 1112238405Sjkim 1113238405Sjkim *) Disable MD2 in the default configuration. 1114238405Sjkim [Steve Henson] 1115238405Sjkim 1116238405Sjkim *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to 1117238405Sjkim indicate the initial BIO being pushed or popped. This makes it possible 1118238405Sjkim to determine whether the BIO is the one explicitly called or as a result 1119238405Sjkim of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so 1120238405Sjkim it handles reference counts correctly and doesn't zero out the I/O bio 1121238405Sjkim when it is not being explicitly popped. WARNING: applications which 1122238405Sjkim included workarounds for the old buggy behaviour will need to be modified 1123238405Sjkim or they could free up already freed BIOs. 1124238405Sjkim [Steve Henson] 1125238405Sjkim 1126238405Sjkim *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni 1127238405Sjkim renaming to all platforms (within the 0.9.8 branch, this was 1128238405Sjkim done conditionally on Netware platforms to avoid a name clash). 1129238405Sjkim [Guenter <lists@gknw.net>] 1130238405Sjkim 1131238405Sjkim *) Add ECDHE and PSK support to DTLS. 1132238405Sjkim [Michael Tuexen <tuexen@fh-muenster.de>] 1133238405Sjkim 1134238405Sjkim *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't 1135238405Sjkim be used on C++. 1136238405Sjkim [Steve Henson] 1137238405Sjkim 1138238405Sjkim *) Add "missing" function EVP_MD_flags() (without this the only way to 1139238405Sjkim retrieve a digest flags is by accessing the structure directly. Update 1140238405Sjkim EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest 1141238405Sjkim or cipher is registered as in the "from" argument. Print out all 1142238405Sjkim registered digests in the dgst usage message instead of manually 1143238405Sjkim attempting to work them out. 1144238405Sjkim [Steve Henson] 1145238405Sjkim 1146238405Sjkim *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello: 1147238405Sjkim this allows the use of compression and extensions. Change default cipher 1148238405Sjkim string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2 1149238405Sjkim by default unless an application cipher string requests it. 1150238405Sjkim [Steve Henson] 1151238405Sjkim 1152238405Sjkim *) Alter match criteria in PKCS12_parse(). It used to try to use local 1153238405Sjkim key ids to find matching certificates and keys but some PKCS#12 files 1154238405Sjkim don't follow the (somewhat unwritten) rules and this strategy fails. 1155238405Sjkim Now just gather all certificates together and the first private key 1156238405Sjkim then look for the first certificate that matches the key. 1157238405Sjkim [Steve Henson] 1158238405Sjkim 1159238405Sjkim *) Support use of registered digest and cipher names for dgst and cipher 1160238405Sjkim commands instead of having to add each one as a special case. So now 1161238405Sjkim you can do: 1162238405Sjkim 1163238405Sjkim openssl sha256 foo 1164238405Sjkim 1165238405Sjkim as well as: 1166238405Sjkim 1167238405Sjkim openssl dgst -sha256 foo 1168238405Sjkim 1169238405Sjkim and this works for ENGINE based algorithms too. 1170238405Sjkim 1171238405Sjkim [Steve Henson] 1172238405Sjkim 1173238405Sjkim *) Update Gost ENGINE to support parameter files. 1174238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>] 1175238405Sjkim 1176238405Sjkim *) Support GeneralizedTime in ca utility. 1177238405Sjkim [Oliver Martin <oliver@volatilevoid.net>, Steve Henson] 1178238405Sjkim 1179238405Sjkim *) Enhance the hash format used for certificate directory links. The new 1180238405Sjkim form uses the canonical encoding (meaning equivalent names will work 1181238405Sjkim even if they aren't identical) and uses SHA1 instead of MD5. This form 1182238405Sjkim is incompatible with the older format and as a result c_rehash should 1183238405Sjkim be used to rebuild symbolic links. 1184238405Sjkim [Steve Henson] 1185238405Sjkim 1186238405Sjkim *) Make PKCS#8 the default write format for private keys, replacing the 1187238405Sjkim traditional format. This form is standardised, more secure and doesn't 1188238405Sjkim include an implicit MD5 dependency. 1189238405Sjkim [Steve Henson] 1190238405Sjkim 1191238405Sjkim *) Add a $gcc_devteam_warn option to Configure. The idea is that any code 1192238405Sjkim committed to OpenSSL should pass this lot as a minimum. 1193238405Sjkim [Steve Henson] 1194238405Sjkim 1195238405Sjkim *) Add session ticket override functionality for use by EAP-FAST. 1196238405Sjkim [Jouni Malinen <j@w1.fi>] 1197238405Sjkim 1198238405Sjkim *) Modify HMAC functions to return a value. Since these can be implemented 1199238405Sjkim in an ENGINE errors can occur. 1200238405Sjkim [Steve Henson] 1201238405Sjkim 1202238405Sjkim *) Type-checked OBJ_bsearch_ex. 1203238405Sjkim [Ben Laurie] 1204238405Sjkim 1205238405Sjkim *) Type-checked OBJ_bsearch. Also some constification necessitated 1206238405Sjkim by type-checking. Still to come: TXT_DB, bsearch(?), 1207238405Sjkim OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING, 1208238405Sjkim CONF_VALUE. 1209238405Sjkim [Ben Laurie] 1210238405Sjkim 1211238405Sjkim *) New function OPENSSL_gmtime_adj() to add a specific number of days and 1212238405Sjkim seconds to a tm structure directly, instead of going through OS 1213238405Sjkim specific date routines. This avoids any issues with OS routines such 1214238405Sjkim as the year 2038 bug. New *_adj() functions for ASN1 time structures 1215238405Sjkim and X509_time_adj_ex() to cover the extended range. The existing 1216238405Sjkim X509_time_adj() is still usable and will no longer have any date issues. 1217238405Sjkim [Steve Henson] 1218238405Sjkim 1219238405Sjkim *) Delta CRL support. New use deltas option which will attempt to locate 1220238405Sjkim and search any appropriate delta CRLs available. 1221238405Sjkim 1222238405Sjkim This work was sponsored by Google. 1223238405Sjkim [Steve Henson] 1224238405Sjkim 1225238405Sjkim *) Support for CRLs partitioned by reason code. Reorganise CRL processing 1226238405Sjkim code and add additional score elements. Validate alternate CRL paths 1227238405Sjkim as part of the CRL checking and indicate a new error "CRL path validation 1228238405Sjkim error" in this case. Applications wanting additional details can use 1229238405Sjkim the verify callback and check the new "parent" field. If this is not 1230238405Sjkim NULL CRL path validation is taking place. Existing applications wont 1231238405Sjkim see this because it requires extended CRL support which is off by 1232238405Sjkim default. 1233238405Sjkim 1234238405Sjkim This work was sponsored by Google. 1235238405Sjkim [Steve Henson] 1236238405Sjkim 1237238405Sjkim *) Support for freshest CRL extension. 1238238405Sjkim 1239238405Sjkim This work was sponsored by Google. 1240238405Sjkim [Steve Henson] 1241238405Sjkim 1242238405Sjkim *) Initial indirect CRL support. Currently only supported in the CRLs 1243238405Sjkim passed directly and not via lookup. Process certificate issuer 1244238405Sjkim CRL entry extension and lookup CRL entries by bother issuer name 1245238405Sjkim and serial number. Check and process CRL issuer entry in IDP extension. 1246238405Sjkim 1247238405Sjkim This work was sponsored by Google. 1248238405Sjkim [Steve Henson] 1249238405Sjkim 1250238405Sjkim *) Add support for distinct certificate and CRL paths. The CRL issuer 1251238405Sjkim certificate is validated separately in this case. Only enabled if 1252238405Sjkim an extended CRL support flag is set: this flag will enable additional 1253238405Sjkim CRL functionality in future. 1254238405Sjkim 1255238405Sjkim This work was sponsored by Google. 1256238405Sjkim [Steve Henson] 1257238405Sjkim 1258238405Sjkim *) Add support for policy mappings extension. 1259238405Sjkim 1260238405Sjkim This work was sponsored by Google. 1261238405Sjkim [Steve Henson] 1262238405Sjkim 1263238405Sjkim *) Fixes to pathlength constraint, self issued certificate handling, 1264238405Sjkim policy processing to align with RFC3280 and PKITS tests. 1265238405Sjkim 1266238405Sjkim This work was sponsored by Google. 1267238405Sjkim [Steve Henson] 1268238405Sjkim 1269238405Sjkim *) Support for name constraints certificate extension. DN, email, DNS 1270238405Sjkim and URI types are currently supported. 1271238405Sjkim 1272238405Sjkim This work was sponsored by Google. 1273238405Sjkim [Steve Henson] 1274238405Sjkim 1275238405Sjkim *) To cater for systems that provide a pointer-based thread ID rather 1276238405Sjkim than numeric, deprecate the current numeric thread ID mechanism and 1277238405Sjkim replace it with a structure and associated callback type. This 1278238405Sjkim mechanism allows a numeric "hash" to be extracted from a thread ID in 1279238405Sjkim either case, and on platforms where pointers are larger than 'long', 1280238405Sjkim mixing is done to help ensure the numeric 'hash' is usable even if it 1281238405Sjkim can't be guaranteed unique. The default mechanism is to use "&errno" 1282238405Sjkim as a pointer-based thread ID to distinguish between threads. 1283238405Sjkim 1284238405Sjkim Applications that want to provide their own thread IDs should now use 1285238405Sjkim CRYPTO_THREADID_set_callback() to register a callback that will call 1286238405Sjkim either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer(). 1287238405Sjkim 1288238405Sjkim Note that ERR_remove_state() is now deprecated, because it is tied 1289238405Sjkim to the assumption that thread IDs are numeric. ERR_remove_state(0) 1290238405Sjkim to free the current thread's error state should be replaced by 1291238405Sjkim ERR_remove_thread_state(NULL). 1292238405Sjkim 1293238405Sjkim (This new approach replaces the functions CRYPTO_set_idptr_callback(), 1294238405Sjkim CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in 1295238405Sjkim OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an 1296238405Sjkim application was previously providing a numeric thread callback that 1297238405Sjkim was inappropriate for distinguishing threads, then uniqueness might 1298238405Sjkim have been obtained with &errno that happened immediately in the 1299238405Sjkim intermediate development versions of OpenSSL; this is no longer the 1300238405Sjkim case, the numeric thread callback will now override the automatic use 1301238405Sjkim of &errno.) 1302238405Sjkim [Geoff Thorpe, with help from Bodo Moeller] 1303238405Sjkim 1304238405Sjkim *) Initial support for different CRL issuing certificates. This covers a 1305238405Sjkim simple case where the self issued certificates in the chain exist and 1306238405Sjkim the real CRL issuer is higher in the existing chain. 1307238405Sjkim 1308238405Sjkim This work was sponsored by Google. 1309238405Sjkim [Steve Henson] 1310238405Sjkim 1311238405Sjkim *) Removed effectively defunct crypto/store from the build. 1312238405Sjkim [Ben Laurie] 1313238405Sjkim 1314238405Sjkim *) Revamp of STACK to provide stronger type-checking. Still to come: 1315238405Sjkim TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE, 1316238405Sjkim ASN1_STRING, CONF_VALUE. 1317238405Sjkim [Ben Laurie] 1318238405Sjkim 1319238405Sjkim *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer 1320238405Sjkim RAM on SSL connections. This option can save about 34k per idle SSL. 1321238405Sjkim [Nick Mathewson] 1322238405Sjkim 1323238405Sjkim *) Revamp of LHASH to provide stronger type-checking. Still to come: 1324238405Sjkim STACK, TXT_DB, bsearch, qsort. 1325238405Sjkim [Ben Laurie] 1326238405Sjkim 1327238405Sjkim *) Initial support for Cryptographic Message Syntax (aka CMS) based 1328238405Sjkim on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility, 1329238405Sjkim support for data, signedData, compressedData, digestedData and 1330238405Sjkim encryptedData, envelopedData types included. Scripts to check against 1331238405Sjkim RFC4134 examples draft and interop and consistency checks of many 1332238405Sjkim content types and variants. 1333238405Sjkim [Steve Henson] 1334238405Sjkim 1335238405Sjkim *) Add options to enc utility to support use of zlib compression BIO. 1336238405Sjkim [Steve Henson] 1337238405Sjkim 1338238405Sjkim *) Extend mk1mf to support importing of options and assembly language 1339238405Sjkim files from Configure script, currently only included in VC-WIN32. 1340238405Sjkim The assembly language rules can now optionally generate the source 1341238405Sjkim files from the associated perl scripts. 1342238405Sjkim [Steve Henson] 1343238405Sjkim 1344238405Sjkim *) Implement remaining functionality needed to support GOST ciphersuites. 1345238405Sjkim Interop testing has been performed using CryptoPro implementations. 1346238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>] 1347238405Sjkim 1348238405Sjkim *) s390x assembler pack. 1349238405Sjkim [Andy Polyakov] 1350238405Sjkim 1351238405Sjkim *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU 1352238405Sjkim "family." 1353238405Sjkim [Andy Polyakov] 1354238405Sjkim 1355238405Sjkim *) Implement Opaque PRF Input TLS extension as specified in 1356238405Sjkim draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an 1357238405Sjkim official specification yet and no extension type assignment by 1358238405Sjkim IANA exists, this extension (for now) will have to be explicitly 1359238405Sjkim enabled when building OpenSSL by providing the extension number 1360238405Sjkim to use. For example, specify an option 1361238405Sjkim 1362238405Sjkim -DTLSEXT_TYPE_opaque_prf_input=0x9527 1363238405Sjkim 1364238405Sjkim to the "config" or "Configure" script to enable the extension, 1365238405Sjkim assuming extension number 0x9527 (which is a completely arbitrary 1366238405Sjkim and unofficial assignment based on the MD5 hash of the Internet 1367238405Sjkim Draft). Note that by doing so, you potentially lose 1368238405Sjkim interoperability with other TLS implementations since these might 1369238405Sjkim be using the same extension number for other purposes. 1370238405Sjkim 1371238405Sjkim SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the 1372238405Sjkim opaque PRF input value to use in the handshake. This will create 1373238405Sjkim an interal copy of the length-'len' string at 'src', and will 1374238405Sjkim return non-zero for success. 1375238405Sjkim 1376238405Sjkim To get more control and flexibility, provide a callback function 1377238405Sjkim by using 1378238405Sjkim 1379238405Sjkim SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb) 1380238405Sjkim SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg) 1381238405Sjkim 1382238405Sjkim where 1383238405Sjkim 1384238405Sjkim int (*cb)(SSL *, void *peerinput, size_t len, void *arg); 1385238405Sjkim void *arg; 1386238405Sjkim 1387238405Sjkim Callback function 'cb' will be called in handshakes, and is 1388238405Sjkim expected to use SSL_set_tlsext_opaque_prf_input() as appropriate. 1389238405Sjkim Argument 'arg' is for application purposes (the value as given to 1390238405Sjkim SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly 1391238405Sjkim be provided to the callback function). The callback function 1392238405Sjkim has to return non-zero to report success: usually 1 to use opaque 1393238405Sjkim PRF input just if possible, or 2 to enforce use of the opaque PRF 1394238405Sjkim input. In the latter case, the library will abort the handshake 1395238405Sjkim if opaque PRF input is not successfully negotiated. 1396238405Sjkim 1397238405Sjkim Arguments 'peerinput' and 'len' given to the callback function 1398238405Sjkim will always be NULL and 0 in the case of a client. A server will 1399238405Sjkim see the client's opaque PRF input through these variables if 1400238405Sjkim available (NULL and 0 otherwise). Note that if the server 1401238405Sjkim provides an opaque PRF input, the length must be the same as the 1402238405Sjkim length of the client's opaque PRF input. 1403238405Sjkim 1404238405Sjkim Note that the callback function will only be called when creating 1405238405Sjkim a new session (session resumption can resume whatever was 1406238405Sjkim previously negotiated), and will not be called in SSL 2.0 1407238405Sjkim handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or 1408238405Sjkim SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended 1409238405Sjkim for applications that need to enforce opaque PRF input. 1410238405Sjkim 1411238405Sjkim [Bodo Moeller] 1412238405Sjkim 1413238405Sjkim *) Update ssl code to support digests other than SHA1+MD5 for handshake 1414238405Sjkim MAC. 1415238405Sjkim 1416238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>] 1417238405Sjkim 1418238405Sjkim *) Add RFC4507 support to OpenSSL. This includes the corrections in 1419238405Sjkim RFC4507bis. The encrypted ticket format is an encrypted encoded 1420238405Sjkim SSL_SESSION structure, that way new session features are automatically 1421238405Sjkim supported. 1422238405Sjkim 1423238405Sjkim If a client application caches session in an SSL_SESSION structure 1424238405Sjkim support is transparent because tickets are now stored in the encoded 1425238405Sjkim SSL_SESSION. 1426238405Sjkim 1427238405Sjkim The SSL_CTX structure automatically generates keys for ticket 1428238405Sjkim protection in servers so again support should be possible 1429238405Sjkim with no application modification. 1430238405Sjkim 1431238405Sjkim If a client or server wishes to disable RFC4507 support then the option 1432238405Sjkim SSL_OP_NO_TICKET can be set. 1433238405Sjkim 1434238405Sjkim Add a TLS extension debugging callback to allow the contents of any client 1435238405Sjkim or server extensions to be examined. 1436238405Sjkim 1437238405Sjkim This work was sponsored by Google. 1438238405Sjkim [Steve Henson] 1439238405Sjkim 1440238405Sjkim *) Final changes to avoid use of pointer pointer casts in OpenSSL. 1441238405Sjkim OpenSSL should now compile cleanly on gcc 4.2 1442238405Sjkim [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson] 1443238405Sjkim 1444238405Sjkim *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC 1445238405Sjkim support including streaming MAC support: this is required for GOST 1446238405Sjkim ciphersuite support. 1447238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson] 1448238405Sjkim 1449238405Sjkim *) Add option -stream to use PKCS#7 streaming in smime utility. New 1450238405Sjkim function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream() 1451238405Sjkim to output in BER and PEM format. 1452238405Sjkim [Steve Henson] 1453238405Sjkim 1454238405Sjkim *) Experimental support for use of HMAC via EVP_PKEY interface. This 1455238405Sjkim allows HMAC to be handled via the EVP_DigestSign*() interface. The 1456238405Sjkim EVP_PKEY "key" in this case is the HMAC key, potentially allowing 1457238405Sjkim ENGINE support for HMAC keys which are unextractable. New -mac and 1458238405Sjkim -macopt options to dgst utility. 1459238405Sjkim [Steve Henson] 1460238405Sjkim 1461238405Sjkim *) New option -sigopt to dgst utility. Update dgst to use 1462238405Sjkim EVP_Digest{Sign,Verify}*. These two changes make it possible to use 1463238405Sjkim alternative signing paramaters such as X9.31 or PSS in the dgst 1464238405Sjkim utility. 1465238405Sjkim [Steve Henson] 1466238405Sjkim 1467238405Sjkim *) Change ssl_cipher_apply_rule(), the internal function that does 1468238405Sjkim the work each time a ciphersuite string requests enabling 1469238405Sjkim ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or 1470238405Sjkim removing ("!foo+bar") a class of ciphersuites: Now it maintains 1471238405Sjkim the order of disabled ciphersuites such that those ciphersuites 1472238405Sjkim that most recently went from enabled to disabled not only stay 1473238405Sjkim in order with respect to each other, but also have higher priority 1474238405Sjkim than other disabled ciphersuites the next time ciphersuites are 1475238405Sjkim enabled again. 1476238405Sjkim 1477238405Sjkim This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable 1478238405Sjkim the same ciphersuites as with "HIGH" alone, but in a specific 1479238405Sjkim order where the PSK ciphersuites come first (since they are the 1480238405Sjkim most recently disabled ciphersuites when "HIGH" is parsed). 1481238405Sjkim 1482238405Sjkim Also, change ssl_create_cipher_list() (using this new 1483238405Sjkim funcionality) such that between otherwise identical 1484238405Sjkim cihpersuites, ephemeral ECDH is preferred over ephemeral DH in 1485238405Sjkim the default order. 1486238405Sjkim [Bodo Moeller] 1487238405Sjkim 1488238405Sjkim *) Change ssl_create_cipher_list() so that it automatically 1489238405Sjkim arranges the ciphersuites in reasonable order before starting 1490238405Sjkim to process the rule string. Thus, the definition for "DEFAULT" 1491238405Sjkim (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but 1492238405Sjkim remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH". 1493238405Sjkim This makes it much easier to arrive at a reasonable default order 1494238405Sjkim in applications for which anonymous ciphers are OK (meaning 1495238405Sjkim that you can't actually use DEFAULT). 1496238405Sjkim [Bodo Moeller; suggested by Victor Duchovni] 1497238405Sjkim 1498238405Sjkim *) Split the SSL/TLS algorithm mask (as used for ciphersuite string 1499238405Sjkim processing) into multiple integers instead of setting 1500238405Sjkim "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK", 1501238405Sjkim "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer. 1502238405Sjkim (These masks as well as the individual bit definitions are hidden 1503238405Sjkim away into the non-exported interface ssl/ssl_locl.h, so this 1504238405Sjkim change to the definition of the SSL_CIPHER structure shouldn't 1505238405Sjkim affect applications.) This give us more bits for each of these 1506238405Sjkim categories, so there is no longer a need to coagulate AES128 and 1507238405Sjkim AES256 into a single algorithm bit, and to coagulate Camellia128 1508238405Sjkim and Camellia256 into a single algorithm bit, which has led to all 1509238405Sjkim kinds of kludges. 1510238405Sjkim 1511238405Sjkim Thus, among other things, the kludge introduced in 0.9.7m and 1512238405Sjkim 0.9.8e for masking out AES256 independently of AES128 or masking 1513238405Sjkim out Camellia256 independently of AES256 is not needed here in 0.9.9. 1514238405Sjkim 1515238405Sjkim With the change, we also introduce new ciphersuite aliases that 1516238405Sjkim so far were missing: "AES128", "AES256", "CAMELLIA128", and 1517238405Sjkim "CAMELLIA256". 1518238405Sjkim [Bodo Moeller] 1519238405Sjkim 1520238405Sjkim *) Add support for dsa-with-SHA224 and dsa-with-SHA256. 1521238405Sjkim Use the leftmost N bytes of the signature input if the input is 1522238405Sjkim larger than the prime q (with N being the size in bytes of q). 1523238405Sjkim [Nils Larsch] 1524238405Sjkim 1525238405Sjkim *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses 1526238405Sjkim it yet and it is largely untested. 1527238405Sjkim [Steve Henson] 1528238405Sjkim 1529238405Sjkim *) Add support for the ecdsa-with-SHA224/256/384/512 signature types. 1530238405Sjkim [Nils Larsch] 1531238405Sjkim 1532238405Sjkim *) Initial incomplete changes to avoid need for function casts in OpenSSL 1533238405Sjkim some compilers (gcc 4.2 and later) reject their use. Safestack is 1534238405Sjkim reimplemented. Update ASN1 to avoid use of legacy functions. 1535238405Sjkim [Steve Henson] 1536238405Sjkim 1537238405Sjkim *) Win32/64 targets are linked with Winsock2. 1538238405Sjkim [Andy Polyakov] 1539238405Sjkim 1540238405Sjkim *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected 1541238405Sjkim to external functions. This can be used to increase CRL handling 1542238405Sjkim efficiency especially when CRLs are very large by (for example) storing 1543238405Sjkim the CRL revoked certificates in a database. 1544238405Sjkim [Steve Henson] 1545238405Sjkim 1546238405Sjkim *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so 1547238405Sjkim new CRLs added to a directory can be used. New command line option 1548238405Sjkim -verify_return_error to s_client and s_server. This causes real errors 1549238405Sjkim to be returned by the verify callback instead of carrying on no matter 1550238405Sjkim what. This reflects the way a "real world" verify callback would behave. 1551238405Sjkim [Steve Henson] 1552238405Sjkim 1553238405Sjkim *) GOST engine, supporting several GOST algorithms and public key formats. 1554238405Sjkim Kindly donated by Cryptocom. 1555238405Sjkim [Cryptocom] 1556238405Sjkim 1557238405Sjkim *) Partial support for Issuing Distribution Point CRL extension. CRLs 1558238405Sjkim partitioned by DP are handled but no indirect CRL or reason partitioning 1559238405Sjkim (yet). Complete overhaul of CRL handling: now the most suitable CRL is 1560238405Sjkim selected via a scoring technique which handles IDP and AKID in CRLs. 1561238405Sjkim [Steve Henson] 1562238405Sjkim 1563238405Sjkim *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which 1564238405Sjkim will ultimately be used for all verify operations: this will remove the 1565238405Sjkim X509_STORE dependency on certificate verification and allow alternative 1566238405Sjkim lookup methods. X509_STORE based implementations of these two callbacks. 1567238405Sjkim [Steve Henson] 1568238405Sjkim 1569238405Sjkim *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names. 1570238405Sjkim Modify get_crl() to find a valid (unexpired) CRL if possible. 1571238405Sjkim [Steve Henson] 1572238405Sjkim 1573238405Sjkim *) New function X509_CRL_match() to check if two CRLs are identical. Normally 1574238405Sjkim this would be called X509_CRL_cmp() but that name is already used by 1575238405Sjkim a function that just compares CRL issuer names. Cache several CRL 1576238405Sjkim extensions in X509_CRL structure and cache CRLDP in X509. 1577238405Sjkim [Steve Henson] 1578238405Sjkim 1579238405Sjkim *) Store a "canonical" representation of X509_NAME structure (ASN1 Name) 1580238405Sjkim this maps equivalent X509_NAME structures into a consistent structure. 1581238405Sjkim Name comparison can then be performed rapidly using memcmp(). 1582238405Sjkim [Steve Henson] 1583238405Sjkim 1584238405Sjkim *) Non-blocking OCSP request processing. Add -timeout option to ocsp 1585238405Sjkim utility. 1586238405Sjkim [Steve Henson] 1587238405Sjkim 1588238405Sjkim *) Allow digests to supply their own micalg string for S/MIME type using 1589238405Sjkim the ctrl EVP_MD_CTRL_MICALG. 1590238405Sjkim [Steve Henson] 1591238405Sjkim 1592238405Sjkim *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the 1593238405Sjkim EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN 1594238405Sjkim ctrl. It can then customise the structure before and/or after signing 1595238405Sjkim if necessary. 1596238405Sjkim [Steve Henson] 1597238405Sjkim 1598238405Sjkim *) New function OBJ_add_sigid() to allow application defined signature OIDs 1599238405Sjkim to be added to OpenSSLs internal tables. New function OBJ_sigid_free() 1600238405Sjkim to free up any added signature OIDs. 1601238405Sjkim [Steve Henson] 1602238405Sjkim 1603238405Sjkim *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(), 1604238405Sjkim EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal 1605238405Sjkim digest and cipher tables. New options added to openssl utility: 1606238405Sjkim list-message-digest-algorithms and list-cipher-algorithms. 1607238405Sjkim [Steve Henson] 1608238405Sjkim 1609238405Sjkim *) Change the array representation of binary polynomials: the list 1610238405Sjkim of degrees of non-zero coefficients is now terminated with -1. 1611238405Sjkim Previously it was terminated with 0, which was also part of the 1612238405Sjkim value; thus, the array representation was not applicable to 1613238405Sjkim polynomials where t^0 has coefficient zero. This change makes 1614238405Sjkim the array representation useful in a more general context. 1615238405Sjkim [Douglas Stebila] 1616238405Sjkim 1617238405Sjkim *) Various modifications and fixes to SSL/TLS cipher string 1618238405Sjkim handling. For ECC, the code now distinguishes between fixed ECDH 1619238405Sjkim with RSA certificates on the one hand and with ECDSA certificates 1620238405Sjkim on the other hand, since these are separate ciphersuites. The 1621238405Sjkim unused code for Fortezza ciphersuites has been removed. 1622238405Sjkim 1623238405Sjkim For consistency with EDH, ephemeral ECDH is now called "EECDH" 1624238405Sjkim (not "ECDHE"). For consistency with the code for DH 1625238405Sjkim certificates, use of ECDH certificates is now considered ECDH 1626238405Sjkim authentication, not RSA or ECDSA authentication (the latter is 1627238405Sjkim merely the CA's signing algorithm and not actively used in the 1628238405Sjkim protocol). 1629238405Sjkim 1630238405Sjkim The temporary ciphersuite alias "ECCdraft" is no longer 1631238405Sjkim available, and ECC ciphersuites are no longer excluded from "ALL" 1632238405Sjkim and "DEFAULT". The following aliases now exist for RFC 4492 1633238405Sjkim ciphersuites, most of these by analogy with the DH case: 1634238405Sjkim 1635238405Sjkim kECDHr - ECDH cert, signed with RSA 1636238405Sjkim kECDHe - ECDH cert, signed with ECDSA 1637238405Sjkim kECDH - ECDH cert (signed with either RSA or ECDSA) 1638238405Sjkim kEECDH - ephemeral ECDH 1639238405Sjkim ECDH - ECDH cert or ephemeral ECDH 1640238405Sjkim 1641238405Sjkim aECDH - ECDH cert 1642238405Sjkim aECDSA - ECDSA cert 1643238405Sjkim ECDSA - ECDSA cert 1644238405Sjkim 1645238405Sjkim AECDH - anonymous ECDH 1646238405Sjkim EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH") 1647238405Sjkim 1648238405Sjkim [Bodo Moeller] 1649238405Sjkim 1650238405Sjkim *) Add additional S/MIME capabilities for AES and GOST ciphers if supported. 1651238405Sjkim Use correct micalg parameters depending on digest(s) in signed message. 1652238405Sjkim [Steve Henson] 1653238405Sjkim 1654238405Sjkim *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process 1655238405Sjkim an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code. 1656238405Sjkim [Steve Henson] 1657238405Sjkim 1658238405Sjkim *) Initial engine support for EVP_PKEY_METHOD. New functions to permit 1659238405Sjkim an engine to register a method. Add ENGINE lookups for methods and 1660238405Sjkim functional reference processing. 1661238405Sjkim [Steve Henson] 1662238405Sjkim 1663238405Sjkim *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of 1664238405Sjkim EVP_{Sign,Verify}* which allow an application to customise the signature 1665238405Sjkim process. 1666238405Sjkim [Steve Henson] 1667238405Sjkim 1668238405Sjkim *) New -resign option to smime utility. This adds one or more signers 1669238405Sjkim to an existing PKCS#7 signedData structure. Also -md option to use an 1670238405Sjkim alternative message digest algorithm for signing. 1671238405Sjkim [Steve Henson] 1672238405Sjkim 1673238405Sjkim *) Tidy up PKCS#7 routines and add new functions to make it easier to 1674238405Sjkim create PKCS7 structures containing multiple signers. Update smime 1675238405Sjkim application to support multiple signers. 1676238405Sjkim [Steve Henson] 1677238405Sjkim 1678238405Sjkim *) New -macalg option to pkcs12 utility to allow setting of an alternative 1679238405Sjkim digest MAC. 1680238405Sjkim [Steve Henson] 1681238405Sjkim 1682238405Sjkim *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC. 1683238405Sjkim Reorganize PBE internals to lookup from a static table using NIDs, 1684238405Sjkim add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl: 1685238405Sjkim EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative 1686238405Sjkim PRF which will be automatically used with PBES2. 1687238405Sjkim [Steve Henson] 1688238405Sjkim 1689238405Sjkim *) Replace the algorithm specific calls to generate keys in "req" with the 1690238405Sjkim new API. 1691238405Sjkim [Steve Henson] 1692238405Sjkim 1693238405Sjkim *) Update PKCS#7 enveloped data routines to use new API. This is now 1694238405Sjkim supported by any public key method supporting the encrypt operation. A 1695238405Sjkim ctrl is added to allow the public key algorithm to examine or modify 1696238405Sjkim the PKCS#7 RecipientInfo structure if it needs to: for RSA this is 1697238405Sjkim a no op. 1698238405Sjkim [Steve Henson] 1699238405Sjkim 1700238405Sjkim *) Add a ctrl to asn1 method to allow a public key algorithm to express 1701238405Sjkim a default digest type to use. In most cases this will be SHA1 but some 1702238405Sjkim algorithms (such as GOST) need to specify an alternative digest. The 1703238405Sjkim return value indicates how strong the prefernce is 1 means optional and 1704238405Sjkim 2 is mandatory (that is it is the only supported type). Modify 1705238405Sjkim ASN1_item_sign() to accept a NULL digest argument to indicate it should 1706238405Sjkim use the default md. Update openssl utilities to use the default digest 1707238405Sjkim type for signing if it is not explicitly indicated. 1708238405Sjkim [Steve Henson] 1709238405Sjkim 1710238405Sjkim *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New 1711238405Sjkim EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant 1712238405Sjkim signing method from the key type. This effectively removes the link 1713238405Sjkim between digests and public key types. 1714238405Sjkim [Steve Henson] 1715238405Sjkim 1716238405Sjkim *) Add an OID cross reference table and utility functions. Its purpose is to 1717238405Sjkim translate between signature OIDs such as SHA1WithrsaEncryption and SHA1, 1718238405Sjkim rsaEncryption. This will allow some of the algorithm specific hackery 1719238405Sjkim needed to use the correct OID to be removed. 1720238405Sjkim [Steve Henson] 1721238405Sjkim 1722238405Sjkim *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO 1723238405Sjkim structures for PKCS7_sign(). They are now set up by the relevant public 1724238405Sjkim key ASN1 method. 1725238405Sjkim [Steve Henson] 1726238405Sjkim 1727238405Sjkim *) Add provisional EC pkey method with support for ECDSA and ECDH. 1728238405Sjkim [Steve Henson] 1729238405Sjkim 1730238405Sjkim *) Add support for key derivation (agreement) in the API, DH method and 1731238405Sjkim pkeyutl. 1732238405Sjkim [Steve Henson] 1733238405Sjkim 1734238405Sjkim *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support 1735238405Sjkim public and private key formats. As a side effect these add additional 1736238405Sjkim command line functionality not previously available: DSA signatures can be 1737238405Sjkim generated and verified using pkeyutl and DH key support and generation in 1738238405Sjkim pkey, genpkey. 1739238405Sjkim [Steve Henson] 1740238405Sjkim 1741238405Sjkim *) BeOS support. 1742238405Sjkim [Oliver Tappe <zooey@hirschkaefer.de>] 1743238405Sjkim 1744238405Sjkim *) New make target "install_html_docs" installs HTML renditions of the 1745238405Sjkim manual pages. 1746238405Sjkim [Oliver Tappe <zooey@hirschkaefer.de>] 1747238405Sjkim 1748238405Sjkim *) New utility "genpkey" this is analagous to "genrsa" etc except it can 1749238405Sjkim generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to 1750238405Sjkim support key and parameter generation and add initial key generation 1751238405Sjkim functionality for RSA. 1752238405Sjkim [Steve Henson] 1753238405Sjkim 1754238405Sjkim *) Add functions for main EVP_PKEY_method operations. The undocumented 1755238405Sjkim functions EVP_PKEY_{encrypt,decrypt} have been renamed to 1756238405Sjkim EVP_PKEY_{encrypt,decrypt}_old. 1757238405Sjkim [Steve Henson] 1758238405Sjkim 1759238405Sjkim *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public 1760238405Sjkim key API, doesn't do much yet. 1761238405Sjkim [Steve Henson] 1762238405Sjkim 1763238405Sjkim *) New function EVP_PKEY_asn1_get0_info() to retrieve information about 1764238405Sjkim public key algorithms. New option to openssl utility: 1765238405Sjkim "list-public-key-algorithms" to print out info. 1766238405Sjkim [Steve Henson] 1767238405Sjkim 1768238405Sjkim *) Implement the Supported Elliptic Curves Extension for 1769238405Sjkim ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 1770238405Sjkim [Douglas Stebila] 1771238405Sjkim 1772238405Sjkim *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or 1773238405Sjkim EVP_CIPHER structures to avoid later problems in EVP_cleanup(). 1774238405Sjkim [Steve Henson] 1775238405Sjkim 1776238405Sjkim *) New utilities pkey and pkeyparam. These are similar to algorithm specific 1777238405Sjkim utilities such as rsa, dsa, dsaparam etc except they process any key 1778238405Sjkim type. 1779238405Sjkim [Steve Henson] 1780238405Sjkim 1781238405Sjkim *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New 1782238405Sjkim functions EVP_PKEY_print_public(), EVP_PKEY_print_private(), 1783238405Sjkim EVP_PKEY_print_param() to print public key data from an EVP_PKEY 1784238405Sjkim structure. 1785238405Sjkim [Steve Henson] 1786238405Sjkim 1787238405Sjkim *) Initial support for pluggable public key ASN1. 1788238405Sjkim De-spaghettify the public key ASN1 handling. Move public and private 1789238405Sjkim key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate 1790238405Sjkim algorithm specific handling to a single module within the relevant 1791238405Sjkim algorithm directory. Add functions to allow (near) opaque processing 1792238405Sjkim of public and private key structures. 1793238405Sjkim [Steve Henson] 1794238405Sjkim 1795238405Sjkim *) Implement the Supported Point Formats Extension for 1796238405Sjkim ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 1797238405Sjkim [Douglas Stebila] 1798238405Sjkim 1799238405Sjkim *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members 1800238405Sjkim for the psk identity [hint] and the psk callback functions to the 1801238405Sjkim SSL_SESSION, SSL and SSL_CTX structure. 1802238405Sjkim 1803238405Sjkim New ciphersuites: 1804238405Sjkim PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA, 1805238405Sjkim PSK-AES256-CBC-SHA 1806238405Sjkim 1807238405Sjkim New functions: 1808238405Sjkim SSL_CTX_use_psk_identity_hint 1809238405Sjkim SSL_get_psk_identity_hint 1810238405Sjkim SSL_get_psk_identity 1811238405Sjkim SSL_use_psk_identity_hint 1812238405Sjkim 1813238405Sjkim [Mika Kousa and Pasi Eronen of Nokia Corporation] 1814238405Sjkim 1815238405Sjkim *) Add RFC 3161 compliant time stamp request creation, response generation 1816238405Sjkim and response verification functionality. 1817238405Sjkim [Zolt�n Gl�zik <zglozik@opentsa.org>, The OpenTSA Project] 1818238405Sjkim 1819238405Sjkim *) Add initial support for TLS extensions, specifically for the server_name 1820238405Sjkim extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 1821238405Sjkim have new members for a host name. The SSL data structure has an 1822238405Sjkim additional member SSL_CTX *initial_ctx so that new sessions can be 1823238405Sjkim stored in that context to allow for session resumption, even after the 1824238405Sjkim SSL has been switched to a new SSL_CTX in reaction to a client's 1825238405Sjkim server_name extension. 1826238405Sjkim 1827238405Sjkim New functions (subject to change): 1828238405Sjkim 1829238405Sjkim SSL_get_servername() 1830238405Sjkim SSL_get_servername_type() 1831238405Sjkim SSL_set_SSL_CTX() 1832238405Sjkim 1833238405Sjkim New CTRL codes and macros (subject to change): 1834238405Sjkim 1835238405Sjkim SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 1836238405Sjkim - SSL_CTX_set_tlsext_servername_callback() 1837238405Sjkim SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 1838238405Sjkim - SSL_CTX_set_tlsext_servername_arg() 1839238405Sjkim SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 1840238405Sjkim 1841238405Sjkim openssl s_client has a new '-servername ...' option. 1842238405Sjkim 1843238405Sjkim openssl s_server has new options '-servername_host ...', '-cert2 ...', 1844238405Sjkim '-key2 ...', '-servername_fatal' (subject to change). This allows 1845238405Sjkim testing the HostName extension for a specific single host name ('-cert' 1846238405Sjkim and '-key' remain fallbacks for handshakes without HostName 1847238405Sjkim negotiation). If the unrecogninzed_name alert has to be sent, this by 1848238405Sjkim default is a warning; it becomes fatal with the '-servername_fatal' 1849238405Sjkim option. 1850238405Sjkim 1851238405Sjkim [Peter Sylvester, Remy Allais, Christophe Renou] 1852238405Sjkim 1853238405Sjkim *) Whirlpool hash implementation is added. 1854238405Sjkim [Andy Polyakov] 1855238405Sjkim 1856238405Sjkim *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to 1857238405Sjkim bn(64,32). Because of instruction set limitations it doesn't have 1858238405Sjkim any negative impact on performance. This was done mostly in order 1859238405Sjkim to make it possible to share assembler modules, such as bn_mul_mont 1860238405Sjkim implementations, between 32- and 64-bit builds without hassle. 1861238405Sjkim [Andy Polyakov] 1862238405Sjkim 1863238405Sjkim *) Move code previously exiled into file crypto/ec/ec2_smpt.c 1864238405Sjkim to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP 1865238405Sjkim macro. 1866238405Sjkim [Bodo Moeller] 1867238405Sjkim 1868238405Sjkim *) New candidate for BIGNUM assembler implementation, bn_mul_mont, 1869238405Sjkim dedicated Montgomery multiplication procedure, is introduced. 1870238405Sjkim BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher 1871238405Sjkim "64-bit" performance on certain 32-bit targets. 1872238405Sjkim [Andy Polyakov] 1873238405Sjkim 1874238405Sjkim *) New option SSL_OP_NO_COMP to disable use of compression selectively 1875238405Sjkim in SSL structures. New SSL ctrl to set maximum send fragment size. 1876238405Sjkim Save memory by seeting the I/O buffer sizes dynamically instead of 1877238405Sjkim using the maximum available value. 1878238405Sjkim [Steve Henson] 1879238405Sjkim 1880238405Sjkim *) New option -V for 'openssl ciphers'. This prints the ciphersuite code 1881238405Sjkim in addition to the text details. 1882238405Sjkim [Bodo Moeller] 1883238405Sjkim 1884238405Sjkim *) Very, very preliminary EXPERIMENTAL support for printing of general 1885238405Sjkim ASN1 structures. This currently produces rather ugly output and doesn't 1886238405Sjkim handle several customised structures at all. 1887238405Sjkim [Steve Henson] 1888238405Sjkim 1889238405Sjkim *) Integrated support for PVK file format and some related formats such 1890238405Sjkim as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support 1891238405Sjkim these in the 'rsa' and 'dsa' utilities. 1892238405Sjkim [Steve Henson] 1893238405Sjkim 1894238405Sjkim *) Support for PKCS#1 RSAPublicKey format on rsa utility command line. 1895238405Sjkim [Steve Henson] 1896238405Sjkim 1897238405Sjkim *) Remove the ancient ASN1_METHOD code. This was only ever used in one 1898238405Sjkim place for the (very old) "NETSCAPE" format certificates which are now 1899238405Sjkim handled using new ASN1 code equivalents. 1900238405Sjkim [Steve Henson] 1901238405Sjkim 1902238405Sjkim *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD 1903238405Sjkim pointer and make the SSL_METHOD parameter in SSL_CTX_new, 1904238405Sjkim SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'. 1905238405Sjkim [Nils Larsch] 1906238405Sjkim 1907238405Sjkim *) Modify CRL distribution points extension code to print out previously 1908238405Sjkim unsupported fields. Enhance extension setting code to allow setting of 1909238405Sjkim all fields. 1910238405Sjkim [Steve Henson] 1911238405Sjkim 1912238405Sjkim *) Add print and set support for Issuing Distribution Point CRL extension. 1913238405Sjkim [Steve Henson] 1914238405Sjkim 1915238405Sjkim *) Change 'Configure' script to enable Camellia by default. 1916238405Sjkim [NTT] 1917238405Sjkim 1918206046Ssimon Changes between 0.9.8m and 0.9.8n [24 Mar 2010] 1919206046Ssimon 1920206046Ssimon *) When rejecting SSL/TLS records due to an incorrect version number, never 1921206046Ssimon update s->server with a new major version number. As of 1922206046Ssimon - OpenSSL 0.9.8m if 'short' is a 16-bit type, 1923206046Ssimon - OpenSSL 0.9.8f if 'short' is longer than 16 bits, 1924206046Ssimon the previous behavior could result in a read attempt at NULL when 1925206046Ssimon receiving specific incorrect SSL/TLS records once record payload 1926206046Ssimon protection is active. (CVE-2010-0740) 1927206046Ssimon [Bodo Moeller, Adam Langley <agl@chromium.org>] 1928206046Ssimon 1929206046Ssimon *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL 1930206046Ssimon could be crashed if the relevant tables were not present (e.g. chrooted). 1931206046Ssimon [Tomas Hoger <thoger@redhat.com>] 1932206046Ssimon 1933205128Ssimon Changes between 0.9.8l and 0.9.8m [25 Feb 2010] 1934205128Ssimon 1935205128Ssimon *) Always check bn_wexpend() return values for failure. (CVE-2009-3245) 1936205128Ssimon [Martin Olsson, Neel Mehta] 1937205128Ssimon 1938205128Ssimon *) Fix X509_STORE locking: Every 'objs' access requires a lock (to 1939205128Ssimon accommodate for stack sorting, always a write lock!). 1940205128Ssimon [Bodo Moeller] 1941205128Ssimon 1942205128Ssimon *) On some versions of WIN32 Heap32Next is very slow. This can cause 1943205128Ssimon excessive delays in the RAND_poll(): over a minute. As a workaround 1944205128Ssimon include a time check in the inner Heap32Next loop too. 1945205128Ssimon [Steve Henson] 1946205128Ssimon 1947205128Ssimon *) The code that handled flushing of data in SSL/TLS originally used the 1948205128Ssimon BIO_CTRL_INFO ctrl to see if any data was pending first. This caused 1949205128Ssimon the problem outlined in PR#1949. The fix suggested there however can 1950205128Ssimon trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions 1951205128Ssimon of Apache). So instead simplify the code to flush unconditionally. 1952205128Ssimon This should be fine since flushing with no data to flush is a no op. 1953205128Ssimon [Steve Henson] 1954205128Ssimon 1955205128Ssimon *) Handle TLS versions 2.0 and later properly and correctly use the 1956205128Ssimon highest version of TLS/SSL supported. Although TLS >= 2.0 is some way 1957205128Ssimon off ancient servers have a habit of sticking around for a while... 1958205128Ssimon [Steve Henson] 1959205128Ssimon 1960205128Ssimon *) Modify compression code so it frees up structures without using the 1961205128Ssimon ex_data callbacks. This works around a problem where some applications 1962205128Ssimon call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when 1963205128Ssimon restarting) then use compression (e.g. SSL with compression) later. 1964205128Ssimon This results in significant per-connection memory leaks and 1965205128Ssimon has caused some security issues including CVE-2008-1678 and 1966205128Ssimon CVE-2009-4355. 1967205128Ssimon [Steve Henson] 1968205128Ssimon 1969205128Ssimon *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't 1970205128Ssimon change when encrypting or decrypting. 1971205128Ssimon [Bodo Moeller] 1972205128Ssimon 1973205128Ssimon *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to 1974205128Ssimon connect and renegotiate with servers which do not support RI. 1975205128Ssimon Until RI is more widely deployed this option is enabled by default. 1976205128Ssimon [Steve Henson] 1977205128Ssimon 1978205128Ssimon *) Add "missing" ssl ctrls to clear options and mode. 1979205128Ssimon [Steve Henson] 1980205128Ssimon 1981205128Ssimon *) If client attempts to renegotiate and doesn't support RI respond with 1982205128Ssimon a no_renegotiation alert as required by RFC5746. Some renegotiating 1983205128Ssimon TLS clients will continue a connection gracefully when they receive 1984205128Ssimon the alert. Unfortunately OpenSSL mishandled this alert and would hang 1985205128Ssimon waiting for a server hello which it will never receive. Now we treat a 1986205128Ssimon received no_renegotiation alert as a fatal error. This is because 1987205128Ssimon applications requesting a renegotiation might well expect it to succeed 1988205128Ssimon and would have no code in place to handle the server denying it so the 1989205128Ssimon only safe thing to do is to terminate the connection. 1990205128Ssimon [Steve Henson] 1991205128Ssimon 1992205128Ssimon *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if 1993205128Ssimon peer supports secure renegotiation and 0 otherwise. Print out peer 1994205128Ssimon renegotiation support in s_client/s_server. 1995205128Ssimon [Steve Henson] 1996205128Ssimon 1997205128Ssimon *) Replace the highly broken and deprecated SPKAC certification method with 1998205128Ssimon the updated NID creation version. This should correctly handle UTF8. 1999205128Ssimon [Steve Henson] 2000205128Ssimon 2001205128Ssimon *) Implement RFC5746. Re-enable renegotiation but require the extension 2002205128Ssimon as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION 2003205128Ssimon turns out to be a bad idea. It has been replaced by 2004205128Ssimon SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with 2005205128Ssimon SSL_CTX_set_options(). This is really not recommended unless you 2006205128Ssimon know what you are doing. 2007205128Ssimon [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson] 2008205128Ssimon 2009205128Ssimon *) Fixes to stateless session resumption handling. Use initial_ctx when 2010205128Ssimon issuing and attempting to decrypt tickets in case it has changed during 2011205128Ssimon servername handling. Use a non-zero length session ID when attempting 2012205128Ssimon stateless session resumption: this makes it possible to determine if 2013205128Ssimon a resumption has occurred immediately after receiving server hello 2014205128Ssimon (several places in OpenSSL subtly assume this) instead of later in 2015205128Ssimon the handshake. 2016205128Ssimon [Steve Henson] 2017205128Ssimon 2018205128Ssimon *) The functions ENGINE_ctrl(), OPENSSL_isservice(), 2019205128Ssimon CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error 2020205128Ssimon fixes for a few places where the return code is not checked 2021205128Ssimon correctly. 2022205128Ssimon [Julia Lawall <julia@diku.dk>] 2023205128Ssimon 2024205128Ssimon *) Add --strict-warnings option to Configure script to include devteam 2025205128Ssimon warnings in other configurations. 2026205128Ssimon [Steve Henson] 2027205128Ssimon 2028205128Ssimon *) Add support for --libdir option and LIBDIR variable in makefiles. This 2029205128Ssimon makes it possible to install openssl libraries in locations which 2030205128Ssimon have names other than "lib", for example "/usr/lib64" which some 2031205128Ssimon systems need. 2032205128Ssimon [Steve Henson, based on patch from Jeremy Utley] 2033205128Ssimon 2034205128Ssimon *) Don't allow the use of leading 0x80 in OIDs. This is a violation of 2035205128Ssimon X690 8.9.12 and can produce some misleading textual output of OIDs. 2036205128Ssimon [Steve Henson, reported by Dan Kaminsky] 2037205128Ssimon 2038205128Ssimon *) Delete MD2 from algorithm tables. This follows the recommendation in 2039205128Ssimon several standards that it is not used in new applications due to 2040205128Ssimon several cryptographic weaknesses. For binary compatibility reasons 2041205128Ssimon the MD2 API is still compiled in by default. 2042205128Ssimon [Steve Henson] 2043205128Ssimon 2044205128Ssimon *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved 2045205128Ssimon and restored. 2046205128Ssimon [Steve Henson] 2047205128Ssimon 2048205128Ssimon *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and 2049205128Ssimon OPENSSL_asc2uni conditionally on Netware platforms to avoid a name 2050205128Ssimon clash. 2051205128Ssimon [Guenter <lists@gknw.net>] 2052205128Ssimon 2053205128Ssimon *) Fix the server certificate chain building code to use X509_verify_cert(), 2054205128Ssimon it used to have an ad-hoc builder which was unable to cope with anything 2055205128Ssimon other than a simple chain. 2056205128Ssimon [David Woodhouse <dwmw2@infradead.org>, Steve Henson] 2057205128Ssimon 2058205128Ssimon *) Don't check self signed certificate signatures in X509_verify_cert() 2059205128Ssimon by default (a flag can override this): it just wastes time without 2060205128Ssimon adding any security. As a useful side effect self signed root CAs 2061205128Ssimon with non-FIPS digests are now usable in FIPS mode. 2062205128Ssimon [Steve Henson] 2063205128Ssimon 2064205128Ssimon *) In dtls1_process_out_of_seq_message() the check if the current message 2065205128Ssimon is already buffered was missing. For every new message was memory 2066205128Ssimon allocated, allowing an attacker to perform an denial of service attack 2067205128Ssimon with sending out of seq handshake messages until there is no memory 2068205128Ssimon left. Additionally every future messege was buffered, even if the 2069205128Ssimon sequence number made no sense and would be part of another handshake. 2070205128Ssimon So only messages with sequence numbers less than 10 in advance will be 2071205128Ssimon buffered. (CVE-2009-1378) 2072205128Ssimon [Robin Seggelmann, discovered by Daniel Mentz] 2073205128Ssimon 2074205128Ssimon *) Records are buffered if they arrive with a future epoch to be 2075205128Ssimon processed after finishing the corresponding handshake. There is 2076205128Ssimon currently no limitation to this buffer allowing an attacker to perform 2077205128Ssimon a DOS attack with sending records with future epochs until there is no 2078205128Ssimon memory left. This patch adds the pqueue_size() function to detemine 2079205128Ssimon the size of a buffer and limits the record buffer to 100 entries. 2080205128Ssimon (CVE-2009-1377) 2081205128Ssimon [Robin Seggelmann, discovered by Daniel Mentz] 2082205128Ssimon 2083205128Ssimon *) Keep a copy of frag->msg_header.frag_len so it can be used after the 2084205128Ssimon parent structure is freed. (CVE-2009-1379) 2085205128Ssimon [Daniel Mentz] 2086205128Ssimon 2087205128Ssimon *) Handle non-blocking I/O properly in SSL_shutdown() call. 2088205128Ssimon [Darryl Miles <darryl-mailinglists@netbauds.net>] 2089205128Ssimon 2090205128Ssimon *) Add 2.5.4.* OIDs 2091205128Ssimon [Ilya O. <vrghost@gmail.com>] 2092205128Ssimon 2093205128Ssimon Changes between 0.9.8k and 0.9.8l [5 Nov 2009] 2094205128Ssimon 2095205128Ssimon *) Disable renegotiation completely - this fixes a severe security 2096205128Ssimon problem (CVE-2009-3555) at the cost of breaking all 2097205128Ssimon renegotiation. Renegotiation can be re-enabled by setting 2098205128Ssimon SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at 2099205128Ssimon run-time. This is really not recommended unless you know what 2100205128Ssimon you're doing. 2101205128Ssimon [Ben Laurie] 2102205128Ssimon 2103194206Ssimon Changes between 0.9.8j and 0.9.8k [25 Mar 2009] 2104167612Ssimon 2105194206Ssimon *) Don't set val to NULL when freeing up structures, it is freed up by 2106194206Ssimon underlying code. If sizeof(void *) > sizeof(long) this can result in 2107194206Ssimon zeroing past the valid field. (CVE-2009-0789) 2108194206Ssimon [Paolo Ganci <Paolo.Ganci@AdNovum.CH>] 2109194206Ssimon 2110194206Ssimon *) Fix bug where return value of CMS_SignerInfo_verify_content() was not 2111194206Ssimon checked correctly. This would allow some invalid signed attributes to 2112194206Ssimon appear to verify correctly. (CVE-2009-0591) 2113194206Ssimon [Ivan Nestlerode <inestlerode@us.ibm.com>] 2114194206Ssimon 2115194206Ssimon *) Reject UniversalString and BMPString types with invalid lengths. This 2116194206Ssimon prevents a crash in ASN1_STRING_print_ex() which assumes the strings have 2117194206Ssimon a legal length. (CVE-2009-0590) 2118194206Ssimon [Steve Henson] 2119194206Ssimon 2120194206Ssimon *) Set S/MIME signing as the default purpose rather than setting it 2121194206Ssimon unconditionally. This allows applications to override it at the store 2122194206Ssimon level. 2123194206Ssimon [Steve Henson] 2124194206Ssimon 2125194206Ssimon *) Permit restricted recursion of ASN1 strings. This is needed in practice 2126194206Ssimon to handle some structures. 2127194206Ssimon [Steve Henson] 2128194206Ssimon 2129194206Ssimon *) Improve efficiency of mem_gets: don't search whole buffer each time 2130194206Ssimon for a '\n' 2131194206Ssimon [Jeremy Shapiro <jnshapir@us.ibm.com>] 2132194206Ssimon 2133194206Ssimon *) New -hex option for openssl rand. 2134194206Ssimon [Matthieu Herrb] 2135194206Ssimon 2136194206Ssimon *) Print out UTF8String and NumericString when parsing ASN1. 2137194206Ssimon [Steve Henson] 2138194206Ssimon 2139194206Ssimon *) Support NumericString type for name components. 2140194206Ssimon [Steve Henson] 2141194206Ssimon 2142194206Ssimon *) Allow CC in the environment to override the automatically chosen 2143194206Ssimon compiler. Note that nothing is done to ensure flags work with the 2144194206Ssimon chosen compiler. 2145194206Ssimon [Ben Laurie] 2146194206Ssimon 2147194206Ssimon Changes between 0.9.8i and 0.9.8j [07 Jan 2009] 2148194206Ssimon 2149194206Ssimon *) Properly check EVP_VerifyFinal() and similar return values 2150194206Ssimon (CVE-2008-5077). 2151194206Ssimon [Ben Laurie, Bodo Moeller, Google Security Team] 2152194206Ssimon 2153194206Ssimon *) Enable TLS extensions by default. 2154194206Ssimon [Ben Laurie] 2155194206Ssimon 2156194206Ssimon *) Allow the CHIL engine to be loaded, whether the application is 2157194206Ssimon multithreaded or not. (This does not release the developer from the 2158194206Ssimon obligation to set up the dynamic locking callbacks.) 2159194206Ssimon [Sander Temme <sander@temme.net>] 2160194206Ssimon 2161194206Ssimon *) Use correct exit code if there is an error in dgst command. 2162194206Ssimon [Steve Henson; problem pointed out by Roland Dirlewanger] 2163194206Ssimon 2164194206Ssimon *) Tweak Configure so that you need to say "experimental-jpake" to enable 2165194206Ssimon JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications. 2166194206Ssimon [Bodo Moeller] 2167194206Ssimon 2168194206Ssimon *) Add experimental JPAKE support, including demo authentication in 2169194206Ssimon s_client and s_server. 2170194206Ssimon [Ben Laurie] 2171194206Ssimon 2172194206Ssimon *) Set the comparison function in v3_addr_canonize(). 2173194206Ssimon [Rob Austein <sra@hactrn.net>] 2174194206Ssimon 2175194206Ssimon *) Add support for XMPP STARTTLS in s_client. 2176194206Ssimon [Philip Paeps <philip@freebsd.org>] 2177194206Ssimon 2178194206Ssimon *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior 2179194206Ssimon to ensure that even with this option, only ciphersuites in the 2180194206Ssimon server's preference list will be accepted. (Note that the option 2181194206Ssimon applies only when resuming a session, so the earlier behavior was 2182194206Ssimon just about the algorithm choice for symmetric cryptography.) 2183194206Ssimon [Bodo Moeller] 2184194206Ssimon 2185194206Ssimon Changes between 0.9.8h and 0.9.8i [15 Sep 2008] 2186194206Ssimon 2187205128Ssimon *) Fix NULL pointer dereference if a DTLS server received 2188205128Ssimon ChangeCipherSpec as first record (CVE-2009-1386). 2189205128Ssimon [PR #1679] 2190205128Ssimon 2191194206Ssimon *) Fix a state transitition in s3_srvr.c and d1_srvr.c 2192194206Ssimon (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...). 2193194206Ssimon [Nagendra Modadugu] 2194194206Ssimon 2195194206Ssimon *) The fix in 0.9.8c that supposedly got rid of unsafe 2196194206Ssimon double-checked locking was incomplete for RSA blinding, 2197194206Ssimon addressing just one layer of what turns out to have been 2198194206Ssimon doubly unsafe triple-checked locking. 2199194206Ssimon 2200194206Ssimon So now fix this for real by retiring the MONT_HELPER macro 2201194206Ssimon in crypto/rsa/rsa_eay.c. 2202194206Ssimon 2203194206Ssimon [Bodo Moeller; problem pointed out by Marius Schilder] 2204194206Ssimon 2205194206Ssimon *) Various precautionary measures: 2206194206Ssimon 2207194206Ssimon - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h). 2208194206Ssimon 2209194206Ssimon - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c). 2210194206Ssimon (NB: This would require knowledge of the secret session ticket key 2211194206Ssimon to exploit, in which case you'd be SOL either way.) 2212194206Ssimon 2213194206Ssimon - Change bn_nist.c so that it will properly handle input BIGNUMs 2214194206Ssimon outside the expected range. 2215194206Ssimon 2216194206Ssimon - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG 2217194206Ssimon builds. 2218194206Ssimon 2219194206Ssimon [Neel Mehta, Bodo Moeller] 2220194206Ssimon 2221194206Ssimon *) Allow engines to be "soft loaded" - i.e. optionally don't die if 2222194206Ssimon the load fails. Useful for distros. 2223194206Ssimon [Ben Laurie and the FreeBSD team] 2224194206Ssimon 2225194206Ssimon *) Add support for Local Machine Keyset attribute in PKCS#12 files. 2226194206Ssimon [Steve Henson] 2227194206Ssimon 2228194206Ssimon *) Fix BN_GF2m_mod_arr() top-bit cleanup code. 2229194206Ssimon [Huang Ying] 2230194206Ssimon 2231194206Ssimon *) Expand ENGINE to support engine supplied SSL client certificate functions. 2232194206Ssimon 2233194206Ssimon This work was sponsored by Logica. 2234194206Ssimon [Steve Henson] 2235194206Ssimon 2236194206Ssimon *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows 2237194206Ssimon keystores. Support for SSL/TLS client authentication too. 2238194206Ssimon Not compiled unless enable-capieng specified to Configure. 2239194206Ssimon 2240194206Ssimon This work was sponsored by Logica. 2241194206Ssimon [Steve Henson] 2242194206Ssimon 2243194206Ssimon *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using 2244194206Ssimon ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain 2245194206Ssimon attribute creation routines such as certifcate requests and PKCS#12 2246194206Ssimon files. 2247194206Ssimon [Steve Henson] 2248194206Ssimon 2249194206Ssimon Changes between 0.9.8g and 0.9.8h [28 May 2008] 2250194206Ssimon 2251194206Ssimon *) Fix flaw if 'Server Key exchange message' is omitted from a TLS 2252194206Ssimon handshake which could lead to a cilent crash as found using the 2253194206Ssimon Codenomicon TLS test suite (CVE-2008-1672) 2254194206Ssimon [Steve Henson, Mark Cox] 2255194206Ssimon 2256194206Ssimon *) Fix double free in TLS server name extensions which could lead to 2257194206Ssimon a remote crash found by Codenomicon TLS test suite (CVE-2008-0891) 2258194206Ssimon [Joe Orton] 2259194206Ssimon 2260194206Ssimon *) Clear error queue in SSL_CTX_use_certificate_chain_file() 2261194206Ssimon 2262194206Ssimon Clear the error queue to ensure that error entries left from 2263194206Ssimon older function calls do not interfere with the correct operation. 2264194206Ssimon [Lutz Jaenicke, Erik de Castro Lopo] 2265194206Ssimon 2266194206Ssimon *) Remove root CA certificates of commercial CAs: 2267194206Ssimon 2268194206Ssimon The OpenSSL project does not recommend any specific CA and does not 2269194206Ssimon have any policy with respect to including or excluding any CA. 2270194206Ssimon Therefore it does not make any sense to ship an arbitrary selection 2271194206Ssimon of root CA certificates with the OpenSSL software. 2272194206Ssimon [Lutz Jaenicke] 2273194206Ssimon 2274194206Ssimon *) RSA OAEP patches to fix two separate invalid memory reads. 2275194206Ssimon The first one involves inputs when 'lzero' is greater than 2276194206Ssimon 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes 2277194206Ssimon before the beginning of from). The second one involves inputs where 2278194206Ssimon the 'db' section contains nothing but zeroes (there is a one-byte 2279194206Ssimon invalid read after the end of 'db'). 2280194206Ssimon [Ivan Nestlerode <inestlerode@us.ibm.com>] 2281194206Ssimon 2282194206Ssimon *) Partial backport from 0.9.9-dev: 2283194206Ssimon 2284194206Ssimon Introduce bn_mul_mont (dedicated Montgomery multiplication 2285194206Ssimon procedure) as a candidate for BIGNUM assembler implementation. 2286194206Ssimon While 0.9.9-dev uses assembler for various architectures, only 2287194206Ssimon x86_64 is available by default here in the 0.9.8 branch, and 2288194206Ssimon 32-bit x86 is available through a compile-time setting. 2289194206Ssimon 2290194206Ssimon To try the 32-bit x86 assembler implementation, use Configure 2291194206Ssimon option "enable-montasm" (which exists only for this backport). 2292194206Ssimon 2293194206Ssimon As "enable-montasm" for 32-bit x86 disclaims code stability 2294194206Ssimon anyway, in this constellation we activate additional code 2295194206Ssimon backported from 0.9.9-dev for further performance improvements, 2296194206Ssimon namely BN_from_montgomery_word. (To enable this otherwise, 2297194206Ssimon e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".) 2298194206Ssimon 2299194206Ssimon [Andy Polyakov (backport partially by Bodo Moeller)] 2300194206Ssimon 2301194206Ssimon *) Add TLS session ticket callback. This allows an application to set 2302194206Ssimon TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed 2303194206Ssimon values. This is useful for key rollover for example where several key 2304194206Ssimon sets may exist with different names. 2305194206Ssimon [Steve Henson] 2306194206Ssimon 2307194206Ssimon *) Reverse ENGINE-internal logic for caching default ENGINE handles. 2308194206Ssimon This was broken until now in 0.9.8 releases, such that the only way 2309194206Ssimon a registered ENGINE could be used (assuming it initialises 2310194206Ssimon successfully on the host) was to explicitly set it as the default 2311194206Ssimon for the relevant algorithms. This is in contradiction with 0.9.7 2312194206Ssimon behaviour and the documentation. With this fix, when an ENGINE is 2313194206Ssimon registered into a given algorithm's table of implementations, the 2314194206Ssimon 'uptodate' flag is reset so that auto-discovery will be used next 2315194206Ssimon time a new context for that algorithm attempts to select an 2316194206Ssimon implementation. 2317194206Ssimon [Ian Lister (tweaked by Geoff Thorpe)] 2318194206Ssimon 2319194206Ssimon *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9 2320194206Ssimon implemention in the following ways: 2321194206Ssimon 2322194206Ssimon Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be 2323194206Ssimon hard coded. 2324194206Ssimon 2325194206Ssimon Lack of BER streaming support means one pass streaming processing is 2326194206Ssimon only supported if data is detached: setting the streaming flag is 2327194206Ssimon ignored for embedded content. 2328194206Ssimon 2329194206Ssimon CMS support is disabled by default and must be explicitly enabled 2330194206Ssimon with the enable-cms configuration option. 2331194206Ssimon [Steve Henson] 2332194206Ssimon 2333194206Ssimon *) Update the GMP engine glue to do direct copies between BIGNUM and 2334194206Ssimon mpz_t when openssl and GMP use the same limb size. Otherwise the 2335194206Ssimon existing "conversion via a text string export" trick is still used. 2336194206Ssimon [Paul Sheer <paulsheer@gmail.com>] 2337194206Ssimon 2338194206Ssimon *) Zlib compression BIO. This is a filter BIO which compressed and 2339194206Ssimon uncompresses any data passed through it. 2340194206Ssimon [Steve Henson] 2341194206Ssimon 2342194206Ssimon *) Add AES_wrap_key() and AES_unwrap_key() functions to implement 2343194206Ssimon RFC3394 compatible AES key wrapping. 2344194206Ssimon [Steve Henson] 2345194206Ssimon 2346194206Ssimon *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0(): 2347194206Ssimon sets string data without copying. X509_ALGOR_set0() and 2348194206Ssimon X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier) 2349194206Ssimon data. Attribute function X509at_get0_data_by_OBJ(): retrieves data 2350194206Ssimon from an X509_ATTRIBUTE structure optionally checking it occurs only 2351194206Ssimon once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied 2352194206Ssimon data. 2353194206Ssimon [Steve Henson] 2354194206Ssimon 2355194206Ssimon *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set() 2356194206Ssimon to get the expected BN_FLG_CONSTTIME behavior. 2357194206Ssimon [Bodo Moeller (Google)] 2358194206Ssimon 2359194206Ssimon *) Netware support: 2360194206Ssimon 2361194206Ssimon - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets 2362194206Ssimon - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT) 2363194206Ssimon - added some more tests to do_tests.pl 2364194206Ssimon - fixed RunningProcess usage so that it works with newer LIBC NDKs too 2365194206Ssimon - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency 2366194206Ssimon - added new Configure targets netware-clib-bsdsock, netware-clib-gcc, 2367194206Ssimon netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc 2368194206Ssimon - various changes to netware.pl to enable gcc-cross builds on Win32 2369194206Ssimon platform 2370194206Ssimon - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD) 2371194206Ssimon - various changes to fix missing prototype warnings 2372194206Ssimon - fixed x86nasm.pl to create correct asm files for NASM COFF output 2373194206Ssimon - added AES, WHIRLPOOL and CPUID assembler code to build files 2374194206Ssimon - added missing AES assembler make rules to mk1mf.pl 2375194206Ssimon - fixed order of includes in apps/ocsp.c so that e_os.h settings apply 2376194206Ssimon [Guenter Knauf <eflash@gmx.net>] 2377194206Ssimon 2378194206Ssimon *) Implement certificate status request TLS extension defined in RFC3546. 2379194206Ssimon A client can set the appropriate parameters and receive the encoded 2380194206Ssimon OCSP response via a callback. A server can query the supplied parameters 2381194206Ssimon and set the encoded OCSP response in the callback. Add simplified examples 2382194206Ssimon to s_client and s_server. 2383194206Ssimon [Steve Henson] 2384194206Ssimon 2385194206Ssimon Changes between 0.9.8f and 0.9.8g [19 Oct 2007] 2386194206Ssimon 2387194206Ssimon *) Fix various bugs: 2388194206Ssimon + Binary incompatibility of ssl_ctx_st structure 2389194206Ssimon + DTLS interoperation with non-compliant servers 2390194206Ssimon + Don't call get_session_cb() without proposed session 2391194206Ssimon + Fix ia64 assembler code 2392194206Ssimon [Andy Polyakov, Steve Henson] 2393194206Ssimon 2394194206Ssimon Changes between 0.9.8e and 0.9.8f [11 Oct 2007] 2395194206Ssimon 2396194206Ssimon *) DTLS Handshake overhaul. There were longstanding issues with 2397194206Ssimon OpenSSL DTLS implementation, which were making it impossible for 2398194206Ssimon RFC 4347 compliant client to communicate with OpenSSL server. 2399194206Ssimon Unfortunately just fixing these incompatibilities would "cut off" 2400194206Ssimon pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e 2401194206Ssimon server keeps tolerating non RFC compliant syntax. The opposite is 2402194206Ssimon not true, 0.9.8f client can not communicate with earlier server. 2403194206Ssimon This update even addresses CVE-2007-4995. 2404194206Ssimon [Andy Polyakov] 2405194206Ssimon 2406194206Ssimon *) Changes to avoid need for function casts in OpenSSL: some compilers 2407194206Ssimon (gcc 4.2 and later) reject their use. 2408194206Ssimon [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>, 2409194206Ssimon Steve Henson] 2410194206Ssimon 2411194206Ssimon *) Add RFC4507 support to OpenSSL. This includes the corrections in 2412194206Ssimon RFC4507bis. The encrypted ticket format is an encrypted encoded 2413194206Ssimon SSL_SESSION structure, that way new session features are automatically 2414194206Ssimon supported. 2415194206Ssimon 2416194206Ssimon If a client application caches session in an SSL_SESSION structure 2417194206Ssimon support is transparent because tickets are now stored in the encoded 2418194206Ssimon SSL_SESSION. 2419194206Ssimon 2420194206Ssimon The SSL_CTX structure automatically generates keys for ticket 2421194206Ssimon protection in servers so again support should be possible 2422194206Ssimon with no application modification. 2423194206Ssimon 2424194206Ssimon If a client or server wishes to disable RFC4507 support then the option 2425194206Ssimon SSL_OP_NO_TICKET can be set. 2426194206Ssimon 2427194206Ssimon Add a TLS extension debugging callback to allow the contents of any client 2428194206Ssimon or server extensions to be examined. 2429194206Ssimon 2430194206Ssimon This work was sponsored by Google. 2431194206Ssimon [Steve Henson] 2432194206Ssimon 2433194206Ssimon *) Add initial support for TLS extensions, specifically for the server_name 2434194206Ssimon extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 2435194206Ssimon have new members for a host name. The SSL data structure has an 2436194206Ssimon additional member SSL_CTX *initial_ctx so that new sessions can be 2437194206Ssimon stored in that context to allow for session resumption, even after the 2438194206Ssimon SSL has been switched to a new SSL_CTX in reaction to a client's 2439194206Ssimon server_name extension. 2440194206Ssimon 2441194206Ssimon New functions (subject to change): 2442194206Ssimon 2443194206Ssimon SSL_get_servername() 2444194206Ssimon SSL_get_servername_type() 2445194206Ssimon SSL_set_SSL_CTX() 2446194206Ssimon 2447194206Ssimon New CTRL codes and macros (subject to change): 2448194206Ssimon 2449194206Ssimon SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 2450194206Ssimon - SSL_CTX_set_tlsext_servername_callback() 2451194206Ssimon SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 2452194206Ssimon - SSL_CTX_set_tlsext_servername_arg() 2453194206Ssimon SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 2454194206Ssimon 2455194206Ssimon openssl s_client has a new '-servername ...' option. 2456194206Ssimon 2457194206Ssimon openssl s_server has new options '-servername_host ...', '-cert2 ...', 2458194206Ssimon '-key2 ...', '-servername_fatal' (subject to change). This allows 2459194206Ssimon testing the HostName extension for a specific single host name ('-cert' 2460194206Ssimon and '-key' remain fallbacks for handshakes without HostName 2461194206Ssimon negotiation). If the unrecogninzed_name alert has to be sent, this by 2462194206Ssimon default is a warning; it becomes fatal with the '-servername_fatal' 2463194206Ssimon option. 2464194206Ssimon 2465194206Ssimon [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson] 2466194206Ssimon 2467194206Ssimon *) Add AES and SSE2 assembly language support to VC++ build. 2468194206Ssimon [Steve Henson] 2469194206Ssimon 2470194206Ssimon *) Mitigate attack on final subtraction in Montgomery reduction. 2471194206Ssimon [Andy Polyakov] 2472194206Ssimon 2473194206Ssimon *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0 2474194206Ssimon (which previously caused an internal error). 2475194206Ssimon [Bodo Moeller] 2476194206Ssimon 2477194206Ssimon *) Squeeze another 10% out of IGE mode when in != out. 2478194206Ssimon [Ben Laurie] 2479194206Ssimon 2480194206Ssimon *) AES IGE mode speedup. 2481194206Ssimon [Dean Gaudet (Google)] 2482194206Ssimon 2483194206Ssimon *) Add the Korean symmetric 128-bit cipher SEED (see 2484194206Ssimon http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and 2485194206Ssimon add SEED ciphersuites from RFC 4162: 2486194206Ssimon 2487194206Ssimon TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA" 2488194206Ssimon TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA" 2489194206Ssimon TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA" 2490194206Ssimon TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA" 2491194206Ssimon 2492194206Ssimon To minimize changes between patchlevels in the OpenSSL 0.9.8 2493194206Ssimon series, SEED remains excluded from compilation unless OpenSSL 2494194206Ssimon is configured with 'enable-seed'. 2495194206Ssimon [KISA, Bodo Moeller] 2496194206Ssimon 2497194206Ssimon *) Mitigate branch prediction attacks, which can be practical if a 2498194206Ssimon single processor is shared, allowing a spy process to extract 2499194206Ssimon information. For detailed background information, see 2500194206Ssimon http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron, 2501194206Ssimon J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL 2502194206Ssimon and Necessary Software Countermeasures"). The core of the change 2503194206Ssimon are new versions BN_div_no_branch() and 2504194206Ssimon BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(), 2505194206Ssimon respectively, which are slower, but avoid the security-relevant 2506194206Ssimon conditional branches. These are automatically called by BN_div() 2507194206Ssimon and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one 2508194206Ssimon of the input BIGNUMs. Also, BN_is_bit_set() has been changed to 2509194206Ssimon remove a conditional branch. 2510194206Ssimon 2511194206Ssimon BN_FLG_CONSTTIME is the new name for the previous 2512194206Ssimon BN_FLG_EXP_CONSTTIME flag, since it now affects more than just 2513194206Ssimon modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag 2514194206Ssimon in the exponent causes BN_mod_exp_mont() to use the alternative 2515194206Ssimon implementation in BN_mod_exp_mont_consttime().) The old name 2516194206Ssimon remains as a deprecated alias. 2517194206Ssimon 2518194206Ssimon Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general 2519194206Ssimon RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses 2520194206Ssimon constant-time implementations for more than just exponentiation. 2521194206Ssimon Here too the old name is kept as a deprecated alias. 2522194206Ssimon 2523194206Ssimon BN_BLINDING_new() will now use BN_dup() for the modulus so that 2524194206Ssimon the BN_BLINDING structure gets an independent copy of the 2525194206Ssimon modulus. This means that the previous "BIGNUM *m" argument to 2526194206Ssimon BN_BLINDING_new() and to BN_BLINDING_create_param() now 2527194206Ssimon essentially becomes "const BIGNUM *m", although we can't actually 2528194206Ssimon change this in the header file before 0.9.9. It allows 2529194206Ssimon RSA_setup_blinding() to use BN_with_flags() on the modulus to 2530194206Ssimon enable BN_FLG_CONSTTIME. 2531194206Ssimon 2532194206Ssimon [Matthew D Wood (Intel Corp)] 2533194206Ssimon 2534194206Ssimon *) In the SSL/TLS server implementation, be strict about session ID 2535194206Ssimon context matching (which matters if an application uses a single 2536194206Ssimon external cache for different purposes). Previously, 2537194206Ssimon out-of-context reuse was forbidden only if SSL_VERIFY_PEER was 2538194206Ssimon set. This did ensure strict client verification, but meant that, 2539194206Ssimon with applications using a single external cache for quite 2540194206Ssimon different requirements, clients could circumvent ciphersuite 2541194206Ssimon restrictions for a given session ID context by starting a session 2542194206Ssimon in a different context. 2543194206Ssimon [Bodo Moeller] 2544194206Ssimon 2545167612Ssimon *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 2546167612Ssimon a ciphersuite string such as "DEFAULT:RSA" cannot enable 2547167612Ssimon authentication-only ciphersuites. 2548167612Ssimon [Bodo Moeller] 2549167612Ssimon 2550194206Ssimon *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was 2551194206Ssimon not complete and could lead to a possible single byte overflow 2552194206Ssimon (CVE-2007-5135) [Ben Laurie] 2553194206Ssimon 2554194206Ssimon Changes between 0.9.8d and 0.9.8e [23 Feb 2007] 2555194206Ssimon 2556167612Ssimon *) Since AES128 and AES256 (and similarly Camellia128 and 2557167612Ssimon Camellia256) share a single mask bit in the logic of 2558167612Ssimon ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 2559167612Ssimon kludge to work properly if AES128 is available and AES256 isn't 2560167612Ssimon (or if Camellia128 is available and Camellia256 isn't). 2561167612Ssimon [Victor Duchovni] 2562167612Ssimon 2563167612Ssimon *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c 2564167612Ssimon (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters): 2565167612Ssimon When a point or a seed is encoded in a BIT STRING, we need to 2566167612Ssimon prevent the removal of trailing zero bits to get the proper DER 2567167612Ssimon encoding. (By default, crypto/asn1/a_bitstr.c assumes the case 2568167612Ssimon of a NamedBitList, for which trailing 0 bits need to be removed.) 2569167612Ssimon [Bodo Moeller] 2570167612Ssimon 2571167612Ssimon *) Have SSL/TLS server implementation tolerate "mismatched" record 2572167612Ssimon protocol version while receiving ClientHello even if the 2573167612Ssimon ClientHello is fragmented. (The server can't insist on the 2574167612Ssimon particular protocol version it has chosen before the ServerHello 2575167612Ssimon message has informed the client about his choice.) 2576167612Ssimon [Bodo Moeller] 2577167612Ssimon 2578167612Ssimon *) Add RFC 3779 support. 2579167612Ssimon [Rob Austein for ARIN, Ben Laurie] 2580167612Ssimon 2581167612Ssimon *) Load error codes if they are not already present instead of using a 2582167612Ssimon static variable. This allows them to be cleanly unloaded and reloaded. 2583167612Ssimon Improve header file function name parsing. 2584167612Ssimon [Steve Henson] 2585167612Ssimon 2586167612Ssimon *) extend SMTP and IMAP protocol emulation in s_client to use EHLO 2587167612Ssimon or CAPABILITY handshake as required by RFCs. 2588167612Ssimon [Goetz Babin-Ebell] 2589167612Ssimon 2590162911Ssimon Changes between 0.9.8c and 0.9.8d [28 Sep 2006] 2591162911Ssimon 2592162911Ssimon *) Introduce limits to prevent malicious keys being able to 2593162911Ssimon cause a denial of service. (CVE-2006-2940) 2594162911Ssimon [Steve Henson, Bodo Moeller] 2595162911Ssimon 2596162911Ssimon *) Fix ASN.1 parsing of certain invalid structures that can result 2597162911Ssimon in a denial of service. (CVE-2006-2937) [Steve Henson] 2598162911Ssimon 2599162911Ssimon *) Fix buffer overflow in SSL_get_shared_ciphers() function. 2600162911Ssimon (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 2601162911Ssimon 2602162911Ssimon *) Fix SSL client code which could crash if connecting to a 2603162911Ssimon malicious SSLv2 server. (CVE-2006-4343) 2604162911Ssimon [Tavis Ormandy and Will Drewry, Google Security Team] 2605162911Ssimon 2606162911Ssimon *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites 2607162911Ssimon match only those. Before that, "AES256-SHA" would be interpreted 2608162911Ssimon as a pattern and match "AES128-SHA" too (since AES128-SHA got 2609162911Ssimon the same strength classification in 0.9.7h) as we currently only 2610162911Ssimon have a single AES bit in the ciphersuite description bitmap. 2611162911Ssimon That change, however, also applied to ciphersuite strings such as 2612162911Ssimon "RC4-MD5" that intentionally matched multiple ciphersuites -- 2613162911Ssimon namely, SSL 2.0 ciphersuites in addition to the more common ones 2614162911Ssimon from SSL 3.0/TLS 1.0. 2615162911Ssimon 2616162911Ssimon So we change the selection algorithm again: Naming an explicit 2617162911Ssimon ciphersuite selects this one ciphersuite, and any other similar 2618162911Ssimon ciphersuite (same bitmap) from *other* protocol versions. 2619162911Ssimon Thus, "RC4-MD5" again will properly select both the SSL 2.0 2620162911Ssimon ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite. 2621162911Ssimon 2622162911Ssimon Since SSL 2.0 does not have any ciphersuites for which the 2623162911Ssimon 128/256 bit distinction would be relevant, this works for now. 2624162911Ssimon The proper fix will be to use different bits for AES128 and 2625162911Ssimon AES256, which would have avoided the problems from the beginning; 2626162911Ssimon however, bits are scarce, so we can only do this in a new release 2627162911Ssimon (not just a patchlevel) when we can change the SSL_CIPHER 2628162911Ssimon definition to split the single 'unsigned long mask' bitmap into 2629162911Ssimon multiple values to extend the available space. 2630162911Ssimon 2631162911Ssimon [Bodo Moeller] 2632162911Ssimon 2633162911Ssimon Changes between 0.9.8b and 0.9.8c [05 Sep 2006] 2634162911Ssimon 2635162911Ssimon *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 2636162911Ssimon (CVE-2006-4339) [Ben Laurie and Google Security Team] 2637162911Ssimon 2638162911Ssimon *) Add AES IGE and biIGE modes. 2639162911Ssimon [Ben Laurie] 2640162911Ssimon 2641162911Ssimon *) Change the Unix randomness entropy gathering to use poll() when 2642162911Ssimon possible instead of select(), since the latter has some 2643162911Ssimon undesirable limitations. 2644162911Ssimon [Darryl Miles via Richard Levitte and Bodo Moeller] 2645162911Ssimon 2646162911Ssimon *) Disable "ECCdraft" ciphersuites more thoroughly. Now special 2647162911Ssimon treatment in ssl/ssl_ciph.s makes sure that these ciphersuites 2648162911Ssimon cannot be implicitly activated as part of, e.g., the "AES" alias. 2649162911Ssimon However, please upgrade to OpenSSL 0.9.9[-dev] for 2650162911Ssimon non-experimental use of the ECC ciphersuites to get TLS extension 2651162911Ssimon support, which is required for curve and point format negotiation 2652162911Ssimon to avoid potential handshake problems. 2653162911Ssimon [Bodo Moeller] 2654162911Ssimon 2655162911Ssimon *) Disable rogue ciphersuites: 2656162911Ssimon 2657162911Ssimon - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 2658162911Ssimon - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 2659162911Ssimon - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 2660162911Ssimon 2661162911Ssimon The latter two were purportedly from 2662162911Ssimon draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 2663162911Ssimon appear there. 2664162911Ssimon 2665167612Ssimon Also deactivate the remaining ciphersuites from 2666162911Ssimon draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 2667162911Ssimon unofficial, and the ID has long expired. 2668162911Ssimon [Bodo Moeller] 2669162911Ssimon 2670162911Ssimon *) Fix RSA blinding Heisenbug (problems sometimes occured on 2671162911Ssimon dual-core machines) and other potential thread-safety issues. 2672162911Ssimon [Bodo Moeller] 2673162911Ssimon 2674162911Ssimon *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key 2675162911Ssimon versions), which is now available for royalty-free use 2676162911Ssimon (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html). 2677162911Ssimon Also, add Camellia TLS ciphersuites from RFC 4132. 2678162911Ssimon 2679162911Ssimon To minimize changes between patchlevels in the OpenSSL 0.9.8 2680162911Ssimon series, Camellia remains excluded from compilation unless OpenSSL 2681162911Ssimon is configured with 'enable-camellia'. 2682162911Ssimon [NTT] 2683162911Ssimon 2684162911Ssimon *) Disable the padding bug check when compression is in use. The padding 2685162911Ssimon bug check assumes the first packet is of even length, this is not 2686162911Ssimon necessarily true if compresssion is enabled and can result in false 2687162911Ssimon positives causing handshake failure. The actual bug test is ancient 2688162911Ssimon code so it is hoped that implementations will either have fixed it by 2689162911Ssimon now or any which still have the bug do not support compression. 2690162911Ssimon [Steve Henson] 2691162911Ssimon 2692160814Ssimon Changes between 0.9.8a and 0.9.8b [04 May 2006] 2693160814Ssimon 2694160814Ssimon *) When applying a cipher rule check to see if string match is an explicit 2695160814Ssimon cipher suite and only match that one cipher suite if it is. 2696160814Ssimon [Steve Henson] 2697160814Ssimon 2698160814Ssimon *) Link in manifests for VC++ if needed. 2699160814Ssimon [Austin Ziegler <halostatue@gmail.com>] 2700160814Ssimon 2701160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 2702160814Ssimon draft-ietf-tls-ecc-12.txt with proposed changes (but without 2703160814Ssimon TLS extensions, which are supported starting with the 0.9.9 2704160814Ssimon branch, not in the OpenSSL 0.9.8 branch). 2705160814Ssimon [Douglas Stebila] 2706160814Ssimon 2707160814Ssimon *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support 2708160814Ssimon opaque EVP_CIPHER_CTX handling. 2709160814Ssimon [Steve Henson] 2710160814Ssimon 2711160814Ssimon *) Fixes and enhancements to zlib compression code. We now only use 2712160814Ssimon "zlib1.dll" and use the default __cdecl calling convention on Win32 2713160814Ssimon to conform with the standards mentioned here: 2714160814Ssimon http://www.zlib.net/DLL_FAQ.txt 2715160814Ssimon Static zlib linking now works on Windows and the new --with-zlib-include 2716160814Ssimon --with-zlib-lib options to Configure can be used to supply the location 2717160814Ssimon of the headers and library. Gracefully handle case where zlib library 2718160814Ssimon can't be loaded. 2719160814Ssimon [Steve Henson] 2720160814Ssimon 2721160814Ssimon *) Several fixes and enhancements to the OID generation code. The old code 2722160814Ssimon sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't 2723160814Ssimon handle numbers larger than ULONG_MAX, truncated printing and had a 2724160814Ssimon non standard OBJ_obj2txt() behaviour. 2725160814Ssimon [Steve Henson] 2726160814Ssimon 2727160814Ssimon *) Add support for building of engines under engine/ as shared libraries 2728160814Ssimon under VC++ build system. 2729160814Ssimon [Steve Henson] 2730160814Ssimon 2731160814Ssimon *) Corrected the numerous bugs in the Win32 path splitter in DSO. 2732160814Ssimon Hopefully, we will not see any false combination of paths any more. 2733160814Ssimon [Richard Levitte] 2734160814Ssimon 2735160814Ssimon Changes between 0.9.8 and 0.9.8a [11 Oct 2005] 2736160814Ssimon 2737160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 2738160814Ssimon (part of SSL_OP_ALL). This option used to disable the 2739160814Ssimon countermeasure against man-in-the-middle protocol-version 2740160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 2741160814Ssimon idea. (CVE-2005-2969) 2742160814Ssimon 2743160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 2744160814Ssimon for Information Security, National Institute of Advanced Industrial 2745160814Ssimon Science and Technology [AIST], Japan)] 2746160814Ssimon 2747160814Ssimon *) Add two function to clear and return the verify parameter flags. 2748160814Ssimon [Steve Henson] 2749160814Ssimon 2750160814Ssimon *) Keep cipherlists sorted in the source instead of sorting them at 2751160814Ssimon runtime, thus removing the need for a lock. 2752160814Ssimon [Nils Larsch] 2753160814Ssimon 2754160814Ssimon *) Avoid some small subgroup attacks in Diffie-Hellman. 2755160814Ssimon [Nick Mathewson and Ben Laurie] 2756160814Ssimon 2757160814Ssimon *) Add functions for well-known primes. 2758160814Ssimon [Nick Mathewson] 2759160814Ssimon 2760160814Ssimon *) Extended Windows CE support. 2761160814Ssimon [Satoshi Nakamura and Andy Polyakov] 2762160814Ssimon 2763160814Ssimon *) Initialize SSL_METHOD structures at compile time instead of during 2764160814Ssimon runtime, thus removing the need for a lock. 2765160814Ssimon [Steve Henson] 2766160814Ssimon 2767160814Ssimon *) Make PKCS7_decrypt() work even if no certificate is supplied by 2768160814Ssimon attempting to decrypt each encrypted key in turn. Add support to 2769160814Ssimon smime utility. 2770160814Ssimon [Steve Henson] 2771160814Ssimon 2772160814Ssimon Changes between 0.9.7h and 0.9.8 [05 Jul 2005] 2773160814Ssimon 2774162911Ssimon [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after 2775162911Ssimon OpenSSL 0.9.8.] 2776162911Ssimon 2777160814Ssimon *) Add libcrypto.pc and libssl.pc for those who feel they need them. 2778160814Ssimon [Richard Levitte] 2779160814Ssimon 2780160814Ssimon *) Change CA.sh and CA.pl so they don't bundle the CSR and the private 2781160814Ssimon key into the same file any more. 2782160814Ssimon [Richard Levitte] 2783160814Ssimon 2784160814Ssimon *) Add initial support for Win64, both IA64 and AMD64/x64 flavors. 2785160814Ssimon [Andy Polyakov] 2786160814Ssimon 2787160814Ssimon *) Add -utf8 command line and config file option to 'ca'. 2788160814Ssimon [Stefan <stf@udoma.org] 2789160814Ssimon 2790160814Ssimon *) Removed the macro des_crypt(), as it seems to conflict with some 2791160814Ssimon libraries. Use DES_crypt(). 2792160814Ssimon [Richard Levitte] 2793160814Ssimon 2794160814Ssimon *) Correct naming of the 'chil' and '4758cca' ENGINEs. This 2795160814Ssimon involves renaming the source and generated shared-libs for 2796160814Ssimon both. The engines will accept the corrected or legacy ids 2797160814Ssimon ('ncipher' and '4758_cca' respectively) when binding. NB, 2798160814Ssimon this only applies when building 'shared'. 2799160814Ssimon [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe] 2800160814Ssimon 2801160814Ssimon *) Add attribute functions to EVP_PKEY structure. Modify 2802160814Ssimon PKCS12_create() to recognize a CSP name attribute and 2803160814Ssimon use it. Make -CSP option work again in pkcs12 utility. 2804160814Ssimon [Steve Henson] 2805160814Ssimon 2806160814Ssimon *) Add new functionality to the bn blinding code: 2807160814Ssimon - automatic re-creation of the BN_BLINDING parameters after 2808160814Ssimon a fixed number of uses (currently 32) 2809160814Ssimon - add new function for parameter creation 2810160814Ssimon - introduce flags to control the update behaviour of the 2811160814Ssimon BN_BLINDING parameters 2812160814Ssimon - hide BN_BLINDING structure 2813160814Ssimon Add a second BN_BLINDING slot to the RSA structure to improve 2814160814Ssimon performance when a single RSA object is shared among several 2815160814Ssimon threads. 2816160814Ssimon [Nils Larsch] 2817160814Ssimon 2818160814Ssimon *) Add support for DTLS. 2819160814Ssimon [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie] 2820160814Ssimon 2821160814Ssimon *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1) 2822160814Ssimon to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file() 2823160814Ssimon [Walter Goulet] 2824160814Ssimon 2825160814Ssimon *) Remove buggy and incompletet DH cert support from 2826160814Ssimon ssl/ssl_rsa.c and ssl/s3_both.c 2827160814Ssimon [Nils Larsch] 2828160814Ssimon 2829160814Ssimon *) Use SHA-1 instead of MD5 as the default digest algorithm for 2830160814Ssimon the apps/openssl applications. 2831160814Ssimon [Nils Larsch] 2832160814Ssimon 2833160814Ssimon *) Compile clean with "-Wall -Wmissing-prototypes 2834160814Ssimon -Wstrict-prototypes -Wmissing-declarations -Werror". Currently 2835160814Ssimon DEBUG_SAFESTACK must also be set. 2836160814Ssimon [Ben Laurie] 2837160814Ssimon 2838160814Ssimon *) Change ./Configure so that certain algorithms can be disabled by default. 2839160814Ssimon The new counterpiece to "no-xxx" is "enable-xxx". 2840160814Ssimon 2841160814Ssimon The patented RC5 and MDC2 algorithms will now be disabled unless 2842160814Ssimon "enable-rc5" and "enable-mdc2", respectively, are specified. 2843160814Ssimon 2844160814Ssimon (IDEA remains enabled despite being patented. This is because IDEA 2845160814Ssimon is frequently required for interoperability, and there is no license 2846160814Ssimon fee for non-commercial use. As before, "no-idea" can be used to 2847160814Ssimon avoid this algorithm.) 2848160814Ssimon 2849160814Ssimon [Bodo Moeller] 2850160814Ssimon 2851160814Ssimon *) Add processing of proxy certificates (see RFC 3820). This work was 2852160814Ssimon sponsored by KTH (The Royal Institute of Technology in Stockholm) and 2853160814Ssimon EGEE (Enabling Grids for E-science in Europe). 2854160814Ssimon [Richard Levitte] 2855160814Ssimon 2856160814Ssimon *) RC4 performance overhaul on modern architectures/implementations, such 2857160814Ssimon as Intel P4, IA-64 and AMD64. 2858160814Ssimon [Andy Polyakov] 2859160814Ssimon 2860160814Ssimon *) New utility extract-section.pl. This can be used specify an alternative 2861160814Ssimon section number in a pod file instead of having to treat each file as 2862160814Ssimon a separate case in Makefile. This can be done by adding two lines to the 2863160814Ssimon pod file: 2864160814Ssimon 2865160814Ssimon =for comment openssl_section:XXX 2866160814Ssimon 2867160814Ssimon The blank line is mandatory. 2868160814Ssimon 2869160814Ssimon [Steve Henson] 2870160814Ssimon 2871160814Ssimon *) New arguments -certform, -keyform and -pass for s_client and s_server 2872160814Ssimon to allow alternative format key and certificate files and passphrase 2873160814Ssimon sources. 2874160814Ssimon [Steve Henson] 2875160814Ssimon 2876160814Ssimon *) New structure X509_VERIFY_PARAM which combines current verify parameters, 2877160814Ssimon update associated structures and add various utility functions. 2878160814Ssimon 2879160814Ssimon Add new policy related verify parameters, include policy checking in 2880160814Ssimon standard verify code. Enhance 'smime' application with extra parameters 2881160814Ssimon to support policy checking and print out. 2882160814Ssimon [Steve Henson] 2883160814Ssimon 2884160814Ssimon *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3 2885160814Ssimon Nehemiah processors. These extensions support AES encryption in hardware 2886160814Ssimon as well as RNG (though RNG support is currently disabled). 2887160814Ssimon [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov] 2888160814Ssimon 2889160814Ssimon *) Deprecate BN_[get|set]_params() functions (they were ignored internally). 2890160814Ssimon [Geoff Thorpe] 2891160814Ssimon 2892160814Ssimon *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented. 2893160814Ssimon [Andy Polyakov and a number of other people] 2894160814Ssimon 2895160814Ssimon *) Improved PowerPC platform support. Most notably BIGNUM assembler 2896160814Ssimon implementation contributed by IBM. 2897160814Ssimon [Suresh Chari, Peter Waltenberg, Andy Polyakov] 2898160814Ssimon 2899160814Ssimon *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public 2900160814Ssimon exponent rather than 'unsigned long'. There is a corresponding change to 2901160814Ssimon the new 'rsa_keygen' element of the RSA_METHOD structure. 2902160814Ssimon [Jelte Jansen, Geoff Thorpe] 2903160814Ssimon 2904160814Ssimon *) Functionality for creating the initial serial number file is now 2905160814Ssimon moved from CA.pl to the 'ca' utility with a new option -create_serial. 2906160814Ssimon 2907160814Ssimon (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial 2908160814Ssimon number file to 1, which is bound to cause problems. To avoid 2909160814Ssimon the problems while respecting compatibility between different 0.9.7 2910160814Ssimon patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in 2911160814Ssimon CA.pl for serial number initialization. With the new release 0.9.8, 2912160814Ssimon we can fix the problem directly in the 'ca' utility.) 2913160814Ssimon [Steve Henson] 2914160814Ssimon 2915160814Ssimon *) Reduced header interdepencies by declaring more opaque objects in 2916160814Ssimon ossl_typ.h. As a consequence, including some headers (eg. engine.h) will 2917160814Ssimon give fewer recursive includes, which could break lazy source code - so 2918160814Ssimon this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always, 2919160814Ssimon developers should define this symbol when building and using openssl to 2920160814Ssimon ensure they track the recommended behaviour, interfaces, [etc], but 2921160814Ssimon backwards-compatible behaviour prevails when this isn't defined. 2922160814Ssimon [Geoff Thorpe] 2923160814Ssimon 2924160814Ssimon *) New function X509_POLICY_NODE_print() which prints out policy nodes. 2925160814Ssimon [Steve Henson] 2926160814Ssimon 2927160814Ssimon *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality. 2928160814Ssimon This will generate a random key of the appropriate length based on the 2929160814Ssimon cipher context. The EVP_CIPHER can provide its own random key generation 2930160814Ssimon routine to support keys of a specific form. This is used in the des and 2931160814Ssimon 3des routines to generate a key of the correct parity. Update S/MIME 2932160814Ssimon code to use new functions and hence generate correct parity DES keys. 2933160814Ssimon Add EVP_CHECK_DES_KEY #define to return an error if the key is not 2934160814Ssimon valid (weak or incorrect parity). 2935160814Ssimon [Steve Henson] 2936160814Ssimon 2937160814Ssimon *) Add a local set of CRLs that can be used by X509_verify_cert() as well 2938160814Ssimon as looking them up. This is useful when the verified structure may contain 2939160814Ssimon CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs 2940160814Ssimon present unless the new PKCS7_NO_CRL flag is asserted. 2941160814Ssimon [Steve Henson] 2942160814Ssimon 2943160814Ssimon *) Extend ASN1 oid configuration module. It now additionally accepts the 2944160814Ssimon syntax: 2945160814Ssimon 2946160814Ssimon shortName = some long name, 1.2.3.4 2947160814Ssimon [Steve Henson] 2948160814Ssimon 2949160814Ssimon *) Reimplemented the BN_CTX implementation. There is now no more static 2950160814Ssimon limitation on the number of variables it can handle nor the depth of the 2951160814Ssimon "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack 2952160814Ssimon information can now expand as required, and rather than having a single 2953160814Ssimon static array of bignums, BN_CTX now uses a linked-list of such arrays 2954160814Ssimon allowing it to expand on demand whilst maintaining the usefulness of 2955160814Ssimon BN_CTX's "bundling". 2956160814Ssimon [Geoff Thorpe] 2957160814Ssimon 2958160814Ssimon *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD 2959160814Ssimon to allow all RSA operations to function using a single BN_CTX. 2960160814Ssimon [Geoff Thorpe] 2961160814Ssimon 2962160814Ssimon *) Preliminary support for certificate policy evaluation and checking. This 2963160814Ssimon is initially intended to pass the tests outlined in "Conformance Testing 2964160814Ssimon of Relying Party Client Certificate Path Processing Logic" v1.07. 2965160814Ssimon [Steve Henson] 2966160814Ssimon 2967160814Ssimon *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and 2968160814Ssimon remained unused and not that useful. A variety of other little bignum 2969160814Ssimon tweaks and fixes have also been made continuing on from the audit (see 2970160814Ssimon below). 2971160814Ssimon [Geoff Thorpe] 2972160814Ssimon 2973160814Ssimon *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with 2974160814Ssimon associated ASN1, EVP and SSL functions and old ASN1 macros. 2975160814Ssimon [Richard Levitte] 2976160814Ssimon 2977160814Ssimon *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results, 2978160814Ssimon and this should never fail. So the return value from the use of 2979160814Ssimon BN_set_word() (which can fail due to needless expansion) is now deprecated; 2980160814Ssimon if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro. 2981160814Ssimon [Geoff Thorpe] 2982160814Ssimon 2983160814Ssimon *) BN_CTX_get() should return zero-valued bignums, providing the same 2984160814Ssimon initialised value as BN_new(). 2985160814Ssimon [Geoff Thorpe, suggested by Ulf M�ller] 2986160814Ssimon 2987160814Ssimon *) Support for inhibitAnyPolicy certificate extension. 2988160814Ssimon [Steve Henson] 2989160814Ssimon 2990160814Ssimon *) An audit of the BIGNUM code is underway, for which debugging code is 2991160814Ssimon enabled when BN_DEBUG is defined. This makes stricter enforcements on what 2992160814Ssimon is considered valid when processing BIGNUMs, and causes execution to 2993160814Ssimon assert() when a problem is discovered. If BN_DEBUG_RAND is defined, 2994160814Ssimon further steps are taken to deliberately pollute unused data in BIGNUM 2995160814Ssimon structures to try and expose faulty code further on. For now, openssl will 2996160814Ssimon (in its default mode of operation) continue to tolerate the inconsistent 2997160814Ssimon forms that it has tolerated in the past, but authors and packagers should 2998160814Ssimon consider trying openssl and their own applications when compiled with 2999160814Ssimon these debugging symbols defined. It will help highlight potential bugs in 3000160814Ssimon their own code, and will improve the test coverage for OpenSSL itself. At 3001160814Ssimon some point, these tighter rules will become openssl's default to improve 3002160814Ssimon maintainability, though the assert()s and other overheads will remain only 3003160814Ssimon in debugging configurations. See bn.h for more details. 3004160814Ssimon [Geoff Thorpe, Nils Larsch, Ulf M�ller] 3005160814Ssimon 3006160814Ssimon *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure 3007160814Ssimon that can only be obtained through BN_CTX_new() (which implicitly 3008160814Ssimon initialises it). The presence of this function only made it possible 3009160814Ssimon to overwrite an existing structure (and cause memory leaks). 3010160814Ssimon [Geoff Thorpe] 3011160814Ssimon 3012160814Ssimon *) Because of the callback-based approach for implementing LHASH as a 3013160814Ssimon template type, lh_insert() adds opaque objects to hash-tables and 3014160814Ssimon lh_doall() or lh_doall_arg() are typically used with a destructor callback 3015160814Ssimon to clean up those corresponding objects before destroying the hash table 3016160814Ssimon (and losing the object pointers). So some over-zealous constifications in 3017160814Ssimon LHASH have been relaxed so that lh_insert() does not take (nor store) the 3018160814Ssimon objects as "const" and the lh_doall[_arg] callback wrappers are not 3019160814Ssimon prototyped to have "const" restrictions on the object pointers they are 3020160814Ssimon given (and so aren't required to cast them away any more). 3021160814Ssimon [Geoff Thorpe] 3022160814Ssimon 3023160814Ssimon *) The tmdiff.h API was so ugly and minimal that our own timing utility 3024160814Ssimon (speed) prefers to use its own implementation. The two implementations 3025160814Ssimon haven't been consolidated as yet (volunteers?) but the tmdiff API has had 3026160814Ssimon its object type properly exposed (MS_TM) instead of casting to/from "char 3027160814Ssimon *". This may still change yet if someone realises MS_TM and "ms_time_***" 3028160814Ssimon aren't necessarily the greatest nomenclatures - but this is what was used 3029160814Ssimon internally to the implementation so I've used that for now. 3030160814Ssimon [Geoff Thorpe] 3031160814Ssimon 3032160814Ssimon *) Ensure that deprecated functions do not get compiled when 3033160814Ssimon OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of 3034160814Ssimon the self-tests were still using deprecated key-generation functions so 3035160814Ssimon these have been updated also. 3036160814Ssimon [Geoff Thorpe] 3037160814Ssimon 3038160814Ssimon *) Reorganise PKCS#7 code to separate the digest location functionality 3039160814Ssimon into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest(). 3040160814Ssimon New function PKCS7_set_digest() to set the digest type for PKCS#7 3041160814Ssimon digestedData type. Add additional code to correctly generate the 3042160814Ssimon digestedData type and add support for this type in PKCS7 initialization 3043160814Ssimon functions. 3044160814Ssimon [Steve Henson] 3045160814Ssimon 3046160814Ssimon *) New function PKCS7_set0_type_other() this initializes a PKCS7 3047160814Ssimon structure of type "other". 3048160814Ssimon [Steve Henson] 3049160814Ssimon 3050160814Ssimon *) Fix prime generation loop in crypto/bn/bn_prime.pl by making 3051160814Ssimon sure the loop does correctly stop and breaking ("division by zero") 3052160814Ssimon modulus operations are not performed. The (pre-generated) prime 3053160814Ssimon table crypto/bn/bn_prime.h was already correct, but it could not be 3054160814Ssimon re-generated on some platforms because of the "division by zero" 3055160814Ssimon situation in the script. 3056160814Ssimon [Ralf S. Engelschall] 3057160814Ssimon 3058160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 3059160814Ssimon draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with 3060160814Ssimon SHA-1 now is only used for "small" curves (where the 3061160814Ssimon representation of a field element takes up to 24 bytes); for 3062160814Ssimon larger curves, the field element resulting from ECDH is directly 3063160814Ssimon used as premaster secret. 3064160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 3065160814Ssimon 3066160814Ssimon *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2 3067160814Ssimon curve secp160r1 to the tests. 3068160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 3069160814Ssimon 3070160814Ssimon *) Add the possibility to load symbols globally with DSO. 3071160814Ssimon [G�tz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte] 3072160814Ssimon 3073160814Ssimon *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better 3074160814Ssimon control of the error stack. 3075160814Ssimon [Richard Levitte] 3076160814Ssimon 3077160814Ssimon *) Add support for STORE in ENGINE. 3078160814Ssimon [Richard Levitte] 3079160814Ssimon 3080160814Ssimon *) Add the STORE type. The intention is to provide a common interface 3081160814Ssimon to certificate and key stores, be they simple file-based stores, or 3082160814Ssimon HSM-type store, or LDAP stores, or... 3083160814Ssimon NOTE: The code is currently UNTESTED and isn't really used anywhere. 3084160814Ssimon [Richard Levitte] 3085160814Ssimon 3086160814Ssimon *) Add a generic structure called OPENSSL_ITEM. This can be used to 3087160814Ssimon pass a list of arguments to any function as well as provide a way 3088160814Ssimon for a function to pass data back to the caller. 3089160814Ssimon [Richard Levitte] 3090160814Ssimon 3091160814Ssimon *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup() 3092160814Ssimon works like BUF_strdup() but can be used to duplicate a portion of 3093160814Ssimon a string. The copy gets NUL-terminated. BUF_memdup() duplicates 3094160814Ssimon a memory area. 3095160814Ssimon [Richard Levitte] 3096160814Ssimon 3097160814Ssimon *) Add the function sk_find_ex() which works like sk_find(), but will 3098160814Ssimon return an index to an element even if an exact match couldn't be 3099160814Ssimon found. The index is guaranteed to point at the element where the 3100160814Ssimon searched-for key would be inserted to preserve sorting order. 3101160814Ssimon [Richard Levitte] 3102160814Ssimon 3103160814Ssimon *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but 3104160814Ssimon takes an extra flags argument for optional functionality. Currently, 3105160814Ssimon the following flags are defined: 3106160814Ssimon 3107160814Ssimon OBJ_BSEARCH_VALUE_ON_NOMATCH 3108160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 3109160814Ssimon element where the comparing function returns a negative or zero 3110160814Ssimon number. 3111160814Ssimon 3112160814Ssimon OBJ_BSEARCH_FIRST_VALUE_ON_MATCH 3113160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 3114160814Ssimon element where the comparing function returns zero. This is useful 3115160814Ssimon if there are more than one element where the comparing function 3116160814Ssimon returns zero. 3117160814Ssimon [Richard Levitte] 3118160814Ssimon 3119160814Ssimon *) Make it possible to create self-signed certificates with 'openssl ca' 3120160814Ssimon in such a way that the self-signed certificate becomes part of the 3121160814Ssimon CA database and uses the same mechanisms for serial number generation 3122160814Ssimon as all other certificate signing. The new flag '-selfsign' enables 3123160814Ssimon this functionality. Adapt CA.sh and CA.pl.in. 3124160814Ssimon [Richard Levitte] 3125160814Ssimon 3126160814Ssimon *) Add functionality to check the public key of a certificate request 3127160814Ssimon against a given private. This is useful to check that a certificate 3128160814Ssimon request can be signed by that key (self-signing). 3129160814Ssimon [Richard Levitte] 3130160814Ssimon 3131160814Ssimon *) Make it possible to have multiple active certificates with the same 3132160814Ssimon subject in the CA index file. This is done only if the keyword 3133160814Ssimon 'unique_subject' is set to 'no' in the main CA section (default 3134160814Ssimon if 'CA_default') of the configuration file. The value is saved 3135160814Ssimon with the database itself in a separate index attribute file, 3136160814Ssimon named like the index file with '.attr' appended to the name. 3137160814Ssimon [Richard Levitte] 3138160814Ssimon 3139160814Ssimon *) Generate muti valued AVAs using '+' notation in config files for 3140160814Ssimon req and dirName. 3141160814Ssimon [Steve Henson] 3142160814Ssimon 3143160814Ssimon *) Support for nameConstraints certificate extension. 3144160814Ssimon [Steve Henson] 3145160814Ssimon 3146160814Ssimon *) Support for policyConstraints certificate extension. 3147160814Ssimon [Steve Henson] 3148160814Ssimon 3149160814Ssimon *) Support for policyMappings certificate extension. 3150160814Ssimon [Steve Henson] 3151160814Ssimon 3152160814Ssimon *) Make sure the default DSA_METHOD implementation only uses its 3153160814Ssimon dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL, 3154160814Ssimon and change its own handlers to be NULL so as to remove unnecessary 3155160814Ssimon indirection. This lets alternative implementations fallback to the 3156160814Ssimon default implementation more easily. 3157160814Ssimon [Geoff Thorpe] 3158160814Ssimon 3159160814Ssimon *) Support for directoryName in GeneralName related extensions 3160160814Ssimon in config files. 3161160814Ssimon [Steve Henson] 3162160814Ssimon 3163160814Ssimon *) Make it possible to link applications using Makefile.shared. 3164160814Ssimon Make that possible even when linking against static libraries! 3165160814Ssimon [Richard Levitte] 3166160814Ssimon 3167160814Ssimon *) Support for single pass processing for S/MIME signing. This now 3168160814Ssimon means that S/MIME signing can be done from a pipe, in addition 3169160814Ssimon cleartext signing (multipart/signed type) is effectively streaming 3170160814Ssimon and the signed data does not need to be all held in memory. 3171160814Ssimon 3172160814Ssimon This is done with a new flag PKCS7_STREAM. When this flag is set 3173160814Ssimon PKCS7_sign() only initializes the PKCS7 structure and the actual signing 3174160814Ssimon is done after the data is output (and digests calculated) in 3175160814Ssimon SMIME_write_PKCS7(). 3176160814Ssimon [Steve Henson] 3177160814Ssimon 3178160814Ssimon *) Add full support for -rpath/-R, both in shared libraries and 3179160814Ssimon applications, at least on the platforms where it's known how 3180160814Ssimon to do it. 3181160814Ssimon [Richard Levitte] 3182160814Ssimon 3183160814Ssimon *) In crypto/ec/ec_mult.c, implement fast point multiplication with 3184160814Ssimon precomputation, based on wNAF splitting: EC_GROUP_precompute_mult() 3185160814Ssimon will now compute a table of multiples of the generator that 3186160814Ssimon makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul() 3187160814Ssimon faster (notably in the case of a single point multiplication, 3188160814Ssimon scalar * generator). 3189160814Ssimon [Nils Larsch, Bodo Moeller] 3190160814Ssimon 3191160814Ssimon *) IPv6 support for certificate extensions. The various extensions 3192160814Ssimon which use the IP:a.b.c.d can now take IPv6 addresses using the 3193160814Ssimon formats of RFC1884 2.2 . IPv6 addresses are now also displayed 3194160814Ssimon correctly. 3195160814Ssimon [Steve Henson] 3196160814Ssimon 3197160814Ssimon *) Added an ENGINE that implements RSA by performing private key 3198160814Ssimon exponentiations with the GMP library. The conversions to and from 3199160814Ssimon GMP's mpz_t format aren't optimised nor are any montgomery forms 3200160814Ssimon cached, and on x86 it appears OpenSSL's own performance has caught up. 3201160814Ssimon However there are likely to be other architectures where GMP could 3202160814Ssimon provide a boost. This ENGINE is not built in by default, but it can be 3203160814Ssimon specified at Configure time and should be accompanied by the necessary 3204160814Ssimon linker additions, eg; 3205160814Ssimon ./config -DOPENSSL_USE_GMP -lgmp 3206160814Ssimon [Geoff Thorpe] 3207160814Ssimon 3208160814Ssimon *) "openssl engine" will not display ENGINE/DSO load failure errors when 3209160814Ssimon testing availability of engines with "-t" - the old behaviour is 3210160814Ssimon produced by increasing the feature's verbosity with "-tt". 3211160814Ssimon [Geoff Thorpe] 3212160814Ssimon 3213160814Ssimon *) ECDSA routines: under certain error conditions uninitialized BN objects 3214160814Ssimon could be freed. Solution: make sure initialization is performed early 3215160814Ssimon enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de> 3216160814Ssimon via PR#459) 3217160814Ssimon [Lutz Jaenicke] 3218160814Ssimon 3219160814Ssimon *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD 3220160814Ssimon and DH_METHOD (eg. by ENGINE implementations) to override the normal 3221160814Ssimon software implementations. For DSA and DH, parameter generation can 3222160814Ssimon also be overriden by providing the appropriate method callbacks. 3223160814Ssimon [Geoff Thorpe] 3224160814Ssimon 3225160814Ssimon *) Change the "progress" mechanism used in key-generation and 3226160814Ssimon primality testing to functions that take a new BN_GENCB pointer in 3227160814Ssimon place of callback/argument pairs. The new API functions have "_ex" 3228160814Ssimon postfixes and the older functions are reimplemented as wrappers for 3229160814Ssimon the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide 3230160814Ssimon declarations of the old functions to help (graceful) attempts to 3231160814Ssimon migrate to the new functions. Also, the new key-generation API 3232160814Ssimon functions operate on a caller-supplied key-structure and return 3233160814Ssimon success/failure rather than returning a key or NULL - this is to 3234160814Ssimon help make "keygen" another member function of RSA_METHOD etc. 3235160814Ssimon 3236160814Ssimon Example for using the new callback interface: 3237160814Ssimon 3238160814Ssimon int (*my_callback)(int a, int b, BN_GENCB *cb) = ...; 3239160814Ssimon void *my_arg = ...; 3240160814Ssimon BN_GENCB my_cb; 3241160814Ssimon 3242160814Ssimon BN_GENCB_set(&my_cb, my_callback, my_arg); 3243160814Ssimon 3244160814Ssimon return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb); 3245160814Ssimon /* For the meaning of a, b in calls to my_callback(), see the 3246160814Ssimon * documentation of the function that calls the callback. 3247160814Ssimon * cb will point to my_cb; my_arg can be retrieved as cb->arg. 3248160814Ssimon * my_callback should return 1 if it wants BN_is_prime_ex() 3249160814Ssimon * to continue, or 0 to stop. 3250160814Ssimon */ 3251160814Ssimon 3252160814Ssimon [Geoff Thorpe] 3253160814Ssimon 3254160814Ssimon *) Change the ZLIB compression method to be stateful, and make it 3255160814Ssimon available to TLS with the number defined in 3256160814Ssimon draft-ietf-tls-compression-04.txt. 3257160814Ssimon [Richard Levitte] 3258160814Ssimon 3259160814Ssimon *) Add the ASN.1 structures and functions for CertificatePair, which 3260160814Ssimon is defined as follows (according to X.509_4thEditionDraftV6.pdf): 3261160814Ssimon 3262160814Ssimon CertificatePair ::= SEQUENCE { 3263160814Ssimon forward [0] Certificate OPTIONAL, 3264160814Ssimon reverse [1] Certificate OPTIONAL, 3265160814Ssimon -- at least one of the pair shall be present -- } 3266160814Ssimon 3267160814Ssimon Also implement the PEM functions to read and write certificate 3268160814Ssimon pairs, and defined the PEM tag as "CERTIFICATE PAIR". 3269160814Ssimon 3270160814Ssimon This needed to be defined, mostly for the sake of the LDAP 3271160814Ssimon attribute crossCertificatePair, but may prove useful elsewhere as 3272160814Ssimon well. 3273160814Ssimon [Richard Levitte] 3274160814Ssimon 3275160814Ssimon *) Make it possible to inhibit symlinking of shared libraries in 3276160814Ssimon Makefile.shared, for Cygwin's sake. 3277160814Ssimon [Richard Levitte] 3278160814Ssimon 3279160814Ssimon *) Extend the BIGNUM API by creating a function 3280160814Ssimon void BN_set_negative(BIGNUM *a, int neg); 3281160814Ssimon and a macro that behave like 3282160814Ssimon int BN_is_negative(const BIGNUM *a); 3283160814Ssimon 3284160814Ssimon to avoid the need to access 'a->neg' directly in applications. 3285160814Ssimon [Nils Larsch] 3286160814Ssimon 3287160814Ssimon *) Implement fast modular reduction for pseudo-Mersenne primes 3288160814Ssimon used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c). 3289160814Ssimon EC_GROUP_new_curve_GFp() will now automatically use this 3290160814Ssimon if applicable. 3291160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3292160814Ssimon 3293160814Ssimon *) Add new lock type (CRYPTO_LOCK_BN). 3294160814Ssimon [Bodo Moeller] 3295160814Ssimon 3296160814Ssimon *) Change the ENGINE framework to automatically load engines 3297160814Ssimon dynamically from specific directories unless they could be 3298160814Ssimon found to already be built in or loaded. Move all the 3299160814Ssimon current engines except for the cryptodev one to a new 3300160814Ssimon directory engines/. 3301160814Ssimon The engines in engines/ are built as shared libraries if 3302160814Ssimon the "shared" options was given to ./Configure or ./config. 3303160814Ssimon Otherwise, they are inserted in libcrypto.a. 3304160814Ssimon /usr/local/ssl/engines is the default directory for dynamic 3305160814Ssimon engines, but that can be overriden at configure time through 3306160814Ssimon the usual use of --prefix and/or --openssldir, and at run 3307160814Ssimon time with the environment variable OPENSSL_ENGINES. 3308160814Ssimon [Geoff Thorpe and Richard Levitte] 3309160814Ssimon 3310160814Ssimon *) Add Makefile.shared, a helper makefile to build shared 3311160814Ssimon libraries. Addapt Makefile.org. 3312160814Ssimon [Richard Levitte] 3313160814Ssimon 3314160814Ssimon *) Add version info to Win32 DLLs. 3315160814Ssimon [Peter 'Luna' Runestig" <peter@runestig.com>] 3316160814Ssimon 3317160814Ssimon *) Add new 'medium level' PKCS#12 API. Certificates and keys 3318160814Ssimon can be added using this API to created arbitrary PKCS#12 3319160814Ssimon files while avoiding the low level API. 3320160814Ssimon 3321160814Ssimon New options to PKCS12_create(), key or cert can be NULL and 3322160814Ssimon will then be omitted from the output file. The encryption 3323160814Ssimon algorithm NIDs can be set to -1 for no encryption, the mac 3324160814Ssimon iteration count can be set to 0 to omit the mac. 3325160814Ssimon 3326160814Ssimon Enhance pkcs12 utility by making the -nokeys and -nocerts 3327160814Ssimon options work when creating a PKCS#12 file. New option -nomac 3328160814Ssimon to omit the mac, NONE can be set for an encryption algorithm. 3329160814Ssimon New code is modified to use the enhanced PKCS12_create() 3330160814Ssimon instead of the low level API. 3331160814Ssimon [Steve Henson] 3332160814Ssimon 3333160814Ssimon *) Extend ASN1 encoder to support indefinite length constructed 3334160814Ssimon encoding. This can output sequences tags and octet strings in 3335160814Ssimon this form. Modify pk7_asn1.c to support indefinite length 3336160814Ssimon encoding. This is experimental and needs additional code to 3337160814Ssimon be useful, such as an ASN1 bio and some enhanced streaming 3338160814Ssimon PKCS#7 code. 3339160814Ssimon 3340160814Ssimon Extend template encode functionality so that tagging is passed 3341160814Ssimon down to the template encoder. 3342160814Ssimon [Steve Henson] 3343160814Ssimon 3344160814Ssimon *) Let 'openssl req' fail if an argument to '-newkey' is not 3345160814Ssimon recognized instead of using RSA as a default. 3346160814Ssimon [Bodo Moeller] 3347160814Ssimon 3348160814Ssimon *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt. 3349160814Ssimon As these are not official, they are not included in "ALL"; 3350160814Ssimon the "ECCdraft" ciphersuite group alias can be used to select them. 3351160814Ssimon [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)] 3352160814Ssimon 3353160814Ssimon *) Add ECDH engine support. 3354160814Ssimon [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)] 3355160814Ssimon 3356160814Ssimon *) Add ECDH in new directory crypto/ecdh/. 3357160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 3358160814Ssimon 3359160814Ssimon *) Let BN_rand_range() abort with an error after 100 iterations 3360160814Ssimon without success (which indicates a broken PRNG). 3361160814Ssimon [Bodo Moeller] 3362160814Ssimon 3363160814Ssimon *) Change BN_mod_sqrt() so that it verifies that the input value 3364160814Ssimon is really the square of the return value. (Previously, 3365160814Ssimon BN_mod_sqrt would show GIGO behaviour.) 3366160814Ssimon [Bodo Moeller] 3367160814Ssimon 3368160814Ssimon *) Add named elliptic curves over binary fields from X9.62, SECG, 3369160814Ssimon and WAP/WTLS; add OIDs that were still missing. 3370160814Ssimon 3371160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3372160814Ssimon (Sun Microsystems Laboratories)] 3373160814Ssimon 3374160814Ssimon *) Extend the EC library for elliptic curves over binary fields 3375160814Ssimon (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/). 3376160814Ssimon New EC_METHOD: 3377160814Ssimon 3378160814Ssimon EC_GF2m_simple_method 3379160814Ssimon 3380160814Ssimon New API functions: 3381160814Ssimon 3382160814Ssimon EC_GROUP_new_curve_GF2m 3383160814Ssimon EC_GROUP_set_curve_GF2m 3384160814Ssimon EC_GROUP_get_curve_GF2m 3385160814Ssimon EC_POINT_set_affine_coordinates_GF2m 3386160814Ssimon EC_POINT_get_affine_coordinates_GF2m 3387160814Ssimon EC_POINT_set_compressed_coordinates_GF2m 3388160814Ssimon 3389160814Ssimon Point compression for binary fields is disabled by default for 3390160814Ssimon patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to 3391160814Ssimon enable it). 3392160814Ssimon 3393160814Ssimon As binary polynomials are represented as BIGNUMs, various members 3394160814Ssimon of the EC_GROUP and EC_POINT data structures can be shared 3395160814Ssimon between the implementations for prime fields and binary fields; 3396160814Ssimon the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m) 3397160814Ssimon are essentially identical to their ..._GFp counterparts. 3398160814Ssimon (For simplicity, the '..._GFp' prefix has been dropped from 3399160814Ssimon various internal method names.) 3400160814Ssimon 3401160814Ssimon An internal 'field_div' method (similar to 'field_mul' and 3402160814Ssimon 'field_sqr') has been added; this is used only for binary fields. 3403160814Ssimon 3404160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3405160814Ssimon (Sun Microsystems Laboratories)] 3406160814Ssimon 3407160814Ssimon *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult() 3408160814Ssimon through methods ('mul', 'precompute_mult'). 3409160814Ssimon 3410160814Ssimon The generic implementations (now internally called 'ec_wNAF_mul' 3411160814Ssimon and 'ec_wNAF_precomputed_mult') remain the default if these 3412160814Ssimon methods are undefined. 3413160814Ssimon 3414160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3415160814Ssimon (Sun Microsystems Laboratories)] 3416160814Ssimon 3417160814Ssimon *) New function EC_GROUP_get_degree, which is defined through 3418160814Ssimon EC_METHOD. For curves over prime fields, this returns the bit 3419160814Ssimon length of the modulus. 3420160814Ssimon 3421160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3422160814Ssimon (Sun Microsystems Laboratories)] 3423160814Ssimon 3424160814Ssimon *) New functions EC_GROUP_dup, EC_POINT_dup. 3425160814Ssimon (These simply call ..._new and ..._copy). 3426160814Ssimon 3427160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3428160814Ssimon (Sun Microsystems Laboratories)] 3429160814Ssimon 3430160814Ssimon *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c. 3431160814Ssimon Polynomials are represented as BIGNUMs (where the sign bit is not 3432160814Ssimon used) in the following functions [macros]: 3433160814Ssimon 3434160814Ssimon BN_GF2m_add 3435160814Ssimon BN_GF2m_sub [= BN_GF2m_add] 3436160814Ssimon BN_GF2m_mod [wrapper for BN_GF2m_mod_arr] 3437160814Ssimon BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr] 3438160814Ssimon BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr] 3439160814Ssimon BN_GF2m_mod_inv 3440160814Ssimon BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr] 3441160814Ssimon BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr] 3442160814Ssimon BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr] 3443160814Ssimon BN_GF2m_cmp [= BN_ucmp] 3444160814Ssimon 3445160814Ssimon (Note that only the 'mod' functions are actually for fields GF(2^m). 3446160814Ssimon BN_GF2m_add() is misnomer, but this is for the sake of consistency.) 3447160814Ssimon 3448160814Ssimon For some functions, an the irreducible polynomial defining a 3449160814Ssimon field can be given as an 'unsigned int[]' with strictly 3450160814Ssimon decreasing elements giving the indices of those bits that are set; 3451160814Ssimon i.e., p[] represents the polynomial 3452160814Ssimon f(t) = t^p[0] + t^p[1] + ... + t^p[k] 3453160814Ssimon where 3454160814Ssimon p[0] > p[1] > ... > p[k] = 0. 3455160814Ssimon This applies to the following functions: 3456160814Ssimon 3457160814Ssimon BN_GF2m_mod_arr 3458160814Ssimon BN_GF2m_mod_mul_arr 3459160814Ssimon BN_GF2m_mod_sqr_arr 3460160814Ssimon BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv] 3461160814Ssimon BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div] 3462160814Ssimon BN_GF2m_mod_exp_arr 3463160814Ssimon BN_GF2m_mod_sqrt_arr 3464160814Ssimon BN_GF2m_mod_solve_quad_arr 3465160814Ssimon BN_GF2m_poly2arr 3466160814Ssimon BN_GF2m_arr2poly 3467160814Ssimon 3468160814Ssimon Conversion can be performed by the following functions: 3469160814Ssimon 3470160814Ssimon BN_GF2m_poly2arr 3471160814Ssimon BN_GF2m_arr2poly 3472160814Ssimon 3473160814Ssimon bntest.c has additional tests for binary polynomial arithmetic. 3474160814Ssimon 3475160814Ssimon Two implementations for BN_GF2m_mod_div() are available. 3476160814Ssimon The default algorithm simply uses BN_GF2m_mod_inv() and 3477160814Ssimon BN_GF2m_mod_mul(). The alternative algorithm is compiled in only 3478160814Ssimon if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the 3479160814Ssimon copyright notice in crypto/bn/bn_gf2m.c before enabling it). 3480160814Ssimon 3481160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3482160814Ssimon (Sun Microsystems Laboratories)] 3483160814Ssimon 3484160814Ssimon *) Add new error code 'ERR_R_DISABLED' that can be used when some 3485160814Ssimon functionality is disabled at compile-time. 3486160814Ssimon [Douglas Stebila <douglas.stebila@sun.com>] 3487160814Ssimon 3488160814Ssimon *) Change default behaviour of 'openssl asn1parse' so that more 3489160814Ssimon information is visible when viewing, e.g., a certificate: 3490160814Ssimon 3491160814Ssimon Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump' 3492160814Ssimon mode the content of non-printable OCTET STRINGs is output in a 3493160814Ssimon style similar to INTEGERs, but with '[HEX DUMP]' prepended to 3494160814Ssimon avoid the appearance of a printable string. 3495160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3496160814Ssimon 3497160814Ssimon *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access 3498160814Ssimon functions 3499160814Ssimon EC_GROUP_set_asn1_flag() 3500160814Ssimon EC_GROUP_get_asn1_flag() 3501160814Ssimon EC_GROUP_set_point_conversion_form() 3502160814Ssimon EC_GROUP_get_point_conversion_form() 3503160814Ssimon These control ASN1 encoding details: 3504160814Ssimon - Curves (i.e., groups) are encoded explicitly unless asn1_flag 3505160814Ssimon has been set to OPENSSL_EC_NAMED_CURVE. 3506160814Ssimon - Points are encoded in uncompressed form by default; options for 3507160814Ssimon asn1_for are as for point2oct, namely 3508160814Ssimon POINT_CONVERSION_COMPRESSED 3509160814Ssimon POINT_CONVERSION_UNCOMPRESSED 3510160814Ssimon POINT_CONVERSION_HYBRID 3511160814Ssimon 3512160814Ssimon Also add 'seed' and 'seed_len' members to EC_GROUP with access 3513160814Ssimon functions 3514160814Ssimon EC_GROUP_set_seed() 3515160814Ssimon EC_GROUP_get0_seed() 3516160814Ssimon EC_GROUP_get_seed_len() 3517160814Ssimon This is used only for ASN1 purposes (so far). 3518160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3519160814Ssimon 3520160814Ssimon *) Add 'field_type' member to EC_METHOD, which holds the NID 3521160814Ssimon of the appropriate field type OID. The new function 3522160814Ssimon EC_METHOD_get_field_type() returns this value. 3523160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3524160814Ssimon 3525160814Ssimon *) Add functions 3526160814Ssimon EC_POINT_point2bn() 3527160814Ssimon EC_POINT_bn2point() 3528160814Ssimon EC_POINT_point2hex() 3529160814Ssimon EC_POINT_hex2point() 3530160814Ssimon providing useful interfaces to EC_POINT_point2oct() and 3531160814Ssimon EC_POINT_oct2point(). 3532160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3533160814Ssimon 3534160814Ssimon *) Change internals of the EC library so that the functions 3535160814Ssimon EC_GROUP_set_generator() 3536160814Ssimon EC_GROUP_get_generator() 3537160814Ssimon EC_GROUP_get_order() 3538160814Ssimon EC_GROUP_get_cofactor() 3539160814Ssimon are implemented directly in crypto/ec/ec_lib.c and not dispatched 3540160814Ssimon to methods, which would lead to unnecessary code duplication when 3541160814Ssimon adding different types of curves. 3542160814Ssimon [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller] 3543160814Ssimon 3544160814Ssimon *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM 3545160814Ssimon arithmetic, and such that modified wNAFs are generated 3546160814Ssimon (which avoid length expansion in many cases). 3547160814Ssimon [Bodo Moeller] 3548160814Ssimon 3549160814Ssimon *) Add a function EC_GROUP_check_discriminant() (defined via 3550160814Ssimon EC_METHOD) that verifies that the curve discriminant is non-zero. 3551160814Ssimon 3552160814Ssimon Add a function EC_GROUP_check() that makes some sanity tests 3553160814Ssimon on a EC_GROUP, its generator and order. This includes 3554160814Ssimon EC_GROUP_check_discriminant(). 3555160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3556160814Ssimon 3557160814Ssimon *) Add ECDSA in new directory crypto/ecdsa/. 3558160814Ssimon 3559160814Ssimon Add applications 'openssl ecparam' and 'openssl ecdsa' 3560160814Ssimon (these are based on 'openssl dsaparam' and 'openssl dsa'). 3561160814Ssimon 3562160814Ssimon ECDSA support is also included in various other files across the 3563160814Ssimon library. Most notably, 3564160814Ssimon - 'openssl req' now has a '-newkey ecdsa:file' option; 3565160814Ssimon - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA; 3566160814Ssimon - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and 3567160814Ssimon d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make 3568160814Ssimon them suitable for ECDSA where domain parameters must be 3569160814Ssimon extracted before the specific public key; 3570160814Ssimon - ECDSA engine support has been added. 3571160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3572160814Ssimon 3573160814Ssimon *) Include some named elliptic curves, and add OIDs from X9.62, 3574160814Ssimon SECG, and WAP/WTLS. Each curve can be obtained from the new 3575160814Ssimon function 3576160814Ssimon EC_GROUP_new_by_curve_name(), 3577160814Ssimon and the list of available named curves can be obtained with 3578160814Ssimon EC_get_builtin_curves(). 3579160814Ssimon Also add a 'curve_name' member to EC_GROUP objects, which can be 3580160814Ssimon accessed via 3581160814Ssimon EC_GROUP_set_curve_name() 3582160814Ssimon EC_GROUP_get_curve_name() 3583160814Ssimon [Nils Larsch <larsch@trustcenter.de, Bodo Moeller] 3584160814Ssimon 3585160814Ssimon *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 3586160814Ssimon was actually never needed) and in BN_mul(). The removal in BN_mul() 3587160814Ssimon required a small change in bn_mul_part_recursive() and the addition 3588160814Ssimon of the functions bn_cmp_part_words(), bn_sub_part_words() and 3589160814Ssimon bn_add_part_words(), which do the same thing as bn_cmp_words(), 3590160814Ssimon bn_sub_words() and bn_add_words() except they take arrays with 3591160814Ssimon differing sizes. 3592160814Ssimon [Richard Levitte] 3593160814Ssimon 3594194206Ssimon Changes between 0.9.7l and 0.9.7m [23 Feb 2007] 3595194206Ssimon 3596167612Ssimon *) Cleanse PEM buffers before freeing them since they may contain 3597167612Ssimon sensitive data. 3598167612Ssimon [Benjamin Bennett <ben@psc.edu>] 3599167612Ssimon 3600167612Ssimon *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 3601167612Ssimon a ciphersuite string such as "DEFAULT:RSA" cannot enable 3602167612Ssimon authentication-only ciphersuites. 3603167612Ssimon [Bodo Moeller] 3604167612Ssimon 3605167612Ssimon *) Since AES128 and AES256 share a single mask bit in the logic of 3606167612Ssimon ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 3607167612Ssimon kludge to work properly if AES128 is available and AES256 isn't. 3608167612Ssimon [Victor Duchovni] 3609167612Ssimon 3610194206Ssimon *) Expand security boundary to match 1.1.1 module. 3611194206Ssimon [Steve Henson] 3612194206Ssimon 3613194206Ssimon *) Remove redundant features: hash file source, editing of test vectors 3614194206Ssimon modify fipsld to use external fips_premain.c signature. 3615194206Ssimon [Steve Henson] 3616194206Ssimon 3617194206Ssimon *) New perl script mkfipsscr.pl to create shell scripts or batch files to 3618194206Ssimon run algorithm test programs. 3619194206Ssimon [Steve Henson] 3620194206Ssimon 3621194206Ssimon *) Make algorithm test programs more tolerant of whitespace. 3622194206Ssimon [Steve Henson] 3623194206Ssimon 3624167612Ssimon *) Have SSL/TLS server implementation tolerate "mismatched" record 3625167612Ssimon protocol version while receiving ClientHello even if the 3626167612Ssimon ClientHello is fragmented. (The server can't insist on the 3627167612Ssimon particular protocol version it has chosen before the ServerHello 3628167612Ssimon message has informed the client about his choice.) 3629167612Ssimon [Bodo Moeller] 3630167612Ssimon 3631167612Ssimon *) Load error codes if they are not already present instead of using a 3632167612Ssimon static variable. This allows them to be cleanly unloaded and reloaded. 3633167612Ssimon [Steve Henson] 3634167612Ssimon 3635167612Ssimon Changes between 0.9.7k and 0.9.7l [28 Sep 2006] 3636167612Ssimon 3637167612Ssimon *) Introduce limits to prevent malicious keys being able to 3638167612Ssimon cause a denial of service. (CVE-2006-2940) 3639167612Ssimon [Steve Henson, Bodo Moeller] 3640167612Ssimon 3641167612Ssimon *) Fix ASN.1 parsing of certain invalid structures that can result 3642167612Ssimon in a denial of service. (CVE-2006-2937) [Steve Henson] 3643167612Ssimon 3644167612Ssimon *) Fix buffer overflow in SSL_get_shared_ciphers() function. 3645167612Ssimon (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 3646167612Ssimon 3647167612Ssimon *) Fix SSL client code which could crash if connecting to a 3648167612Ssimon malicious SSLv2 server. (CVE-2006-4343) 3649167612Ssimon [Tavis Ormandy and Will Drewry, Google Security Team] 3650167612Ssimon 3651162911Ssimon *) Change ciphersuite string processing so that an explicit 3652162911Ssimon ciphersuite selects this one ciphersuite (so that "AES256-SHA" 3653162911Ssimon will no longer include "AES128-SHA"), and any other similar 3654162911Ssimon ciphersuite (same bitmap) from *other* protocol versions (so that 3655162911Ssimon "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the 3656162911Ssimon SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining 3657162911Ssimon changes from 0.9.8b and 0.9.8d. 3658162911Ssimon [Bodo Moeller] 3659162911Ssimon 3660162911Ssimon Changes between 0.9.7j and 0.9.7k [05 Sep 2006] 3661162911Ssimon 3662162911Ssimon *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 3663162911Ssimon (CVE-2006-4339) [Ben Laurie and Google Security Team] 3664162911Ssimon 3665162911Ssimon *) Change the Unix randomness entropy gathering to use poll() when 3666162911Ssimon possible instead of select(), since the latter has some 3667162911Ssimon undesirable limitations. 3668162911Ssimon [Darryl Miles via Richard Levitte and Bodo Moeller] 3669162911Ssimon 3670162911Ssimon *) Disable rogue ciphersuites: 3671162911Ssimon 3672162911Ssimon - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 3673162911Ssimon - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 3674162911Ssimon - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 3675162911Ssimon 3676162911Ssimon The latter two were purportedly from 3677162911Ssimon draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 3678162911Ssimon appear there. 3679162911Ssimon 3680162911Ssimon Also deactive the remaining ciphersuites from 3681162911Ssimon draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 3682162911Ssimon unofficial, and the ID has long expired. 3683162911Ssimon [Bodo Moeller] 3684162911Ssimon 3685162911Ssimon *) Fix RSA blinding Heisenbug (problems sometimes occured on 3686162911Ssimon dual-core machines) and other potential thread-safety issues. 3687162911Ssimon [Bodo Moeller] 3688162911Ssimon 3689162911Ssimon Changes between 0.9.7i and 0.9.7j [04 May 2006] 3690162911Ssimon 3691162911Ssimon *) Adapt fipsld and the build system to link against the validated FIPS 3692162911Ssimon module in FIPS mode. 3693162911Ssimon [Steve Henson] 3694162911Ssimon 3695162911Ssimon *) Fixes for VC++ 2005 build under Windows. 3696162911Ssimon [Steve Henson] 3697162911Ssimon 3698162911Ssimon *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make 3699162911Ssimon from a Windows bash shell such as MSYS. It is autodetected from the 3700162911Ssimon "config" script when run from a VC++ environment. Modify standard VC++ 3701162911Ssimon build to use fipscanister.o from the GNU make build. 3702162911Ssimon [Steve Henson] 3703162911Ssimon 3704160814Ssimon Changes between 0.9.7h and 0.9.7i [14 Oct 2005] 3705160814Ssimon 3706160814Ssimon *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS. 3707160814Ssimon The value now differs depending on if you build for FIPS or not. 3708160814Ssimon BEWARE! A program linked with a shared FIPSed libcrypto can't be 3709160814Ssimon safely run with a non-FIPSed libcrypto, as it may crash because of 3710160814Ssimon the difference induced by this change. 3711160814Ssimon [Andy Polyakov] 3712160814Ssimon 3713160814Ssimon Changes between 0.9.7g and 0.9.7h [11 Oct 2005] 3714160814Ssimon 3715160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 3716160814Ssimon (part of SSL_OP_ALL). This option used to disable the 3717160814Ssimon countermeasure against man-in-the-middle protocol-version 3718160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 3719160814Ssimon idea. (CVE-2005-2969) 3720160814Ssimon 3721160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 3722160814Ssimon for Information Security, National Institute of Advanced Industrial 3723160814Ssimon Science and Technology [AIST], Japan)] 3724160814Ssimon 3725160814Ssimon *) Minimal support for X9.31 signatures and PSS padding modes. This is 3726160814Ssimon mainly for FIPS compliance and not fully integrated at this stage. 3727160814Ssimon [Steve Henson] 3728160814Ssimon 3729160814Ssimon *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform 3730160814Ssimon the exponentiation using a fixed-length exponent. (Otherwise, 3731160814Ssimon the information leaked through timing could expose the secret key 3732160814Ssimon after many signatures; cf. Bleichenbacher's attack on DSA with 3733160814Ssimon biased k.) 3734160814Ssimon [Bodo Moeller] 3735160814Ssimon 3736160814Ssimon *) Make a new fixed-window mod_exp implementation the default for 3737160814Ssimon RSA, DSA, and DH private-key operations so that the sequence of 3738160814Ssimon squares and multiplies and the memory access pattern are 3739160814Ssimon independent of the particular secret key. This will mitigate 3740160814Ssimon cache-timing and potential related attacks. 3741160814Ssimon 3742160814Ssimon BN_mod_exp_mont_consttime() is the new exponentiation implementation, 3743160814Ssimon and this is automatically used by BN_mod_exp_mont() if the new flag 3744160814Ssimon BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH 3745160814Ssimon will use this BN flag for private exponents unless the flag 3746160814Ssimon RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or 3747160814Ssimon DH_FLAG_NO_EXP_CONSTTIME, respectively, is set. 3748160814Ssimon 3749160814Ssimon [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller] 3750160814Ssimon 3751160814Ssimon *) Change the client implementation for SSLv23_method() and 3752160814Ssimon SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0 3753160814Ssimon Client Hello message format if the SSL_OP_NO_SSLv2 option is set. 3754160814Ssimon (Previously, the SSL 2.0 backwards compatible Client Hello 3755160814Ssimon message format would be used even with SSL_OP_NO_SSLv2.) 3756160814Ssimon [Bodo Moeller] 3757160814Ssimon 3758160814Ssimon *) Add support for smime-type MIME parameter in S/MIME messages which some 3759160814Ssimon clients need. 3760160814Ssimon [Steve Henson] 3761160814Ssimon 3762160814Ssimon *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in 3763160814Ssimon a threadsafe manner. Modify rsa code to use new function and add calls 3764160814Ssimon to dsa and dh code (which had race conditions before). 3765160814Ssimon [Steve Henson] 3766160814Ssimon 3767160814Ssimon *) Include the fixed error library code in the C error file definitions 3768160814Ssimon instead of fixing them up at runtime. This keeps the error code 3769160814Ssimon structures constant. 3770160814Ssimon [Steve Henson] 3771160814Ssimon 3772160814Ssimon Changes between 0.9.7f and 0.9.7g [11 Apr 2005] 3773160814Ssimon 3774160814Ssimon [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after 3775160814Ssimon OpenSSL 0.9.8.] 3776160814Ssimon 3777160814Ssimon *) Fixes for newer kerberos headers. NB: the casts are needed because 3778160814Ssimon the 'length' field is signed on one version and unsigned on another 3779160814Ssimon with no (?) obvious way to tell the difference, without these VC++ 3780160814Ssimon complains. Also the "definition" of FAR (blank) is no longer included 3781160814Ssimon nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up 3782160814Ssimon some needed definitions. 3783160814Ssimon [Steve Henson] 3784160814Ssimon 3785160814Ssimon *) Undo Cygwin change. 3786160814Ssimon [Ulf M�ller] 3787160814Ssimon 3788160814Ssimon *) Added support for proxy certificates according to RFC 3820. 3789160814Ssimon Because they may be a security thread to unaware applications, 3790160814Ssimon they must be explicitely allowed in run-time. See 3791160814Ssimon docs/HOWTO/proxy_certificates.txt for further information. 3792160814Ssimon [Richard Levitte] 3793160814Ssimon 3794160814Ssimon Changes between 0.9.7e and 0.9.7f [22 Mar 2005] 3795160814Ssimon 3796160814Ssimon *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating 3797160814Ssimon server and client random values. Previously 3798160814Ssimon (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in 3799160814Ssimon less random data when sizeof(time_t) > 4 (some 64 bit platforms). 3800160814Ssimon 3801160814Ssimon This change has negligible security impact because: 3802160814Ssimon 3803160814Ssimon 1. Server and client random values still have 24 bytes of pseudo random 3804160814Ssimon data. 3805160814Ssimon 3806160814Ssimon 2. Server and client random values are sent in the clear in the initial 3807160814Ssimon handshake. 3808160814Ssimon 3809160814Ssimon 3. The master secret is derived using the premaster secret (48 bytes in 3810160814Ssimon size for static RSA ciphersuites) as well as client server and random 3811160814Ssimon values. 3812160814Ssimon 3813160814Ssimon The OpenSSL team would like to thank the UK NISCC for bringing this issue 3814160814Ssimon to our attention. 3815160814Ssimon 3816160814Ssimon [Stephen Henson, reported by UK NISCC] 3817160814Ssimon 3818160814Ssimon *) Use Windows randomness collection on Cygwin. 3819160814Ssimon [Ulf M�ller] 3820160814Ssimon 3821160814Ssimon *) Fix hang in EGD/PRNGD query when communication socket is closed 3822160814Ssimon prematurely by EGD/PRNGD. 3823160814Ssimon [Darren Tucker <dtucker@zip.com.au> via Lutz J�nicke, resolves #1014] 3824160814Ssimon 3825160814Ssimon *) Prompt for pass phrases when appropriate for PKCS12 input format. 3826160814Ssimon [Steve Henson] 3827160814Ssimon 3828160814Ssimon *) Back-port of selected performance improvements from development 3829160814Ssimon branch, as well as improved support for PowerPC platforms. 3830160814Ssimon [Andy Polyakov] 3831160814Ssimon 3832160814Ssimon *) Add lots of checks for memory allocation failure, error codes to indicate 3833160814Ssimon failure and freeing up memory if a failure occurs. 3834160814Ssimon [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson] 3835160814Ssimon 3836160814Ssimon *) Add new -passin argument to dgst. 3837160814Ssimon [Steve Henson] 3838160814Ssimon 3839160814Ssimon *) Perform some character comparisons of different types in X509_NAME_cmp: 3840160814Ssimon this is needed for some certificates that reencode DNs into UTF8Strings 3841160814Ssimon (in violation of RFC3280) and can't or wont issue name rollover 3842160814Ssimon certificates. 3843160814Ssimon [Steve Henson] 3844160814Ssimon 3845160814Ssimon *) Make an explicit check during certificate validation to see that 3846160814Ssimon the CA setting in each certificate on the chain is correct. As a 3847160814Ssimon side effect always do the following basic checks on extensions, 3848160814Ssimon not just when there's an associated purpose to the check: 3849160814Ssimon 3850160814Ssimon - if there is an unhandled critical extension (unless the user 3851160814Ssimon has chosen to ignore this fault) 3852160814Ssimon - if the path length has been exceeded (if one is set at all) 3853160814Ssimon - that certain extensions fit the associated purpose (if one has 3854160814Ssimon been given) 3855160814Ssimon [Richard Levitte] 3856160814Ssimon 3857142425Snectar Changes between 0.9.7d and 0.9.7e [25 Oct 2004] 3858142425Snectar 3859142425Snectar *) Avoid a race condition when CRLs are checked in a multi threaded 3860142425Snectar environment. This would happen due to the reordering of the revoked 3861142425Snectar entries during signature checking and serial number lookup. Now the 3862142425Snectar encoding is cached and the serial number sort performed under a lock. 3863142425Snectar Add new STACK function sk_is_sorted(). 3864142425Snectar [Steve Henson] 3865142425Snectar 3866142425Snectar *) Add Delta CRL to the extension code. 3867142425Snectar [Steve Henson] 3868142425Snectar 3869142425Snectar *) Various fixes to s3_pkt.c so alerts are sent properly. 3870142425Snectar [David Holmes <d.holmes@f5.com>] 3871142425Snectar 3872142425Snectar *) Reduce the chances of duplicate issuer name and serial numbers (in 3873142425Snectar violation of RFC3280) using the OpenSSL certificate creation utilities. 3874142425Snectar This is done by creating a random 64 bit value for the initial serial 3875142425Snectar number when a serial number file is created or when a self signed 3876142425Snectar certificate is created using 'openssl req -x509'. The initial serial 3877142425Snectar number file is created using 'openssl x509 -next_serial' in CA.pl 3878142425Snectar rather than being initialized to 1. 3879142425Snectar [Steve Henson] 3880142425Snectar 3881127128Snectar Changes between 0.9.7c and 0.9.7d [17 Mar 2004] 3882127128Snectar 3883127128Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 3884160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 3885127128Snectar [Joe Orton, Steve Henson] 3886127128Snectar 3887127128Snectar *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites 3888160814Ssimon (CVE-2004-0112) 3889127128Snectar [Joe Orton, Steve Henson] 3890127128Snectar 3891127128Snectar *) Make it possible to have multiple active certificates with the same 3892127128Snectar subject in the CA index file. This is done only if the keyword 3893127128Snectar 'unique_subject' is set to 'no' in the main CA section (default 3894127128Snectar if 'CA_default') of the configuration file. The value is saved 3895127128Snectar with the database itself in a separate index attribute file, 3896127128Snectar named like the index file with '.attr' appended to the name. 3897127128Snectar [Richard Levitte] 3898127128Snectar 3899127128Snectar *) X509 verify fixes. Disable broken certificate workarounds when 3900127128Snectar X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if 3901127128Snectar keyUsage extension present. Don't accept CRLs with unhandled critical 3902127128Snectar extensions: since verify currently doesn't process CRL extensions this 3903127128Snectar rejects a CRL with *any* critical extensions. Add new verify error codes 3904127128Snectar for these cases. 3905127128Snectar [Steve Henson] 3906127128Snectar 3907127128Snectar *) When creating an OCSP nonce use an OCTET STRING inside the extnValue. 3908127128Snectar A clarification of RFC2560 will require the use of OCTET STRINGs and 3909127128Snectar some implementations cannot handle the current raw format. Since OpenSSL 3910127128Snectar copies and compares OCSP nonces as opaque blobs without any attempt at 3911127128Snectar parsing them this should not create any compatibility issues. 3912127128Snectar [Steve Henson] 3913127128Snectar 3914127128Snectar *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when 3915127128Snectar calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without 3916127128Snectar this HMAC (and other) operations are several times slower than OpenSSL 3917127128Snectar < 0.9.7. 3918127128Snectar [Steve Henson] 3919127128Snectar 3920127128Snectar *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex(). 3921127128Snectar [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 3922127128Snectar 3923127128Snectar *) Use the correct content when signing type "other". 3924127128Snectar [Steve Henson] 3925127128Snectar 3926120631Snectar Changes between 0.9.7b and 0.9.7c [30 Sep 2003] 3927120631Snectar 3928120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 3929120631Snectar 3930120631Snectar Stop out of bounds reads in the ASN1 code when presented with 3931160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 3932120631Snectar 3933160814Ssimon Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545). 3934120631Snectar 3935120631Snectar If verify callback ignores invalid public key errors don't try to check 3936120631Snectar certificate signature with the NULL public key. 3937120631Snectar 3938120631Snectar [Steve Henson] 3939120631Snectar 3940120631Snectar *) New -ignore_err option in ocsp application to stop the server 3941120631Snectar exiting on the first error in a request. 3942120631Snectar [Steve Henson] 3943120631Snectar 3944120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 3945120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 3946120631Snectar specifications. 3947120631Snectar [Steve Henson] 3948120631Snectar 3949120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 3950120631Snectar extra data after the compression methods not only for TLS 1.0 3951120631Snectar but also for SSL 3.0 (as required by the specification). 3952120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 3953120631Snectar 3954120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 3955120631Snectar when it's 512 *bits* long, not 512 bytes. 3956120631Snectar [Richard Levitte] 3957120631Snectar 3958120631Snectar *) Change AES_cbc_encrypt() so it outputs exact multiple of 3959120631Snectar blocks during encryption. 3960120631Snectar [Richard Levitte] 3961120631Snectar 3962120631Snectar *) Various fixes to base64 BIO and non blocking I/O. On write 3963120631Snectar flushes were not handled properly if the BIO retried. On read 3964120631Snectar data was not being buffered properly and had various logic bugs. 3965120631Snectar This also affects blocking I/O when the data being decoded is a 3966120631Snectar certain size. 3967120631Snectar [Steve Henson] 3968120631Snectar 3969120631Snectar *) Various S/MIME bugfixes and compatibility changes: 3970120631Snectar output correct application/pkcs7 MIME type if 3971120631Snectar PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures. 3972120631Snectar Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening 3973120631Snectar of files as .eml work). Correctly handle very long lines in MIME 3974120631Snectar parser. 3975120631Snectar [Steve Henson] 3976120631Snectar 3977120631Snectar Changes between 0.9.7a and 0.9.7b [10 Apr 2003] 3978120631Snectar 3979120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 3980120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 3981120631Snectar a protocol version number mismatch like a decryption error 3982120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 3983120631Snectar [Bodo Moeller] 3984120631Snectar 3985120631Snectar *) Turn on RSA blinding by default in the default implementation 3986120631Snectar to avoid a timing attack. Applications that don't want it can call 3987120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 3988120631Snectar They would be ill-advised to do so in most cases. 3989120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 3990120631Snectar 3991120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 3992120631Snectar seeded (in this case, the secret RSA exponent is abused as 3993120631Snectar an unpredictable seed -- if it is not unpredictable, there 3994120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 3995120631Snectar by remembering the creator's thread ID in rsa->blinding and 3996120631Snectar having all other threads use local one-time blinding factors 3997120631Snectar (this requires more computation than sharing rsa->blinding, but 3998120631Snectar avoids excessive locking; and if an RSA object is not shared 3999120631Snectar between threads, blinding will still be very fast). 4000120631Snectar [Bodo Moeller] 4001120631Snectar 4002120631Snectar *) Fixed a typo bug that would cause ENGINE_set_default() to set an 4003120631Snectar ENGINE as defaults for all supported algorithms irrespective of 4004120631Snectar the 'flags' parameter. 'flags' is now honoured, so applications 4005120631Snectar should make sure they are passing it correctly. 4006120631Snectar [Geoff Thorpe] 4007120631Snectar 4008120631Snectar *) Target "mingw" now allows native Windows code to be generated in 4009120631Snectar the Cygwin environment as well as with the MinGW compiler. 4010120631Snectar [Ulf Moeller] 4011120631Snectar 4012111147Snectar Changes between 0.9.7 and 0.9.7a [19 Feb 2003] 4013111147Snectar 4014111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 4015111147Snectar via timing by performing a MAC computation even if incorrrect 4016111147Snectar block cipher padding has been found. This is a countermeasure 4017111147Snectar against active attacks where the attacker has to distinguish 4018160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 4019111147Snectar 4020111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 4021111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 4022111147Snectar Martin Vuagnoux (EPFL, Ilion)] 4023111147Snectar 4024111147Snectar *) Make the no-err option work as intended. The intention with no-err 4025111147Snectar is not to have the whole error stack handling routines removed from 4026111147Snectar libcrypto, it's only intended to remove all the function name and 4027111147Snectar reason texts, thereby removing some of the footprint that may not 4028111147Snectar be interesting if those errors aren't displayed anyway. 4029111147Snectar 4030111147Snectar NOTE: it's still possible for any application or module to have it's 4031111147Snectar own set of error texts inserted. The routines are there, just not 4032111147Snectar used by default when no-err is given. 4033111147Snectar [Richard Levitte] 4034111147Snectar 4035111147Snectar *) Add support for FreeBSD on IA64. 4036111147Snectar [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454] 4037111147Snectar 4038111147Snectar *) Adjust DES_cbc_cksum() so it returns the same value as the MIT 4039111147Snectar Kerberos function mit_des_cbc_cksum(). Before this change, 4040111147Snectar the value returned by DES_cbc_cksum() was like the one from 4041111147Snectar mit_des_cbc_cksum(), except the bytes were swapped. 4042111147Snectar [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte] 4043111147Snectar 4044111147Snectar *) Allow an application to disable the automatic SSL chain building. 4045111147Snectar Before this a rather primitive chain build was always performed in 4046111147Snectar ssl3_output_cert_chain(): an application had no way to send the 4047111147Snectar correct chain if the automatic operation produced an incorrect result. 4048111147Snectar 4049111147Snectar Now the chain builder is disabled if either: 4050111147Snectar 4051111147Snectar 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert(). 4052111147Snectar 4053111147Snectar 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set. 4054111147Snectar 4055111147Snectar The reasoning behind this is that an application would not want the 4056111147Snectar auto chain building to take place if extra chain certificates are 4057111147Snectar present and it might also want a means of sending no additional 4058111147Snectar certificates (for example the chain has two certificates and the 4059111147Snectar root is omitted). 4060111147Snectar [Steve Henson] 4061111147Snectar 4062111147Snectar *) Add the possibility to build without the ENGINE framework. 4063111147Snectar [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 4064111147Snectar 4065111147Snectar *) Under Win32 gmtime() can return NULL: check return value in 4066111147Snectar OPENSSL_gmtime(). Add error code for case where gmtime() fails. 4067111147Snectar [Steve Henson] 4068111147Snectar 4069111147Snectar *) DSA routines: under certain error conditions uninitialized BN objects 4070111147Snectar could be freed. Solution: make sure initialization is performed early 4071111147Snectar enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>, 4072111147Snectar Nils Larsch <nla@trustcenter.de> via PR#459) 4073111147Snectar [Lutz Jaenicke] 4074111147Snectar 4075111147Snectar *) Another fix for SSLv2 session ID handling: the session ID was incorrectly 4076111147Snectar checked on reconnect on the client side, therefore session resumption 4077111147Snectar could still fail with a "ssl session id is different" error. This 4078111147Snectar behaviour is masked when SSL_OP_ALL is used due to 4079111147Snectar SSL_OP_MICROSOFT_SESS_ID_BUG being set. 4080111147Snectar Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 4081111147Snectar followup to PR #377. 4082111147Snectar [Lutz Jaenicke] 4083111147Snectar 4084111147Snectar *) IA-32 assembler support enhancements: unified ELF targets, support 4085111147Snectar for SCO/Caldera platforms, fix for Cygwin shared build. 4086111147Snectar [Andy Polyakov] 4087111147Snectar 4088111147Snectar *) Add support for FreeBSD on sparc64. As a consequence, support for 4089111147Snectar FreeBSD on non-x86 processors is separate from x86 processors on 4090111147Snectar the config script, much like the NetBSD support. 4091111147Snectar [Richard Levitte & Kris Kennaway <kris@obsecurity.org>] 4092111147Snectar 4093109998Smarkm Changes between 0.9.6h and 0.9.7 [31 Dec 2002] 4094109998Smarkm 4095120631Snectar [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after 4096120631Snectar OpenSSL 0.9.7.] 4097120631Snectar 4098109998Smarkm *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED 4099109998Smarkm code (06) was taken as the first octet of the session ID and the last 4100109998Smarkm octet was ignored consequently. As a result SSLv2 client side session 4101109998Smarkm caching could not have worked due to the session ID mismatch between 4102109998Smarkm client and server. 4103109998Smarkm Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 4104109998Smarkm PR #377. 4105109998Smarkm [Lutz Jaenicke] 4106109998Smarkm 4107109998Smarkm *) Change the declaration of needed Kerberos libraries to use EX_LIBS 4108109998Smarkm instead of the special (and badly supported) LIBKRB5. LIBKRB5 is 4109109998Smarkm removed entirely. 4110109998Smarkm [Richard Levitte] 4111109998Smarkm 4112109998Smarkm *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it 4113109998Smarkm seems that in spite of existing for more than a year, many application 4114109998Smarkm author have done nothing to provide the necessary callbacks, which 4115109998Smarkm means that this particular engine will not work properly anywhere. 4116109998Smarkm This is a very unfortunate situation which forces us, in the name 4117109998Smarkm of usability, to give the hw_ncipher.c a static lock, which is part 4118109998Smarkm of libcrypto. 4119109998Smarkm NOTE: This is for the 0.9.7 series ONLY. This hack will never 4120109998Smarkm appear in 0.9.8 or later. We EXPECT application authors to have 4121109998Smarkm dealt properly with this when 0.9.8 is released (unless we actually 4122109998Smarkm make such changes in the libcrypto locking code that changes will 4123109998Smarkm have to be made anyway). 4124109998Smarkm [Richard Levitte] 4125109998Smarkm 4126109998Smarkm *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content 4127109998Smarkm octets have been read, EOF or an error occurs. Without this change 4128109998Smarkm some truncated ASN1 structures will not produce an error. 4129109998Smarkm [Steve Henson] 4130109998Smarkm 4131109998Smarkm *) Disable Heimdal support, since it hasn't been fully implemented. 4132109998Smarkm Still give the possibility to force the use of Heimdal, but with 4133109998Smarkm warnings and a request that patches get sent to openssl-dev. 4134109998Smarkm [Richard Levitte] 4135109998Smarkm 4136109998Smarkm *) Add the VC-CE target, introduce the WINCE sysname, and add 4137109998Smarkm INSTALL.WCE and appropriate conditionals to make it build. 4138109998Smarkm [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 4139109998Smarkm 4140109998Smarkm *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and 4141109998Smarkm cygssl-x.y.z.dll, where x, y and z are the major, minor and 4142109998Smarkm edit numbers of the version. 4143109998Smarkm [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 4144109998Smarkm 4145109998Smarkm *) Introduce safe string copy and catenation functions 4146109998Smarkm (BUF_strlcpy() and BUF_strlcat()). 4147109998Smarkm [Ben Laurie (CHATS) and Richard Levitte] 4148109998Smarkm 4149109998Smarkm *) Avoid using fixed-size buffers for one-line DNs. 4150109998Smarkm [Ben Laurie (CHATS)] 4151109998Smarkm 4152109998Smarkm *) Add BUF_MEM_grow_clean() to avoid information leakage when 4153109998Smarkm resizing buffers containing secrets, and use where appropriate. 4154109998Smarkm [Ben Laurie (CHATS)] 4155109998Smarkm 4156109998Smarkm *) Avoid using fixed size buffers for configuration file location. 4157109998Smarkm [Ben Laurie (CHATS)] 4158109998Smarkm 4159109998Smarkm *) Avoid filename truncation for various CA files. 4160109998Smarkm [Ben Laurie (CHATS)] 4161109998Smarkm 4162109998Smarkm *) Use sizeof in preference to magic numbers. 4163109998Smarkm [Ben Laurie (CHATS)] 4164109998Smarkm 4165109998Smarkm *) Avoid filename truncation in cert requests. 4166109998Smarkm [Ben Laurie (CHATS)] 4167109998Smarkm 4168109998Smarkm *) Add assertions to check for (supposedly impossible) buffer 4169109998Smarkm overflows. 4170109998Smarkm [Ben Laurie (CHATS)] 4171109998Smarkm 4172109998Smarkm *) Don't cache truncated DNS entries in the local cache (this could 4173109998Smarkm potentially lead to a spoofing attack). 4174109998Smarkm [Ben Laurie (CHATS)] 4175109998Smarkm 4176109998Smarkm *) Fix various buffers to be large enough for hex/decimal 4177109998Smarkm representations in a platform independent manner. 4178109998Smarkm [Ben Laurie (CHATS)] 4179109998Smarkm 4180109998Smarkm *) Add CRYPTO_realloc_clean() to avoid information leakage when 4181109998Smarkm resizing buffers containing secrets, and use where appropriate. 4182109998Smarkm [Ben Laurie (CHATS)] 4183109998Smarkm 4184109998Smarkm *) Add BIO_indent() to avoid much slightly worrying code to do 4185109998Smarkm indents. 4186109998Smarkm [Ben Laurie (CHATS)] 4187109998Smarkm 4188109998Smarkm *) Convert sprintf()/BIO_puts() to BIO_printf(). 4189109998Smarkm [Ben Laurie (CHATS)] 4190109998Smarkm 4191109998Smarkm *) buffer_gets() could terminate with the buffer only half 4192109998Smarkm full. Fixed. 4193109998Smarkm [Ben Laurie (CHATS)] 4194109998Smarkm 4195109998Smarkm *) Add assertions to prevent user-supplied crypto functions from 4196109998Smarkm overflowing internal buffers by having large block sizes, etc. 4197109998Smarkm [Ben Laurie (CHATS)] 4198109998Smarkm 4199109998Smarkm *) New OPENSSL_assert() macro (similar to assert(), but enabled 4200109998Smarkm unconditionally). 4201109998Smarkm [Ben Laurie (CHATS)] 4202109998Smarkm 4203109998Smarkm *) Eliminate unused copy of key in RC4. 4204109998Smarkm [Ben Laurie (CHATS)] 4205109998Smarkm 4206109998Smarkm *) Eliminate unused and incorrectly sized buffers for IV in pem.h. 4207109998Smarkm [Ben Laurie (CHATS)] 4208109998Smarkm 4209109998Smarkm *) Fix off-by-one error in EGD path. 4210109998Smarkm [Ben Laurie (CHATS)] 4211109998Smarkm 4212109998Smarkm *) If RANDFILE path is too long, ignore instead of truncating. 4213109998Smarkm [Ben Laurie (CHATS)] 4214109998Smarkm 4215109998Smarkm *) Eliminate unused and incorrectly sized X.509 structure 4216109998Smarkm CBCParameter. 4217109998Smarkm [Ben Laurie (CHATS)] 4218109998Smarkm 4219109998Smarkm *) Eliminate unused and dangerous function knumber(). 4220109998Smarkm [Ben Laurie (CHATS)] 4221109998Smarkm 4222109998Smarkm *) Eliminate unused and dangerous structure, KSSL_ERR. 4223109998Smarkm [Ben Laurie (CHATS)] 4224109998Smarkm 4225109998Smarkm *) Protect against overlong session ID context length in an encoded 4226109998Smarkm session object. Since these are local, this does not appear to be 4227109998Smarkm exploitable. 4228109998Smarkm [Ben Laurie (CHATS)] 4229109998Smarkm 4230109998Smarkm *) Change from security patch (see 0.9.6e below) that did not affect 4231109998Smarkm the 0.9.6 release series: 4232109998Smarkm 4233109998Smarkm Remote buffer overflow in SSL3 protocol - an attacker could 4234109998Smarkm supply an oversized master key in Kerberos-enabled versions. 4235160814Ssimon (CVE-2002-0657) 4236109998Smarkm [Ben Laurie (CHATS)] 4237109998Smarkm 4238109998Smarkm *) Change the SSL kerb5 codes to match RFC 2712. 4239109998Smarkm [Richard Levitte] 4240109998Smarkm 4241109998Smarkm *) Make -nameopt work fully for req and add -reqopt switch. 4242109998Smarkm [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson] 4243109998Smarkm 4244109998Smarkm *) The "block size" for block ciphers in CFB and OFB mode should be 1. 4245109998Smarkm [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>] 4246109998Smarkm 4247109998Smarkm *) Make sure tests can be performed even if the corresponding algorithms 4248109998Smarkm have been removed entirely. This was also the last step to make 4249109998Smarkm OpenSSL compilable with DJGPP under all reasonable conditions. 4250109998Smarkm [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>] 4251109998Smarkm 4252109998Smarkm *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT 4253109998Smarkm to allow version independent disabling of normally unselected ciphers, 4254109998Smarkm which may be activated as a side-effect of selecting a single cipher. 4255109998Smarkm 4256109998Smarkm (E.g., cipher list string "RSA" enables ciphersuites that are left 4257109998Smarkm out of "ALL" because they do not provide symmetric encryption. 4258109998Smarkm "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) 4259109998Smarkm [Lutz Jaenicke, Bodo Moeller] 4260109998Smarkm 4261109998Smarkm *) Add appropriate support for separate platform-dependent build 4262109998Smarkm directories. The recommended way to make a platform-dependent 4263109998Smarkm build directory is the following (tested on Linux), maybe with 4264109998Smarkm some local tweaks: 4265109998Smarkm 4266109998Smarkm # Place yourself outside of the OpenSSL source tree. In 4267109998Smarkm # this example, the environment variable OPENSSL_SOURCE 4268109998Smarkm # is assumed to contain the absolute OpenSSL source directory. 4269109998Smarkm mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" 4270109998Smarkm cd objtree/"`uname -s`-`uname -r`-`uname -m`" 4271111147Snectar (cd $OPENSSL_SOURCE; find . -type f) | while read F; do 4272109998Smarkm mkdir -p `dirname $F` 4273109998Smarkm ln -s $OPENSSL_SOURCE/$F $F 4274109998Smarkm done 4275109998Smarkm 4276109998Smarkm To be absolutely sure not to disturb the source tree, a "make clean" 4277109998Smarkm is a good thing. If it isn't successfull, don't worry about it, 4278109998Smarkm it probably means the source directory is very clean. 4279109998Smarkm [Richard Levitte] 4280109998Smarkm 4281109998Smarkm *) Make sure any ENGINE control commands make local copies of string 4282109998Smarkm pointers passed to them whenever necessary. Otherwise it is possible 4283109998Smarkm the caller may have overwritten (or deallocated) the original string 4284109998Smarkm data when a later ENGINE operation tries to use the stored values. 4285109998Smarkm [G�tz Babin-Ebell <babinebell@trustcenter.de>] 4286109998Smarkm 4287109998Smarkm *) Improve diagnostics in file reading and command-line digests. 4288109998Smarkm [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>] 4289109998Smarkm 4290109998Smarkm *) Add AES modes CFB and OFB to the object database. Correct an 4291109998Smarkm error in AES-CFB decryption. 4292109998Smarkm [Richard Levitte] 4293109998Smarkm 4294109998Smarkm *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this 4295109998Smarkm allows existing EVP_CIPHER_CTX structures to be reused after 4296109998Smarkm calling EVP_*Final(). This behaviour is used by encryption 4297109998Smarkm BIOs and some applications. This has the side effect that 4298109998Smarkm applications must explicitly clean up cipher contexts with 4299109998Smarkm EVP_CIPHER_CTX_cleanup() or they will leak memory. 4300109998Smarkm [Steve Henson] 4301109998Smarkm 4302109998Smarkm *) Check the values of dna and dnb in bn_mul_recursive before calling 4303109998Smarkm bn_mul_comba (a non zero value means the a or b arrays do not contain 4304109998Smarkm n2 elements) and fallback to bn_mul_normal if either is not zero. 4305109998Smarkm [Steve Henson] 4306109998Smarkm 4307109998Smarkm *) Fix escaping of non-ASCII characters when using the -subj option 4308109998Smarkm of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) 4309109998Smarkm [Lutz Jaenicke] 4310109998Smarkm 4311109998Smarkm *) Make object definitions compliant to LDAP (RFC2256): SN is the short 4312109998Smarkm form for "surname", serialNumber has no short form. 4313109998Smarkm Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; 4314109998Smarkm therefore remove "mail" short name for "internet 7". 4315109998Smarkm The OID for unique identifiers in X509 certificates is 4316109998Smarkm x500UniqueIdentifier, not uniqueIdentifier. 4317109998Smarkm Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) 4318109998Smarkm [Lutz Jaenicke] 4319109998Smarkm 4320109998Smarkm *) Add an "init" command to the ENGINE config module and auto initialize 4321109998Smarkm ENGINEs. Without any "init" command the ENGINE will be initialized 4322109998Smarkm after all ctrl commands have been executed on it. If init=1 the 4323109998Smarkm ENGINE is initailized at that point (ctrls before that point are run 4324109998Smarkm on the uninitialized ENGINE and after on the initialized one). If 4325109998Smarkm init=0 then the ENGINE will not be iniatialized at all. 4326109998Smarkm [Steve Henson] 4327109998Smarkm 4328109998Smarkm *) Fix the 'app_verify_callback' interface so that the user-defined 4329109998Smarkm argument is actually passed to the callback: In the 4330109998Smarkm SSL_CTX_set_cert_verify_callback() prototype, the callback 4331109998Smarkm declaration has been changed from 4332109998Smarkm int (*cb)() 4333109998Smarkm into 4334109998Smarkm int (*cb)(X509_STORE_CTX *,void *); 4335109998Smarkm in ssl_verify_cert_chain (ssl/ssl_cert.c), the call 4336109998Smarkm i=s->ctx->app_verify_callback(&ctx) 4337109998Smarkm has been changed into 4338109998Smarkm i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). 4339109998Smarkm 4340109998Smarkm To update applications using SSL_CTX_set_cert_verify_callback(), 4341109998Smarkm a dummy argument can be added to their callback functions. 4342109998Smarkm [D. K. Smetters <smetters@parc.xerox.com>] 4343109998Smarkm 4344109998Smarkm *) Added the '4758cca' ENGINE to support IBM 4758 cards. 4345109998Smarkm [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe] 4346109998Smarkm 4347109998Smarkm *) Add and OPENSSL_LOAD_CONF define which will cause 4348109998Smarkm OpenSSL_add_all_algorithms() to load the openssl.cnf config file. 4349109998Smarkm This allows older applications to transparently support certain 4350109998Smarkm OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. 4351109998Smarkm Two new functions OPENSSL_add_all_algorithms_noconf() which will never 4352109998Smarkm load the config file and OPENSSL_add_all_algorithms_conf() which will 4353109998Smarkm always load it have also been added. 4354109998Smarkm [Steve Henson] 4355109998Smarkm 4356109998Smarkm *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. 4357109998Smarkm Adjust NIDs and EVP layer. 4358109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 4359109998Smarkm 4360109998Smarkm *) Config modules support in openssl utility. 4361109998Smarkm 4362109998Smarkm Most commands now load modules from the config file, 4363109998Smarkm though in a few (such as version) this isn't done 4364109998Smarkm because it couldn't be used for anything. 4365109998Smarkm 4366109998Smarkm In the case of ca and req the config file used is 4367109998Smarkm the same as the utility itself: that is the -config 4368109998Smarkm command line option can be used to specify an 4369109998Smarkm alternative file. 4370109998Smarkm [Steve Henson] 4371109998Smarkm 4372109998Smarkm *) Move default behaviour from OPENSSL_config(). If appname is NULL 4373109998Smarkm use "openssl_conf" if filename is NULL use default openssl config file. 4374109998Smarkm [Steve Henson] 4375109998Smarkm 4376109998Smarkm *) Add an argument to OPENSSL_config() to allow the use of an alternative 4377109998Smarkm config section name. Add a new flag to tolerate a missing config file 4378109998Smarkm and move code to CONF_modules_load_file(). 4379109998Smarkm [Steve Henson] 4380109998Smarkm 4381109998Smarkm *) Support for crypto accelerator cards from Accelerated Encryption 4382109998Smarkm Processing, www.aep.ie. (Use engine 'aep') 4383109998Smarkm The support was copied from 0.9.6c [engine] and adapted/corrected 4384109998Smarkm to work with the new engine framework. 4385109998Smarkm [AEP Inc. and Richard Levitte] 4386109998Smarkm 4387109998Smarkm *) Support for SureWare crypto accelerator cards from Baltimore 4388109998Smarkm Technologies. (Use engine 'sureware') 4389109998Smarkm The support was copied from 0.9.6c [engine] and adapted 4390109998Smarkm to work with the new engine framework. 4391109998Smarkm [Richard Levitte] 4392109998Smarkm 4393109998Smarkm *) Have the CHIL engine fork-safe (as defined by nCipher) and actually 4394109998Smarkm make the newer ENGINE framework commands for the CHIL engine work. 4395109998Smarkm [Toomas Kiisk <vix@cyber.ee> and Richard Levitte] 4396109998Smarkm 4397109998Smarkm *) Make it possible to produce shared libraries on ReliantUNIX. 4398109998Smarkm [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte] 4399109998Smarkm 4400109998Smarkm *) Add the configuration target debug-linux-ppro. 4401109998Smarkm Make 'openssl rsa' use the general key loading routines 4402109998Smarkm implemented in apps.c, and make those routines able to 4403109998Smarkm handle the key format FORMAT_NETSCAPE and the variant 4404109998Smarkm FORMAT_IISSGC. 4405109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 4406109998Smarkm 4407109998Smarkm *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 4408109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 4409109998Smarkm 4410109998Smarkm *) Add -keyform to rsautl, and document -engine. 4411109998Smarkm [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>] 4412109998Smarkm 4413109998Smarkm *) Change BIO_new_file (crypto/bio/bss_file.c) to use new 4414109998Smarkm BIO_R_NO_SUCH_FILE error code rather than the generic 4415109998Smarkm ERR_R_SYS_LIB error code if fopen() fails with ENOENT. 4416109998Smarkm [Ben Laurie] 4417109998Smarkm 4418109998Smarkm *) Add new functions 4419109998Smarkm ERR_peek_last_error 4420109998Smarkm ERR_peek_last_error_line 4421109998Smarkm ERR_peek_last_error_line_data. 4422109998Smarkm These are similar to 4423109998Smarkm ERR_peek_error 4424109998Smarkm ERR_peek_error_line 4425109998Smarkm ERR_peek_error_line_data, 4426109998Smarkm but report on the latest error recorded rather than the first one 4427109998Smarkm still in the error queue. 4428109998Smarkm [Ben Laurie, Bodo Moeller] 4429109998Smarkm 4430109998Smarkm *) default_algorithms option in ENGINE config module. This allows things 4431109998Smarkm like: 4432109998Smarkm default_algorithms = ALL 4433109998Smarkm default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS 4434109998Smarkm [Steve Henson] 4435109998Smarkm 4436109998Smarkm *) Prelminary ENGINE config module. 4437109998Smarkm [Steve Henson] 4438109998Smarkm 4439109998Smarkm *) New experimental application configuration code. 4440109998Smarkm [Steve Henson] 4441109998Smarkm 4442109998Smarkm *) Change the AES code to follow the same name structure as all other 4443109998Smarkm symmetric ciphers, and behave the same way. Move everything to 4444109998Smarkm the directory crypto/aes, thereby obsoleting crypto/rijndael. 4445109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 4446109998Smarkm 4447109998Smarkm *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. 4448109998Smarkm [Ben Laurie and Theo de Raadt] 4449109998Smarkm 4450109998Smarkm *) Add option to output public keys in req command. 4451109998Smarkm [Massimiliano Pala madwolf@openca.org] 4452109998Smarkm 4453109998Smarkm *) Use wNAFs in EC_POINTs_mul() for improved efficiency 4454109998Smarkm (up to about 10% better than before for P-192 and P-224). 4455109998Smarkm [Bodo Moeller] 4456109998Smarkm 4457109998Smarkm *) New functions/macros 4458109998Smarkm 4459109998Smarkm SSL_CTX_set_msg_callback(ctx, cb) 4460109998Smarkm SSL_CTX_set_msg_callback_arg(ctx, arg) 4461109998Smarkm SSL_set_msg_callback(ssl, cb) 4462109998Smarkm SSL_set_msg_callback_arg(ssl, arg) 4463109998Smarkm 4464109998Smarkm to request calling a callback function 4465109998Smarkm 4466109998Smarkm void cb(int write_p, int version, int content_type, 4467109998Smarkm const void *buf, size_t len, SSL *ssl, void *arg) 4468109998Smarkm 4469109998Smarkm whenever a protocol message has been completely received 4470109998Smarkm (write_p == 0) or sent (write_p == 1). Here 'version' is the 4471109998Smarkm protocol version according to which the SSL library interprets 4472109998Smarkm the current protocol message (SSL2_VERSION, SSL3_VERSION, or 4473109998Smarkm TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or 4474109998Smarkm the content type as defined in the SSL 3.0/TLS 1.0 protocol 4475109998Smarkm specification (change_cipher_spec(20), alert(21), handshake(22)). 4476109998Smarkm 'buf' and 'len' point to the actual message, 'ssl' to the 4477109998Smarkm SSL object, and 'arg' is the application-defined value set by 4478109998Smarkm SSL[_CTX]_set_msg_callback_arg(). 4479109998Smarkm 4480109998Smarkm 'openssl s_client' and 'openssl s_server' have new '-msg' options 4481109998Smarkm to enable a callback that displays all protocol messages. 4482109998Smarkm [Bodo Moeller] 4483109998Smarkm 4484109998Smarkm *) Change the shared library support so shared libraries are built as 4485109998Smarkm soon as the corresponding static library is finished, and thereby get 4486109998Smarkm openssl and the test programs linked against the shared library. 4487109998Smarkm This still only happens when the keyword "shard" has been given to 4488109998Smarkm the configuration scripts. 4489109998Smarkm 4490109998Smarkm NOTE: shared library support is still an experimental thing, and 4491109998Smarkm backward binary compatibility is still not guaranteed. 4492109998Smarkm ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte] 4493109998Smarkm 4494109998Smarkm *) Add support for Subject Information Access extension. 4495109998Smarkm [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 4496109998Smarkm 4497109998Smarkm *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero 4498109998Smarkm additional bytes when new memory had to be allocated, not just 4499109998Smarkm when reusing an existing buffer. 4500109998Smarkm [Bodo Moeller] 4501109998Smarkm 4502109998Smarkm *) New command line and configuration option 'utf8' for the req command. 4503109998Smarkm This allows field values to be specified as UTF8 strings. 4504109998Smarkm [Steve Henson] 4505109998Smarkm 4506109998Smarkm *) Add -multi and -mr options to "openssl speed" - giving multiple parallel 4507109998Smarkm runs for the former and machine-readable output for the latter. 4508109998Smarkm [Ben Laurie] 4509109998Smarkm 4510109998Smarkm *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion 4511109998Smarkm of the e-mail address in the DN (i.e., it will go into a certificate 4512109998Smarkm extension only). The new configuration file option 'email_in_dn = no' 4513109998Smarkm has the same effect. 4514109998Smarkm [Massimiliano Pala madwolf@openca.org] 4515109998Smarkm 4516109998Smarkm *) Change all functions with names starting with des_ to be starting 4517109998Smarkm with DES_ instead. Add wrappers that are compatible with libdes, 4518109998Smarkm but are named _ossl_old_des_*. Finally, add macros that map the 4519109998Smarkm des_* symbols to the corresponding _ossl_old_des_* if libdes 4520109998Smarkm compatibility is desired. If OpenSSL 0.9.6c compatibility is 4521109998Smarkm desired, the des_* symbols will be mapped to DES_*, with one 4522109998Smarkm exception. 4523109998Smarkm 4524109998Smarkm Since we provide two compatibility mappings, the user needs to 4525109998Smarkm define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes 4526109998Smarkm compatibility is desired. The default (i.e., when that macro 4527109998Smarkm isn't defined) is OpenSSL 0.9.6c compatibility. 4528109998Smarkm 4529109998Smarkm There are also macros that enable and disable the support of old 4530109998Smarkm des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT 4531109998Smarkm and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those 4532109998Smarkm are defined, the default will apply: to support the old des routines. 4533109998Smarkm 4534109998Smarkm In either case, one must include openssl/des.h to get the correct 4535109998Smarkm definitions. Do not try to just include openssl/des_old.h, that 4536109998Smarkm won't work. 4537109998Smarkm 4538109998Smarkm NOTE: This is a major break of an old API into a new one. Software 4539109998Smarkm authors are encouraged to switch to the DES_ style functions. Some 4540109998Smarkm time in the future, des_old.h and the libdes compatibility functions 4541109998Smarkm will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the 4542109998Smarkm default), and then completely removed. 4543109998Smarkm [Richard Levitte] 4544109998Smarkm 4545109998Smarkm *) Test for certificates which contain unsupported critical extensions. 4546109998Smarkm If such a certificate is found during a verify operation it is 4547109998Smarkm rejected by default: this behaviour can be overridden by either 4548109998Smarkm handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or 4549109998Smarkm by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function 4550109998Smarkm X509_supported_extension() has also been added which returns 1 if a 4551109998Smarkm particular extension is supported. 4552109998Smarkm [Steve Henson] 4553109998Smarkm 4554109998Smarkm *) Modify the behaviour of EVP cipher functions in similar way to digests 4555109998Smarkm to retain compatibility with existing code. 4556109998Smarkm [Steve Henson] 4557109998Smarkm 4558109998Smarkm *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain 4559109998Smarkm compatibility with existing code. In particular the 'ctx' parameter does 4560109998Smarkm not have to be to be initialized before the call to EVP_DigestInit() and 4561109998Smarkm it is tidied up after a call to EVP_DigestFinal(). New function 4562109998Smarkm EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function 4563109998Smarkm EVP_MD_CTX_copy() changed to not require the destination to be 4564109998Smarkm initialized valid and new function EVP_MD_CTX_copy_ex() added which 4565109998Smarkm requires the destination to be valid. 4566109998Smarkm 4567109998Smarkm Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), 4568109998Smarkm EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). 4569109998Smarkm [Steve Henson] 4570109998Smarkm 4571109998Smarkm *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it 4572109998Smarkm so that complete 'Handshake' protocol structures are kept in memory 4573109998Smarkm instead of overwriting 'msg_type' and 'length' with 'body' data. 4574109998Smarkm [Bodo Moeller] 4575109998Smarkm 4576109998Smarkm *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. 4577109998Smarkm [Massimo Santin via Richard Levitte] 4578109998Smarkm 4579109998Smarkm *) Major restructuring to the underlying ENGINE code. This includes 4580109998Smarkm reduction of linker bloat, separation of pure "ENGINE" manipulation 4581109998Smarkm (initialisation, etc) from functionality dealing with implementations 4582109998Smarkm of specific crypto iterfaces. This change also introduces integrated 4583109998Smarkm support for symmetric ciphers and digest implementations - so ENGINEs 4584109998Smarkm can now accelerate these by providing EVP_CIPHER and EVP_MD 4585109998Smarkm implementations of their own. This is detailed in crypto/engine/README 4586109998Smarkm as it couldn't be adequately described here. However, there are a few 4587109998Smarkm API changes worth noting - some RSA, DSA, DH, and RAND functions that 4588109998Smarkm were changed in the original introduction of ENGINE code have now 4589109998Smarkm reverted back - the hooking from this code to ENGINE is now a good 4590109998Smarkm deal more passive and at run-time, operations deal directly with 4591109998Smarkm RSA_METHODs, DSA_METHODs (etc) as they did before, rather than 4592109998Smarkm dereferencing through an ENGINE pointer any more. Also, the ENGINE 4593109998Smarkm functions dealing with BN_MOD_EXP[_CRT] handlers have been removed - 4594109998Smarkm they were not being used by the framework as there is no concept of a 4595109998Smarkm BIGNUM_METHOD and they could not be generalised to the new 4596109998Smarkm 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, 4597109998Smarkm ENGINE_cpy() has been removed as it cannot be consistently defined in 4598109998Smarkm the new code. 4599109998Smarkm [Geoff Thorpe] 4600109998Smarkm 4601109998Smarkm *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. 4602109998Smarkm [Steve Henson] 4603109998Smarkm 4604109998Smarkm *) Change mkdef.pl to sort symbols that get the same entry number, 4605109998Smarkm and make sure the automatically generated functions ERR_load_* 4606109998Smarkm become part of libeay.num as well. 4607109998Smarkm [Richard Levitte] 4608109998Smarkm 4609109998Smarkm *) New function SSL_renegotiate_pending(). This returns true once 4610109998Smarkm renegotiation has been requested (either SSL_renegotiate() call 4611109998Smarkm or HelloRequest/ClientHello receveived from the peer) and becomes 4612109998Smarkm false once a handshake has been completed. 4613109998Smarkm (For servers, SSL_renegotiate() followed by SSL_do_handshake() 4614109998Smarkm sends a HelloRequest, but does not ensure that a handshake takes 4615109998Smarkm place. SSL_renegotiate_pending() is useful for checking if the 4616109998Smarkm client has followed the request.) 4617109998Smarkm [Bodo Moeller] 4618109998Smarkm 4619109998Smarkm *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. 4620109998Smarkm By default, clients may request session resumption even during 4621109998Smarkm renegotiation (if session ID contexts permit); with this option, 4622109998Smarkm session resumption is possible only in the first handshake. 4623109998Smarkm 4624109998Smarkm SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes 4625109998Smarkm more bits available for options that should not be part of 4626109998Smarkm SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). 4627109998Smarkm [Bodo Moeller] 4628109998Smarkm 4629109998Smarkm *) Add some demos for certificate and certificate request creation. 4630109998Smarkm [Steve Henson] 4631109998Smarkm 4632109998Smarkm *) Make maximum certificate chain size accepted from the peer application 4633109998Smarkm settable (SSL*_get/set_max_cert_list()), as proposed by 4634109998Smarkm "Douglas E. Engert" <deengert@anl.gov>. 4635109998Smarkm [Lutz Jaenicke] 4636109998Smarkm 4637109998Smarkm *) Add support for shared libraries for Unixware-7 4638109998Smarkm (Boyd Lynn Gerber <gerberb@zenez.com>). 4639109998Smarkm [Lutz Jaenicke] 4640109998Smarkm 4641109998Smarkm *) Add a "destroy" handler to ENGINEs that allows structural cleanup to 4642109998Smarkm be done prior to destruction. Use this to unload error strings from 4643109998Smarkm ENGINEs that load their own error strings. NB: This adds two new API 4644109998Smarkm functions to "get" and "set" this destroy handler in an ENGINE. 4645109998Smarkm [Geoff Thorpe] 4646109998Smarkm 4647109998Smarkm *) Alter all existing ENGINE implementations (except "openssl" and 4648109998Smarkm "openbsd") to dynamically instantiate their own error strings. This 4649109998Smarkm makes them more flexible to be built both as statically-linked ENGINEs 4650109998Smarkm and self-contained shared-libraries loadable via the "dynamic" ENGINE. 4651109998Smarkm Also, add stub code to each that makes building them as self-contained 4652109998Smarkm shared-libraries easier (see README.ENGINE). 4653109998Smarkm [Geoff Thorpe] 4654109998Smarkm 4655109998Smarkm *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE 4656109998Smarkm implementations into applications that are completely implemented in 4657109998Smarkm self-contained shared-libraries. The "dynamic" ENGINE exposes control 4658109998Smarkm commands that can be used to configure what shared-library to load and 4659109998Smarkm to control aspects of the way it is handled. Also, made an update to 4660109998Smarkm the README.ENGINE file that brings its information up-to-date and 4661109998Smarkm provides some information and instructions on the "dynamic" ENGINE 4662109998Smarkm (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). 4663109998Smarkm [Geoff Thorpe] 4664109998Smarkm 4665109998Smarkm *) Make it possible to unload ranges of ERR strings with a new 4666109998Smarkm "ERR_unload_strings" function. 4667109998Smarkm [Geoff Thorpe] 4668109998Smarkm 4669109998Smarkm *) Add a copy() function to EVP_MD. 4670109998Smarkm [Ben Laurie] 4671109998Smarkm 4672109998Smarkm *) Make EVP_MD routines take a context pointer instead of just the 4673109998Smarkm md_data void pointer. 4674109998Smarkm [Ben Laurie] 4675109998Smarkm 4676109998Smarkm *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates 4677109998Smarkm that the digest can only process a single chunk of data 4678109998Smarkm (typically because it is provided by a piece of 4679109998Smarkm hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application 4680109998Smarkm is only going to provide a single chunk of data, and hence the 4681109998Smarkm framework needn't accumulate the data for oneshot drivers. 4682109998Smarkm [Ben Laurie] 4683109998Smarkm 4684109998Smarkm *) As with "ERR", make it possible to replace the underlying "ex_data" 4685109998Smarkm functions. This change also alters the storage and management of global 4686109998Smarkm ex_data state - it's now all inside ex_data.c and all "class" code (eg. 4687109998Smarkm RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class 4688109998Smarkm index counters. The API functions that use this state have been changed 4689109998Smarkm to take a "class_index" rather than pointers to the class's local STACK 4690109998Smarkm and counter, and there is now an API function to dynamically create new 4691109998Smarkm classes. This centralisation allows us to (a) plug a lot of the 4692109998Smarkm thread-safety problems that existed, and (b) makes it possible to clean 4693109998Smarkm up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) 4694109998Smarkm such data would previously have always leaked in application code and 4695109998Smarkm workarounds were in place to make the memory debugging turn a blind eye 4696109998Smarkm to it. Application code that doesn't use this new function will still 4697109998Smarkm leak as before, but their memory debugging output will announce it now 4698109998Smarkm rather than letting it slide. 4699109998Smarkm 4700109998Smarkm Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change 4701109998Smarkm induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now 4702109998Smarkm has a return value to indicate success or failure. 4703109998Smarkm [Geoff Thorpe] 4704109998Smarkm 4705109998Smarkm *) Make it possible to replace the underlying "ERR" functions such that the 4706109998Smarkm global state (2 LHASH tables and 2 locks) is only used by the "default" 4707109998Smarkm implementation. This change also adds two functions to "get" and "set" 4708109998Smarkm the implementation prior to it being automatically set the first time 4709109998Smarkm any other ERR function takes place. Ie. an application can call "get", 4710109998Smarkm pass the return value to a module it has just loaded, and that module 4711109998Smarkm can call its own "set" function using that value. This means the 4712109998Smarkm module's "ERR" operations will use (and modify) the error state in the 4713109998Smarkm application and not in its own statically linked copy of OpenSSL code. 4714109998Smarkm [Geoff Thorpe] 4715109998Smarkm 4716109998Smarkm *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment 4717109998Smarkm reference counts. This performs normal REF_PRINT/REF_CHECK macros on 4718109998Smarkm the operation, and provides a more encapsulated way for external code 4719109998Smarkm (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code 4720109998Smarkm to use these functions rather than manually incrementing the counts. 4721109998Smarkm 4722109998Smarkm Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". 4723109998Smarkm [Geoff Thorpe] 4724109998Smarkm 4725109998Smarkm *) Add EVP test program. 4726109998Smarkm [Ben Laurie] 4727109998Smarkm 4728109998Smarkm *) Add symmetric cipher support to ENGINE. Expect the API to change! 4729109998Smarkm [Ben Laurie] 4730109998Smarkm 4731109998Smarkm *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() 4732109998Smarkm X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), 4733109998Smarkm X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). 4734109998Smarkm These allow a CRL to be built without having to access X509_CRL fields 4735109998Smarkm directly. Modify 'ca' application to use new functions. 4736109998Smarkm [Steve Henson] 4737109998Smarkm 4738109998Smarkm *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended 4739109998Smarkm bug workarounds. Rollback attack detection is a security feature. 4740109998Smarkm The problem will only arise on OpenSSL servers when TLSv1 is not 4741109998Smarkm available (sslv3_server_method() or SSL_OP_NO_TLSv1). 4742109998Smarkm Software authors not wanting to support TLSv1 will have special reasons 4743109998Smarkm for their choice and can explicitly enable this option. 4744109998Smarkm [Bodo Moeller, Lutz Jaenicke] 4745109998Smarkm 4746109998Smarkm *) Rationalise EVP so it can be extended: don't include a union of 4747109998Smarkm cipher/digest structures, add init/cleanup functions for EVP_MD_CTX 4748109998Smarkm (similar to those existing for EVP_CIPHER_CTX). 4749109998Smarkm Usage example: 4750109998Smarkm 4751109998Smarkm EVP_MD_CTX md; 4752109998Smarkm 4753109998Smarkm EVP_MD_CTX_init(&md); /* new function call */ 4754109998Smarkm EVP_DigestInit(&md, EVP_sha1()); 4755109998Smarkm EVP_DigestUpdate(&md, in, len); 4756109998Smarkm EVP_DigestFinal(&md, out, NULL); 4757109998Smarkm EVP_MD_CTX_cleanup(&md); /* new function call */ 4758109998Smarkm 4759109998Smarkm [Ben Laurie] 4760109998Smarkm 4761109998Smarkm *) Make DES key schedule conform to the usual scheme, as well as 4762109998Smarkm correcting its structure. This means that calls to DES functions 4763109998Smarkm now have to pass a pointer to a des_key_schedule instead of a 4764109998Smarkm plain des_key_schedule (which was actually always a pointer 4765109998Smarkm anyway): E.g., 4766109998Smarkm 4767109998Smarkm des_key_schedule ks; 4768109998Smarkm 4769109998Smarkm des_set_key_checked(..., &ks); 4770109998Smarkm des_ncbc_encrypt(..., &ks, ...); 4771109998Smarkm 4772109998Smarkm (Note that a later change renames 'des_...' into 'DES_...'.) 4773109998Smarkm [Ben Laurie] 4774109998Smarkm 4775109998Smarkm *) Initial reduction of linker bloat: the use of some functions, such as 4776109998Smarkm PEM causes large amounts of unused functions to be linked in due to 4777109998Smarkm poor organisation. For example pem_all.c contains every PEM function 4778109998Smarkm which has a knock on effect of linking in large amounts of (unused) 4779109998Smarkm ASN1 code. Grouping together similar functions and splitting unrelated 4780109998Smarkm functions prevents this. 4781109998Smarkm [Steve Henson] 4782109998Smarkm 4783109998Smarkm *) Cleanup of EVP macros. 4784109998Smarkm [Ben Laurie] 4785109998Smarkm 4786109998Smarkm *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the 4787109998Smarkm correct _ecb suffix. 4788109998Smarkm [Ben Laurie] 4789109998Smarkm 4790109998Smarkm *) Add initial OCSP responder support to ocsp application. The 4791109998Smarkm revocation information is handled using the text based index 4792109998Smarkm use by the ca application. The responder can either handle 4793109998Smarkm requests generated internally, supplied in files (for example 4794109998Smarkm via a CGI script) or using an internal minimal server. 4795109998Smarkm [Steve Henson] 4796109998Smarkm 4797109998Smarkm *) Add configuration choices to get zlib compression for TLS. 4798109998Smarkm [Richard Levitte] 4799109998Smarkm 4800109998Smarkm *) Changes to Kerberos SSL for RFC 2712 compliance: 4801109998Smarkm 1. Implemented real KerberosWrapper, instead of just using 4802109998Smarkm KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] 4803109998Smarkm 2. Implemented optional authenticator field of KerberosWrapper. 4804109998Smarkm 4805109998Smarkm Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, 4806109998Smarkm and authenticator structs; see crypto/krb5/. 4807109998Smarkm 4808109998Smarkm Generalized Kerberos calls to support multiple Kerberos libraries. 4809109998Smarkm [Vern Staats <staatsvr@asc.hpc.mil>, 4810109998Smarkm Jeffrey Altman <jaltman@columbia.edu> 4811109998Smarkm via Richard Levitte] 4812109998Smarkm 4813109998Smarkm *) Cause 'openssl speed' to use fully hard-coded DSA keys as it 4814109998Smarkm already does with RSA. testdsa.h now has 'priv_key/pub_key' 4815109998Smarkm values for each of the key sizes rather than having just 4816109998Smarkm parameters (and 'speed' generating keys each time). 4817109998Smarkm [Geoff Thorpe] 4818109998Smarkm 4819109998Smarkm *) Speed up EVP routines. 4820109998Smarkm Before: 4821109998Smarkmencrypt 4822109998Smarkmtype 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 4823109998Smarkmdes-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k 4824109998Smarkmdes-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k 4825109998Smarkmdes-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k 4826109998Smarkmdecrypt 4827109998Smarkmdes-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k 4828109998Smarkmdes-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k 4829109998Smarkmdes-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k 4830109998Smarkm After: 4831109998Smarkmencrypt 4832109998Smarkmdes-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k 4833109998Smarkmdecrypt 4834109998Smarkmdes-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k 4835109998Smarkm [Ben Laurie] 4836109998Smarkm 4837109998Smarkm *) Added the OS2-EMX target. 4838109998Smarkm ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte] 4839109998Smarkm 4840109998Smarkm *) Rewrite apps to use NCONF routines instead of the old CONF. New functions 4841109998Smarkm to support NCONF routines in extension code. New function CONF_set_nconf() 4842109998Smarkm to allow functions which take an NCONF to also handle the old LHASH 4843109998Smarkm structure: this means that the old CONF compatible routines can be 4844109998Smarkm retained (in particular wrt extensions) without having to duplicate the 4845109998Smarkm code. New function X509V3_add_ext_nconf_sk to add extensions to a stack. 4846109998Smarkm [Steve Henson] 4847109998Smarkm 4848109998Smarkm *) Enhance the general user interface with mechanisms for inner control 4849109998Smarkm and with possibilities to have yes/no kind of prompts. 4850109998Smarkm [Richard Levitte] 4851109998Smarkm 4852109998Smarkm *) Change all calls to low level digest routines in the library and 4853109998Smarkm applications to use EVP. Add missing calls to HMAC_cleanup() and 4854109998Smarkm don't assume HMAC_CTX can be copied using memcpy(). 4855109998Smarkm [Verdon Walker <VWalker@novell.com>, Steve Henson] 4856109998Smarkm 4857109998Smarkm *) Add the possibility to control engines through control names but with 4858109998Smarkm arbitrary arguments instead of just a string. 4859109998Smarkm Change the key loaders to take a UI_METHOD instead of a callback 4860109998Smarkm function pointer. NOTE: this breaks binary compatibility with earlier 4861109998Smarkm versions of OpenSSL [engine]. 4862109998Smarkm Adapt the nCipher code for these new conditions and add a card insertion 4863109998Smarkm callback. 4864109998Smarkm [Richard Levitte] 4865109998Smarkm 4866109998Smarkm *) Enhance the general user interface with mechanisms to better support 4867109998Smarkm dialog box interfaces, application-defined prompts, the possibility 4868109998Smarkm to use defaults (for example default passwords from somewhere else) 4869109998Smarkm and interrupts/cancellations. 4870109998Smarkm [Richard Levitte] 4871109998Smarkm 4872109998Smarkm *) Tidy up PKCS#12 attribute handling. Add support for the CSP name 4873109998Smarkm attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. 4874109998Smarkm [Steve Henson] 4875109998Smarkm 4876109998Smarkm *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also 4877109998Smarkm tidy up some unnecessarily weird code in 'sk_new()'). 4878109998Smarkm [Geoff, reported by Diego Tartara <dtartara@novamens.com>] 4879109998Smarkm 4880109998Smarkm *) Change the key loading routines for ENGINEs to use the same kind 4881109998Smarkm callback (pem_password_cb) as all other routines that need this 4882109998Smarkm kind of callback. 4883109998Smarkm [Richard Levitte] 4884109998Smarkm 4885109998Smarkm *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with 4886109998Smarkm 256 bit (=32 byte) keys. Of course seeding with more entropy bytes 4887109998Smarkm than this minimum value is recommended. 4888109998Smarkm [Lutz Jaenicke] 4889109998Smarkm 4890109998Smarkm *) New random seeder for OpenVMS, using the system process statistics 4891109998Smarkm that are easily reachable. 4892109998Smarkm [Richard Levitte] 4893109998Smarkm 4894109998Smarkm *) Windows apparently can't transparently handle global 4895109998Smarkm variables defined in DLLs. Initialisations such as: 4896109998Smarkm 4897109998Smarkm const ASN1_ITEM *it = &ASN1_INTEGER_it; 4898109998Smarkm 4899109998Smarkm wont compile. This is used by the any applications that need to 4900109998Smarkm declare their own ASN1 modules. This was fixed by adding the option 4901109998Smarkm EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly 4902109998Smarkm needed for static libraries under Win32. 4903109998Smarkm [Steve Henson] 4904109998Smarkm 4905109998Smarkm *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle 4906109998Smarkm setting of purpose and trust fields. New X509_STORE trust and 4907109998Smarkm purpose functions and tidy up setting in other SSL functions. 4908109998Smarkm [Steve Henson] 4909109998Smarkm 4910109998Smarkm *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE 4911109998Smarkm structure. These are inherited by X509_STORE_CTX when it is 4912109998Smarkm initialised. This allows various defaults to be set in the 4913109998Smarkm X509_STORE structure (such as flags for CRL checking and custom 4914109998Smarkm purpose or trust settings) for functions which only use X509_STORE_CTX 4915109998Smarkm internally such as S/MIME. 4916109998Smarkm 4917109998Smarkm Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and 4918109998Smarkm trust settings if they are not set in X509_STORE. This allows X509_STORE 4919109998Smarkm purposes and trust (in S/MIME for example) to override any set by default. 4920109998Smarkm 4921109998Smarkm Add command line options for CRL checking to smime, s_client and s_server 4922109998Smarkm applications. 4923109998Smarkm [Steve Henson] 4924109998Smarkm 4925109998Smarkm *) Initial CRL based revocation checking. If the CRL checking flag(s) 4926109998Smarkm are set then the CRL is looked up in the X509_STORE structure and 4927109998Smarkm its validity and signature checked, then if the certificate is found 4928109998Smarkm in the CRL the verify fails with a revoked error. 4929109998Smarkm 4930109998Smarkm Various new CRL related callbacks added to X509_STORE_CTX structure. 4931109998Smarkm 4932109998Smarkm Command line options added to 'verify' application to support this. 4933109998Smarkm 4934109998Smarkm This needs some additional work, such as being able to handle multiple 4935109998Smarkm CRLs with different times, extension based lookup (rather than just 4936109998Smarkm by subject name) and ultimately more complete V2 CRL extension 4937109998Smarkm handling. 4938109998Smarkm [Steve Henson] 4939109998Smarkm 4940109998Smarkm *) Add a general user interface API (crypto/ui/). This is designed 4941109998Smarkm to replace things like des_read_password and friends (backward 4942109998Smarkm compatibility functions using this new API are provided). 4943109998Smarkm The purpose is to remove prompting functions from the DES code 4944109998Smarkm section as well as provide for prompting through dialog boxes in 4945109998Smarkm a window system and the like. 4946109998Smarkm [Richard Levitte] 4947109998Smarkm 4948109998Smarkm *) Add "ex_data" support to ENGINE so implementations can add state at a 4949109998Smarkm per-structure level rather than having to store it globally. 4950109998Smarkm [Geoff] 4951109998Smarkm 4952109998Smarkm *) Make it possible for ENGINE structures to be copied when retrieved by 4953109998Smarkm ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. 4954109998Smarkm This causes the "original" ENGINE structure to act like a template, 4955109998Smarkm analogous to the RSA vs. RSA_METHOD type of separation. Because of this 4956109998Smarkm operational state can be localised to each ENGINE structure, despite the 4957109998Smarkm fact they all share the same "methods". New ENGINE structures returned in 4958109998Smarkm this case have no functional references and the return value is the single 4959109998Smarkm structural reference. This matches the single structural reference returned 4960109998Smarkm by ENGINE_by_id() normally, when it is incremented on the pre-existing 4961109998Smarkm ENGINE structure. 4962109998Smarkm [Geoff] 4963109998Smarkm 4964109998Smarkm *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this 4965109998Smarkm needs to match any other type at all we need to manually clear the 4966109998Smarkm tag cache. 4967109998Smarkm [Steve Henson] 4968109998Smarkm 4969109998Smarkm *) Changes to the "openssl engine" utility to include; 4970109998Smarkm - verbosity levels ('-v', '-vv', and '-vvv') that provide information 4971109998Smarkm about an ENGINE's available control commands. 4972109998Smarkm - executing control commands from command line arguments using the 4973109998Smarkm '-pre' and '-post' switches. '-post' is only used if '-t' is 4974109998Smarkm specified and the ENGINE is successfully initialised. The syntax for 4975109998Smarkm the individual commands are colon-separated, for example; 4976109998Smarkm openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so 4977109998Smarkm [Geoff] 4978109998Smarkm 4979109998Smarkm *) New dynamic control command support for ENGINEs. ENGINEs can now 4980109998Smarkm declare their own commands (numbers), names (strings), descriptions, 4981109998Smarkm and input types for run-time discovery by calling applications. A 4982109998Smarkm subset of these commands are implicitly classed as "executable" 4983109998Smarkm depending on their input type, and only these can be invoked through 4984109998Smarkm the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this 4985109998Smarkm can be based on user input, config files, etc). The distinction is 4986109998Smarkm that "executable" commands cannot return anything other than a boolean 4987109998Smarkm result and can only support numeric or string input, whereas some 4988109998Smarkm discoverable commands may only be for direct use through 4989109998Smarkm ENGINE_ctrl(), eg. supporting the exchange of binary data, function 4990109998Smarkm pointers, or other custom uses. The "executable" commands are to 4991109998Smarkm support parameterisations of ENGINE behaviour that can be 4992109998Smarkm unambiguously defined by ENGINEs and used consistently across any 4993109998Smarkm OpenSSL-based application. Commands have been added to all the 4994109998Smarkm existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow 4995109998Smarkm control over shared-library paths without source code alterations. 4996109998Smarkm [Geoff] 4997109998Smarkm 4998109998Smarkm *) Changed all ENGINE implementations to dynamically allocate their 4999109998Smarkm ENGINEs rather than declaring them statically. Apart from this being 5000109998Smarkm necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, 5001109998Smarkm this also allows the implementations to compile without using the 5002109998Smarkm internal engine_int.h header. 5003109998Smarkm [Geoff] 5004109998Smarkm 5005109998Smarkm *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a 5006109998Smarkm 'const' value. Any code that should be able to modify a RAND_METHOD 5007109998Smarkm should already have non-const pointers to it (ie. they should only 5008109998Smarkm modify their own ones). 5009109998Smarkm [Geoff] 5010109998Smarkm 5011109998Smarkm *) Made a variety of little tweaks to the ENGINE code. 5012109998Smarkm - "atalla" and "ubsec" string definitions were moved from header files 5013109998Smarkm to C code. "nuron" string definitions were placed in variables 5014109998Smarkm rather than hard-coded - allowing parameterisation of these values 5015109998Smarkm later on via ctrl() commands. 5016109998Smarkm - Removed unused "#if 0"'d code. 5017109998Smarkm - Fixed engine list iteration code so it uses ENGINE_free() to release 5018109998Smarkm structural references. 5019109998Smarkm - Constified the RAND_METHOD element of ENGINE structures. 5020109998Smarkm - Constified various get/set functions as appropriate and added 5021109998Smarkm missing functions (including a catch-all ENGINE_cpy that duplicates 5022109998Smarkm all ENGINE values onto a new ENGINE except reference counts/state). 5023109998Smarkm - Removed NULL parameter checks in get/set functions. Setting a method 5024109998Smarkm or function to NULL is a way of cancelling out a previously set 5025109998Smarkm value. Passing a NULL ENGINE parameter is just plain stupid anyway 5026109998Smarkm and doesn't justify the extra error symbols and code. 5027109998Smarkm - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for 5028109998Smarkm flags from engine_int.h to engine.h. 5029109998Smarkm - Changed prototypes for ENGINE handler functions (init(), finish(), 5030109998Smarkm ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. 5031109998Smarkm [Geoff] 5032109998Smarkm 5033109998Smarkm *) Implement binary inversion algorithm for BN_mod_inverse in addition 5034109998Smarkm to the algorithm using long division. The binary algorithm can be 5035109998Smarkm used only if the modulus is odd. On 32-bit systems, it is faster 5036109998Smarkm only for relatively small moduli (roughly 20-30% for 128-bit moduli, 5037109998Smarkm roughly 5-15% for 256-bit moduli), so we use it only for moduli 5038109998Smarkm up to 450 bits. In 64-bit environments, the binary algorithm 5039109998Smarkm appears to be advantageous for much longer moduli; here we use it 5040109998Smarkm for moduli up to 2048 bits. 5041109998Smarkm [Bodo Moeller] 5042109998Smarkm 5043109998Smarkm *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code 5044109998Smarkm could not support the combine flag in choice fields. 5045109998Smarkm [Steve Henson] 5046109998Smarkm 5047109998Smarkm *) Add a 'copy_extensions' option to the 'ca' utility. This copies 5048109998Smarkm extensions from a certificate request to the certificate. 5049109998Smarkm [Steve Henson] 5050109998Smarkm 5051109998Smarkm *) Allow multiple 'certopt' and 'nameopt' options to be separated 5052109998Smarkm by commas. Add 'namopt' and 'certopt' options to the 'ca' config 5053109998Smarkm file: this allows the display of the certificate about to be 5054109998Smarkm signed to be customised, to allow certain fields to be included 5055109998Smarkm or excluded and extension details. The old system didn't display 5056109998Smarkm multicharacter strings properly, omitted fields not in the policy 5057109998Smarkm and couldn't display additional details such as extensions. 5058109998Smarkm [Steve Henson] 5059109998Smarkm 5060109998Smarkm *) Function EC_POINTs_mul for multiple scalar multiplication 5061109998Smarkm of an arbitrary number of elliptic curve points 5062109998Smarkm \sum scalars[i]*points[i], 5063109998Smarkm optionally including the generator defined for the EC_GROUP: 5064109998Smarkm scalar*generator + \sum scalars[i]*points[i]. 5065109998Smarkm 5066109998Smarkm EC_POINT_mul is a simple wrapper function for the typical case 5067109998Smarkm that the point list has just one item (besides the optional 5068109998Smarkm generator). 5069109998Smarkm [Bodo Moeller] 5070109998Smarkm 5071109998Smarkm *) First EC_METHODs for curves over GF(p): 5072109998Smarkm 5073109998Smarkm EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr 5074109998Smarkm operations and provides various method functions that can also 5075109998Smarkm operate with faster implementations of modular arithmetic. 5076109998Smarkm 5077109998Smarkm EC_GFp_mont_method() reuses most functions that are part of 5078109998Smarkm EC_GFp_simple_method, but uses Montgomery arithmetic. 5079109998Smarkm 5080109998Smarkm [Bodo Moeller; point addition and point doubling 5081109998Smarkm implementation directly derived from source code provided by 5082109998Smarkm Lenka Fibikova <fibikova@exp-math.uni-essen.de>] 5083109998Smarkm 5084109998Smarkm *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, 5085109998Smarkm crypto/ec/ec_lib.c): 5086109998Smarkm 5087109998Smarkm Curves are EC_GROUP objects (with an optional group generator) 5088109998Smarkm based on EC_METHODs that are built into the library. 5089109998Smarkm 5090109998Smarkm Points are EC_POINT objects based on EC_GROUP objects. 5091109998Smarkm 5092109998Smarkm Most of the framework would be able to handle curves over arbitrary 5093109998Smarkm finite fields, but as there are no obvious types for fields other 5094109998Smarkm than GF(p), some functions are limited to that for now. 5095109998Smarkm [Bodo Moeller] 5096109998Smarkm 5097109998Smarkm *) Add the -HTTP option to s_server. It is similar to -WWW, but requires 5098109998Smarkm that the file contains a complete HTTP response. 5099109998Smarkm [Richard Levitte] 5100109998Smarkm 5101109998Smarkm *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl 5102109998Smarkm change the def and num file printf format specifier from "%-40sXXX" 5103109998Smarkm to "%-39s XXX". The latter will always guarantee a space after the 5104109998Smarkm field while the former will cause them to run together if the field 5105109998Smarkm is 40 of more characters long. 5106109998Smarkm [Steve Henson] 5107109998Smarkm 5108109998Smarkm *) Constify the cipher and digest 'method' functions and structures 5109109998Smarkm and modify related functions to take constant EVP_MD and EVP_CIPHER 5110109998Smarkm pointers. 5111109998Smarkm [Steve Henson] 5112109998Smarkm 5113109998Smarkm *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them 5114109998Smarkm in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. 5115109998Smarkm [Bodo Moeller] 5116109998Smarkm 5117109998Smarkm *) Modify EVP_Digest*() routines so they now return values. Although the 5118109998Smarkm internal software routines can never fail additional hardware versions 5119109998Smarkm might. 5120109998Smarkm [Steve Henson] 5121109998Smarkm 5122109998Smarkm *) Clean up crypto/err/err.h and change some error codes to avoid conflicts: 5123109998Smarkm 5124109998Smarkm Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 5125109998Smarkm (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. 5126109998Smarkm 5127109998Smarkm ASN1 error codes 5128109998Smarkm ERR_R_NESTED_ASN1_ERROR 5129109998Smarkm ... 5130109998Smarkm ERR_R_MISSING_ASN1_EOS 5131109998Smarkm were 4 .. 9, conflicting with 5132109998Smarkm ERR_LIB_RSA (= ERR_R_RSA_LIB) 5133109998Smarkm ... 5134109998Smarkm ERR_LIB_PEM (= ERR_R_PEM_LIB). 5135109998Smarkm They are now 58 .. 63 (i.e., just below ERR_R_FATAL). 5136109998Smarkm 5137109998Smarkm Add new error code 'ERR_R_INTERNAL_ERROR'. 5138109998Smarkm [Bodo Moeller] 5139109998Smarkm 5140109998Smarkm *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock 5141109998Smarkm suffices. 5142109998Smarkm [Bodo Moeller] 5143109998Smarkm 5144109998Smarkm *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This 5145109998Smarkm sets the subject name for a new request or supersedes the 5146109998Smarkm subject name in a given request. Formats that can be parsed are 5147109998Smarkm 'CN=Some Name, OU=myOU, C=IT' 5148109998Smarkm and 5149109998Smarkm 'CN=Some Name/OU=myOU/C=IT'. 5150109998Smarkm 5151109998Smarkm Add options '-batch' and '-verbose' to 'openssl req'. 5152109998Smarkm [Massimiliano Pala <madwolf@hackmasters.net>] 5153109998Smarkm 5154109998Smarkm *) Introduce the possibility to access global variables through 5155109998Smarkm functions on platform were that's the best way to handle exporting 5156109998Smarkm global variables in shared libraries. To enable this functionality, 5157109998Smarkm one must configure with "EXPORT_VAR_AS_FN" or defined the C macro 5158109998Smarkm "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter 5159109998Smarkm is normally done by Configure or something similar). 5160109998Smarkm 5161109998Smarkm To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL 5162109998Smarkm in the source file (foo.c) like this: 5163109998Smarkm 5164109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; 5165109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(double,bar); 5166109998Smarkm 5167109998Smarkm To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL 5168109998Smarkm and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: 5169109998Smarkm 5170109998Smarkm OPENSSL_DECLARE_GLOBAL(int,foo); 5171109998Smarkm #define foo OPENSSL_GLOBAL_REF(foo) 5172109998Smarkm OPENSSL_DECLARE_GLOBAL(double,bar); 5173109998Smarkm #define bar OPENSSL_GLOBAL_REF(bar) 5174109998Smarkm 5175109998Smarkm The #defines are very important, and therefore so is including the 5176109998Smarkm header file everywhere where the defined globals are used. 5177109998Smarkm 5178109998Smarkm The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition 5179109998Smarkm of ASN.1 items, but that structure is a bit different. 5180109998Smarkm 5181109998Smarkm The largest change is in util/mkdef.pl which has been enhanced with 5182109998Smarkm better and easier to understand logic to choose which symbols should 5183109998Smarkm go into the Windows .def files as well as a number of fixes and code 5184109998Smarkm cleanup (among others, algorithm keywords are now sorted 5185109998Smarkm lexicographically to avoid constant rewrites). 5186109998Smarkm [Richard Levitte] 5187109998Smarkm 5188109998Smarkm *) In BN_div() keep a copy of the sign of 'num' before writing the 5189109998Smarkm result to 'rm' because if rm==num the value will be overwritten 5190109998Smarkm and produce the wrong result if 'num' is negative: this caused 5191109998Smarkm problems with BN_mod() and BN_nnmod(). 5192109998Smarkm [Steve Henson] 5193109998Smarkm 5194109998Smarkm *) Function OCSP_request_verify(). This checks the signature on an 5195109998Smarkm OCSP request and verifies the signer certificate. The signer 5196109998Smarkm certificate is just checked for a generic purpose and OCSP request 5197109998Smarkm trust settings. 5198109998Smarkm [Steve Henson] 5199109998Smarkm 5200109998Smarkm *) Add OCSP_check_validity() function to check the validity of OCSP 5201109998Smarkm responses. OCSP responses are prepared in real time and may only 5202109998Smarkm be a few seconds old. Simply checking that the current time lies 5203109998Smarkm between thisUpdate and nextUpdate max reject otherwise valid responses 5204109998Smarkm caused by either OCSP responder or client clock inaccuracy. Instead 5205109998Smarkm we allow thisUpdate and nextUpdate to fall within a certain period of 5206109998Smarkm the current time. The age of the response can also optionally be 5207109998Smarkm checked. Two new options -validity_period and -status_age added to 5208109998Smarkm ocsp utility. 5209109998Smarkm [Steve Henson] 5210109998Smarkm 5211109998Smarkm *) If signature or public key algorithm is unrecognized print out its 5212109998Smarkm OID rather that just UNKNOWN. 5213109998Smarkm [Steve Henson] 5214109998Smarkm 5215109998Smarkm *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and 5216109998Smarkm OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate 5217109998Smarkm ID to be generated from the issuer certificate alone which can then be 5218109998Smarkm passed to OCSP_id_issuer_cmp(). 5219109998Smarkm [Steve Henson] 5220109998Smarkm 5221109998Smarkm *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new 5222109998Smarkm ASN1 modules to export functions returning ASN1_ITEM pointers 5223109998Smarkm instead of the ASN1_ITEM structures themselves. This adds several 5224109998Smarkm new macros which allow the underlying ASN1 function/structure to 5225109998Smarkm be accessed transparently. As a result code should not use ASN1_ITEM 5226109998Smarkm references directly (such as &X509_it) but instead use the relevant 5227109998Smarkm macros (such as ASN1_ITEM_rptr(X509)). This option is to allow 5228109998Smarkm use of the new ASN1 code on platforms where exporting structures 5229109998Smarkm is problematical (for example in shared libraries) but exporting 5230109998Smarkm functions returning pointers to structures is not. 5231109998Smarkm [Steve Henson] 5232109998Smarkm 5233109998Smarkm *) Add support for overriding the generation of SSL/TLS session IDs. 5234109998Smarkm These callbacks can be registered either in an SSL_CTX or per SSL. 5235109998Smarkm The purpose of this is to allow applications to control, if they wish, 5236109998Smarkm the arbitrary values chosen for use as session IDs, particularly as it 5237109998Smarkm can be useful for session caching in multiple-server environments. A 5238109998Smarkm command-line switch for testing this (and any client code that wishes 5239109998Smarkm to use such a feature) has been added to "s_server". 5240109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 5241109998Smarkm 5242109998Smarkm *) Modify mkdef.pl to recognise and parse preprocessor conditionals 5243109998Smarkm of the form '#if defined(...) || defined(...) || ...' and 5244109998Smarkm '#if !defined(...) && !defined(...) && ...'. This also avoids 5245109998Smarkm the growing number of special cases it was previously handling. 5246109998Smarkm [Richard Levitte] 5247109998Smarkm 5248109998Smarkm *) Make all configuration macros available for application by making 5249109998Smarkm sure they are available in opensslconf.h, by giving them names starting 5250109998Smarkm with "OPENSSL_" to avoid conflicts with other packages and by making 5251109998Smarkm sure e_os2.h will cover all platform-specific cases together with 5252109998Smarkm opensslconf.h. 5253109998Smarkm Additionally, it is now possible to define configuration/platform- 5254109998Smarkm specific names (called "system identities"). In the C code, these 5255109998Smarkm are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another 5256109998Smarkm macro with the name beginning with "OPENSSL_SYS_", which is determined 5257109998Smarkm from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on 5258109998Smarkm what is available. 5259109998Smarkm [Richard Levitte] 5260109998Smarkm 5261109998Smarkm *) New option -set_serial to 'req' and 'x509' this allows the serial 5262109998Smarkm number to use to be specified on the command line. Previously self 5263109998Smarkm signed certificates were hard coded with serial number 0 and the 5264109998Smarkm CA options of 'x509' had to use a serial number in a file which was 5265109998Smarkm auto incremented. 5266109998Smarkm [Steve Henson] 5267109998Smarkm 5268109998Smarkm *) New options to 'ca' utility to support V2 CRL entry extensions. 5269109998Smarkm Currently CRL reason, invalidity date and hold instruction are 5270109998Smarkm supported. Add new CRL extensions to V3 code and some new objects. 5271109998Smarkm [Steve Henson] 5272109998Smarkm 5273109998Smarkm *) New function EVP_CIPHER_CTX_set_padding() this is used to 5274109998Smarkm disable standard block padding (aka PKCS#5 padding) in the EVP 5275109998Smarkm API, which was previously mandatory. This means that the data is 5276109998Smarkm not padded in any way and so the total length much be a multiple 5277109998Smarkm of the block size, otherwise an error occurs. 5278109998Smarkm [Steve Henson] 5279109998Smarkm 5280109998Smarkm *) Initial (incomplete) OCSP SSL support. 5281109998Smarkm [Steve Henson] 5282109998Smarkm 5283109998Smarkm *) New function OCSP_parse_url(). This splits up a URL into its host, 5284109998Smarkm port and path components: primarily to parse OCSP URLs. New -url 5285109998Smarkm option to ocsp utility. 5286109998Smarkm [Steve Henson] 5287109998Smarkm 5288109998Smarkm *) New nonce behavior. The return value of OCSP_check_nonce() now 5289109998Smarkm reflects the various checks performed. Applications can decide 5290109998Smarkm whether to tolerate certain situations such as an absent nonce 5291109998Smarkm in a response when one was present in a request: the ocsp application 5292109998Smarkm just prints out a warning. New function OCSP_add1_basic_nonce() 5293109998Smarkm this is to allow responders to include a nonce in a response even if 5294109998Smarkm the request is nonce-less. 5295109998Smarkm [Steve Henson] 5296109998Smarkm 5297109998Smarkm *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are 5298109998Smarkm skipped when using openssl x509 multiple times on a single input file, 5299109998Smarkm e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs". 5300109998Smarkm [Bodo Moeller] 5301109998Smarkm 5302109998Smarkm *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() 5303109998Smarkm set string type: to handle setting ASN1_TIME structures. Fix ca 5304109998Smarkm utility to correctly initialize revocation date of CRLs. 5305109998Smarkm [Steve Henson] 5306109998Smarkm 5307109998Smarkm *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override 5308109998Smarkm the clients preferred ciphersuites and rather use its own preferences. 5309109998Smarkm Should help to work around M$ SGC (Server Gated Cryptography) bug in 5310109998Smarkm Internet Explorer by ensuring unchanged hash method during stepup. 5311109998Smarkm (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) 5312109998Smarkm [Lutz Jaenicke] 5313109998Smarkm 5314109998Smarkm *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael 5315109998Smarkm to aes and add a new 'exist' option to print out symbols that don't 5316109998Smarkm appear to exist. 5317109998Smarkm [Steve Henson] 5318109998Smarkm 5319109998Smarkm *) Additional options to ocsp utility to allow flags to be set and 5320109998Smarkm additional certificates supplied. 5321109998Smarkm [Steve Henson] 5322109998Smarkm 5323109998Smarkm *) Add the option -VAfile to 'openssl ocsp', so the user can give the 5324109998Smarkm OCSP client a number of certificate to only verify the response 5325109998Smarkm signature against. 5326109998Smarkm [Richard Levitte] 5327109998Smarkm 5328109998Smarkm *) Update Rijndael code to version 3.0 and change EVP AES ciphers to 5329109998Smarkm handle the new API. Currently only ECB, CBC modes supported. Add new 5330109998Smarkm AES OIDs. 5331109998Smarkm 5332109998Smarkm Add TLS AES ciphersuites as described in RFC3268, "Advanced 5333109998Smarkm Encryption Standard (AES) Ciphersuites for Transport Layer 5334109998Smarkm Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were 5335109998Smarkm not enabled by default and were not part of the "ALL" ciphersuite 5336109998Smarkm alias because they were not yet official; they could be 5337109998Smarkm explicitly requested by specifying the "AESdraft" ciphersuite 5338109998Smarkm group alias. In the final release of OpenSSL 0.9.7, the group 5339109998Smarkm alias is called "AES" and is part of "ALL".) 5340109998Smarkm [Ben Laurie, Steve Henson, Bodo Moeller] 5341109998Smarkm 5342109998Smarkm *) New function OCSP_copy_nonce() to copy nonce value (if present) from 5343109998Smarkm request to response. 5344109998Smarkm [Steve Henson] 5345109998Smarkm 5346109998Smarkm *) Functions for OCSP responders. OCSP_request_onereq_count(), 5347109998Smarkm OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() 5348109998Smarkm extract information from a certificate request. OCSP_response_create() 5349109998Smarkm creates a response and optionally adds a basic response structure. 5350109998Smarkm OCSP_basic_add1_status() adds a complete single response to a basic 5351109998Smarkm response and returns the OCSP_SINGLERESP structure just added (to allow 5352109998Smarkm extensions to be included for example). OCSP_basic_add1_cert() adds a 5353109998Smarkm certificate to a basic response and OCSP_basic_sign() signs a basic 5354109998Smarkm response with various flags. New helper functions ASN1_TIME_check() 5355109998Smarkm (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() 5356109998Smarkm (converts ASN1_TIME to GeneralizedTime). 5357109998Smarkm [Steve Henson] 5358109998Smarkm 5359109998Smarkm *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() 5360109998Smarkm in a single operation. X509_get0_pubkey_bitstr() extracts the public_key 5361109998Smarkm structure from a certificate. X509_pubkey_digest() digests the public_key 5362109998Smarkm contents: this is used in various key identifiers. 5363109998Smarkm [Steve Henson] 5364109998Smarkm 5365109998Smarkm *) Make sk_sort() tolerate a NULL argument. 5366109998Smarkm [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>] 5367109998Smarkm 5368109998Smarkm *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates 5369109998Smarkm passed by the function are trusted implicitly. If any of them signed the 5370109998Smarkm response then it is assumed to be valid and is not verified. 5371109998Smarkm [Steve Henson] 5372109998Smarkm 5373109998Smarkm *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT 5374109998Smarkm to data. This was previously part of the PKCS7 ASN1 code. This 5375109998Smarkm was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. 5376109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 5377109998Smarkm <support@securenetterm.com>] 5378109998Smarkm 5379109998Smarkm *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 5380109998Smarkm routines: without these tracing memory leaks is very painful. 5381109998Smarkm Fix leaks in PKCS12 and PKCS7 routines. 5382109998Smarkm [Steve Henson] 5383109998Smarkm 5384109998Smarkm *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). 5385109998Smarkm Previously it initialised the 'type' argument to V_ASN1_UTCTIME which 5386109998Smarkm effectively meant GeneralizedTime would never be used. Now it 5387109998Smarkm is initialised to -1 but X509_time_adj() now has to check the value 5388109998Smarkm and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or 5389109998Smarkm V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. 5390109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 5391109998Smarkm <support@securenetterm.com>] 5392109998Smarkm 5393109998Smarkm *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously 5394109998Smarkm result in a zero length in the ASN1_INTEGER structure which was 5395109998Smarkm not consistent with the structure when d2i_ASN1_INTEGER() was used 5396109998Smarkm and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() 5397109998Smarkm to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() 5398109998Smarkm where it did not print out a minus for negative ASN1_INTEGER. 5399109998Smarkm [Steve Henson] 5400109998Smarkm 5401109998Smarkm *) Add summary printout to ocsp utility. The various functions which 5402109998Smarkm convert status values to strings have been renamed to: 5403109998Smarkm OCSP_response_status_str(), OCSP_cert_status_str() and 5404109998Smarkm OCSP_crl_reason_str() and are no longer static. New options 5405109998Smarkm to verify nonce values and to disable verification. OCSP response 5406109998Smarkm printout format cleaned up. 5407109998Smarkm [Steve Henson] 5408109998Smarkm 5409109998Smarkm *) Add additional OCSP certificate checks. These are those specified 5410109998Smarkm in RFC2560. This consists of two separate checks: the CA of the 5411109998Smarkm certificate being checked must either be the OCSP signer certificate 5412109998Smarkm or the issuer of the OCSP signer certificate. In the latter case the 5413109998Smarkm OCSP signer certificate must contain the OCSP signing extended key 5414109998Smarkm usage. This check is performed by attempting to match the OCSP 5415109998Smarkm signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash 5416109998Smarkm in the OCSP_CERTID structures of the response. 5417109998Smarkm [Steve Henson] 5418109998Smarkm 5419109998Smarkm *) Initial OCSP certificate verification added to OCSP_basic_verify() 5420109998Smarkm and related routines. This uses the standard OpenSSL certificate 5421109998Smarkm verify routines to perform initial checks (just CA validity) and 5422109998Smarkm to obtain the certificate chain. Then additional checks will be 5423109998Smarkm performed on the chain. Currently the root CA is checked to see 5424109998Smarkm if it is explicitly trusted for OCSP signing. This is used to set 5425109998Smarkm a root CA as a global signing root: that is any certificate that 5426109998Smarkm chains to that CA is an acceptable OCSP signing certificate. 5427109998Smarkm [Steve Henson] 5428109998Smarkm 5429109998Smarkm *) New '-extfile ...' option to 'openssl ca' for reading X.509v3 5430109998Smarkm extensions from a separate configuration file. 5431109998Smarkm As when reading extensions from the main configuration file, 5432109998Smarkm the '-extensions ...' option may be used for specifying the 5433109998Smarkm section to use. 5434109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 5435109998Smarkm 5436109998Smarkm *) New OCSP utility. Allows OCSP requests to be generated or 5437109998Smarkm read. The request can be sent to a responder and the output 5438109998Smarkm parsed, outputed or printed in text form. Not complete yet: 5439109998Smarkm still needs to check the OCSP response validity. 5440109998Smarkm [Steve Henson] 5441109998Smarkm 5442109998Smarkm *) New subcommands for 'openssl ca': 5443109998Smarkm 'openssl ca -status <serial>' prints the status of the cert with 5444109998Smarkm the given serial number (according to the index file). 5445109998Smarkm 'openssl ca -updatedb' updates the expiry status of certificates 5446109998Smarkm in the index file. 5447109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 5448109998Smarkm 5449109998Smarkm *) New '-newreq-nodes' command option to CA.pl. This is like 5450109998Smarkm '-newreq', but calls 'openssl req' with the '-nodes' option 5451109998Smarkm so that the resulting key is not encrypted. 5452109998Smarkm [Damien Miller <djm@mindrot.org>] 5453109998Smarkm 5454109998Smarkm *) New configuration for the GNU Hurd. 5455109998Smarkm [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte] 5456109998Smarkm 5457109998Smarkm *) Initial code to implement OCSP basic response verify. This 5458109998Smarkm is currently incomplete. Currently just finds the signer's 5459109998Smarkm certificate and verifies the signature on the response. 5460109998Smarkm [Steve Henson] 5461109998Smarkm 5462109998Smarkm *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in 5463109998Smarkm value of OPENSSLDIR. This is available via the new '-d' option 5464109998Smarkm to 'openssl version', and is also included in 'openssl version -a'. 5465109998Smarkm [Bodo Moeller] 5466109998Smarkm 5467109998Smarkm *) Allowing defining memory allocation callbacks that will be given 5468109998Smarkm file name and line number information in additional arguments 5469109998Smarkm (a const char* and an int). The basic functionality remains, as 5470109998Smarkm well as the original possibility to just replace malloc(), 5471109998Smarkm realloc() and free() by functions that do not know about these 5472109998Smarkm additional arguments. To register and find out the current 5473109998Smarkm settings for extended allocation functions, the following 5474109998Smarkm functions are provided: 5475109998Smarkm 5476109998Smarkm CRYPTO_set_mem_ex_functions 5477109998Smarkm CRYPTO_set_locked_mem_ex_functions 5478109998Smarkm CRYPTO_get_mem_ex_functions 5479109998Smarkm CRYPTO_get_locked_mem_ex_functions 5480109998Smarkm 5481109998Smarkm These work the same way as CRYPTO_set_mem_functions and friends. 5482109998Smarkm CRYPTO_get_[locked_]mem_functions now writes 0 where such an 5483109998Smarkm extended allocation function is enabled. 5484109998Smarkm Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where 5485109998Smarkm a conventional allocation function is enabled. 5486109998Smarkm [Richard Levitte, Bodo Moeller] 5487109998Smarkm 5488109998Smarkm *) Finish off removing the remaining LHASH function pointer casts. 5489109998Smarkm There should no longer be any prototype-casting required when using 5490109998Smarkm the LHASH abstraction, and any casts that remain are "bugs". See 5491109998Smarkm the callback types and macros at the head of lhash.h for details 5492109998Smarkm (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). 5493109998Smarkm [Geoff Thorpe] 5494109998Smarkm 5495109998Smarkm *) Add automatic query of EGD sockets in RAND_poll() for the unix variant. 5496109998Smarkm If /dev/[u]random devices are not available or do not return enough 5497109998Smarkm entropy, EGD style sockets (served by EGD or PRNGD) will automatically 5498109998Smarkm be queried. 5499109998Smarkm The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and 5500109998Smarkm /etc/entropy will be queried once each in this sequence, quering stops 5501109998Smarkm when enough entropy was collected without querying more sockets. 5502109998Smarkm [Lutz Jaenicke] 5503109998Smarkm 5504109998Smarkm *) Change the Unix RAND_poll() variant to be able to poll several 5505109998Smarkm random devices, as specified by DEVRANDOM, until a sufficient amount 5506109998Smarkm of data has been collected. We spend at most 10 ms on each file 5507109998Smarkm (select timeout) and read in non-blocking mode. DEVRANDOM now 5508109998Smarkm defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" 5509109998Smarkm (previously it was just the string "/dev/urandom"), so on typical 5510109998Smarkm platforms the 10 ms delay will never occur. 5511109998Smarkm Also separate out the Unix variant to its own file, rand_unix.c. 5512109998Smarkm For VMS, there's a currently-empty rand_vms.c. 5513109998Smarkm [Richard Levitte] 5514109998Smarkm 5515109998Smarkm *) Move OCSP client related routines to ocsp_cl.c. These 5516109998Smarkm provide utility functions which an application needing 5517109998Smarkm to issue a request to an OCSP responder and analyse the 5518109998Smarkm response will typically need: as opposed to those which an 5519109998Smarkm OCSP responder itself would need which will be added later. 5520109998Smarkm 5521109998Smarkm OCSP_request_sign() signs an OCSP request with an API similar 5522109998Smarkm to PKCS7_sign(). OCSP_response_status() returns status of OCSP 5523109998Smarkm response. OCSP_response_get1_basic() extracts basic response 5524109998Smarkm from response. OCSP_resp_find_status(): finds and extracts status 5525109998Smarkm information from an OCSP_CERTID structure (which will be created 5526109998Smarkm when the request structure is built). These are built from lower 5527109998Smarkm level functions which work on OCSP_SINGLERESP structures but 5528109998Smarkm wont normally be used unless the application wishes to examine 5529109998Smarkm extensions in the OCSP response for example. 5530109998Smarkm 5531109998Smarkm Replace nonce routines with a pair of functions. 5532109998Smarkm OCSP_request_add1_nonce() adds a nonce value and optionally 5533109998Smarkm generates a random value. OCSP_check_nonce() checks the 5534109998Smarkm validity of the nonce in an OCSP response. 5535109998Smarkm [Steve Henson] 5536109998Smarkm 5537109998Smarkm *) Change function OCSP_request_add() to OCSP_request_add0_id(). 5538109998Smarkm This doesn't copy the supplied OCSP_CERTID and avoids the 5539109998Smarkm need to free up the newly created id. Change return type 5540109998Smarkm to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. 5541109998Smarkm This can then be used to add extensions to the request. 5542109998Smarkm Deleted OCSP_request_new(), since most of its functionality 5543109998Smarkm is now in OCSP_REQUEST_new() (and the case insensitive name 5544109998Smarkm clash) apart from the ability to set the request name which 5545109998Smarkm will be added elsewhere. 5546109998Smarkm [Steve Henson] 5547109998Smarkm 5548109998Smarkm *) Update OCSP API. Remove obsolete extensions argument from 5549109998Smarkm various functions. Extensions are now handled using the new 5550109998Smarkm OCSP extension code. New simple OCSP HTTP function which 5551109998Smarkm can be used to send requests and parse the response. 5552109998Smarkm [Steve Henson] 5553109998Smarkm 5554109998Smarkm *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new 5555109998Smarkm ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN 5556109998Smarkm uses the special reorder version of SET OF to sort the attributes 5557109998Smarkm and reorder them to match the encoded order. This resolves a long 5558109998Smarkm standing problem: a verify on a PKCS7 structure just after signing 5559109998Smarkm it used to fail because the attribute order did not match the 5560109998Smarkm encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: 5561109998Smarkm it uses the received order. This is necessary to tolerate some broken 5562109998Smarkm software that does not order SET OF. This is handled by encoding 5563109998Smarkm as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) 5564109998Smarkm to produce the required SET OF. 5565109998Smarkm [Steve Henson] 5566109998Smarkm 5567109998Smarkm *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and 5568109998Smarkm OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header 5569109998Smarkm files to get correct declarations of the ASN.1 item variables. 5570109998Smarkm [Richard Levitte] 5571109998Smarkm 5572109998Smarkm *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many 5573109998Smarkm PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: 5574109998Smarkm asn1_check_tlen() would sometimes attempt to use 'ctx' when it was 5575109998Smarkm NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). 5576109998Smarkm New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant 5577109998Smarkm ASN1_ITEM and no wrapper functions. 5578109998Smarkm [Steve Henson] 5579109998Smarkm 5580109998Smarkm *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These 5581109998Smarkm replace the old function pointer based I/O routines. Change most of 5582109998Smarkm the *_d2i_bio() and *_d2i_fp() functions to use these. 5583109998Smarkm [Steve Henson] 5584109998Smarkm 5585109998Smarkm *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor 5586109998Smarkm lines, recognice more "algorithms" that can be deselected, and make 5587109998Smarkm it complain about algorithm deselection that isn't recognised. 5588109998Smarkm [Richard Levitte] 5589109998Smarkm 5590109998Smarkm *) New ASN1 functions to handle dup, sign, verify, digest, pack and 5591109998Smarkm unpack operations in terms of ASN1_ITEM. Modify existing wrappers 5592109998Smarkm to use new functions. Add NO_ASN1_OLD which can be set to remove 5593109998Smarkm some old style ASN1 functions: this can be used to determine if old 5594109998Smarkm code will still work when these eventually go away. 5595109998Smarkm [Steve Henson] 5596109998Smarkm 5597109998Smarkm *) New extension functions for OCSP structures, these follow the 5598109998Smarkm same conventions as certificates and CRLs. 5599109998Smarkm [Steve Henson] 5600109998Smarkm 5601109998Smarkm *) New function X509V3_add1_i2d(). This automatically encodes and 5602109998Smarkm adds an extension. Its behaviour can be customised with various 5603109998Smarkm flags to append, replace or delete. Various wrappers added for 5604109998Smarkm certifcates and CRLs. 5605109998Smarkm [Steve Henson] 5606109998Smarkm 5607109998Smarkm *) Fix to avoid calling the underlying ASN1 print routine when 5608109998Smarkm an extension cannot be parsed. Correct a typo in the 5609109998Smarkm OCSP_SERVICELOC extension. Tidy up print OCSP format. 5610109998Smarkm [Steve Henson] 5611109998Smarkm 5612109998Smarkm *) Make mkdef.pl parse some of the ASN1 macros and add apropriate 5613109998Smarkm entries for variables. 5614109998Smarkm [Steve Henson] 5615109998Smarkm 5616109998Smarkm *) Add functionality to apps/openssl.c for detecting locking 5617109998Smarkm problems: As the program is single-threaded, all we have 5618109998Smarkm to do is register a locking callback using an array for 5619109998Smarkm storing which locks are currently held by the program. 5620109998Smarkm [Bodo Moeller] 5621109998Smarkm 5622109998Smarkm *) Use a lock around the call to CRYPTO_get_ex_new_index() in 5623109998Smarkm SSL_get_ex_data_X509_STORE_idx(), which is used in 5624109998Smarkm ssl_verify_cert_chain() and thus can be called at any time 5625109998Smarkm during TLS/SSL handshakes so that thread-safety is essential. 5626109998Smarkm Unfortunately, the ex_data design is not at all suited 5627109998Smarkm for multi-threaded use, so it probably should be abolished. 5628109998Smarkm [Bodo Moeller] 5629109998Smarkm 5630109998Smarkm *) Added Broadcom "ubsec" ENGINE to OpenSSL. 5631109998Smarkm [Broadcom, tweaked and integrated by Geoff Thorpe] 5632109998Smarkm 5633109998Smarkm *) Move common extension printing code to new function 5634109998Smarkm X509V3_print_extensions(). Reorganise OCSP print routines and 5635109998Smarkm implement some needed OCSP ASN1 functions. Add OCSP extensions. 5636109998Smarkm [Steve Henson] 5637109998Smarkm 5638109998Smarkm *) New function X509_signature_print() to remove duplication in some 5639109998Smarkm print routines. 5640109998Smarkm [Steve Henson] 5641109998Smarkm 5642109998Smarkm *) Add a special meaning when SET OF and SEQUENCE OF flags are both 5643109998Smarkm set (this was treated exactly the same as SET OF previously). This 5644109998Smarkm is used to reorder the STACK representing the structure to match the 5645109998Smarkm encoding. This will be used to get round a problem where a PKCS7 5646109998Smarkm structure which was signed could not be verified because the STACK 5647109998Smarkm order did not reflect the encoded order. 5648109998Smarkm [Steve Henson] 5649109998Smarkm 5650109998Smarkm *) Reimplement the OCSP ASN1 module using the new code. 5651109998Smarkm [Steve Henson] 5652109998Smarkm 5653109998Smarkm *) Update the X509V3 code to permit the use of an ASN1_ITEM structure 5654109998Smarkm for its ASN1 operations. The old style function pointers still exist 5655109998Smarkm for now but they will eventually go away. 5656109998Smarkm [Steve Henson] 5657109998Smarkm 5658109998Smarkm *) Merge in replacement ASN1 code from the ASN1 branch. This almost 5659109998Smarkm completely replaces the old ASN1 functionality with a table driven 5660109998Smarkm encoder and decoder which interprets an ASN1_ITEM structure describing 5661109998Smarkm the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is 5662109998Smarkm largely maintained. Almost all of the old asn1_mac.h macro based ASN1 5663109998Smarkm has also been converted to the new form. 5664109998Smarkm [Steve Henson] 5665109998Smarkm 5666109998Smarkm *) Change BN_mod_exp_recp so that negative moduli are tolerated 5667109998Smarkm (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set 5668109998Smarkm so that BN_mod_exp_mont and BN_mod_exp_mont_word work 5669109998Smarkm for negative moduli. 5670109998Smarkm [Bodo Moeller] 5671109998Smarkm 5672109998Smarkm *) Fix BN_uadd and BN_usub: Always return non-negative results instead 5673109998Smarkm of not touching the result's sign bit. 5674109998Smarkm [Bodo Moeller] 5675109998Smarkm 5676109998Smarkm *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be 5677109998Smarkm set. 5678109998Smarkm [Bodo Moeller] 5679109998Smarkm 5680109998Smarkm *) Changed the LHASH code to use prototypes for callbacks, and created 5681109998Smarkm macros to declare and implement thin (optionally static) functions 5682109998Smarkm that provide type-safety and avoid function pointer casting for the 5683109998Smarkm type-specific callbacks. 5684109998Smarkm [Geoff Thorpe] 5685109998Smarkm 5686109998Smarkm *) Added Kerberos Cipher Suites to be used with TLS, as written in 5687109998Smarkm RFC 2712. 5688109998Smarkm [Veers Staats <staatsvr@asc.hpc.mil>, 5689109998Smarkm Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte] 5690109998Smarkm 5691109998Smarkm *) Reformat the FAQ so the different questions and answers can be divided 5692109998Smarkm in sections depending on the subject. 5693109998Smarkm [Richard Levitte] 5694109998Smarkm 5695109998Smarkm *) Have the zlib compression code load ZLIB.DLL dynamically under 5696109998Smarkm Windows. 5697109998Smarkm [Richard Levitte] 5698109998Smarkm 5699109998Smarkm *) New function BN_mod_sqrt for computing square roots modulo a prime 5700109998Smarkm (using the probabilistic Tonelli-Shanks algorithm unless 5701109998Smarkm p == 3 (mod 4) or p == 5 (mod 8), which are cases that can 5702109998Smarkm be handled deterministically). 5703109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 5704109998Smarkm 5705109998Smarkm *) Make BN_mod_inverse faster by explicitly handling small quotients 5706109998Smarkm in the Euclid loop. (Speed gain about 20% for small moduli [256 or 5707109998Smarkm 512 bits], about 30% for larger ones [1024 or 2048 bits].) 5708109998Smarkm [Bodo Moeller] 5709109998Smarkm 5710109998Smarkm *) New function BN_kronecker. 5711109998Smarkm [Bodo Moeller] 5712109998Smarkm 5713109998Smarkm *) Fix BN_gcd so that it works on negative inputs; the result is 5714109998Smarkm positive unless both parameters are zero. 5715109998Smarkm Previously something reasonably close to an infinite loop was 5716109998Smarkm possible because numbers could be growing instead of shrinking 5717109998Smarkm in the implementation of Euclid's algorithm. 5718109998Smarkm [Bodo Moeller] 5719109998Smarkm 5720109998Smarkm *) Fix BN_is_word() and BN_is_one() macros to take into account the 5721109998Smarkm sign of the number in question. 5722109998Smarkm 5723109998Smarkm Fix BN_is_word(a,w) to work correctly for w == 0. 5724109998Smarkm 5725109998Smarkm The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) 5726109998Smarkm because its test if the absolute value of 'a' equals 'w'. 5727109998Smarkm Note that BN_abs_is_word does *not* handle w == 0 reliably; 5728109998Smarkm it exists mostly for use in the implementations of BN_is_zero(), 5729109998Smarkm BN_is_one(), and BN_is_word(). 5730109998Smarkm [Bodo Moeller] 5731109998Smarkm 5732109998Smarkm *) New function BN_swap. 5733109998Smarkm [Bodo Moeller] 5734109998Smarkm 5735109998Smarkm *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that 5736109998Smarkm the exponentiation functions are more likely to produce reasonable 5737109998Smarkm results on negative inputs. 5738109998Smarkm [Bodo Moeller] 5739109998Smarkm 5740109998Smarkm *) Change BN_mod_mul so that the result is always non-negative. 5741109998Smarkm Previously, it could be negative if one of the factors was negative; 5742109998Smarkm I don't think anyone really wanted that behaviour. 5743109998Smarkm [Bodo Moeller] 5744109998Smarkm 5745109998Smarkm *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c 5746109998Smarkm (except for exponentiation, which stays in crypto/bn/bn_exp.c, 5747109998Smarkm and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c) 5748109998Smarkm and add new functions: 5749109998Smarkm 5750109998Smarkm BN_nnmod 5751109998Smarkm BN_mod_sqr 5752109998Smarkm BN_mod_add 5753109998Smarkm BN_mod_add_quick 5754109998Smarkm BN_mod_sub 5755109998Smarkm BN_mod_sub_quick 5756109998Smarkm BN_mod_lshift1 5757109998Smarkm BN_mod_lshift1_quick 5758109998Smarkm BN_mod_lshift 5759109998Smarkm BN_mod_lshift_quick 5760109998Smarkm 5761109998Smarkm These functions always generate non-negative results. 5762109998Smarkm 5763109998Smarkm BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r 5764109998Smarkm such that |m| < r < 0, BN_nnmod will output rem + |m| instead). 5765109998Smarkm 5766109998Smarkm BN_mod_XXX_quick(r, a, [b,] m) generates the same result as 5767109998Smarkm BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b] 5768109998Smarkm be reduced modulo m. 5769109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 5770109998Smarkm 5771111147Snectar#if 0 5772111147Snectar The following entry accidentily appeared in the CHANGES file 5773111147Snectar distributed with OpenSSL 0.9.7. The modifications described in 5774111147Snectar it do *not* apply to OpenSSL 0.9.7. 5775111147Snectar 5776109998Smarkm *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 5777109998Smarkm was actually never needed) and in BN_mul(). The removal in BN_mul() 5778109998Smarkm required a small change in bn_mul_part_recursive() and the addition 5779109998Smarkm of the functions bn_cmp_part_words(), bn_sub_part_words() and 5780109998Smarkm bn_add_part_words(), which do the same thing as bn_cmp_words(), 5781109998Smarkm bn_sub_words() and bn_add_words() except they take arrays with 5782109998Smarkm differing sizes. 5783109998Smarkm [Richard Levitte] 5784111147Snectar#endif 5785109998Smarkm 5786109998Smarkm *) In 'openssl passwd', verify passwords read from the terminal 5787109998Smarkm unless the '-salt' option is used (which usually means that 5788109998Smarkm verification would just waste user's time since the resulting 5789109998Smarkm hash is going to be compared with some given password hash) 5790109998Smarkm or the new '-noverify' option is used. 5791109998Smarkm 5792109998Smarkm This is an incompatible change, but it does not affect 5793109998Smarkm non-interactive use of 'openssl passwd' (passwords on the command 5794109998Smarkm line, '-stdin' option, '-in ...' option) and thus should not 5795109998Smarkm cause any problems. 5796109998Smarkm [Bodo Moeller] 5797109998Smarkm 5798109998Smarkm *) Remove all references to RSAref, since there's no more need for it. 5799109998Smarkm [Richard Levitte] 5800109998Smarkm 5801109998Smarkm *) Make DSO load along a path given through an environment variable 5802109998Smarkm (SHLIB_PATH) with shl_load(). 5803109998Smarkm [Richard Levitte] 5804109998Smarkm 5805109998Smarkm *) Constify the ENGINE code as a result of BIGNUM constification. 5806109998Smarkm Also constify the RSA code and most things related to it. In a 5807109998Smarkm few places, most notable in the depth of the ASN.1 code, ugly 5808109998Smarkm casts back to non-const were required (to be solved at a later 5809109998Smarkm time) 5810109998Smarkm [Richard Levitte] 5811109998Smarkm 5812109998Smarkm *) Make it so the openssl application has all engines loaded by default. 5813109998Smarkm [Richard Levitte] 5814109998Smarkm 5815109998Smarkm *) Constify the BIGNUM routines a little more. 5816109998Smarkm [Richard Levitte] 5817109998Smarkm 5818109998Smarkm *) Add the following functions: 5819109998Smarkm 5820109998Smarkm ENGINE_load_cswift() 5821109998Smarkm ENGINE_load_chil() 5822109998Smarkm ENGINE_load_atalla() 5823109998Smarkm ENGINE_load_nuron() 5824109998Smarkm ENGINE_load_builtin_engines() 5825109998Smarkm 5826109998Smarkm That way, an application can itself choose if external engines that 5827109998Smarkm are built-in in OpenSSL shall ever be used or not. The benefit is 5828109998Smarkm that applications won't have to be linked with libdl or other dso 5829109998Smarkm libraries unless it's really needed. 5830109998Smarkm 5831109998Smarkm Changed 'openssl engine' to load all engines on demand. 5832109998Smarkm Changed the engine header files to avoid the duplication of some 5833109998Smarkm declarations (they differed!). 5834109998Smarkm [Richard Levitte] 5835109998Smarkm 5836109998Smarkm *) 'openssl engine' can now list capabilities. 5837109998Smarkm [Richard Levitte] 5838109998Smarkm 5839109998Smarkm *) Better error reporting in 'openssl engine'. 5840109998Smarkm [Richard Levitte] 5841109998Smarkm 5842109998Smarkm *) Never call load_dh_param(NULL) in s_server. 5843109998Smarkm [Bodo Moeller] 5844109998Smarkm 5845109998Smarkm *) Add engine application. It can currently list engines by name and 5846109998Smarkm identity, and test if they are actually available. 5847109998Smarkm [Richard Levitte] 5848109998Smarkm 5849109998Smarkm *) Improve RPM specification file by forcing symbolic linking and making 5850109998Smarkm sure the installed documentation is also owned by root.root. 5851109998Smarkm [Damien Miller <djm@mindrot.org>] 5852109998Smarkm 5853109998Smarkm *) Give the OpenSSL applications more possibilities to make use of 5854109998Smarkm keys (public as well as private) handled by engines. 5855109998Smarkm [Richard Levitte] 5856109998Smarkm 5857109998Smarkm *) Add OCSP code that comes from CertCo. 5858109998Smarkm [Richard Levitte] 5859109998Smarkm 5860109998Smarkm *) Add VMS support for the Rijndael code. 5861109998Smarkm [Richard Levitte] 5862109998Smarkm 5863109998Smarkm *) Added untested support for Nuron crypto accelerator. 5864109998Smarkm [Ben Laurie] 5865109998Smarkm 5866109998Smarkm *) Add support for external cryptographic devices. This code was 5867109998Smarkm previously distributed separately as the "engine" branch. 5868109998Smarkm [Geoff Thorpe, Richard Levitte] 5869109998Smarkm 5870109998Smarkm *) Rework the filename-translation in the DSO code. It is now possible to 5871109998Smarkm have far greater control over how a "name" is turned into a filename 5872109998Smarkm depending on the operating environment and any oddities about the 5873109998Smarkm different shared library filenames on each system. 5874109998Smarkm [Geoff Thorpe] 5875109998Smarkm 5876109998Smarkm *) Support threads on FreeBSD-elf in Configure. 5877109998Smarkm [Richard Levitte] 5878109998Smarkm 5879109998Smarkm *) Fix for SHA1 assembly problem with MASM: it produces 5880109998Smarkm warnings about corrupt line number information when assembling 5881109998Smarkm with debugging information. This is caused by the overlapping 5882109998Smarkm of two sections. 5883109998Smarkm [Bernd Matthes <mainbug@celocom.de>, Steve Henson] 5884109998Smarkm 5885109998Smarkm *) NCONF changes. 5886109998Smarkm NCONF_get_number() has no error checking at all. As a replacement, 5887109998Smarkm NCONF_get_number_e() is defined (_e for "error checking") and is 5888109998Smarkm promoted strongly. The old NCONF_get_number is kept around for 5889109998Smarkm binary backward compatibility. 5890109998Smarkm Make it possible for methods to load from something other than a BIO, 5891109998Smarkm by providing a function pointer that is given a name instead of a BIO. 5892109998Smarkm For example, this could be used to load configuration data from an 5893109998Smarkm LDAP server. 5894109998Smarkm [Richard Levitte] 5895109998Smarkm 5896109998Smarkm *) Fix for non blocking accept BIOs. Added new I/O special reason 5897109998Smarkm BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs 5898109998Smarkm with non blocking I/O was not possible because no retry code was 5899109998Smarkm implemented. Also added new SSL code SSL_WANT_ACCEPT to cover 5900109998Smarkm this case. 5901109998Smarkm [Steve Henson] 5902109998Smarkm 5903109998Smarkm *) Added the beginnings of Rijndael support. 5904109998Smarkm [Ben Laurie] 5905109998Smarkm 5906109998Smarkm *) Fix for bug in DirectoryString mask setting. Add support for 5907109998Smarkm X509_NAME_print_ex() in 'req' and X509_print_ex() function 5908109998Smarkm to allow certificate printing to more controllable, additional 5909109998Smarkm 'certopt' option to 'x509' to allow new printing options to be 5910109998Smarkm set. 5911109998Smarkm [Steve Henson] 5912109998Smarkm 5913109998Smarkm *) Clean old EAY MD5 hack from e_os.h. 5914109998Smarkm [Richard Levitte] 5915109998Smarkm 5916142425Snectar Changes between 0.9.6l and 0.9.6m [17 Mar 2004] 5917142425Snectar 5918142425Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 5919160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 5920142425Snectar [Joe Orton, Steve Henson] 5921142425Snectar 5922142425Snectar Changes between 0.9.6k and 0.9.6l [04 Nov 2003] 5923142425Snectar 5924142425Snectar *) Fix additional bug revealed by the NISCC test suite: 5925142425Snectar 5926142425Snectar Stop bug triggering large recursion when presented with 5927160814Ssimon certain ASN.1 tags (CVE-2003-0851) 5928142425Snectar [Steve Henson] 5929142425Snectar 5930120631Snectar Changes between 0.9.6j and 0.9.6k [30 Sep 2003] 5931120631Snectar 5932120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 5933120631Snectar 5934120631Snectar Stop out of bounds reads in the ASN1 code when presented with 5935160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 5936120631Snectar 5937120631Snectar If verify callback ignores invalid public key errors don't try to check 5938120631Snectar certificate signature with the NULL public key. 5939120631Snectar 5940120631Snectar [Steve Henson] 5941120631Snectar 5942120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 5943120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 5944120631Snectar specifications. 5945120631Snectar [Steve Henson] 5946120631Snectar 5947120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 5948120631Snectar extra data after the compression methods not only for TLS 1.0 5949120631Snectar but also for SSL 3.0 (as required by the specification). 5950120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 5951120631Snectar 5952120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 5953120631Snectar when it's 512 *bits* long, not 512 bytes. 5954120631Snectar [Richard Levitte] 5955120631Snectar 5956120631Snectar Changes between 0.9.6i and 0.9.6j [10 Apr 2003] 5957120631Snectar 5958120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 5959120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 5960120631Snectar a protocol version number mismatch like a decryption error 5961120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 5962120631Snectar [Bodo Moeller] 5963120631Snectar 5964120631Snectar *) Turn on RSA blinding by default in the default implementation 5965120631Snectar to avoid a timing attack. Applications that don't want it can call 5966120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 5967120631Snectar They would be ill-advised to do so in most cases. 5968120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 5969120631Snectar 5970120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 5971120631Snectar seeded (in this case, the secret RSA exponent is abused as 5972120631Snectar an unpredictable seed -- if it is not unpredictable, there 5973120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 5974120631Snectar by remembering the creator's thread ID in rsa->blinding and 5975120631Snectar having all other threads use local one-time blinding factors 5976120631Snectar (this requires more computation than sharing rsa->blinding, but 5977120631Snectar avoids excessive locking; and if an RSA object is not shared 5978120631Snectar between threads, blinding will still be very fast). 5979120631Snectar [Bodo Moeller] 5980120631Snectar 5981111147Snectar Changes between 0.9.6h and 0.9.6i [19 Feb 2003] 5982111147Snectar 5983111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 5984111147Snectar via timing by performing a MAC computation even if incorrrect 5985111147Snectar block cipher padding has been found. This is a countermeasure 5986111147Snectar against active attacks where the attacker has to distinguish 5987160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 5988111147Snectar 5989111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 5990111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 5991111147Snectar Martin Vuagnoux (EPFL, Ilion)] 5992111147Snectar 5993109998Smarkm Changes between 0.9.6g and 0.9.6h [5 Dec 2002] 5994109998Smarkm 5995109998Smarkm *) New function OPENSSL_cleanse(), which is used to cleanse a section of 5996109998Smarkm memory from it's contents. This is done with a counter that will 5997109998Smarkm place alternating values in each byte. This can be used to solve 5998109998Smarkm two issues: 1) the removal of calls to memset() by highly optimizing 5999109998Smarkm compilers, and 2) cleansing with other values than 0, since those can 6000109998Smarkm be read through on certain media, for example a swap space on disk. 6001109998Smarkm [Geoff Thorpe] 6002109998Smarkm 6003109998Smarkm *) Bugfix: client side session caching did not work with external caching, 6004109998Smarkm because the session->cipher setting was not restored when reloading 6005109998Smarkm from the external cache. This problem was masked, when 6006109998Smarkm SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. 6007109998Smarkm (Found by Steve Haslam <steve@araqnid.ddts.net>.) 6008109998Smarkm [Lutz Jaenicke] 6009109998Smarkm 6010109998Smarkm *) Fix client_certificate (ssl/s2_clnt.c): The permissible total 6011109998Smarkm length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. 6012109998Smarkm [Zeev Lieber <zeev-l@yahoo.com>] 6013109998Smarkm 6014109998Smarkm *) Undo an undocumented change introduced in 0.9.6e which caused 6015109998Smarkm repeated calls to OpenSSL_add_all_ciphers() and 6016109998Smarkm OpenSSL_add_all_digests() to be ignored, even after calling 6017109998Smarkm EVP_cleanup(). 6018109998Smarkm [Richard Levitte] 6019109998Smarkm 6020109998Smarkm *) Change the default configuration reader to deal with last line not 6021109998Smarkm being properly terminated. 6022109998Smarkm [Richard Levitte] 6023109998Smarkm 6024109998Smarkm *) Change X509_NAME_cmp() so it applies the special rules on handling 6025109998Smarkm DN values that are of type PrintableString, as well as RDNs of type 6026109998Smarkm emailAddress where the value has the type ia5String. 6027109998Smarkm [stefank@valicert.com via Richard Levitte] 6028109998Smarkm 6029109998Smarkm *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half 6030109998Smarkm the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently 6031109998Smarkm doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be 6032109998Smarkm the bitwise-OR of the two for use by the majority of applications 6033109998Smarkm wanting this behaviour, and update the docs. The documented 6034109998Smarkm behaviour and actual behaviour were inconsistent and had been 6035109998Smarkm changing anyway, so this is more a bug-fix than a behavioural 6036109998Smarkm change. 6037109998Smarkm [Geoff Thorpe, diagnosed by Nadav Har'El] 6038109998Smarkm 6039109998Smarkm *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c 6040109998Smarkm (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). 6041109998Smarkm [Bodo Moeller] 6042109998Smarkm 6043109998Smarkm *) Fix initialization code race conditions in 6044109998Smarkm SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), 6045109998Smarkm SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), 6046109998Smarkm SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), 6047109998Smarkm TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), 6048109998Smarkm ssl2_get_cipher_by_char(), 6049109998Smarkm ssl3_get_cipher_by_char(). 6050109998Smarkm [Patrick McCormick <patrick@tellme.com>, Bodo Moeller] 6051109998Smarkm 6052109998Smarkm *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after 6053109998Smarkm the cached sessions are flushed, as the remove_cb() might use ex_data 6054109998Smarkm contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> 6055109998Smarkm (see [openssl.org #212]). 6056109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 6057109998Smarkm 6058109998Smarkm *) Fix typo in OBJ_txt2obj which incorrectly passed the content 6059109998Smarkm length, instead of the encoding length to d2i_ASN1_OBJECT. 6060109998Smarkm [Steve Henson] 6061109998Smarkm 6062101618Snectar Changes between 0.9.6f and 0.9.6g [9 Aug 2002] 6063101618Snectar 6064101618Snectar *) [In 0.9.6g-engine release:] 6065101618Snectar Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall'). 6066101618Snectar [Lynn Gazis <lgazis@rainbow.com>] 6067101618Snectar 6068101613Snectar Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 6069101613Snectar 6070101613Snectar *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 6071101613Snectar and get fix the header length calculation. 6072101613Snectar [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 6073101613Snectar Alon Kantor <alonk@checkpoint.com> (and others), 6074101613Snectar Steve Henson] 6075101613Snectar 6076101613Snectar *) Use proper error handling instead of 'assertions' in buffer 6077101613Snectar overflow checks added in 0.9.6e. This prevents DoS (the 6078101613Snectar assertions could call abort()). 6079101613Snectar [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] 6080101613Snectar 6081100936Snectar Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 6082100936Snectar 6083109998Smarkm *) Add various sanity checks to asn1_get_length() to reject 6084109998Smarkm the ASN1 length bytes if they exceed sizeof(long), will appear 6085109998Smarkm negative or the content length exceeds the length of the 6086109998Smarkm supplied buffer. 6087109998Smarkm [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 6088109998Smarkm 6089100936Snectar *) Fix cipher selection routines: ciphers without encryption had no flags 6090100936Snectar for the cipher strength set and where therefore not handled correctly 6091100936Snectar by the selection routines (PR #130). 6092100936Snectar [Lutz Jaenicke] 6093100936Snectar 6094100936Snectar *) Fix EVP_dsa_sha macro. 6095100936Snectar [Nils Larsch] 6096100936Snectar 6097100936Snectar *) New option 6098100936Snectar SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 6099100936Snectar for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 6100100936Snectar that was added in OpenSSL 0.9.6d. 6101100936Snectar 6102100936Snectar As the countermeasure turned out to be incompatible with some 6103100936Snectar broken SSL implementations, the new option is part of SSL_OP_ALL. 6104100936Snectar SSL_OP_ALL is usually employed when compatibility with weird SSL 6105100936Snectar implementations is desired (e.g. '-bugs' option to 's_client' and 6106100936Snectar 's_server'), so the new option is automatically set in many 6107100936Snectar applications. 6108100936Snectar [Bodo Moeller] 6109100936Snectar 6110100936Snectar *) Changes in security patch: 6111100936Snectar 6112100936Snectar Changes marked "(CHATS)" were sponsored by the Defense Advanced 6113100936Snectar Research Projects Agency (DARPA) and Air Force Research Laboratory, 6114100936Snectar Air Force Materiel Command, USAF, under agreement number 6115100936Snectar F30602-01-2-0537. 6116100936Snectar 6117100936Snectar *) Add various sanity checks to asn1_get_length() to reject 6118100936Snectar the ASN1 length bytes if they exceed sizeof(long), will appear 6119100936Snectar negative or the content length exceeds the length of the 6120160814Ssimon supplied buffer. (CVE-2002-0659) 6121100936Snectar [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 6122100936Snectar 6123100936Snectar *) Assertions for various potential buffer overflows, not known to 6124100936Snectar happen in practice. 6125100936Snectar [Ben Laurie (CHATS)] 6126100936Snectar 6127100936Snectar *) Various temporary buffers to hold ASCII versions of integers were 6128160814Ssimon too small for 64 bit platforms. (CVE-2002-0655) 6129100936Snectar [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 6130100936Snectar 6131100936Snectar *) Remote buffer overflow in SSL3 protocol - an attacker could 6132160814Ssimon supply an oversized session ID to a client. (CVE-2002-0656) 6133100936Snectar [Ben Laurie (CHATS)] 6134100936Snectar 6135100936Snectar *) Remote buffer overflow in SSL2 protocol - an attacker could 6136160814Ssimon supply an oversized client master key. (CVE-2002-0656) 6137100936Snectar [Ben Laurie (CHATS)] 6138100936Snectar 6139100928Snectar Changes between 0.9.6c and 0.9.6d [9 May 2002] 6140100928Snectar 6141100928Snectar *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 6142100928Snectar encoded as NULL) with id-dsa-with-sha1. 6143100928Snectar [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 6144100928Snectar 6145100928Snectar *) Check various X509_...() return values in apps/req.c. 6146100928Snectar [Nils Larsch <nla@trustcenter.de>] 6147100928Snectar 6148100928Snectar *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 6149100928Snectar an end-of-file condition would erronously be flagged, when the CRLF 6150100928Snectar was just at the end of a processed block. The bug was discovered when 6151100928Snectar processing data through a buffering memory BIO handing the data to a 6152100928Snectar BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 6153100928Snectar <ptsekov@syntrex.com> and Nedelcho Stanev. 6154100928Snectar [Lutz Jaenicke] 6155100928Snectar 6156100928Snectar *) Implement a countermeasure against a vulnerability recently found 6157100928Snectar in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 6158100928Snectar before application data chunks to avoid the use of known IVs 6159100928Snectar with data potentially chosen by the attacker. 6160100928Snectar [Bodo Moeller] 6161100928Snectar 6162100928Snectar *) Fix length checks in ssl3_get_client_hello(). 6163100928Snectar [Bodo Moeller] 6164100928Snectar 6165100928Snectar *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 6166100928Snectar to prevent ssl3_read_internal() from incorrectly assuming that 6167100928Snectar ssl3_read_bytes() found application data while handshake 6168100928Snectar processing was enabled when in fact s->s3->in_read_app_data was 6169100928Snectar merely automatically cleared during the initial handshake. 6170100928Snectar [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 6171100928Snectar 6172100928Snectar *) Fix object definitions for Private and Enterprise: they were not 6173100928Snectar recognized in their shortname (=lowercase) representation. Extend 6174100928Snectar obj_dat.pl to issue an error when using undefined keywords instead 6175100928Snectar of silently ignoring the problem (Svenning Sorensen 6176100928Snectar <sss@sss.dnsalias.net>). 6177100928Snectar [Lutz Jaenicke] 6178100928Snectar 6179100928Snectar *) Fix DH_generate_parameters() so that it works for 'non-standard' 6180100928Snectar generators, i.e. generators other than 2 and 5. (Previously, the 6181100928Snectar code did not properly initialise the 'add' and 'rem' values to 6182100928Snectar BN_generate_prime().) 6183100928Snectar 6184100928Snectar In the new general case, we do not insist that 'generator' is 6185100928Snectar actually a primitive root: This requirement is rather pointless; 6186100928Snectar a generator of the order-q subgroup is just as good, if not 6187100928Snectar better. 6188100928Snectar [Bodo Moeller] 6189100928Snectar 6190100928Snectar *) Map new X509 verification errors to alerts. Discovered and submitted by 6191100928Snectar Tom Wu <tom@arcot.com>. 6192100928Snectar [Lutz Jaenicke] 6193100928Snectar 6194100928Snectar *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 6195100928Snectar returning non-zero before the data has been completely received 6196100928Snectar when using non-blocking I/O. 6197100928Snectar [Bodo Moeller; problem pointed out by John Hughes] 6198100928Snectar 6199100928Snectar *) Some of the ciphers missed the strength entry (SSL_LOW etc). 6200100928Snectar [Ben Laurie, Lutz Jaenicke] 6201100928Snectar 6202100928Snectar *) Fix bug in SSL_clear(): bad sessions were not removed (found by 6203100928Snectar Yoram Zahavi <YoramZ@gilian.com>). 6204100928Snectar [Lutz Jaenicke] 6205100928Snectar 6206100928Snectar *) Add information about CygWin 1.3 and on, and preserve proper 6207100928Snectar configuration for the versions before that. 6208100928Snectar [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 6209100928Snectar 6210100928Snectar *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 6211100928Snectar check whether we deal with a copy of a session and do not delete from 6212100928Snectar the cache in this case. Problem reported by "Izhar Shoshani Levi" 6213100928Snectar <izhar@checkpoint.com>. 6214100928Snectar [Lutz Jaenicke] 6215100928Snectar 6216100928Snectar *) Do not store session data into the internal session cache, if it 6217100928Snectar is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 6218100928Snectar flag is set). Proposed by Aslam <aslam@funk.com>. 6219100928Snectar [Lutz Jaenicke] 6220100928Snectar 6221100928Snectar *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 6222100928Snectar value is 0. 6223100928Snectar [Richard Levitte] 6224100928Snectar 6225109998Smarkm *) [In 0.9.6d-engine release:] 6226109998Smarkm Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 6227100928Snectar [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 6228100928Snectar 6229100928Snectar *) Add the configuration target linux-s390x. 6230100928Snectar [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 6231100928Snectar 6232100928Snectar *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 6233100928Snectar ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 6234100928Snectar variable as an indication that a ClientHello message has been 6235100928Snectar received. As the flag value will be lost between multiple 6236100928Snectar invocations of ssl3_accept when using non-blocking I/O, the 6237100928Snectar function may not be aware that a handshake has actually taken 6238100928Snectar place, thus preventing a new session from being added to the 6239100928Snectar session cache. 6240100928Snectar 6241100928Snectar To avoid this problem, we now set s->new_session to 2 instead of 6242100928Snectar using a local variable. 6243100928Snectar [Lutz Jaenicke, Bodo Moeller] 6244100928Snectar 6245100928Snectar *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 6246100928Snectar if the SSL_R_LENGTH_MISMATCH error is detected. 6247100928Snectar [Geoff Thorpe, Bodo Moeller] 6248100928Snectar 6249100928Snectar *) New 'shared_ldflag' column in Configure platform table. 6250100928Snectar [Richard Levitte] 6251100928Snectar 6252100928Snectar *) Fix EVP_CIPHER_mode macro. 6253100928Snectar ["Dan S. Camper" <dan@bti.net>] 6254100928Snectar 6255100928Snectar *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 6256100928Snectar type, we must throw them away by setting rr->length to 0. 6257100928Snectar [D P Chang <dpc@qualys.com>] 6258100928Snectar 625989837Skris Changes between 0.9.6b and 0.9.6c [21 dec 2001] 626089837Skris 626189837Skris *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 626289837Skris <Dominikus.Scherkl@biodata.com>. (The previous implementation 626389837Skris worked incorrectly for those cases where range = 10..._2 and 626489837Skris 3*range is two bits longer than range.) 626589837Skris [Bodo Moeller] 626689837Skris 626789837Skris *) Only add signing time to PKCS7 structures if it is not already 626889837Skris present. 626989837Skris [Steve Henson] 627089837Skris 627189837Skris *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 627289837Skris OBJ_ld_ce should be OBJ_id_ce. 627389837Skris Also some ip-pda OIDs in crypto/objects/objects.txt were 627489837Skris incorrect (cf. RFC 3039). 627589837Skris [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 627689837Skris 627789837Skris *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 627889837Skris returns early because it has nothing to do. 627989837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 628089837Skris 628189837Skris *) [In 0.9.6c-engine release:] 628289837Skris Fix mutex callback return values in crypto/engine/hw_ncipher.c. 628389837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 628489837Skris 628589837Skris *) [In 0.9.6c-engine release:] 628689837Skris Add support for Cryptographic Appliance's keyserver technology. 628789837Skris (Use engine 'keyclient') 628889837Skris [Cryptographic Appliances and Geoff Thorpe] 628989837Skris 629089837Skris *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 629189837Skris is called via tools/c89.sh because arguments have to be 629289837Skris rearranged (all '-L' options must appear before the first object 629389837Skris modules). 629489837Skris [Richard Shapiro <rshapiro@abinitio.com>] 629589837Skris 629689837Skris *) [In 0.9.6c-engine release:] 629789837Skris Add support for Broadcom crypto accelerator cards, backported 629889837Skris from 0.9.7. 629989837Skris [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 630089837Skris 630189837Skris *) [In 0.9.6c-engine release:] 630289837Skris Add support for SureWare crypto accelerator cards from 630389837Skris Baltimore Technologies. (Use engine 'sureware') 630489837Skris [Baltimore Technologies and Mark Cox] 630589837Skris 630689837Skris *) [In 0.9.6c-engine release:] 630789837Skris Add support for crypto accelerator cards from Accelerated 630889837Skris Encryption Processing, www.aep.ie. (Use engine 'aep') 630989837Skris [AEP Inc. and Mark Cox] 631089837Skris 631189837Skris *) Add a configuration entry for gcc on UnixWare. 631289837Skris [Gary Benson <gbenson@redhat.com>] 631389837Skris 631489837Skris *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 631589837Skris messages are stored in a single piece (fixed-length part and 631689837Skris variable-length part combined) and fix various bugs found on the way. 631789837Skris [Bodo Moeller] 631889837Skris 631989837Skris *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 632089837Skris instead. BIO_gethostbyname() does not know what timeouts are 632189837Skris appropriate, so entries would stay in cache even when they have 632289837Skris become invalid. 632389837Skris [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 632489837Skris 632589837Skris *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 632689837Skris faced with a pathologically small ClientHello fragment that does 632789837Skris not contain client_version: Instead of aborting with an error, 632889837Skris simply choose the highest available protocol version (i.e., 632989837Skris TLS 1.0 unless it is disabled). In practice, ClientHello 633089837Skris messages are never sent like this, but this change gives us 633189837Skris strictly correct behaviour at least for TLS. 633289837Skris [Bodo Moeller] 633389837Skris 633489837Skris *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 633589837Skris never resets s->method to s->ctx->method when called from within 633689837Skris one of the SSL handshake functions. 633789837Skris [Bodo Moeller; problem pointed out by Niko Baric] 633889837Skris 633989837Skris *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 634089837Skris (sent using the client's version number) if client_version is 634189837Skris smaller than the protocol version in use. Also change 634289837Skris ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 634389837Skris the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 634489837Skris the client will at least see that alert. 634589837Skris [Bodo Moeller] 634689837Skris 634789837Skris *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 634889837Skris correctly. 634989837Skris [Bodo Moeller] 635089837Skris 635189837Skris *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 635289837Skris client receives HelloRequest while in a handshake. 635389837Skris [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 635489837Skris 635589837Skris *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 635689837Skris should end in 'break', not 'goto end' which circuments various 635789837Skris cleanups done in state SSL_ST_OK. But session related stuff 635889837Skris must be disabled for SSL_ST_OK in the case that we just sent a 635989837Skris HelloRequest. 636089837Skris 636189837Skris Also avoid some overhead by not calling ssl_init_wbio_buffer() 636289837Skris before just sending a HelloRequest. 636389837Skris [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 636489837Skris 636589837Skris *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 636689837Skris reveal whether illegal block cipher padding was found or a MAC 636789837Skris verification error occured. (Neither SSLerr() codes nor alerts 636889837Skris are directly visible to potential attackers, but the information 636989837Skris may leak via logfiles.) 637089837Skris 637189837Skris Similar changes are not required for the SSL 2.0 implementation 637289837Skris because the number of padding bytes is sent in clear for SSL 2.0, 637389837Skris and the extra bytes are just ignored. However ssl/s2_pkt.c 637489837Skris failed to verify that the purported number of padding bytes is in 637589837Skris the legal range. 637689837Skris [Bodo Moeller] 637789837Skris 637889837Skris *) Add OpenUNIX-8 support including shared libraries 637989837Skris (Boyd Lynn Gerber <gerberb@zenez.com>). 638089837Skris [Lutz Jaenicke] 638189837Skris 638289837Skris *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 638389837Skris 'wristwatch attack' using huge encoding parameters (cf. 638489837Skris James H. Manger's CRYPTO 2001 paper). Note that the 638589837Skris RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 638689837Skris encoding parameters and hence was not vulnerable. 638789837Skris [Bodo Moeller] 638889837Skris 638989837Skris *) BN_sqr() bug fix. 639089837Skris [Ulf M�ller, reported by Jim Ellis <jim.ellis@cavium.com>] 639189837Skris 639289837Skris *) Rabin-Miller test analyses assume uniformly distributed witnesses, 639389837Skris so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 639489837Skris followed by modular reduction. 639589837Skris [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 639689837Skris 639789837Skris *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 639889837Skris equivalent based on BN_pseudo_rand() instead of BN_rand(). 639989837Skris [Bodo Moeller] 640089837Skris 640189837Skris *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 640289837Skris This function was broken, as the check for a new client hello message 640389837Skris to handle SGC did not allow these large messages. 640489837Skris (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 640589837Skris [Lutz Jaenicke] 640689837Skris 640789837Skris *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 640889837Skris [Lutz Jaenicke] 640989837Skris 641089837Skris *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 641189837Skris for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 641289837Skris [Lutz Jaenicke] 641389837Skris 641489837Skris *) Rework the configuration and shared library support for Tru64 Unix. 641589837Skris The configuration part makes use of modern compiler features and 641689837Skris still retains old compiler behavior for those that run older versions 641789837Skris of the OS. The shared library support part includes a variant that 641889837Skris uses the RPATH feature, and is available through the special 641989837Skris configuration target "alpha-cc-rpath", which will never be selected 642089837Skris automatically. 642189837Skris [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 642289837Skris 642389837Skris *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 642489837Skris with the same message size as in ssl3_get_certificate_request(). 642589837Skris Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 642689837Skris messages might inadvertently be reject as too long. 642789837Skris [Petr Lampa <lampa@fee.vutbr.cz>] 642889837Skris 642989837Skris *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 643089837Skris [Andy Polyakov] 643189837Skris 643289837Skris *) Modified SSL library such that the verify_callback that has been set 643389837Skris specificly for an SSL object with SSL_set_verify() is actually being 643489837Skris used. Before the change, a verify_callback set with this function was 643589837Skris ignored and the verify_callback() set in the SSL_CTX at the time of 643689837Skris the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 643789837Skris to allow the necessary settings. 643889837Skris [Lutz Jaenicke] 643989837Skris 644089837Skris *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 644189837Skris explicitly to NULL, as at least on Solaris 8 this seems not always to be 644289837Skris done automatically (in contradiction to the requirements of the C 644389837Skris standard). This made problems when used from OpenSSH. 644489837Skris [Lutz Jaenicke] 644589837Skris 644689837Skris *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 644789837Skris dh->length and always used 644889837Skris 644989837Skris BN_rand_range(priv_key, dh->p). 645089837Skris 645189837Skris BN_rand_range() is not necessary for Diffie-Hellman, and this 645289837Skris specific range makes Diffie-Hellman unnecessarily inefficient if 645389837Skris dh->length (recommended exponent length) is much smaller than the 645489837Skris length of dh->p. We could use BN_rand_range() if the order of 645589837Skris the subgroup was stored in the DH structure, but we only have 645689837Skris dh->length. 645789837Skris 645889837Skris So switch back to 645989837Skris 646089837Skris BN_rand(priv_key, l, ...) 646189837Skris 646289837Skris where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 646389837Skris otherwise. 646489837Skris [Bodo Moeller] 646589837Skris 646689837Skris *) In 646789837Skris 646889837Skris RSA_eay_public_encrypt 646989837Skris RSA_eay_private_decrypt 647089837Skris RSA_eay_private_encrypt (signing) 647189837Skris RSA_eay_public_decrypt (signature verification) 647289837Skris 647389837Skris (default implementations for RSA_public_encrypt, 647489837Skris RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 647589837Skris always reject numbers >= n. 647689837Skris [Bodo Moeller] 647789837Skris 647889837Skris *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 647989837Skris to synchronize access to 'locking_thread'. This is necessary on 648089837Skris systems where access to 'locking_thread' (an 'unsigned long' 648189837Skris variable) is not atomic. 648289837Skris [Bodo Moeller] 648389837Skris 648489837Skris *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 648589837Skris *before* setting the 'crypto_lock_rand' flag. The previous code had 648689837Skris a race condition if 0 is a valid thread ID. 648789837Skris [Travis Vitek <vitek@roguewave.com>] 648889837Skris 648989837Skris *) Add support for shared libraries under Irix. 649089837Skris [Albert Chin-A-Young <china@thewrittenword.com>] 649189837Skris 649289837Skris *) Add configuration option to build on Linux on both big-endian and 649389837Skris little-endian MIPS. 649489837Skris [Ralf Baechle <ralf@uni-koblenz.de>] 649589837Skris 649689837Skris *) Add the possibility to create shared libraries on HP-UX. 649789837Skris [Richard Levitte] 649889837Skris 649979998Skris Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 650079998Skris 650179998Skris *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 650279998Skris to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 650379998Skris Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 650479998Skris PRNG state recovery was possible based on the output of 650579998Skris one PRNG request appropriately sized to gain knowledge on 650679998Skris 'md' followed by enough consecutive 1-byte PRNG requests 650779998Skris to traverse all of 'state'. 650879998Skris 650979998Skris 1. When updating 'md_local' (the current thread's copy of 'md') 651079998Skris during PRNG output generation, hash all of the previous 651179998Skris 'md_local' value, not just the half used for PRNG output. 651279998Skris 651379998Skris 2. Make the number of bytes from 'state' included into the hash 651479998Skris independent from the number of PRNG bytes requested. 651579998Skris 651679998Skris The first measure alone would be sufficient to avoid 651779998Skris Markku-Juhani's attack. (Actually it had never occurred 651879998Skris to me that the half of 'md_local' used for chaining was the 651979998Skris half from which PRNG output bytes were taken -- I had always 652079998Skris assumed that the secret half would be used.) The second 652179998Skris measure makes sure that additional data from 'state' is never 652279998Skris mixed into 'md_local' in small portions; this heuristically 652379998Skris further strengthens the PRNG. 652479998Skris [Bodo Moeller] 652579998Skris 652679998Skris *) Fix crypto/bn/asm/mips3.s. 652779998Skris [Andy Polyakov] 652879998Skris 652979998Skris *) When only the key is given to "enc", the IV is undefined. Print out 653079998Skris an error message in this case. 653179998Skris [Lutz Jaenicke] 653279998Skris 653379998Skris *) Handle special case when X509_NAME is empty in X509 printing routines. 653479998Skris [Steve Henson] 653579998Skris 653679998Skris *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 653779998Skris positive and less than q. 653879998Skris [Bodo Moeller] 653979998Skris 654079998Skris *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 654179998Skris used: it isn't thread safe and the add_lock_callback should handle 654279998Skris that itself. 654379998Skris [Paul Rose <Paul.Rose@bridge.com>] 654479998Skris 654579998Skris *) Verify that incoming data obeys the block size in 654679998Skris ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 654779998Skris [Bodo Moeller] 654879998Skris 654979998Skris *) Fix OAEP check. 655079998Skris [Ulf M�ller, Bodo M�ller] 655179998Skris 655279998Skris *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 655389837Skris RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 655479998Skris when fixing the server behaviour for backwards-compatible 'client 655579998Skris hello' messages. (Note that the attack is impractical against 655679998Skris SSL 3.0 and TLS 1.0 anyway because length and version checking 655779998Skris means that the probability of guessing a valid ciphertext is 655879998Skris around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 655979998Skris paper.) 656079998Skris 656179998Skris Before 0.9.5, the countermeasure (hide the error by generating a 656279998Skris random 'decryption result') did not work properly because 656379998Skris ERR_clear_error() was missing, meaning that SSL_get_error() would 656479998Skris detect the supposedly ignored error. 656579998Skris 656679998Skris Both problems are now fixed. 656779998Skris [Bodo Moeller] 656879998Skris 656979998Skris *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 657079998Skris (previously it was 1024). 657179998Skris [Bodo Moeller] 657279998Skris 657379998Skris *) Fix for compatibility mode trust settings: ignore trust settings 657479998Skris unless some valid trust or reject settings are present. 657579998Skris [Steve Henson] 657679998Skris 657779998Skris *) Fix for blowfish EVP: its a variable length cipher. 657879998Skris [Steve Henson] 657979998Skris 658079998Skris *) Fix various bugs related to DSA S/MIME verification. Handle missing 658179998Skris parameters in DSA public key structures and return an error in the 658279998Skris DSA routines if parameters are absent. 658379998Skris [Steve Henson] 658479998Skris 658579998Skris *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 658679998Skris in the current directory if neither $RANDFILE nor $HOME was set. 658779998Skris RAND_file_name() in 0.9.6a returned NULL in this case. This has 658879998Skris caused some confusion to Windows users who haven't defined $HOME. 658979998Skris Thus RAND_file_name() is changed again: e_os.h can define a 659079998Skris DEFAULT_HOME, which will be used if $HOME is not set. 659179998Skris For Windows, we use "C:"; on other platforms, we still require 659279998Skris environment variables. 659379998Skris 659479998Skris *) Move 'if (!initialized) RAND_poll()' into regions protected by 659579998Skris CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 659679998Skris having multiple threads call RAND_poll() concurrently. 659779998Skris [Bodo Moeller] 659879998Skris 659979998Skris *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 660079998Skris combination of a flag and a thread ID variable. 660179998Skris Otherwise while one thread is in ssleay_rand_bytes (which sets the 660279998Skris flag), *other* threads can enter ssleay_add_bytes without obeying 660389837Skris the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 660479998Skris that they do not hold after the first thread unsets add_do_not_lock). 660579998Skris [Bodo Moeller] 660679998Skris 660779998Skris *) Change bctest again: '-x' expressions are not available in all 660879998Skris versions of 'test'. 660979998Skris [Bodo Moeller] 661079998Skris 661176866Skris Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 661272613Skris 661376866Skris *) Fix a couple of memory leaks in PKCS7_dataDecode() 661476866Skris [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 661576866Skris 661676866Skris *) Change Configure and Makefiles to provide EXE_EXT, which will contain 661776866Skris the default extension for executables, if any. Also, make the perl 661876866Skris scripts that use symlink() to test if it really exists and use "cp" 661976866Skris if it doesn't. All this made OpenSSL compilable and installable in 662076866Skris CygWin. 662176866Skris [Richard Levitte] 662276866Skris 662376866Skris *) Fix for asn1_GetSequence() for indefinite length constructed data. 662476866Skris If SEQUENCE is length is indefinite just set c->slen to the total 662576866Skris amount of data available. 662676866Skris [Steve Henson, reported by shige@FreeBSD.org] 662776866Skris [This change does not apply to 0.9.7.] 662876866Skris 662976866Skris *) Change bctest to avoid here-documents inside command substitution 663076866Skris (workaround for FreeBSD /bin/sh bug). 663176866Skris For compatibility with Ultrix, avoid shell functions (introduced 663276866Skris in the bctest version that searches along $PATH). 663376866Skris [Bodo Moeller] 663476866Skris 663576866Skris *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 663676866Skris with des_encrypt() defined on some operating systems, like Solaris 663776866Skris and UnixWare. 663876866Skris [Richard Levitte] 663976866Skris 664076866Skris *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 664176866Skris On the Importance of Eliminating Errors in Cryptographic 664276866Skris Computations, J. Cryptology 14 (2001) 2, 101-119, 664376866Skris http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 664476866Skris [Ulf Moeller] 664576866Skris 664676866Skris *) MIPS assembler BIGNUM division bug fix. 664776866Skris [Andy Polyakov] 664876866Skris 664976866Skris *) Disabled incorrect Alpha assembler code. 665076866Skris [Richard Levitte] 665176866Skris 665276866Skris *) Fix PKCS#7 decode routines so they correctly update the length 665376866Skris after reading an EOC for the EXPLICIT tag. 665476866Skris [Steve Henson] 665576866Skris [This change does not apply to 0.9.7.] 665676866Skris 665776866Skris *) Fix bug in PKCS#12 key generation routines. This was triggered 665876866Skris if a 3DES key was generated with a 0 initial byte. Include 665976866Skris PKCS12_BROKEN_KEYGEN compilation option to retain the old 666076866Skris (but broken) behaviour. 666176866Skris [Steve Henson] 666276866Skris 666376866Skris *) Enhance bctest to search for a working bc along $PATH and print 666476866Skris it when found. 666576866Skris [Tim Rice <tim@multitalents.net> via Richard Levitte] 666676866Skris 666776866Skris *) Fix memory leaks in err.c: free err_data string if necessary; 666876866Skris don't write to the wrong index in ERR_set_error_data. 666976866Skris [Bodo Moeller] 667076866Skris 667176866Skris *) Implement ssl23_peek (analogous to ssl23_read), which previously 667276866Skris did not exist. 667376866Skris [Bodo Moeller] 667476866Skris 667576866Skris *) Replace rdtsc with _emit statements for VC++ version 5. 667676866Skris [Jeremy Cooper <jeremy@baymoo.org>] 667776866Skris 667876866Skris *) Make it possible to reuse SSLv2 sessions. 667976866Skris [Richard Levitte] 668076866Skris 668176866Skris *) In copy_email() check for >= 0 as a return value for 668276866Skris X509_NAME_get_index_by_NID() since 0 is a valid index. 668376866Skris [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 668476866Skris 668576866Skris *) Avoid coredump with unsupported or invalid public keys by checking if 668676866Skris X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 668776866Skris PKCS7_verify() fails with non detached data. 668876866Skris [Steve Henson] 668976866Skris 669076866Skris *) Don't use getenv in library functions when run as setuid/setgid. 669176866Skris New function OPENSSL_issetugid(). 669276866Skris [Ulf Moeller] 669376866Skris 669476866Skris *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 669576866Skris due to incorrect handling of multi-threading: 669676866Skris 669776866Skris 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 669876866Skris 669976866Skris 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 670076866Skris 670176866Skris 3. Count how many times MemCheck_off() has been called so that 670276866Skris nested use can be treated correctly. This also avoids 670376866Skris inband-signalling in the previous code (which relied on the 670476866Skris assumption that thread ID 0 is impossible). 670576866Skris [Bodo Moeller] 670676866Skris 670776866Skris *) Add "-rand" option also to s_client and s_server. 670876866Skris [Lutz Jaenicke] 670976866Skris 671076866Skris *) Fix CPU detection on Irix 6.x. 671176866Skris [Kurt Hockenbury <khockenb@stevens-tech.edu> and 671276866Skris "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 671376866Skris 671476866Skris *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 671576866Skris was empty. 671676866Skris [Steve Henson] 671776866Skris [This change does not apply to 0.9.7.] 671876866Skris 671976866Skris *) Use the cached encoding of an X509_NAME structure rather than 672076866Skris copying it. This is apparently the reason for the libsafe "errors" 672176866Skris but the code is actually correct. 672276866Skris [Steve Henson] 672376866Skris 672472613Skris *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 672572613Skris Bleichenbacher's DSA attack. 672676866Skris Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 672776866Skris to be set and top=0 forces the highest bit to be set; top=-1 is new 672876866Skris and leaves the highest bit random. 672976866Skris [Ulf Moeller, Bodo Moeller] 673072613Skris 673172613Skris *) In the NCONF_...-based implementations for CONF_... queries 673272613Skris (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 673372613Skris a temporary CONF structure with the data component set to NULL 673472613Skris (which gives segmentation faults in lh_retrieve). 673572613Skris Instead, use NULL for the CONF pointer in CONF_get_string and 673672613Skris CONF_get_number (which may use environment variables) and directly 673772613Skris return NULL from CONF_get_section. 673872613Skris [Bodo Moeller] 673972613Skris 674072613Skris *) Fix potential buffer overrun for EBCDIC. 674172613Skris [Ulf Moeller] 674272613Skris 674372613Skris *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 674472613Skris keyUsage if basicConstraints absent for a CA. 674572613Skris [Steve Henson] 674672613Skris 674772613Skris *) Make SMIME_write_PKCS7() write mail header values with a format that 674872613Skris is more generally accepted (no spaces before the semicolon), since 674972613Skris some programs can't parse those values properly otherwise. Also make 675072613Skris sure BIO's that break lines after each write do not create invalid 675172613Skris headers. 675272613Skris [Richard Levitte] 675372613Skris 675472613Skris *) Make the CRL encoding routines work with empty SEQUENCE OF. The 675572613Skris macros previously used would not encode an empty SEQUENCE OF 675672613Skris and break the signature. 675772613Skris [Steve Henson] 675876866Skris [This change does not apply to 0.9.7.] 675972613Skris 676072613Skris *) Zero the premaster secret after deriving the master secret in 676172613Skris DH ciphersuites. 676272613Skris [Steve Henson] 676372613Skris 676472613Skris *) Add some EVP_add_digest_alias registrations (as found in 676572613Skris OpenSSL_add_all_digests()) to SSL_library_init() 676672613Skris aka OpenSSL_add_ssl_algorithms(). This provides improved 676772613Skris compatibility with peers using X.509 certificates 676872613Skris with unconventional AlgorithmIdentifier OIDs. 676972613Skris [Bodo Moeller] 677072613Skris 677172613Skris *) Fix for Irix with NO_ASM. 677272613Skris ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 677372613Skris 677472613Skris *) ./config script fixes. 677572613Skris [Ulf Moeller, Richard Levitte] 677672613Skris 677772613Skris *) Fix 'openssl passwd -1'. 677872613Skris [Bodo Moeller] 677972613Skris 678072613Skris *) Change PKCS12_key_gen_asc() so it can cope with non null 678172613Skris terminated strings whose length is passed in the passlen 678272613Skris parameter, for example from PEM callbacks. This was done 678372613Skris by adding an extra length parameter to asc2uni(). 678472613Skris [Steve Henson, reported by <oddissey@samsung.co.kr>] 678572613Skris 678672613Skris *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 678772613Skris call failed, free the DSA structure. 678872613Skris [Bodo Moeller] 678972613Skris 679072613Skris *) Fix to uni2asc() to cope with zero length Unicode strings. 679172613Skris These are present in some PKCS#12 files. 679272613Skris [Steve Henson] 679372613Skris 679472613Skris *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 679572613Skris Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 679672613Skris when writing a 32767 byte record. 679772613Skris [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 679872613Skris 679972613Skris *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 680072613Skris obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 680172613Skris 680272613Skris (RSA objects have a reference count access to which is protected 680372613Skris by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 680472613Skris so they are meant to be shared between threads.) 680572613Skris [Bodo Moeller, Geoff Thorpe; original patch submitted by 680672613Skris "Reddie, Steven" <Steven.Reddie@ca.com>] 680772613Skris 680872613Skris *) Fix a deadlock in CRYPTO_mem_leaks(). 680972613Skris [Bodo Moeller] 681072613Skris 681176866Skris *) Use better test patterns in bntest. 681276866Skris [Ulf M�ller] 681376866Skris 681472613Skris *) rand_win.c fix for Borland C. 681572613Skris [Ulf M�ller] 681672613Skris 681772613Skris *) BN_rshift bugfix for n == 0. 681872613Skris [Bodo Moeller] 681972613Skris 682076866Skris *) Add a 'bctest' script that checks for some known 'bc' bugs 682176866Skris so that 'make test' does not abort just because 'bc' is broken. 682276866Skris [Bodo Moeller] 682376866Skris 682472613Skris *) Store verify_result within SSL_SESSION also for client side to 682572613Skris avoid potential security hole. (Re-used sessions on the client side 682672613Skris always resulted in verify_result==X509_V_OK, not using the original 682772613Skris result of the server certificate verification.) 682872613Skris [Lutz Jaenicke] 682972613Skris 683072613Skris *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 683172613Skris SSL3_RT_APPLICATION_DATA, return 0. 683272613Skris Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 683372613Skris [Bodo Moeller] 683472613Skris 683572613Skris *) Fix SSL_peek: 683672613Skris Both ssl2_peek and ssl3_peek, which were totally broken in earlier 683772613Skris releases, have been re-implemented by renaming the previous 683872613Skris implementations of ssl2_read and ssl3_read to ssl2_read_internal 683972613Skris and ssl3_read_internal, respectively, and adding 'peek' parameters 684072613Skris to them. The new ssl[23]_{read,peek} functions are calls to 684172613Skris ssl[23]_read_internal with the 'peek' flag set appropriately. 684272613Skris A 'peek' parameter has also been added to ssl3_read_bytes, which 684372613Skris does the actual work for ssl3_read_internal. 684472613Skris [Bodo Moeller] 684572613Skris 684676866Skris *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 684776866Skris the method-specific "init()" handler. Also clean up ex_data after 684876866Skris calling the method-specific "finish()" handler. Previously, this was 684976866Skris happening the other way round. 685076866Skris [Geoff Thorpe] 685176866Skris 685272613Skris *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 685372613Skris The previous value, 12, was not always sufficient for BN_mod_exp(). 685472613Skris [Bodo Moeller] 685572613Skris 685676866Skris *) Make sure that shared libraries get the internal name engine with 685776866Skris the full version number and not just 0. This should mark the 685876866Skris shared libraries as not backward compatible. Of course, this should 685976866Skris be changed again when we can guarantee backward binary compatibility. 686076866Skris [Richard Levitte] 686176866Skris 686272613Skris *) Fix typo in get_cert_by_subject() in by_dir.c 686372613Skris [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 686472613Skris 686576866Skris *) Rework the system to generate shared libraries: 686676866Skris 686776866Skris - Make note of the expected extension for the shared libraries and 686876866Skris if there is a need for symbolic links from for example libcrypto.so.0 686976866Skris to libcrypto.so.0.9.7. There is extended info in Configure for 687076866Skris that. 687176866Skris 687276866Skris - Make as few rebuilds of the shared libraries as possible. 687376866Skris 687476866Skris - Still avoid linking the OpenSSL programs with the shared libraries. 687576866Skris 687676866Skris - When installing, install the shared libraries separately from the 687776866Skris static ones. 687876866Skris [Richard Levitte] 687976866Skris 688072613Skris *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 688172613Skris 688272613Skris Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 688372613Skris and not in SSL_clear because the latter is also used by the 688472613Skris accept/connect functions; previously, the settings made by 688572613Skris SSL_set_read_ahead would be lost during the handshake. 688672613Skris [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 688772613Skris 688872613Skris *) Correct util/mkdef.pl to be selective about disabled algorithms. 688972613Skris Previously, it would create entries for disableed algorithms no 689072613Skris matter what. 689172613Skris [Richard Levitte] 689272613Skris 689372613Skris *) Added several new manual pages for SSL_* function. 689472613Skris [Lutz Jaenicke] 689572613Skris 689668651Skris Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 689768651Skris 689868651Skris *) In ssl23_get_client_hello, generate an error message when faced 689968651Skris with an initial SSL 3.0/TLS record that is too small to contain the 690068651Skris first two bytes of the ClientHello message, i.e. client_version. 690168651Skris (Note that this is a pathologic case that probably has never happened 690268651Skris in real life.) The previous approach was to use the version number 690368651Skris from the record header as a substitute; but our protocol choice 690468651Skris should not depend on that one because it is not authenticated 690568651Skris by the Finished messages. 690668651Skris [Bodo Moeller] 690768651Skris 690868651Skris *) More robust randomness gathering functions for Windows. 690968651Skris [Jeffrey Altman <jaltman@columbia.edu>] 691068651Skris 691168651Skris *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 691268651Skris not set then we don't setup the error code for issuer check errors 691368651Skris to avoid possibly overwriting other errors which the callback does 691468651Skris handle. If an application does set the flag then we assume it knows 691568651Skris what it is doing and can handle the new informational codes 691668651Skris appropriately. 691768651Skris [Steve Henson] 691868651Skris 691968651Skris *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 692068651Skris a general "ANY" type, as such it should be able to decode anything 692168651Skris including tagged types. However it didn't check the class so it would 692268651Skris wrongly interpret tagged types in the same way as their universal 692368651Skris counterpart and unknown types were just rejected. Changed so that the 692468651Skris tagged and unknown types are handled in the same way as a SEQUENCE: 692568651Skris that is the encoding is stored intact. There is also a new type 692668651Skris "V_ASN1_OTHER" which is used when the class is not universal, in this 692768651Skris case we have no idea what the actual type is so we just lump them all 692868651Skris together. 692968651Skris [Steve Henson] 693068651Skris 693168651Skris *) On VMS, stdout may very well lead to a file that is written to 693268651Skris in a record-oriented fashion. That means that every write() will 693368651Skris write a separate record, which will be read separately by the 693468651Skris programs trying to read from it. This can be very confusing. 693568651Skris 693668651Skris The solution is to put a BIO filter in the way that will buffer 693768651Skris text until a linefeed is reached, and then write everything a 693868651Skris line at a time, so every record written will be an actual line, 693968651Skris not chunks of lines and not (usually doesn't happen, but I've 694068651Skris seen it once) several lines in one record. BIO_f_linebuffer() is 694168651Skris the answer. 694268651Skris 694368651Skris Currently, it's a VMS-only method, because that's where it has 694468651Skris been tested well enough. 694568651Skris [Richard Levitte] 694668651Skris 694768651Skris *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 694868651Skris it can return incorrect results. 694968651Skris (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 695068651Skris but it was in 0.9.6-beta[12].) 695168651Skris [Bodo Moeller] 695268651Skris 695368651Skris *) Disable the check for content being present when verifying detached 695468651Skris signatures in pk7_smime.c. Some versions of Netscape (wrongly) 695568651Skris include zero length content when signing messages. 695668651Skris [Steve Henson] 695768651Skris 695868651Skris *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 695968651Skris BIO_ctrl (for BIO pairs). 696068651Skris [Bodo M�ller] 696168651Skris 696268651Skris *) Add DSO method for VMS. 696368651Skris [Richard Levitte] 696468651Skris 696568651Skris *) Bug fix: Montgomery multiplication could produce results with the 696668651Skris wrong sign. 696768651Skris [Ulf M�ller] 696868651Skris 696968651Skris *) Add RPM specification openssl.spec and modify it to build three 697068651Skris packages. The default package contains applications, application 697168651Skris documentation and run-time libraries. The devel package contains 697268651Skris include files, static libraries and function documentation. The 697368651Skris doc package contains the contents of the doc directory. The original 697468651Skris openssl.spec was provided by Damien Miller <djm@mindrot.org>. 697568651Skris [Richard Levitte] 697668651Skris 697768651Skris *) Add a large number of documentation files for many SSL routines. 697868651Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 697968651Skris 698068651Skris *) Add a configuration entry for Sony News 4. 698168651Skris [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 698268651Skris 698368651Skris *) Don't set the two most significant bits to one when generating a 698468651Skris random number < q in the DSA library. 698568651Skris [Ulf M�ller] 698668651Skris 698768651Skris *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 698868651Skris behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 698968651Skris the underlying transport is blocking) if a handshake took place. 699068651Skris (The default behaviour is needed by applications such as s_client 699168651Skris and s_server that use select() to determine when to use SSL_read; 699268651Skris but for applications that know in advance when to expect data, it 699368651Skris just makes things more complicated.) 699468651Skris [Bodo Moeller] 699568651Skris 699668651Skris *) Add RAND_egd_bytes(), which gives control over the number of bytes read 699768651Skris from EGD. 699868651Skris [Ben Laurie] 699968651Skris 700068651Skris *) Add a few more EBCDIC conditionals that make `req' and `x509' 700168651Skris work better on such systems. 700268651Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 700368651Skris 700468651Skris *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 700568651Skris Update PKCS12_parse() so it copies the friendlyName and the 700668651Skris keyid to the certificates aux info. 700768651Skris [Steve Henson] 700868651Skris 700968651Skris *) Fix bug in PKCS7_verify() which caused an infinite loop 701068651Skris if there was more than one signature. 701168651Skris [Sven Uszpelkat <su@celocom.de>] 701268651Skris 701368651Skris *) Major change in util/mkdef.pl to include extra information 701468651Skris about each symbol, as well as presentig variables as well 701568651Skris as functions. This change means that there's n more need 701668651Skris to rebuild the .num files when some algorithms are excluded. 701768651Skris [Richard Levitte] 701868651Skris 701968651Skris *) Allow the verify time to be set by an application, 702068651Skris rather than always using the current time. 702168651Skris [Steve Henson] 702268651Skris 702368651Skris *) Phase 2 verify code reorganisation. The certificate 702468651Skris verify code now looks up an issuer certificate by a 702568651Skris number of criteria: subject name, authority key id 702668651Skris and key usage. It also verifies self signed certificates 702768651Skris by the same criteria. The main comparison function is 702868651Skris X509_check_issued() which performs these checks. 702968651Skris 703068651Skris Lot of changes were necessary in order to support this 703168651Skris without completely rewriting the lookup code. 703268651Skris 703368651Skris Authority and subject key identifier are now cached. 703468651Skris 703568651Skris The LHASH 'certs' is X509_STORE has now been replaced 703668651Skris by a STACK_OF(X509_OBJECT). This is mainly because an 703768651Skris LHASH can't store or retrieve multiple objects with 703868651Skris the same hash value. 703968651Skris 704068651Skris As a result various functions (which were all internal 704168651Skris use only) have changed to handle the new X509_STORE 704268651Skris structure. This will break anything that messed round 704368651Skris with X509_STORE internally. 704468651Skris 704568651Skris The functions X509_STORE_add_cert() now checks for an 704668651Skris exact match, rather than just subject name. 704768651Skris 704868651Skris The X509_STORE API doesn't directly support the retrieval 704968651Skris of multiple certificates matching a given criteria, however 705068651Skris this can be worked round by performing a lookup first 705168651Skris (which will fill the cache with candidate certificates) 705268651Skris and then examining the cache for matches. This is probably 705368651Skris the best we can do without throwing out X509_LOOKUP 705468651Skris entirely (maybe later...). 705568651Skris 705668651Skris The X509_VERIFY_CTX structure has been enhanced considerably. 705768651Skris 705868651Skris All certificate lookup operations now go via a get_issuer() 705968651Skris callback. Although this currently uses an X509_STORE it 706068651Skris can be replaced by custom lookups. This is a simple way 706168651Skris to bypass the X509_STORE hackery necessary to make this 706268651Skris work and makes it possible to use more efficient techniques 706368651Skris in future. A very simple version which uses a simple 706468651Skris STACK for its trusted certificate store is also provided 706568651Skris using X509_STORE_CTX_trusted_stack(). 706668651Skris 706768651Skris The verify_cb() and verify() callbacks now have equivalents 706868651Skris in the X509_STORE_CTX structure. 706968651Skris 707068651Skris X509_STORE_CTX also has a 'flags' field which can be used 707168651Skris to customise the verify behaviour. 707268651Skris [Steve Henson] 707368651Skris 707468651Skris *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 707568651Skris excludes S/MIME capabilities. 707668651Skris [Steve Henson] 707768651Skris 707868651Skris *) When a certificate request is read in keep a copy of the 707968651Skris original encoding of the signed data and use it when outputing 708068651Skris again. Signatures then use the original encoding rather than 708168651Skris a decoded, encoded version which may cause problems if the 708268651Skris request is improperly encoded. 708368651Skris [Steve Henson] 708468651Skris 708568651Skris *) For consistency with other BIO_puts implementations, call 708668651Skris buffer_write(b, ...) directly in buffer_puts instead of calling 708768651Skris BIO_write(b, ...). 708868651Skris 708968651Skris In BIO_puts, increment b->num_write as in BIO_write. 709068651Skris [Peter.Sylvester@EdelWeb.fr] 709168651Skris 709268651Skris *) Fix BN_mul_word for the case where the word is 0. (We have to use 709368651Skris BN_zero, we may not return a BIGNUM with an array consisting of 709468651Skris words set to zero.) 709568651Skris [Bodo Moeller] 709668651Skris 709768651Skris *) Avoid calling abort() from within the library when problems are 709868651Skris detected, except if preprocessor symbols have been defined 709968651Skris (such as REF_CHECK, BN_DEBUG etc.). 710068651Skris [Bodo Moeller] 710168651Skris 710268651Skris *) New openssl application 'rsautl'. This utility can be 710368651Skris used for low level RSA operations. DER public key 710468651Skris BIO/fp routines also added. 710568651Skris [Steve Henson] 710668651Skris 710768651Skris *) New Configure entry and patches for compiling on QNX 4. 710868651Skris [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 710968651Skris 711068651Skris *) A demo state-machine implementation was sponsored by 711168651Skris Nuron (http://www.nuron.com/) and is now available in 711268651Skris demos/state_machine. 711368651Skris [Ben Laurie] 711468651Skris 711568651Skris *) New options added to the 'dgst' utility for signature 711668651Skris generation and verification. 711768651Skris [Steve Henson] 711868651Skris 711968651Skris *) Unrecognized PKCS#7 content types are now handled via a 712068651Skris catch all ASN1_TYPE structure. This allows unsupported 712168651Skris types to be stored as a "blob" and an application can 712268651Skris encode and decode it manually. 712368651Skris [Steve Henson] 712468651Skris 712568651Skris *) Fix various signed/unsigned issues to make a_strex.c 712668651Skris compile under VC++. 712768651Skris [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 712868651Skris 712968651Skris *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 713068651Skris length if passed a buffer. ASN1_INTEGER_to_BN failed 713168651Skris if passed a NULL BN and its argument was negative. 713268651Skris [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 713368651Skris 713468651Skris *) Modification to PKCS#7 encoding routines to output definite 713568651Skris length encoding. Since currently the whole structures are in 713668651Skris memory there's not real point in using indefinite length 713768651Skris constructed encoding. However if OpenSSL is compiled with 713868651Skris the flag PKCS7_INDEFINITE_ENCODING the old form is used. 713968651Skris [Steve Henson] 714068651Skris 714168651Skris *) Added BIO_vprintf() and BIO_vsnprintf(). 714268651Skris [Richard Levitte] 714368651Skris 714468651Skris *) Added more prefixes to parse for in the the strings written 714568651Skris through a logging bio, to cover all the levels that are available 714668651Skris through syslog. The prefixes are now: 714768651Skris 714868651Skris PANIC, EMERG, EMR => LOG_EMERG 714968651Skris ALERT, ALR => LOG_ALERT 715068651Skris CRIT, CRI => LOG_CRIT 715168651Skris ERROR, ERR => LOG_ERR 715268651Skris WARNING, WARN, WAR => LOG_WARNING 715368651Skris NOTICE, NOTE, NOT => LOG_NOTICE 715468651Skris INFO, INF => LOG_INFO 715568651Skris DEBUG, DBG => LOG_DEBUG 715668651Skris 715768651Skris and as before, if none of those prefixes are present at the 715868651Skris beginning of the string, LOG_ERR is chosen. 715968651Skris 716068651Skris On Win32, the LOG_* levels are mapped according to this: 716168651Skris 716268651Skris LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 716368651Skris LOG_WARNING => EVENTLOG_WARNING_TYPE 716468651Skris LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 716568651Skris 716668651Skris [Richard Levitte] 716768651Skris 716868651Skris *) Made it possible to reconfigure with just the configuration 716968651Skris argument "reconf" or "reconfigure". The command line arguments 717068651Skris are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 717168651Skris and are retrieved from there when reconfiguring. 717268651Skris [Richard Levitte] 717368651Skris 717468651Skris *) MD4 implemented. 717568651Skris [Assar Westerlund <assar@sics.se>, Richard Levitte] 717668651Skris 717768651Skris *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 717868651Skris [Richard Levitte] 717968651Skris 718068651Skris *) The obj_dat.pl script was messing up the sorting of object 718168651Skris names. The reason was that it compared the quoted version 718268651Skris of strings as a result "OCSP" > "OCSP Signing" because 718368651Skris " > SPACE. Changed script to store unquoted versions of 718468651Skris names and add quotes on output. It was also omitting some 718568651Skris names from the lookup table if they were given a default 718668651Skris value (that is if SN is missing it is given the same 718768651Skris value as LN and vice versa), these are now added on the 718868651Skris grounds that if an object has a name we should be able to 718968651Skris look it up. Finally added warning output when duplicate 719068651Skris short or long names are found. 719168651Skris [Steve Henson] 719268651Skris 719368651Skris *) Changes needed for Tandem NSK. 719468651Skris [Scott Uroff <scott@xypro.com>] 719568651Skris 719668651Skris *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 719768651Skris RSA_padding_check_SSLv23(), special padding was never detected 719868651Skris and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 719968651Skris version rollback attacks was not effective. 720068651Skris 720168651Skris In s23_clnt.c, don't use special rollback-attack detection padding 720268651Skris (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 720368651Skris client; similarly, in s23_srvr.c, don't do the rollback check if 720468651Skris SSL 2.0 is the only protocol enabled in the server. 720568651Skris [Bodo Moeller] 720668651Skris 720768651Skris *) Make it possible to get hexdumps of unprintable data with 'openssl 720868651Skris asn1parse'. By implication, the functions ASN1_parse_dump() and 720968651Skris BIO_dump_indent() are added. 721068651Skris [Richard Levitte] 721168651Skris 721268651Skris *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 721368651Skris these print out strings and name structures based on various 721468651Skris flags including RFC2253 support and proper handling of 721568651Skris multibyte characters. Added options to the 'x509' utility 721668651Skris to allow the various flags to be set. 721768651Skris [Steve Henson] 721868651Skris 721968651Skris *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 722068651Skris Also change the functions X509_cmp_current_time() and 722168651Skris X509_gmtime_adj() work with an ASN1_TIME structure, 722268651Skris this will enable certificates using GeneralizedTime in validity 722368651Skris dates to be checked. 722468651Skris [Steve Henson] 722568651Skris 722668651Skris *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 722768651Skris negative public key encodings) on by default, 722868651Skris NO_NEG_PUBKEY_BUG can be set to disable it. 722968651Skris [Steve Henson] 723068651Skris 723168651Skris *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 723268651Skris content octets. An i2c_ASN1_OBJECT is unnecessary because 723368651Skris the encoding can be trivially obtained from the structure. 723468651Skris [Steve Henson] 723568651Skris 723668651Skris *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 723768651Skris not read locks (CRYPTO_r_[un]lock). 723868651Skris [Bodo Moeller] 723968651Skris 724068651Skris *) A first attempt at creating official support for shared 724168651Skris libraries through configuration. I've kept it so the 724268651Skris default is static libraries only, and the OpenSSL programs 724368651Skris are always statically linked for now, but there are 724468651Skris preparations for dynamic linking in place. 724589837Skris This has been tested on Linux and Tru64. 724668651Skris [Richard Levitte] 724768651Skris 724868651Skris *) Randomness polling function for Win9x, as described in: 724968651Skris Peter Gutmann, Software Generation of Practically Strong 725068651Skris Random Numbers. 725168651Skris [Ulf M�ller] 725268651Skris 725368651Skris *) Fix so PRNG is seeded in req if using an already existing 725468651Skris DSA key. 725568651Skris [Steve Henson] 725668651Skris 725768651Skris *) New options to smime application. -inform and -outform 725868651Skris allow alternative formats for the S/MIME message including 725968651Skris PEM and DER. The -content option allows the content to be 726068651Skris specified separately. This should allow things like Netscape 726168651Skris form signing output easier to verify. 726268651Skris [Steve Henson] 726368651Skris 726468651Skris *) Fix the ASN1 encoding of tags using the 'long form'. 726568651Skris [Steve Henson] 726668651Skris 726768651Skris *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 726868651Skris STRING types. These convert content octets to and from the 726968651Skris underlying type. The actual tag and length octets are 727068651Skris already assumed to have been read in and checked. These 727168651Skris are needed because all other string types have virtually 727268651Skris identical handling apart from the tag. By having versions 727368651Skris of the ASN1 functions that just operate on content octets 727468651Skris IMPLICIT tagging can be handled properly. It also allows 727568651Skris the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 727668651Skris and ASN1_INTEGER are identical apart from the tag. 727768651Skris [Steve Henson] 727868651Skris 727968651Skris *) Change the handling of OID objects as follows: 728068651Skris 728168651Skris - New object identifiers are inserted in objects.txt, following 728268651Skris the syntax given in objects.README. 728368651Skris - objects.pl is used to process obj_mac.num and create a new 728468651Skris obj_mac.h. 728568651Skris - obj_dat.pl is used to create a new obj_dat.h, using the data in 728668651Skris obj_mac.h. 728768651Skris 728868651Skris This is currently kind of a hack, and the perl code in objects.pl 728968651Skris isn't very elegant, but it works as I intended. The simplest way 729068651Skris to check that it worked correctly is to look in obj_dat.h and 729168651Skris check the array nid_objs and make sure the objects haven't moved 729268651Skris around (this is important!). Additions are OK, as well as 729368651Skris consistent name changes. 729468651Skris [Richard Levitte] 729568651Skris 729668651Skris *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 729768651Skris [Bodo Moeller] 729868651Skris 729968651Skris *) Addition of the command line parameter '-rand file' to 'openssl req'. 730068651Skris The given file adds to whatever has already been seeded into the 730168651Skris random pool through the RANDFILE configuration file option or 730268651Skris environment variable, or the default random state file. 730368651Skris [Richard Levitte] 730468651Skris 730568651Skris *) mkstack.pl now sorts each macro group into lexical order. 730668651Skris Previously the output order depended on the order the files 730768651Skris appeared in the directory, resulting in needless rewriting 730868651Skris of safestack.h . 730968651Skris [Steve Henson] 731068651Skris 731168651Skris *) Patches to make OpenSSL compile under Win32 again. Mostly 731268651Skris work arounds for the VC++ problem that it treats func() as 731368651Skris func(void). Also stripped out the parts of mkdef.pl that 731468651Skris added extra typesafe functions: these no longer exist. 731568651Skris [Steve Henson] 731668651Skris 731768651Skris *) Reorganisation of the stack code. The macros are now all 731868651Skris collected in safestack.h . Each macro is defined in terms of 731968651Skris a "stack macro" of the form SKM_<name>(type, a, b). The 732068651Skris DEBUG_SAFESTACK is now handled in terms of function casts, 732168651Skris this has the advantage of retaining type safety without the 732268651Skris use of additional functions. If DEBUG_SAFESTACK is not defined 732368651Skris then the non typesafe macros are used instead. Also modified the 732468651Skris mkstack.pl script to handle the new form. Needs testing to see 732568651Skris if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 732668651Skris the default if no major problems. Similar behaviour for ASN1_SET_OF 732768651Skris and PKCS12_STACK_OF. 732868651Skris [Steve Henson] 732968651Skris 733068651Skris *) When some versions of IIS use the 'NET' form of private key the 733168651Skris key derivation algorithm is different. Normally MD5(password) is 733268651Skris used as a 128 bit RC4 key. In the modified case 733368651Skris MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 733468651Skris new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 733568651Skris as the old Netscape_RSA functions except they have an additional 733668651Skris 'sgckey' parameter which uses the modified algorithm. Also added 733768651Skris an -sgckey command line option to the rsa utility. Thanks to 733868651Skris Adrian Peck <bertie@ncipher.com> for posting details of the modified 733968651Skris algorithm to openssl-dev. 734068651Skris [Steve Henson] 734168651Skris 734268651Skris *) The evp_local.h macros were using 'c.##kname' which resulted in 734368651Skris invalid expansion on some systems (SCO 5.0.5 for example). 734468651Skris Corrected to 'c.kname'. 734568651Skris [Phillip Porch <root@theporch.com>] 734668651Skris 734768651Skris *) New X509_get1_email() and X509_REQ_get1_email() functions that return 734868651Skris a STACK of email addresses from a certificate or request, these look 734968651Skris in the subject name and the subject alternative name extensions and 735068651Skris omit any duplicate addresses. 735168651Skris [Steve Henson] 735268651Skris 735368651Skris *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 735468651Skris This makes DSA verification about 2 % faster. 735568651Skris [Bodo Moeller] 735668651Skris 735768651Skris *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 735868651Skris (meaning that now 2^5 values will be precomputed, which is only 4 KB 735968651Skris plus overhead for 1024 bit moduli). 736068651Skris This makes exponentiations about 0.5 % faster for 1024 bit 736168651Skris exponents (as measured by "openssl speed rsa2048"). 736268651Skris [Bodo Moeller] 736368651Skris 736468651Skris *) Rename memory handling macros to avoid conflicts with other 736568651Skris software: 736668651Skris Malloc => OPENSSL_malloc 736768651Skris Malloc_locked => OPENSSL_malloc_locked 736868651Skris Realloc => OPENSSL_realloc 736968651Skris Free => OPENSSL_free 737068651Skris [Richard Levitte] 737168651Skris 737268651Skris *) New function BN_mod_exp_mont_word for small bases (roughly 15% 737368651Skris faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 737468651Skris [Bodo Moeller] 737568651Skris 737668651Skris *) CygWin32 support. 737768651Skris [John Jarvie <jjarvie@newsguy.com>] 737868651Skris 737968651Skris *) The type-safe stack code has been rejigged. It is now only compiled 738068651Skris in when OpenSSL is configured with the DEBUG_SAFESTACK option and 738168651Skris by default all type-specific stack functions are "#define"d back to 738268651Skris standard stack functions. This results in more streamlined output 738368651Skris but retains the type-safety checking possibilities of the original 738468651Skris approach. 738568651Skris [Geoff Thorpe] 738668651Skris 738768651Skris *) The STACK code has been cleaned up, and certain type declarations 738868651Skris that didn't make a lot of sense have been brought in line. This has 738968651Skris also involved a cleanup of sorts in safestack.h to more correctly 739068651Skris map type-safe stack functions onto their plain stack counterparts. 739168651Skris This work has also resulted in a variety of "const"ifications of 739268651Skris lots of the code, especially "_cmp" operations which should normally 739368651Skris be prototyped with "const" parameters anyway. 739468651Skris [Geoff Thorpe] 739568651Skris 739668651Skris *) When generating bytes for the first time in md_rand.c, 'stir the pool' 739768651Skris by seeding with STATE_SIZE dummy bytes (with zero entropy count). 739868651Skris (The PRNG state consists of two parts, the large pool 'state' and 'md', 739968651Skris where all of 'md' is used each time the PRNG is used, but 'state' 740068651Skris is used only indexed by a cyclic counter. As entropy may not be 740168651Skris well distributed from the beginning, 'md' is important as a 740268651Skris chaining variable. However, the output function chains only half 740368651Skris of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 740468651Skris all of 'md', and seeding with STATE_SIZE dummy bytes will result 740568651Skris in all of 'state' being rewritten, with the new values depending 740668651Skris on virtually all of 'md'. This overcomes the 80 bit limitation.) 740768651Skris [Bodo Moeller] 740868651Skris 740968651Skris *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 741068651Skris the handshake is continued after ssl_verify_cert_chain(); 741168651Skris otherwise, if SSL_VERIFY_NONE is set, remaining error codes 741268651Skris can lead to 'unexplainable' connection aborts later. 741368651Skris [Bodo Moeller; problem tracked down by Lutz Jaenicke] 741468651Skris 741568651Skris *) Major EVP API cipher revision. 741668651Skris Add hooks for extra EVP features. This allows various cipher 741768651Skris parameters to be set in the EVP interface. Support added for variable 741868651Skris key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 741968651Skris setting of RC2 and RC5 parameters. 742068651Skris 742168651Skris Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 742268651Skris ciphers. 742368651Skris 742468651Skris Remove lots of duplicated code from the EVP library. For example *every* 742568651Skris cipher init() function handles the 'iv' in the same way according to the 742668651Skris cipher mode. They also all do nothing if the 'key' parameter is NULL and 742768651Skris for CFB and OFB modes they zero ctx->num. 742868651Skris 742968651Skris New functionality allows removal of S/MIME code RC2 hack. 743068651Skris 743168651Skris Most of the routines have the same form and so can be declared in terms 743268651Skris of macros. 743368651Skris 743468651Skris By shifting this to the top level EVP_CipherInit() it can be removed from 743568651Skris all individual ciphers. If the cipher wants to handle IVs or keys 743668651Skris differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 743768651Skris flags. 743868651Skris 743968651Skris Change lots of functions like EVP_EncryptUpdate() to now return a 744068651Skris value: although software versions of the algorithms cannot fail 744168651Skris any installed hardware versions can. 744268651Skris [Steve Henson] 744368651Skris 744468651Skris *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 744568651Skris this option is set, tolerate broken clients that send the negotiated 744668651Skris protocol version number instead of the requested protocol version 744768651Skris number. 744868651Skris [Bodo Moeller] 744968651Skris 745068651Skris *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 745168651Skris i.e. non-zero for export ciphersuites, zero otherwise. 745268651Skris Previous versions had this flag inverted, inconsistent with 745368651Skris rsa_tmp_cb (..._TMP_RSA_CB). 745468651Skris [Bodo Moeller; problem reported by Amit Chopra] 745568651Skris 745668651Skris *) Add missing DSA library text string. Work around for some IIS 745768651Skris key files with invalid SEQUENCE encoding. 745868651Skris [Steve Henson] 745968651Skris 746068651Skris *) Add a document (doc/standards.txt) that list all kinds of standards 746168651Skris and so on that are implemented in OpenSSL. 746268651Skris [Richard Levitte] 746368651Skris 746468651Skris *) Enhance c_rehash script. Old version would mishandle certificates 746568651Skris with the same subject name hash and wouldn't handle CRLs at all. 746668651Skris Added -fingerprint option to crl utility, to support new c_rehash 746768651Skris features. 746868651Skris [Steve Henson] 746968651Skris 747068651Skris *) Eliminate non-ANSI declarations in crypto.h and stack.h. 747168651Skris [Ulf M�ller] 747268651Skris 747368651Skris *) Fix for SSL server purpose checking. Server checking was 747468651Skris rejecting certificates which had extended key usage present 747568651Skris but no ssl client purpose. 747668651Skris [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 747768651Skris 747868651Skris *) Make PKCS#12 code work with no password. The PKCS#12 spec 747968651Skris is a little unclear about how a blank password is handled. 748068651Skris Since the password in encoded as a BMPString with terminating 748168651Skris double NULL a zero length password would end up as just the 748268651Skris double NULL. However no password at all is different and is 748368651Skris handled differently in the PKCS#12 key generation code. NS 748468651Skris treats a blank password as zero length. MSIE treats it as no 748568651Skris password on export: but it will try both on import. We now do 748668651Skris the same: PKCS12_parse() tries zero length and no password if 748768651Skris the password is set to "" or NULL (NULL is now a valid password: 748868651Skris it wasn't before) as does the pkcs12 application. 748968651Skris [Steve Henson] 749068651Skris 749168651Skris *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 749268651Skris perror when PEM_read_bio_X509_REQ fails, the error message must 749368651Skris be obtained from the error queue. 749468651Skris [Bodo Moeller] 749568651Skris 749668651Skris *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 749768651Skris it in ERR_remove_state if appropriate, and change ERR_get_state 749868651Skris accordingly to avoid race conditions (this is necessary because 749968651Skris thread_hash is no longer constant once set). 750068651Skris [Bodo Moeller] 750168651Skris 750268651Skris *) Bugfix for linux-elf makefile.one. 750368651Skris [Ulf M�ller] 750468651Skris 750568651Skris *) RSA_get_default_method() will now cause a default 750668651Skris RSA_METHOD to be chosen if one doesn't exist already. 750768651Skris Previously this was only set during a call to RSA_new() 750868651Skris or RSA_new_method(NULL) meaning it was possible for 750968651Skris RSA_get_default_method() to return NULL. 751068651Skris [Geoff Thorpe] 751168651Skris 751268651Skris *) Added native name translation to the existing DSO code 751368651Skris that will convert (if the flag to do so is set) filenames 751468651Skris that are sufficiently small and have no path information 751568651Skris into a canonical native form. Eg. "blah" converted to 751668651Skris "libblah.so" or "blah.dll" etc. 751768651Skris [Geoff Thorpe] 751868651Skris 751968651Skris *) New function ERR_error_string_n(e, buf, len) which is like 752068651Skris ERR_error_string(e, buf), but writes at most 'len' bytes 752168651Skris including the 0 terminator. For ERR_error_string_n, 'buf' 752268651Skris may not be NULL. 752368651Skris [Damien Miller <djm@mindrot.org>, Bodo Moeller] 752468651Skris 752568651Skris *) CONF library reworked to become more general. A new CONF 752668651Skris configuration file reader "class" is implemented as well as a 752768651Skris new functions (NCONF_*, for "New CONF") to handle it. The now 752868651Skris old CONF_* functions are still there, but are reimplemented to 752968651Skris work in terms of the new functions. Also, a set of functions 753068651Skris to handle the internal storage of the configuration data is 753168651Skris provided to make it easier to write new configuration file 753268651Skris reader "classes" (I can definitely see something reading a 753368651Skris configuration file in XML format, for example), called _CONF_*, 753468651Skris or "the configuration storage API"... 753568651Skris 753668651Skris The new configuration file reading functions are: 753768651Skris 753868651Skris NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 753968651Skris NCONF_get_section, NCONF_get_string, NCONF_get_numbre 754068651Skris 754168651Skris NCONF_default, NCONF_WIN32 754268651Skris 754368651Skris NCONF_dump_fp, NCONF_dump_bio 754468651Skris 754568651Skris NCONF_default and NCONF_WIN32 are method (or "class") choosers, 754668651Skris NCONF_new creates a new CONF object. This works in the same way 754768651Skris as other interfaces in OpenSSL, like the BIO interface. 754868651Skris NCONF_dump_* dump the internal storage of the configuration file, 754968651Skris which is useful for debugging. All other functions take the same 755068651Skris arguments as the old CONF_* functions wth the exception of the 755168651Skris first that must be a `CONF *' instead of a `LHASH *'. 755268651Skris 755368651Skris To make it easer to use the new classes with the old CONF_* functions, 755468651Skris the function CONF_set_default_method is provided. 755568651Skris [Richard Levitte] 755668651Skris 755768651Skris *) Add '-tls1' option to 'openssl ciphers', which was already 755868651Skris mentioned in the documentation but had not been implemented. 755968651Skris (This option is not yet really useful because even the additional 756068651Skris experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 756168651Skris [Bodo Moeller] 756268651Skris 756368651Skris *) Initial DSO code added into libcrypto for letting OpenSSL (and 756468651Skris OpenSSL-based applications) load shared libraries and bind to 756568651Skris them in a portable way. 756668651Skris [Geoff Thorpe, with contributions from Richard Levitte] 756768651Skris 756859191Skris Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 756959191Skris 757059191Skris *) Make sure _lrotl and _lrotr are only used with MSVC. 757159191Skris 757259191Skris *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 757359191Skris (the default implementation of RAND_status). 757459191Skris 757559191Skris *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 757659191Skris to '-clrext' (= clear extensions), as intended and documented. 757759191Skris [Bodo Moeller; inconsistency pointed out by Michael Attili 757859191Skris <attili@amaxo.com>] 757959191Skris 758059191Skris *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 758159191Skris was larger than the MD block size. 758259191Skris [Steve Henson, pointed out by Yost William <YostW@tce.com>] 758359191Skris 758459191Skris *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 758559191Skris fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 758659191Skris using the passed key: if the passed key was a private key the result 758759191Skris of X509_print(), for example, would be to print out all the private key 758859191Skris components. 758959191Skris [Steve Henson] 759059191Skris 759159191Skris *) des_quad_cksum() byte order bug fix. 759259191Skris [Ulf M�ller, using the problem description in krb4-0.9.7, where 759359191Skris the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 759459191Skris 759559191Skris *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 759659191Skris discouraged. 759759191Skris [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 759859191Skris 759959191Skris *) For easily testing in shell scripts whether some command 760059191Skris 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 760159191Skris returns with exit code 0 iff no command of the given name is available. 760259191Skris 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 760359191Skris the output goes to stdout and nothing is printed to stderr. 760459191Skris Additional arguments are always ignored. 760559191Skris 760659191Skris Since for each cipher there is a command of the same name, 760759191Skris the 'no-cipher' compilation switches can be tested this way. 760859191Skris 760959191Skris ('openssl no-XXX' is not able to detect pseudo-commands such 761059191Skris as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 761159191Skris [Bodo Moeller] 761259191Skris 761359191Skris *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 761459191Skris [Bodo Moeller] 761559191Skris 761659191Skris *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 761759191Skris is set; it will be thrown away anyway because each handshake creates 761859191Skris its own key. 761959191Skris ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 762059191Skris to parameters -- in previous versions (since OpenSSL 0.9.3) the 762159191Skris 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 762259191Skris you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 762359191Skris [Bodo Moeller] 762459191Skris 762559191Skris *) New s_client option -ign_eof: EOF at stdin is ignored, and 762659191Skris 'Q' and 'R' lose their special meanings (quit/renegotiate). 762759191Skris This is part of what -quiet does; unlike -quiet, -ign_eof 762859191Skris does not suppress any output. 762959191Skris [Richard Levitte] 763059191Skris 763159191Skris *) Add compatibility options to the purpose and trust code. The 763259191Skris purpose X509_PURPOSE_ANY is "any purpose" which automatically 763359191Skris accepts a certificate or CA, this was the previous behaviour, 763459191Skris with all the associated security issues. 763559191Skris 763659191Skris X509_TRUST_COMPAT is the old trust behaviour: only and 763759191Skris automatically trust self signed roots in certificate store. A 763859191Skris new trust setting X509_TRUST_DEFAULT is used to specify that 763959191Skris a purpose has no associated trust setting and it should instead 764059191Skris use the value in the default purpose. 764159191Skris [Steve Henson] 764259191Skris 764359191Skris *) Fix the PKCS#8 DSA private key code so it decodes keys again 764459191Skris and fix a memory leak. 764559191Skris [Steve Henson] 764659191Skris 764759191Skris *) In util/mkerr.pl (which implements 'make errors'), preserve 764859191Skris reason strings from the previous version of the .c file, as 764959191Skris the default to have only downcase letters (and digits) in 765059191Skris automatically generated reasons codes is not always appropriate. 765159191Skris [Bodo Moeller] 765259191Skris 765359191Skris *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 765459191Skris using strerror. Previously, ERR_reason_error_string() returned 765559191Skris library names as reason strings for SYSerr; but SYSerr is a special 765659191Skris case where small numbers are errno values, not library numbers. 765759191Skris [Bodo Moeller] 765859191Skris 765959191Skris *) Add '-dsaparam' option to 'openssl dhparam' application. This 766059191Skris converts DSA parameters into DH parameters. (When creating parameters, 766159191Skris DSA_generate_parameters is used.) 766259191Skris [Bodo Moeller] 766359191Skris 766459191Skris *) Include 'length' (recommended exponent length) in C code generated 766559191Skris by 'openssl dhparam -C'. 766659191Skris [Bodo Moeller] 766759191Skris 766859191Skris *) The second argument to set_label in perlasm was already being used 766959191Skris so couldn't be used as a "file scope" flag. Moved to third argument 767059191Skris which was free. 767159191Skris [Steve Henson] 767259191Skris 767359191Skris *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 767459191Skris instead of RAND_bytes for encryption IVs and salts. 767559191Skris [Bodo Moeller] 767659191Skris 767759191Skris *) Include RAND_status() into RAND_METHOD instead of implementing 767859191Skris it only for md_rand.c Otherwise replacing the PRNG by calling 767959191Skris RAND_set_rand_method would be impossible. 768059191Skris [Bodo Moeller] 768159191Skris 768259191Skris *) Don't let DSA_generate_key() enter an infinite loop if the random 768359191Skris number generation fails. 768459191Skris [Bodo Moeller] 768559191Skris 768659191Skris *) New 'rand' application for creating pseudo-random output. 768759191Skris [Bodo Moeller] 768859191Skris 768959191Skris *) Added configuration support for Linux/IA64 769059191Skris [Rolf Haberrecker <rolf@suse.de>] 769159191Skris 769259191Skris *) Assembler module support for Mingw32. 769359191Skris [Ulf M�ller] 769459191Skris 769559191Skris *) Shared library support for HPUX (in shlib/). 769659191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 769759191Skris 769859191Skris *) Shared library support for Solaris gcc. 769959191Skris [Lutz Behnke <behnke@trustcenter.de>] 770059191Skris 770159191Skris Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 770259191Skris 770359191Skris *) PKCS7_encrypt() was adding text MIME headers twice because they 770459191Skris were added manually and by SMIME_crlf_copy(). 770559191Skris [Steve Henson] 770659191Skris 770759191Skris *) In bntest.c don't call BN_rand with zero bits argument. 770859191Skris [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 770959191Skris 771059191Skris *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 771159191Skris case was implemented. This caused BN_div_recp() to fail occasionally. 771259191Skris [Ulf M�ller] 771359191Skris 771459191Skris *) Add an optional second argument to the set_label() in the perl 771559191Skris assembly language builder. If this argument exists and is set 771659191Skris to 1 it signals that the assembler should use a symbol whose 771759191Skris scope is the entire file, not just the current function. This 771859191Skris is needed with MASM which uses the format label:: for this scope. 771959191Skris [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 772059191Skris 772159191Skris *) Change the ASN1 types so they are typedefs by default. Before 772259191Skris almost all types were #define'd to ASN1_STRING which was causing 772359191Skris STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 772459191Skris for example. 772559191Skris [Steve Henson] 772659191Skris 772759191Skris *) Change names of new functions to the new get1/get0 naming 772859191Skris convention: After 'get1', the caller owns a reference count 772959191Skris and has to call ..._free; 'get0' returns a pointer to some 773059191Skris data structure without incrementing reference counters. 773159191Skris (Some of the existing 'get' functions increment a reference 773259191Skris counter, some don't.) 773359191Skris Similarly, 'set1' and 'add1' functions increase reference 773459191Skris counters or duplicate objects. 773559191Skris [Steve Henson] 773659191Skris 773759191Skris *) Allow for the possibility of temp RSA key generation failure: 773859191Skris the code used to assume it always worked and crashed on failure. 773959191Skris [Steve Henson] 774059191Skris 774159191Skris *) Fix potential buffer overrun problem in BIO_printf(). 774259191Skris [Ulf M�ller, using public domain code by Patrick Powell; problem 774359191Skris pointed out by David Sacerdote <das33@cornell.edu>] 774459191Skris 774559191Skris *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 774659191Skris RAND_egd() and RAND_status(). In the command line application, 774759191Skris the EGD socket can be specified like a seed file using RANDFILE 774859191Skris or -rand. 774959191Skris [Ulf M�ller] 775059191Skris 775159191Skris *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 775259191Skris Some CAs (e.g. Verisign) distribute certificates in this form. 775359191Skris [Steve Henson] 775459191Skris 775559191Skris *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 775659191Skris list to exclude them. This means that no special compilation option 775759191Skris is needed to use anonymous DH: it just needs to be included in the 775859191Skris cipher list. 775959191Skris [Steve Henson] 776059191Skris 776159191Skris *) Change the EVP_MD_CTX_type macro so its meaning consistent with 776259191Skris EVP_MD_type. The old functionality is available in a new macro called 776359191Skris EVP_MD_md(). Change code that uses it and update docs. 776459191Skris [Steve Henson] 776559191Skris 776659191Skris *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 776759191Skris where the 'void *' argument is replaced by a function pointer argument. 776859191Skris Previously 'void *' was abused to point to functions, which works on 776959191Skris many platforms, but is not correct. As these functions are usually 777059191Skris called by macros defined in OpenSSL header files, most source code 777159191Skris should work without changes. 777259191Skris [Richard Levitte] 777359191Skris 777459191Skris *) <openssl/opensslconf.h> (which is created by Configure) now contains 777559191Skris sections with information on -D... compiler switches used for 777659191Skris compiling the library so that applications can see them. To enable 777759191Skris one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 777859191Skris must be defined. E.g., 777959191Skris #define OPENSSL_ALGORITHM_DEFINES 778059191Skris #include <openssl/opensslconf.h> 778159191Skris defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 778259191Skris [Richard Levitte, Ulf and Bodo M�ller] 778359191Skris 778459191Skris *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 778559191Skris record layer. 778659191Skris [Bodo Moeller] 778759191Skris 778859191Skris *) Change the 'other' type in certificate aux info to a STACK_OF 778959191Skris X509_ALGOR. Although not an AlgorithmIdentifier as such it has 779059191Skris the required ASN1 format: arbitrary types determined by an OID. 779159191Skris [Steve Henson] 779259191Skris 779359191Skris *) Add some PEM_write_X509_REQ_NEW() functions and a command line 779459191Skris argument to 'req'. This is not because the function is newer or 779559191Skris better than others it just uses the work 'NEW' in the certificate 779659191Skris request header lines. Some software needs this. 779759191Skris [Steve Henson] 779859191Skris 779959191Skris *) Reorganise password command line arguments: now passwords can be 780059191Skris obtained from various sources. Delete the PEM_cb function and make 780159191Skris it the default behaviour: i.e. if the callback is NULL and the 780259191Skris usrdata argument is not NULL interpret it as a null terminated pass 780359191Skris phrase. If usrdata and the callback are NULL then the pass phrase 780459191Skris is prompted for as usual. 780559191Skris [Steve Henson] 780659191Skris 780759191Skris *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 780859191Skris the support is automatically enabled. The resulting binaries will 780959191Skris autodetect the card and use it if present. 781059191Skris [Ben Laurie and Compaq Inc.] 781159191Skris 781259191Skris *) Work around for Netscape hang bug. This sends certificate request 781359191Skris and server done in one record. Since this is perfectly legal in the 781459191Skris SSL/TLS protocol it isn't a "bug" option and is on by default. See 781559191Skris the bugs/SSLv3 entry for more info. 781659191Skris [Steve Henson] 781759191Skris 781859191Skris *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 781959191Skris [Andy Polyakov] 782059191Skris 782159191Skris *) Add -rand argument to smime and pkcs12 applications and read/write 782259191Skris of seed file. 782359191Skris [Steve Henson] 782459191Skris 782559191Skris *) New 'passwd' tool for crypt(3) and apr1 password hashes. 782659191Skris [Bodo Moeller] 782759191Skris 782859191Skris *) Add command line password options to the remaining applications. 782959191Skris [Steve Henson] 783059191Skris 783159191Skris *) Bug fix for BN_div_recp() for numerators with an even number of 783259191Skris bits. 783359191Skris [Ulf M�ller] 783459191Skris 783559191Skris *) More tests in bntest.c, and changed test_bn output. 783659191Skris [Ulf M�ller] 783759191Skris 783859191Skris *) ./config recognizes MacOS X now. 783959191Skris [Andy Polyakov] 784059191Skris 784159191Skris *) Bug fix for BN_div() when the first words of num and divsor are 784259191Skris equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 784359191Skris [Ulf M�ller] 784459191Skris 784559191Skris *) Add support for various broken PKCS#8 formats, and command line 784659191Skris options to produce them. 784759191Skris [Steve Henson] 784859191Skris 784959191Skris *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 785059191Skris get temporary BIGNUMs from a BN_CTX. 785159191Skris [Ulf M�ller] 785259191Skris 785359191Skris *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 785459191Skris for p == 0. 785559191Skris [Ulf M�ller] 785659191Skris 785759191Skris *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 785859191Skris include a #define from the old name to the new. The original intent 785959191Skris was that statically linked binaries could for example just call 786059191Skris SSLeay_add_all_ciphers() to just add ciphers to the table and not 786159191Skris link with digests. This never worked becayse SSLeay_add_all_digests() 786259191Skris and SSLeay_add_all_ciphers() were in the same source file so calling 786359191Skris one would link with the other. They are now in separate source files. 786459191Skris [Steve Henson] 786559191Skris 786659191Skris *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 786759191Skris [Steve Henson] 786859191Skris 786959191Skris *) Use a less unusual form of the Miller-Rabin primality test (it used 787059191Skris a binary algorithm for exponentiation integrated into the Miller-Rabin 787159191Skris loop, our standard modexp algorithms are faster). 787259191Skris [Bodo Moeller] 787359191Skris 787459191Skris *) Support for the EBCDIC character set completed. 787559191Skris [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 787659191Skris 787759191Skris *) Source code cleanups: use const where appropriate, eliminate casts, 787859191Skris use void * instead of char * in lhash. 787959191Skris [Ulf M�ller] 788059191Skris 788159191Skris *) Bugfix: ssl3_send_server_key_exchange was not restartable 788259191Skris (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 788359191Skris this the server could overwrite ephemeral keys that the client 788459191Skris has already seen). 788559191Skris [Bodo Moeller] 788659191Skris 788759191Skris *) Turn DSA_is_prime into a macro that calls BN_is_prime, 788859191Skris using 50 iterations of the Rabin-Miller test. 788959191Skris 789059191Skris DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 789159191Skris iterations of the Rabin-Miller test as required by the appendix 789259191Skris to FIPS PUB 186[-1]) instead of DSA_is_prime. 789359191Skris As BN_is_prime_fasttest includes trial division, DSA parameter 789459191Skris generation becomes much faster. 789559191Skris 789659191Skris This implies a change for the callback functions in DSA_is_prime 789759191Skris and DSA_generate_parameters: The callback function is called once 789859191Skris for each positive witness in the Rabin-Miller test, not just 789959191Skris occasionally in the inner loop; and the parameters to the 790059191Skris callback function now provide an iteration count for the outer 790159191Skris loop rather than for the current invocation of the inner loop. 790259191Skris DSA_generate_parameters additionally can call the callback 790359191Skris function with an 'iteration count' of -1, meaning that a 790459191Skris candidate has passed the trial division test (when q is generated 790559191Skris from an application-provided seed, trial division is skipped). 790659191Skris [Bodo Moeller] 790759191Skris 790859191Skris *) New function BN_is_prime_fasttest that optionally does trial 790959191Skris division before starting the Rabin-Miller test and has 791059191Skris an additional BN_CTX * argument (whereas BN_is_prime always 791159191Skris has to allocate at least one BN_CTX). 791259191Skris 'callback(1, -1, cb_arg)' is called when a number has passed the 791359191Skris trial division stage. 791459191Skris [Bodo Moeller] 791559191Skris 791659191Skris *) Fix for bug in CRL encoding. The validity dates weren't being handled 791759191Skris as ASN1_TIME. 791859191Skris [Steve Henson] 791959191Skris 792059191Skris *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 792159191Skris [Steve Henson] 792259191Skris 792359191Skris *) New function BN_pseudo_rand(). 792459191Skris [Ulf M�ller] 792559191Skris 792659191Skris *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 792759191Skris bignum version of BN_from_montgomery() with the working code from 792859191Skris SSLeay 0.9.0 (the word based version is faster anyway), and clean up 792959191Skris the comments. 793059191Skris [Ulf M�ller] 793159191Skris 793259191Skris *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 793359191Skris made it impossible to use the same SSL_SESSION data structure in 793459191Skris SSL2 clients in multiple threads. 793559191Skris [Bodo Moeller] 793659191Skris 793759191Skris *) The return value of RAND_load_file() no longer counts bytes obtained 793859191Skris by stat(). RAND_load_file(..., -1) is new and uses the complete file 793959191Skris to seed the PRNG (previously an explicit byte count was required). 794059191Skris [Ulf M�ller, Bodo M�ller] 794159191Skris 794259191Skris *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 794359191Skris used (char *) instead of (void *) and had casts all over the place. 794459191Skris [Steve Henson] 794559191Skris 794659191Skris *) Make BN_generate_prime() return NULL on error if ret!=NULL. 794759191Skris [Ulf M�ller] 794859191Skris 794959191Skris *) Retain source code compatibility for BN_prime_checks macro: 795059191Skris BN_is_prime(..., BN_prime_checks, ...) now uses 795159191Skris BN_prime_checks_for_size to determine the appropriate number of 795259191Skris Rabin-Miller iterations. 795359191Skris [Ulf M�ller] 795459191Skris 795559191Skris *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 795659191Skris DH_CHECK_P_NOT_SAFE_PRIME. 795759191Skris (Check if this is true? OpenPGP calls them "strong".) 795859191Skris [Ulf M�ller] 795959191Skris 796059191Skris *) Merge the functionality of "dh" and "gendh" programs into a new program 796159191Skris "dhparam". The old programs are retained for now but will handle DH keys 796259191Skris (instead of parameters) in future. 796359191Skris [Steve Henson] 796459191Skris 796559191Skris *) Make the ciphers, s_server and s_client programs check the return values 796659191Skris when a new cipher list is set. 796759191Skris [Steve Henson] 796859191Skris 796959191Skris *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 797059191Skris ciphers. Before when the 56bit ciphers were enabled the sorting was 797159191Skris wrong. 797259191Skris 797359191Skris The syntax for the cipher sorting has been extended to support sorting by 797459191Skris cipher-strength (using the strength_bits hard coded in the tables). 797559191Skris The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 797659191Skris 797759191Skris Fix a bug in the cipher-command parser: when supplying a cipher command 797859191Skris string with an "undefined" symbol (neither command nor alphanumeric 797959191Skris [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 798059191Skris an error is flagged. 798159191Skris 798259191Skris Due to the strength-sorting extension, the code of the 798359191Skris ssl_create_cipher_list() function was completely rearranged. I hope that 798459191Skris the readability was also increased :-) 798559191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 798659191Skris 798759191Skris *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 798859191Skris for the first serial number and places 2 in the serial number file. This 798959191Skris avoids problems when the root CA is created with serial number zero and 799059191Skris the first user certificate has the same issuer name and serial number 799159191Skris as the root CA. 799259191Skris [Steve Henson] 799359191Skris 799459191Skris *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 799559191Skris the new code. Add documentation for this stuff. 799659191Skris [Steve Henson] 799759191Skris 799859191Skris *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 799959191Skris X509_*() to X509at_*() on the grounds that they don't handle X509 800059191Skris structures and behave in an analagous way to the X509v3 functions: 800159191Skris they shouldn't be called directly but wrapper functions should be used 800259191Skris instead. 800359191Skris 800459191Skris So we also now have some wrapper functions that call the X509at functions 800559191Skris when passed certificate requests. (TO DO: similar things can be done with 800659191Skris PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 800759191Skris things. Some of these need some d2i or i2d and print functionality 800859191Skris because they handle more complex structures.) 800959191Skris [Steve Henson] 801059191Skris 801159191Skris *) Add missing #ifndefs that caused missing symbols when building libssl 801259191Skris as a shared library without RSA. Use #ifndef NO_SSL2 instead of 801359191Skris NO_RSA in ssl/s2*.c. 801459191Skris [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 801559191Skris 801659191Skris *) Precautions against using the PRNG uninitialized: RAND_bytes() now 801759191Skris has a return value which indicates the quality of the random data 801859191Skris (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 801959191Skris error queue. New function RAND_pseudo_bytes() generates output that is 802059191Skris guaranteed to be unique but not unpredictable. RAND_add is like 802159191Skris RAND_seed, but takes an extra argument for an entropy estimate 802259191Skris (RAND_seed always assumes full entropy). 802359191Skris [Ulf M�ller] 802459191Skris 802559191Skris *) Do more iterations of Rabin-Miller probable prime test (specifically, 802659191Skris 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 802759191Skris instead of only 2 for all lengths; see BN_prime_checks_for_size definition 802859191Skris in crypto/bn/bn_prime.c for the complete table). This guarantees a 802959191Skris false-positive rate of at most 2^-80 for random input. 803059191Skris [Bodo Moeller] 803159191Skris 803259191Skris *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 803359191Skris [Bodo Moeller] 803459191Skris 803559191Skris *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 803659191Skris in the 0.9.5 release), this returns the chain 803759191Skris from an X509_CTX structure with a dup of the stack and all 803859191Skris the X509 reference counts upped: so the stack will exist 803959191Skris after X509_CTX_cleanup() has been called. Modify pkcs12.c 804059191Skris to use this. 804159191Skris 804259191Skris Also make SSL_SESSION_print() print out the verify return 804359191Skris code. 804459191Skris [Steve Henson] 804559191Skris 804659191Skris *) Add manpage for the pkcs12 command. Also change the default 804759191Skris behaviour so MAC iteration counts are used unless the new 804859191Skris -nomaciter option is used. This improves file security and 804959191Skris only older versions of MSIE (4.0 for example) need it. 805059191Skris [Steve Henson] 805159191Skris 805259191Skris *) Honor the no-xxx Configure options when creating .DEF files. 805359191Skris [Ulf M�ller] 805459191Skris 805559191Skris *) Add PKCS#10 attributes to field table: challengePassword, 805659191Skris unstructuredName and unstructuredAddress. These are taken from 805759191Skris draft PKCS#9 v2.0 but are compatible with v1.2 provided no 805859191Skris international characters are used. 805959191Skris 806059191Skris More changes to X509_ATTRIBUTE code: allow the setting of types 806159191Skris based on strings. Remove the 'loc' parameter when adding 806259191Skris attributes because these will be a SET OF encoding which is sorted 806359191Skris in ASN1 order. 806459191Skris [Steve Henson] 806559191Skris 806659191Skris *) Initial changes to the 'req' utility to allow request generation 806759191Skris automation. This will allow an application to just generate a template 806859191Skris file containing all the field values and have req construct the 806959191Skris request. 807059191Skris 807159191Skris Initial support for X509_ATTRIBUTE handling. Stacks of these are 807259191Skris used all over the place including certificate requests and PKCS#7 807359191Skris structures. They are currently handled manually where necessary with 807459191Skris some primitive wrappers for PKCS#7. The new functions behave in a 807559191Skris manner analogous to the X509 extension functions: they allow 807659191Skris attributes to be looked up by NID and added. 807759191Skris 807859191Skris Later something similar to the X509V3 code would be desirable to 807959191Skris automatically handle the encoding, decoding and printing of the 808059191Skris more complex types. The string types like challengePassword can 808159191Skris be handled by the string table functions. 808259191Skris 808359191Skris Also modified the multi byte string table handling. Now there is 808459191Skris a 'global mask' which masks out certain types. The table itself 808559191Skris can use the flag STABLE_NO_MASK to ignore the mask setting: this 808659191Skris is useful when for example there is only one permissible type 808759191Skris (as in countryName) and using the mask might result in no valid 808859191Skris types at all. 808959191Skris [Steve Henson] 809059191Skris 809159191Skris *) Clean up 'Finished' handling, and add functions SSL_get_finished and 809259191Skris SSL_get_peer_finished to allow applications to obtain the latest 809359191Skris Finished messages sent to the peer or expected from the peer, 809459191Skris respectively. (SSL_get_peer_finished is usually the Finished message 809559191Skris actually received from the peer, otherwise the protocol will be aborted.) 809659191Skris 809759191Skris As the Finished message are message digests of the complete handshake 809859191Skris (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 809959191Skris be used for external authentication procedures when the authentication 810059191Skris provided by SSL/TLS is not desired or is not enough. 810159191Skris [Bodo Moeller] 810259191Skris 810359191Skris *) Enhanced support for Alpha Linux is added. Now ./config checks if 810459191Skris the host supports BWX extension and if Compaq C is present on the 810559191Skris $PATH. Just exploiting of the BWX extension results in 20-30% 810659191Skris performance kick for some algorithms, e.g. DES and RC4 to mention 810759191Skris a couple. Compaq C in turn generates ~20% faster code for MD5 and 810859191Skris SHA1. 810959191Skris [Andy Polyakov] 811059191Skris 811159191Skris *) Add support for MS "fast SGC". This is arguably a violation of the 811259191Skris SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 811359191Skris weak crypto and after checking the certificate is SGC a second one 811459191Skris with strong crypto. MS SGC stops the first handshake after receiving 811559191Skris the server certificate message and sends a second client hello. Since 811659191Skris a server will typically do all the time consuming operations before 811759191Skris expecting any further messages from the client (server key exchange 811859191Skris is the most expensive) there is little difference between the two. 811959191Skris 812059191Skris To get OpenSSL to support MS SGC we have to permit a second client 812159191Skris hello message after we have sent server done. In addition we have to 812259191Skris reset the MAC if we do get this second client hello. 812359191Skris [Steve Henson] 812459191Skris 812559191Skris *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 812659191Skris if a DER encoded private key is RSA or DSA traditional format. Changed 812759191Skris d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 812859191Skris format DER encoded private key. Newer code should use PKCS#8 format which 812959191Skris has the key type encoded in the ASN1 structure. Added DER private key 813059191Skris support to pkcs8 application. 813159191Skris [Steve Henson] 813259191Skris 813359191Skris *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 813459191Skris ciphersuites has been selected (as required by the SSL 3/TLS 1 813559191Skris specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 813659191Skris is set, we interpret this as a request to violate the specification 813759191Skris (the worst that can happen is a handshake failure, and 'correct' 813859191Skris behaviour would result in a handshake failure anyway). 813959191Skris [Bodo Moeller] 814059191Skris 814159191Skris *) In SSL_CTX_add_session, take into account that there might be multiple 814259191Skris SSL_SESSION structures with the same session ID (e.g. when two threads 814359191Skris concurrently obtain them from an external cache). 814459191Skris The internal cache can handle only one SSL_SESSION with a given ID, 814559191Skris so if there's a conflict, we now throw out the old one to achieve 814659191Skris consistency. 814759191Skris [Bodo Moeller] 814859191Skris 814959191Skris *) Add OIDs for idea and blowfish in CBC mode. This will allow both 815059191Skris to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 815159191Skris some routines that use cipher OIDs: some ciphers do not have OIDs 815259191Skris defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 815359191Skris example. 815459191Skris [Steve Henson] 815559191Skris 815659191Skris *) Simplify the trust setting structure and code. Now we just have 815759191Skris two sequences of OIDs for trusted and rejected settings. These will 815859191Skris typically have values the same as the extended key usage extension 815959191Skris and any application specific purposes. 816059191Skris 816159191Skris The trust checking code now has a default behaviour: it will just 816259191Skris check for an object with the same NID as the passed id. Functions can 816359191Skris be provided to override either the default behaviour or the behaviour 816459191Skris for a given id. SSL client, server and email already have functions 816559191Skris in place for compatibility: they check the NID and also return "trusted" 816659191Skris if the certificate is self signed. 816759191Skris [Steve Henson] 816859191Skris 816959191Skris *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 817059191Skris traditional format into an EVP_PKEY structure. 817159191Skris [Steve Henson] 817259191Skris 817359191Skris *) Add a password callback function PEM_cb() which either prompts for 817459191Skris a password if usr_data is NULL or otherwise assumes it is a null 817559191Skris terminated password. Allow passwords to be passed on command line 817659191Skris environment or config files in a few more utilities. 817759191Skris [Steve Henson] 817859191Skris 817959191Skris *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 818059191Skris keys. Add some short names for PKCS#8 PBE algorithms and allow them 818159191Skris to be specified on the command line for the pkcs8 and pkcs12 utilities. 818259191Skris Update documentation. 818359191Skris [Steve Henson] 818459191Skris 818559191Skris *) Support for ASN1 "NULL" type. This could be handled before by using 818659191Skris ASN1_TYPE but there wasn't any function that would try to read a NULL 818759191Skris and produce an error if it couldn't. For compatibility we also have 818859191Skris ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 818959191Skris don't allocate anything because they don't need to. 819059191Skris [Steve Henson] 819159191Skris 819259191Skris *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 819359191Skris for details. 819459191Skris [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 819559191Skris 819659191Skris *) Rebuild of the memory allocation routines used by OpenSSL code and 819759191Skris possibly others as well. The purpose is to make an interface that 819859191Skris provide hooks so anyone can build a separate set of allocation and 819959191Skris deallocation routines to be used by OpenSSL, for example memory 820059191Skris pool implementations, or something else, which was previously hard 820159191Skris since Malloc(), Realloc() and Free() were defined as macros having 820259191Skris the values malloc, realloc and free, respectively (except for Win32 820359191Skris compilations). The same is provided for memory debugging code. 820459191Skris OpenSSL already comes with functionality to find memory leaks, but 820559191Skris this gives people a chance to debug other memory problems. 820659191Skris 820759191Skris With these changes, a new set of functions and macros have appeared: 820859191Skris 820968651Skris CRYPTO_set_mem_debug_functions() [F] 821068651Skris CRYPTO_get_mem_debug_functions() [F] 821168651Skris CRYPTO_dbg_set_options() [F] 821268651Skris CRYPTO_dbg_get_options() [F] 821368651Skris CRYPTO_malloc_debug_init() [M] 821459191Skris 821559191Skris The memory debug functions are NULL by default, unless the library 821659191Skris is compiled with CRYPTO_MDEBUG or friends is defined. If someone 821759191Skris wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 821859191Skris gives the standard debugging functions that come with OpenSSL) or 821959191Skris CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 822059191Skris provided by the library user) must be used. When the standard 822159191Skris debugging functions are used, CRYPTO_dbg_set_options can be used to 822259191Skris request additional information: 822359191Skris CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 822459191Skris the CRYPTO_MDEBUG_xxx macro when compiling the library. 822559191Skris 822659191Skris Also, things like CRYPTO_set_mem_functions will always give the 822759191Skris expected result (the new set of functions is used for allocation 822859191Skris and deallocation) at all times, regardless of platform and compiler 822959191Skris options. 823059191Skris 823159191Skris To finish it up, some functions that were never use in any other 823259191Skris way than through macros have a new API and new semantic: 823359191Skris 823459191Skris CRYPTO_dbg_malloc() 823559191Skris CRYPTO_dbg_realloc() 823659191Skris CRYPTO_dbg_free() 823759191Skris 823859191Skris All macros of value have retained their old syntax. 823959191Skris [Richard Levitte and Bodo Moeller] 824059191Skris 824159191Skris *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 824259191Skris ordering of SMIMECapabilities wasn't in "strength order" and there 824359191Skris was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 824459191Skris algorithm. 824559191Skris [Steve Henson] 824659191Skris 824759191Skris *) Some ASN1 types with illegal zero length encoding (INTEGER, 824859191Skris ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 824959191Skris [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 825059191Skris 825159191Skris *) Merge in my S/MIME library for OpenSSL. This provides a simple 825259191Skris S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 825359191Skris functionality to handle multipart/signed properly) and a utility 825459191Skris called 'smime' to call all this stuff. This is based on code I 825559191Skris originally wrote for Celo who have kindly allowed it to be 825659191Skris included in OpenSSL. 825759191Skris [Steve Henson] 825859191Skris 825959191Skris *) Add variants des_set_key_checked and des_set_key_unchecked of 826059191Skris des_set_key (aka des_key_sched). Global variable des_check_key 826159191Skris decides which of these is called by des_set_key; this way 826259191Skris des_check_key behaves as it always did, but applications and 826359191Skris the library itself, which was buggy for des_check_key == 1, 826459191Skris have a cleaner way to pick the version they need. 826559191Skris [Bodo Moeller] 826659191Skris 826759191Skris *) New function PKCS12_newpass() which changes the password of a 826859191Skris PKCS12 structure. 826959191Skris [Steve Henson] 827059191Skris 827159191Skris *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 827259191Skris dynamic mix. In both cases the ids can be used as an index into the 827359191Skris table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 827459191Skris functions so they accept a list of the field values and the 827559191Skris application doesn't need to directly manipulate the X509_TRUST 827659191Skris structure. 827759191Skris [Steve Henson] 827859191Skris 827959191Skris *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 828059191Skris need initialising. 828159191Skris [Steve Henson] 828259191Skris 828359191Skris *) Modify the way the V3 extension code looks up extensions. This now 828459191Skris works in a similar way to the object code: we have some "standard" 828559191Skris extensions in a static table which is searched with OBJ_bsearch() 828659191Skris and the application can add dynamic ones if needed. The file 828759191Skris crypto/x509v3/ext_dat.h now has the info: this file needs to be 828859191Skris updated whenever a new extension is added to the core code and kept 828959191Skris in ext_nid order. There is a simple program 'tabtest.c' which checks 829059191Skris this. New extensions are not added too often so this file can readily 829159191Skris be maintained manually. 829259191Skris 829359191Skris There are two big advantages in doing things this way. The extensions 829459191Skris can be looked up immediately and no longer need to be "added" using 829559191Skris X509V3_add_standard_extensions(): this function now does nothing. 829659191Skris [Side note: I get *lots* of email saying the extension code doesn't 829759191Skris work because people forget to call this function] 829859191Skris Also no dynamic allocation is done unless new extensions are added: 829959191Skris so if we don't add custom extensions there is no need to call 830059191Skris X509V3_EXT_cleanup(). 830159191Skris [Steve Henson] 830259191Skris 830359191Skris *) Modify enc utility's salting as follows: make salting the default. Add a 830459191Skris magic header, so unsalted files fail gracefully instead of just decrypting 830559191Skris to garbage. This is because not salting is a big security hole, so people 830659191Skris should be discouraged from doing it. 830759191Skris [Ben Laurie] 830859191Skris 830959191Skris *) Fixes and enhancements to the 'x509' utility. It allowed a message 831059191Skris digest to be passed on the command line but it only used this 831159191Skris parameter when signing a certificate. Modified so all relevant 831259191Skris operations are affected by the digest parameter including the 831359191Skris -fingerprint and -x509toreq options. Also -x509toreq choked if a 831459191Skris DSA key was used because it didn't fix the digest. 831559191Skris [Steve Henson] 831659191Skris 831759191Skris *) Initial certificate chain verify code. Currently tests the untrusted 831859191Skris certificates for consistency with the verify purpose (which is set 831959191Skris when the X509_STORE_CTX structure is set up) and checks the pathlength. 832059191Skris 832159191Skris There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 832259191Skris this is because it will reject chains with invalid extensions whereas 832359191Skris every previous version of OpenSSL and SSLeay made no checks at all. 832459191Skris 832559191Skris Trust code: checks the root CA for the relevant trust settings. Trust 832659191Skris settings have an initial value consistent with the verify purpose: e.g. 832759191Skris if the verify purpose is for SSL client use it expects the CA to be 832859191Skris trusted for SSL client use. However the default value can be changed to 832959191Skris permit custom trust settings: one example of this would be to only trust 833059191Skris certificates from a specific "secure" set of CAs. 833159191Skris 833259191Skris Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 833359191Skris which should be used for version portability: especially since the 833459191Skris verify structure is likely to change more often now. 833559191Skris 833659191Skris SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 833759191Skris to set them. If not set then assume SSL clients will verify SSL servers 833859191Skris and vice versa. 833959191Skris 834059191Skris Two new options to the verify program: -untrusted allows a set of 834159191Skris untrusted certificates to be passed in and -purpose which sets the 834259191Skris intended purpose of the certificate. If a purpose is set then the 834359191Skris new chain verify code is used to check extension consistency. 834459191Skris [Steve Henson] 834559191Skris 834659191Skris *) Support for the authority information access extension. 834759191Skris [Steve Henson] 834859191Skris 834959191Skris *) Modify RSA and DSA PEM read routines to transparently handle 835059191Skris PKCS#8 format private keys. New *_PUBKEY_* functions that handle 835159191Skris public keys in a format compatible with certificate 835259191Skris SubjectPublicKeyInfo structures. Unfortunately there were already 835359191Skris functions called *_PublicKey_* which used various odd formats so 835459191Skris these are retained for compatibility: however the DSA variants were 835559191Skris never in a public release so they have been deleted. Changed dsa/rsa 835659191Skris utilities to handle the new format: note no releases ever handled public 835759191Skris keys so we should be OK. 835859191Skris 835959191Skris The primary motivation for this change is to avoid the same fiasco 836059191Skris that dogs private keys: there are several incompatible private key 836159191Skris formats some of which are standard and some OpenSSL specific and 836259191Skris require various evil hacks to allow partial transparent handling and 836359191Skris even then it doesn't work with DER formats. Given the option anything 836459191Skris other than PKCS#8 should be dumped: but the other formats have to 836559191Skris stay in the name of compatibility. 836659191Skris 836759191Skris With public keys and the benefit of hindsight one standard format 836859191Skris is used which works with EVP_PKEY, RSA or DSA structures: though 836959191Skris it clearly returns an error if you try to read the wrong kind of key. 837059191Skris 837159191Skris Added a -pubkey option to the 'x509' utility to output the public key. 837259191Skris Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 837359191Skris (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 837459191Skris EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 837559191Skris that do the same as the EVP_PKEY_assign_*() except they up the 837659191Skris reference count of the added key (they don't "swallow" the 837759191Skris supplied key). 837859191Skris [Steve Henson] 837959191Skris 838059191Skris *) Fixes to crypto/x509/by_file.c the code to read in certificates and 838159191Skris CRLs would fail if the file contained no certificates or no CRLs: 838259191Skris added a new function to read in both types and return the number 838359191Skris read: this means that if none are read it will be an error. The 838459191Skris DER versions of the certificate and CRL reader would always fail 838559191Skris because it isn't possible to mix certificates and CRLs in DER format 838659191Skris without choking one or the other routine. Changed this to just read 838759191Skris a certificate: this is the best we can do. Also modified the code 838859191Skris in apps/verify.c to take notice of return codes: it was previously 838959191Skris attempting to read in certificates from NULL pointers and ignoring 839059191Skris any errors: this is one reason why the cert and CRL reader seemed 839159191Skris to work. It doesn't check return codes from the default certificate 839259191Skris routines: these may well fail if the certificates aren't installed. 839359191Skris [Steve Henson] 839459191Skris 839559191Skris *) Code to support otherName option in GeneralName. 839659191Skris [Steve Henson] 839759191Skris 839859191Skris *) First update to verify code. Change the verify utility 839959191Skris so it warns if it is passed a self signed certificate: 840059191Skris for consistency with the normal behaviour. X509_verify 840159191Skris has been modified to it will now verify a self signed 840259191Skris certificate if *exactly* the same certificate appears 840359191Skris in the store: it was previously impossible to trust a 840459191Skris single self signed certificate. This means that: 840559191Skris openssl verify ss.pem 840659191Skris now gives a warning about a self signed certificate but 840759191Skris openssl verify -CAfile ss.pem ss.pem 840859191Skris is OK. 840959191Skris [Steve Henson] 841059191Skris 841159191Skris *) For servers, store verify_result in SSL_SESSION data structure 841259191Skris (and add it to external session representation). 841359191Skris This is needed when client certificate verifications fails, 841459191Skris but an application-provided verification callback (set by 841559191Skris SSL_CTX_set_cert_verify_callback) allows accepting the session 841659191Skris anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 841759191Skris but returns 1): When the session is reused, we have to set 841859191Skris ssl->verify_result to the appropriate error code to avoid 841959191Skris security holes. 842059191Skris [Bodo Moeller, problem pointed out by Lutz Jaenicke] 842159191Skris 842259191Skris *) Fix a bug in the new PKCS#7 code: it didn't consider the 842359191Skris case in PKCS7_dataInit() where the signed PKCS7 structure 842459191Skris didn't contain any existing data because it was being created. 842559191Skris [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 842659191Skris 842759191Skris *) Add a salt to the key derivation routines in enc.c. This 842859191Skris forms the first 8 bytes of the encrypted file. Also add a 842959191Skris -S option to allow a salt to be input on the command line. 843059191Skris [Steve Henson] 843159191Skris 843259191Skris *) New function X509_cmp(). Oddly enough there wasn't a function 843359191Skris to compare two certificates. We do this by working out the SHA1 843459191Skris hash and comparing that. X509_cmp() will be needed by the trust 843559191Skris code. 843659191Skris [Steve Henson] 843759191Skris 843859191Skris *) SSL_get1_session() is like SSL_get_session(), but increments 843959191Skris the reference count in the SSL_SESSION returned. 844059191Skris [Geoff Thorpe <geoff@eu.c2.net>] 844159191Skris 844259191Skris *) Fix for 'req': it was adding a null to request attributes. 844359191Skris Also change the X509_LOOKUP and X509_INFO code to handle 844459191Skris certificate auxiliary information. 844559191Skris [Steve Henson] 844659191Skris 844759191Skris *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 844859191Skris the 'enc' command. 844959191Skris [Steve Henson] 845059191Skris 845159191Skris *) Add the possibility to add extra information to the memory leak 845259191Skris detecting output, to form tracebacks, showing from where each 845359191Skris allocation was originated: CRYPTO_push_info("constant string") adds 845459191Skris the string plus current file name and line number to a per-thread 845559191Skris stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 845659191Skris is like calling CYRPTO_pop_info() until the stack is empty. 845759191Skris Also updated memory leak detection code to be multi-thread-safe. 845859191Skris [Richard Levitte] 845959191Skris 846059191Skris *) Add options -text and -noout to pkcs7 utility and delete the 846159191Skris encryption options which never did anything. Update docs. 846259191Skris [Steve Henson] 846359191Skris 846459191Skris *) Add options to some of the utilities to allow the pass phrase 846559191Skris to be included on either the command line (not recommended on 846659191Skris OSes like Unix) or read from the environment. Update the 846759191Skris manpages and fix a few bugs. 846859191Skris [Steve Henson] 846959191Skris 847059191Skris *) Add a few manpages for some of the openssl commands. 847159191Skris [Steve Henson] 847259191Skris 847359191Skris *) Fix the -revoke option in ca. It was freeing up memory twice, 847459191Skris leaking and not finding already revoked certificates. 847559191Skris [Steve Henson] 847659191Skris 847759191Skris *) Extensive changes to support certificate auxiliary information. 847859191Skris This involves the use of X509_CERT_AUX structure and X509_AUX 847959191Skris functions. An X509_AUX function such as PEM_read_X509_AUX() 848059191Skris can still read in a certificate file in the usual way but it 848159191Skris will also read in any additional "auxiliary information". By 848259191Skris doing things this way a fair degree of compatibility can be 848359191Skris retained: existing certificates can have this information added 848459191Skris using the new 'x509' options. 848559191Skris 848659191Skris Current auxiliary information includes an "alias" and some trust 848759191Skris settings. The trust settings will ultimately be used in enhanced 848859191Skris certificate chain verification routines: currently a certificate 848959191Skris can only be trusted if it is self signed and then it is trusted 849059191Skris for all purposes. 849159191Skris [Steve Henson] 849259191Skris 849359191Skris *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 849459191Skris The problem was that one of the replacement routines had not been working 849559191Skris since SSLeay releases. For now the offending routine has been replaced 849659191Skris with non-optimised assembler. Even so, this now gives around 95% 849759191Skris performance improvement for 1024 bit RSA signs. 849859191Skris [Mark Cox] 849959191Skris 850059191Skris *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 850159191Skris handling. Most clients have the effective key size in bits equal to 850259191Skris the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 850359191Skris A few however don't do this and instead use the size of the decrypted key 850459191Skris to determine the RC2 key length and the AlgorithmIdentifier to determine 850559191Skris the effective key length. In this case the effective key length can still 850659191Skris be 40 bits but the key length can be 168 bits for example. This is fixed 850759191Skris by manually forcing an RC2 key into the EVP_PKEY structure because the 850859191Skris EVP code can't currently handle unusual RC2 key sizes: it always assumes 850959191Skris the key length and effective key length are equal. 851059191Skris [Steve Henson] 851159191Skris 851259191Skris *) Add a bunch of functions that should simplify the creation of 851359191Skris X509_NAME structures. Now you should be able to do: 851459191Skris X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 851559191Skris and have it automatically work out the correct field type and fill in 851659191Skris the structures. The more adventurous can try: 851759191Skris X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 851859191Skris and it will (hopefully) work out the correct multibyte encoding. 851959191Skris [Steve Henson] 852059191Skris 852159191Skris *) Change the 'req' utility to use the new field handling and multibyte 852259191Skris copy routines. Before the DN field creation was handled in an ad hoc 852359191Skris way in req, ca, and x509 which was rather broken and didn't support 852459191Skris BMPStrings or UTF8Strings. Since some software doesn't implement 852559191Skris BMPStrings or UTF8Strings yet, they can be enabled using the config file 852659191Skris using the dirstring_type option. See the new comment in the default 852759191Skris openssl.cnf for more info. 852859191Skris [Steve Henson] 852959191Skris 853059191Skris *) Make crypto/rand/md_rand.c more robust: 853159191Skris - Assure unique random numbers after fork(). 853259191Skris - Make sure that concurrent threads access the global counter and 853359191Skris md serializably so that we never lose entropy in them 853459191Skris or use exactly the same state in multiple threads. 853559191Skris Access to the large state is not always serializable because 853659191Skris the additional locking could be a performance killer, and 853759191Skris md should be large enough anyway. 853859191Skris [Bodo Moeller] 853959191Skris 854059191Skris *) New file apps/app_rand.c with commonly needed functionality 854159191Skris for handling the random seed file. 854259191Skris 854359191Skris Use the random seed file in some applications that previously did not: 854459191Skris ca, 854559191Skris dsaparam -genkey (which also ignored its '-rand' option), 854659191Skris s_client, 854759191Skris s_server, 854859191Skris x509 (when signing). 854959191Skris Except on systems with /dev/urandom, it is crucial to have a random 855059191Skris seed file at least for key creation, DSA signing, and for DH exchanges; 855159191Skris for RSA signatures we could do without one. 855259191Skris 855359191Skris gendh and gendsa (unlike genrsa) used to read only the first byte 855459191Skris of each file listed in the '-rand' option. The function as previously 855559191Skris found in genrsa is now in app_rand.c and is used by all programs 855659191Skris that support '-rand'. 855759191Skris [Bodo Moeller] 855859191Skris 855959191Skris *) In RAND_write_file, use mode 0600 for creating files; 856059191Skris don't just chmod when it may be too late. 856159191Skris [Bodo Moeller] 856259191Skris 856359191Skris *) Report an error from X509_STORE_load_locations 856459191Skris when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 856559191Skris [Bill Perry] 856659191Skris 856759191Skris *) New function ASN1_mbstring_copy() this copies a string in either 856859191Skris ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 856959191Skris into an ASN1_STRING type. A mask of permissible types is passed 857059191Skris and it chooses the "minimal" type to use or an error if not type 857159191Skris is suitable. 857259191Skris [Steve Henson] 857359191Skris 857459191Skris *) Add function equivalents to the various macros in asn1.h. The old 857559191Skris macros are retained with an M_ prefix. Code inside the library can 857659191Skris use the M_ macros. External code (including the openssl utility) 857759191Skris should *NOT* in order to be "shared library friendly". 857859191Skris [Steve Henson] 857959191Skris 858059191Skris *) Add various functions that can check a certificate's extensions 858159191Skris to see if it usable for various purposes such as SSL client, 858259191Skris server or S/MIME and CAs of these types. This is currently 858359191Skris VERY EXPERIMENTAL but will ultimately be used for certificate chain 858459191Skris verification. Also added a -purpose flag to x509 utility to 858559191Skris print out all the purposes. 858659191Skris [Steve Henson] 858759191Skris 858859191Skris *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 858959191Skris functions. 859059191Skris [Steve Henson] 859159191Skris 859259191Skris *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 859359191Skris for, obtain and decode and extension and obtain its critical flag. 859459191Skris This allows all the necessary extension code to be handled in a 859559191Skris single function call. 859659191Skris [Steve Henson] 859759191Skris 859859191Skris *) RC4 tune-up featuring 30-40% performance improvement on most RISC 859959191Skris platforms. See crypto/rc4/rc4_enc.c for further details. 860059191Skris [Andy Polyakov] 860159191Skris 860259191Skris *) New -noout option to asn1parse. This causes no output to be produced 860359191Skris its main use is when combined with -strparse and -out to extract data 860459191Skris from a file (which may not be in ASN.1 format). 860559191Skris [Steve Henson] 860659191Skris 860759191Skris *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 860859191Skris when producing the local key id. 860959191Skris [Richard Levitte <levitte@stacken.kth.se>] 861059191Skris 861159191Skris *) New option -dhparam in s_server. This allows a DH parameter file to be 861259191Skris stated explicitly. If it is not stated then it tries the first server 861359191Skris certificate file. The previous behaviour hard coded the filename 861459191Skris "server.pem". 861559191Skris [Steve Henson] 861659191Skris 861759191Skris *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 861859191Skris a public key to be input or output. For example: 861959191Skris openssl rsa -in key.pem -pubout -out pubkey.pem 862059191Skris Also added necessary DSA public key functions to handle this. 862159191Skris [Steve Henson] 862259191Skris 862359191Skris *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 862459191Skris in the message. This was handled by allowing 862559191Skris X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 862659191Skris [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 862759191Skris 862859191Skris *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 862959191Skris to the end of the strings whereas this didn't. This would cause problems 863059191Skris if strings read with d2i_ASN1_bytes() were later modified. 863159191Skris [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 863259191Skris 863359191Skris *) Fix for base64 decode bug. When a base64 bio reads only one line of 863459191Skris data and it contains EOF it will end up returning an error. This is 863559191Skris caused by input 46 bytes long. The cause is due to the way base64 863659191Skris BIOs find the start of base64 encoded data. They do this by trying a 863759191Skris trial decode on each line until they find one that works. When they 863859191Skris do a flag is set and it starts again knowing it can pass all the 863959191Skris data directly through the decoder. Unfortunately it doesn't reset 864059191Skris the context it uses. This means that if EOF is reached an attempt 864159191Skris is made to pass two EOFs through the context and this causes the 864259191Skris resulting error. This can also cause other problems as well. As is 864359191Skris usual with these problems it takes *ages* to find and the fix is 864459191Skris trivial: move one line. 864559191Skris [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 864659191Skris 864759191Skris *) Ugly workaround to get s_client and s_server working under Windows. The 864859191Skris old code wouldn't work because it needed to select() on sockets and the 864959191Skris tty (for keypresses and to see if data could be written). Win32 only 865059191Skris supports select() on sockets so we select() with a 1s timeout on the 865159191Skris sockets and then see if any characters are waiting to be read, if none 865259191Skris are present then we retry, we also assume we can always write data to 865359191Skris the tty. This isn't nice because the code then blocks until we've 865459191Skris received a complete line of data and it is effectively polling the 865559191Skris keyboard at 1s intervals: however it's quite a bit better than not 865659191Skris working at all :-) A dedicated Windows application might handle this 865759191Skris with an event loop for example. 865859191Skris [Steve Henson] 865959191Skris 866059191Skris *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 866159191Skris and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 866259191Skris will be called when RSA_sign() and RSA_verify() are used. This is useful 866359191Skris if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 866459191Skris For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 866559191Skris should *not* be used: RSA_sign() and RSA_verify() must be used instead. 866659191Skris This necessitated the support of an extra signature type NID_md5_sha1 866759191Skris for SSL signatures and modifications to the SSL library to use it instead 866859191Skris of calling RSA_public_decrypt() and RSA_private_encrypt(). 866959191Skris [Steve Henson] 867059191Skris 867159191Skris *) Add new -verify -CAfile and -CApath options to the crl program, these 867259191Skris will lookup a CRL issuers certificate and verify the signature in a 867359191Skris similar way to the verify program. Tidy up the crl program so it 867459191Skris no longer accesses structures directly. Make the ASN1 CRL parsing a bit 867559191Skris less strict. It will now permit CRL extensions even if it is not 867659191Skris a V2 CRL: this will allow it to tolerate some broken CRLs. 867759191Skris [Steve Henson] 867859191Skris 867959191Skris *) Initialize all non-automatic variables each time one of the openssl 868059191Skris sub-programs is started (this is necessary as they may be started 868159191Skris multiple times from the "OpenSSL>" prompt). 868259191Skris [Lennart Bang, Bodo Moeller] 868359191Skris 868459191Skris *) Preliminary compilation option RSA_NULL which disables RSA crypto without 868559191Skris removing all other RSA functionality (this is what NO_RSA does). This 868659191Skris is so (for example) those in the US can disable those operations covered 868759191Skris by the RSA patent while allowing storage and parsing of RSA keys and RSA 868859191Skris key generation. 868959191Skris [Steve Henson] 869059191Skris 869159191Skris *) Non-copying interface to BIO pairs. 869259191Skris (still largely untested) 869359191Skris [Bodo Moeller] 869459191Skris 869559191Skris *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 869659191Skris ASCII string. This was handled independently in various places before. 869759191Skris [Steve Henson] 869859191Skris 869959191Skris *) New functions UTF8_getc() and UTF8_putc() that parse and generate 870059191Skris UTF8 strings a character at a time. 870159191Skris [Steve Henson] 870259191Skris 870359191Skris *) Use client_version from client hello to select the protocol 870459191Skris (s23_srvr.c) and for RSA client key exchange verification 870559191Skris (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 870659191Skris [Bodo Moeller] 870759191Skris 870859191Skris *) Add various utility functions to handle SPKACs, these were previously 870959191Skris handled by poking round in the structure internals. Added new function 871059191Skris NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 871159191Skris print, verify and generate SPKACs. Based on an original idea from 871259191Skris Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 871359191Skris [Steve Henson] 871459191Skris 871559191Skris *) RIPEMD160 is operational on all platforms and is back in 'make test'. 871659191Skris [Andy Polyakov] 871759191Skris 871859191Skris *) Allow the config file extension section to be overwritten on the 871959191Skris command line. Based on an original idea from Massimiliano Pala 872059191Skris <madwolf@comune.modena.it>. The new option is called -extensions 872159191Skris and can be applied to ca, req and x509. Also -reqexts to override 872259191Skris the request extensions in req and -crlexts to override the crl extensions 872359191Skris in ca. 872459191Skris [Steve Henson] 872559191Skris 872659191Skris *) Add new feature to the SPKAC handling in ca. Now you can include 872759191Skris the same field multiple times by preceding it by "XXXX." for example: 872859191Skris 1.OU="Unit name 1" 872959191Skris 2.OU="Unit name 2" 873059191Skris this is the same syntax as used in the req config file. 873159191Skris [Steve Henson] 873259191Skris 873359191Skris *) Allow certificate extensions to be added to certificate requests. These 873459191Skris are specified in a 'req_extensions' option of the req section of the 873559191Skris config file. They can be printed out with the -text option to req but 873659191Skris are otherwise ignored at present. 873759191Skris [Steve Henson] 873859191Skris 873959191Skris *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 874059191Skris data read consists of only the final block it would not decrypted because 874159191Skris EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 874259191Skris A misplaced 'break' also meant the decrypted final block might not be 874359191Skris copied until the next read. 874459191Skris [Steve Henson] 874559191Skris 874659191Skris *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 874759191Skris a few extra parameters to the DH structure: these will be useful if 874859191Skris for example we want the value of 'q' or implement X9.42 DH. 874959191Skris [Steve Henson] 875059191Skris 875159191Skris *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 875259191Skris provides hooks that allow the default DSA functions or functions on a 875359191Skris "per key" basis to be replaced. This allows hardware acceleration and 875459191Skris hardware key storage to be handled without major modification to the 875559191Skris library. Also added low level modexp hooks and CRYPTO_EX structure and 875659191Skris associated functions. 875759191Skris [Steve Henson] 875859191Skris 875959191Skris *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 876059191Skris as "read only": it can't be written to and the buffer it points to will 876159191Skris not be freed. Reading from a read only BIO is much more efficient than 876259191Skris a normal memory BIO. This was added because there are several times when 876359191Skris an area of memory needs to be read from a BIO. The previous method was 876459191Skris to create a memory BIO and write the data to it, this results in two 876559191Skris copies of the data and an O(n^2) reading algorithm. There is a new 876659191Skris function BIO_new_mem_buf() which creates a read only memory BIO from 876759191Skris an area of memory. Also modified the PKCS#7 routines to use read only 876859191Skris memory BIOs. 876959191Skris [Steve Henson] 877059191Skris 877159191Skris *) Bugfix: ssl23_get_client_hello did not work properly when called in 877259191Skris state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 877359191Skris a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 877459191Skris but a retry condition occured while trying to read the rest. 877559191Skris [Bodo Moeller] 877659191Skris 877759191Skris *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 877859191Skris NID_pkcs7_encrypted by default: this was wrong since this should almost 877959191Skris always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 878059191Skris the encrypted data type: this is a more sensible place to put it and it 878159191Skris allows the PKCS#12 code to be tidied up that duplicated this 878259191Skris functionality. 878359191Skris [Steve Henson] 878459191Skris 878559191Skris *) Changed obj_dat.pl script so it takes its input and output files on 878659191Skris the command line. This should avoid shell escape redirection problems 878759191Skris under Win32. 878859191Skris [Steve Henson] 878959191Skris 879059191Skris *) Initial support for certificate extension requests, these are included 879159191Skris in things like Xenroll certificate requests. Included functions to allow 879259191Skris extensions to be obtained and added. 879359191Skris [Steve Henson] 879459191Skris 879559191Skris *) -crlf option to s_client and s_server for sending newlines as 879659191Skris CRLF (as required by many protocols). 879759191Skris [Bodo Moeller] 879859191Skris 879955714Skris Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 880055714Skris 880155714Skris *) Install libRSAglue.a when OpenSSL is built with RSAref. 880255714Skris [Ralf S. Engelschall] 880355714Skris 880455714Skris *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 880555714Skris [Andrija Antonijevic <TheAntony2@bigfoot.com>] 880655714Skris 880755714Skris *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 880855714Skris program. 880955714Skris [Steve Henson] 881055714Skris 881155714Skris *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 881255714Skris DH parameters/keys (q is lost during that conversion, but the resulting 881355714Skris DH parameters contain its length). 881455714Skris 881555714Skris For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 881655714Skris much faster than DH_generate_parameters (which creates parameters 881755714Skris where p = 2*q + 1), and also the smaller q makes DH computations 881855714Skris much more efficient (160-bit exponentiation instead of 1024-bit 881955714Skris exponentiation); so this provides a convenient way to support DHE 882055714Skris ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 882155714Skris utter importance to use 882255714Skris SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 882355714Skris or 882455714Skris SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 882555714Skris when such DH parameters are used, because otherwise small subgroup 882655714Skris attacks may become possible! 882755714Skris [Bodo Moeller] 882855714Skris 882955714Skris *) Avoid memory leak in i2d_DHparams. 883055714Skris [Bodo Moeller] 883155714Skris 883255714Skris *) Allow the -k option to be used more than once in the enc program: 883355714Skris this allows the same encrypted message to be read by multiple recipients. 883455714Skris [Steve Henson] 883555714Skris 883655714Skris *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 883755714Skris an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 883855714Skris it will always use the numerical form of the OID, even if it has a short 883955714Skris or long name. 884055714Skris [Steve Henson] 884155714Skris 884255714Skris *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 884355714Skris method only got called if p,q,dmp1,dmq1,iqmp components were present, 884455714Skris otherwise bn_mod_exp was called. In the case of hardware keys for example 884555714Skris no private key components need be present and it might store extra data 884659191Skris in the RSA structure, which cannot be accessed from bn_mod_exp. 884759191Skris By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 884859191Skris private key operations. 884955714Skris [Steve Henson] 885055714Skris 885155714Skris *) Added support for SPARC Linux. 885255714Skris [Andy Polyakov] 885355714Skris 885455714Skris *) pem_password_cb function type incompatibly changed from 885555714Skris typedef int pem_password_cb(char *buf, int size, int rwflag); 885655714Skris to 885755714Skris ....(char *buf, int size, int rwflag, void *userdata); 885855714Skris so that applications can pass data to their callbacks: 885955714Skris The PEM[_ASN1]_{read,write}... functions and macros now take an 886055714Skris additional void * argument, which is just handed through whenever 886155714Skris the password callback is called. 886259191Skris [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 886355714Skris 886455714Skris New function SSL_CTX_set_default_passwd_cb_userdata. 886555714Skris 886655714Skris Compatibility note: As many C implementations push function arguments 886755714Skris onto the stack in reverse order, the new library version is likely to 886855714Skris interoperate with programs that have been compiled with the old 886955714Skris pem_password_cb definition (PEM_whatever takes some data that 887055714Skris happens to be on the stack as its last argument, and the callback 887155714Skris just ignores this garbage); but there is no guarantee whatsoever that 887255714Skris this will work. 887355714Skris 887455714Skris *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 887555714Skris (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 887655714Skris problems not only on Windows, but also on some Unix platforms. 887755714Skris To avoid problematic command lines, these definitions are now in an 887855714Skris auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 887955714Skris for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 888055714Skris [Bodo Moeller] 888155714Skris 888255714Skris *) MIPS III/IV assembler module is reimplemented. 888355714Skris [Andy Polyakov] 888455714Skris 888555714Skris *) More DES library cleanups: remove references to srand/rand and 888655714Skris delete an unused file. 888755714Skris [Ulf M�ller] 888855714Skris 888955714Skris *) Add support for the the free Netwide assembler (NASM) under Win32, 889055714Skris since not many people have MASM (ml) and it can be hard to obtain. 889155714Skris This is currently experimental but it seems to work OK and pass all 889255714Skris the tests. Check out INSTALL.W32 for info. 889355714Skris [Steve Henson] 889455714Skris 889555714Skris *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 889655714Skris without temporary keys kept an extra copy of the server key, 889755714Skris and connections with temporary keys did not free everything in case 889855714Skris of an error. 889955714Skris [Bodo Moeller] 890055714Skris 890155714Skris *) New function RSA_check_key and new openssl rsa option -check 890255714Skris for verifying the consistency of RSA keys. 890355714Skris [Ulf Moeller, Bodo Moeller] 890455714Skris 890555714Skris *) Various changes to make Win32 compile work: 890655714Skris 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 890755714Skris 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 890855714Skris comparison" warnings. 890955714Skris 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 891055714Skris [Steve Henson] 891155714Skris 891255714Skris *) Add a debugging option to PKCS#5 v2 key generation function: when 891355714Skris you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 891455714Skris derived keys are printed to stderr. 891555714Skris [Steve Henson] 891655714Skris 891755714Skris *) Copy the flags in ASN1_STRING_dup(). 891855714Skris [Roman E. Pavlov <pre@mo.msk.ru>] 891955714Skris 892055714Skris *) The x509 application mishandled signing requests containing DSA 892155714Skris keys when the signing key was also DSA and the parameters didn't match. 892255714Skris 892355714Skris It was supposed to omit the parameters when they matched the signing key: 892455714Skris the verifying software was then supposed to automatically use the CA's 892555714Skris parameters if they were absent from the end user certificate. 892655714Skris 892755714Skris Omitting parameters is no longer recommended. The test was also 892855714Skris the wrong way round! This was probably due to unusual behaviour in 892955714Skris EVP_cmp_parameters() which returns 1 if the parameters match. 893055714Skris This meant that parameters were omitted when they *didn't* match and 893155714Skris the certificate was useless. Certificates signed with 'ca' didn't have 893255714Skris this bug. 893355714Skris [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 893455714Skris 893555714Skris *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 893655714Skris The interface is as follows: 893755714Skris Applications can use 893855714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 893955714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 894055714Skris "off" is now the default. 894155714Skris The library internally uses 894255714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 894355714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 894455714Skris to disable memory-checking temporarily. 894555714Skris 894655714Skris Some inconsistent states that previously were possible (and were 894755714Skris even the default) are now avoided. 894855714Skris 894955714Skris -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 895055714Skris with each memory chunk allocated; this is occasionally more helpful 895155714Skris than just having a counter. 895255714Skris 895355714Skris -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 895455714Skris 895555714Skris -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 895655714Skris extensions. 895755714Skris [Bodo Moeller] 895855714Skris 895955714Skris *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 896055714Skris which largely parallels "options", but is for changing API behaviour, 896155714Skris whereas "options" are about protocol behaviour. 896255714Skris Initial "mode" flags are: 896355714Skris 896455714Skris SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 896555714Skris a single record has been written. 896655714Skris SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 896755714Skris retries use the same buffer location. 896855714Skris (But all of the contents must be 896955714Skris copied!) 897055714Skris [Bodo Moeller] 897155714Skris 897279998Skris *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 897355714Skris worked. 897455714Skris 897555714Skris *) Fix problems with no-hmac etc. 897655714Skris [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 897755714Skris 897855714Skris *) New functions RSA_get_default_method(), RSA_set_method() and 897955714Skris RSA_get_method(). These allows replacement of RSA_METHODs without having 898055714Skris to mess around with the internals of an RSA structure. 898155714Skris [Steve Henson] 898255714Skris 898355714Skris *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 898455714Skris Also really enable memory leak checks in openssl.c and in some 898555714Skris test programs. 898655714Skris [Chad C. Mulligan, Bodo Moeller] 898755714Skris 898855714Skris *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 898955714Skris up the length of negative integers. This has now been simplified to just 899055714Skris store the length when it is first determined and use it later, rather 899155714Skris than trying to keep track of where data is copied and updating it to 899255714Skris point to the end. 899355714Skris [Steve Henson, reported by Brien Wheeler 899455714Skris <bwheeler@authentica-security.com>] 899555714Skris 899655714Skris *) Add a new function PKCS7_signatureVerify. This allows the verification 899755714Skris of a PKCS#7 signature but with the signing certificate passed to the 899855714Skris function itself. This contrasts with PKCS7_dataVerify which assumes the 899955714Skris certificate is present in the PKCS#7 structure. This isn't always the 900055714Skris case: certificates can be omitted from a PKCS#7 structure and be 900155714Skris distributed by "out of band" means (such as a certificate database). 900255714Skris [Steve Henson] 900355714Skris 900455714Skris *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 900555714Skris function prototypes in pem.h, also change util/mkdef.pl to add the 900655714Skris necessary function names. 900755714Skris [Steve Henson] 900855714Skris 900955714Skris *) mk1mf.pl (used by Windows builds) did not properly read the 901055714Skris options set by Configure in the top level Makefile, and Configure 901155714Skris was not even able to write more than one option correctly. 901255714Skris Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 901355714Skris [Bodo Moeller] 901455714Skris 901555714Skris *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 901655714Skris file to be loaded from a BIO or FILE pointer. The BIO version will 901755714Skris for example allow memory BIOs to contain config info. 901855714Skris [Steve Henson] 901955714Skris 902055714Skris *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 902155714Skris Whoever hopes to achieve shared-library compatibility across versions 902255714Skris must use this, not the compile-time macro. 902355714Skris (Exercise 0.9.4: Which is the minimum library version required by 902455714Skris such programs?) 902555714Skris Note: All this applies only to multi-threaded programs, others don't 902655714Skris need locks. 902755714Skris [Bodo Moeller] 902855714Skris 902955714Skris *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 903055714Skris through a BIO pair triggered the default case, i.e. 903155714Skris SSLerr(...,SSL_R_UNKNOWN_STATE). 903255714Skris [Bodo Moeller] 903355714Skris 903455714Skris *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 903555714Skris can use the SSL library even if none of the specific BIOs is 903655714Skris appropriate. 903755714Skris [Bodo Moeller] 903855714Skris 903955714Skris *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 904055714Skris for the encoded length. 904155714Skris [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 904255714Skris 904355714Skris *) Add initial documentation of the X509V3 functions. 904455714Skris [Steve Henson] 904555714Skris 904655714Skris *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 904755714Skris PEM_write_bio_PKCS8PrivateKey() that are equivalent to 904855714Skris PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 904955714Skris secure PKCS#8 private key format with a high iteration count. 905055714Skris [Steve Henson] 905155714Skris 905255714Skris *) Fix determination of Perl interpreter: A perl or perl5 905355714Skris _directory_ in $PATH was also accepted as the interpreter. 905455714Skris [Ralf S. Engelschall] 905555714Skris 905655714Skris *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 905755714Skris wrong with it but it was very old and did things like calling 905855714Skris PEM_ASN1_read() directly and used MD5 for the hash not to mention some 905955714Skris unusual formatting. 906055714Skris [Steve Henson] 906155714Skris 906255714Skris *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 906355714Skris to use the new extension code. 906455714Skris [Steve Henson] 906555714Skris 906655714Skris *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 906755714Skris with macros. This should make it easier to change their form, add extra 906855714Skris arguments etc. Fix a few PEM prototypes which didn't have cipher as a 906955714Skris constant. 907055714Skris [Steve Henson] 907155714Skris 907255714Skris *) Add to configuration table a new entry that can specify an alternative 907355714Skris name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 907455714Skris according to Mark Crispin <MRC@Panda.COM>. 907555714Skris [Bodo Moeller] 907655714Skris 907755714Skris#if 0 907855714Skris *) DES CBC did not update the IV. Weird. 907955714Skris [Ben Laurie] 908055714Skris#else 908155714Skris des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 908255714Skris Changing the behaviour of the former might break existing programs -- 908355714Skris where IV updating is needed, des_ncbc_encrypt can be used. 908455714Skris#endif 908555714Skris 908655714Skris *) When bntest is run from "make test" it drives bc to check its 908755714Skris calculations, as well as internally checking them. If an internal check 908855714Skris fails, it needs to cause bc to give a non-zero result or make test carries 908955714Skris on without noticing the failure. Fixed. 909055714Skris [Ben Laurie] 909155714Skris 909255714Skris *) DES library cleanups. 909355714Skris [Ulf M�ller] 909455714Skris 909555714Skris *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 909655714Skris used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 909755714Skris ciphers. NOTE: although the key derivation function has been verified 909855714Skris against some published test vectors it has not been extensively tested 909955714Skris yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 910055714Skris of v2.0. 910155714Skris [Steve Henson] 910255714Skris 910355714Skris *) Instead of "mkdir -p", which is not fully portable, use new 910455714Skris Perl script "util/mkdir-p.pl". 910555714Skris [Bodo Moeller] 910655714Skris 910755714Skris *) Rewrite the way password based encryption (PBE) is handled. It used to 910855714Skris assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 910955714Skris structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 911055714Skris but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 911155714Skris the 'parameter' field of the AlgorithmIdentifier is passed to the 911255714Skris underlying key generation function so it must do its own ASN1 parsing. 911355714Skris This has also changed the EVP_PBE_CipherInit() function which now has a 911455714Skris 'parameter' argument instead of literal salt and iteration count values 911555714Skris and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 911655714Skris [Steve Henson] 911755714Skris 911855714Skris *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 911955714Skris and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 912055714Skris Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 912155714Skris KEY" because this clashed with PKCS#8 unencrypted string. Since this 912255714Skris value was just used as a "magic string" and not used directly its 912355714Skris value doesn't matter. 912455714Skris [Steve Henson] 912555714Skris 912655714Skris *) Introduce some semblance of const correctness to BN. Shame C doesn't 912755714Skris support mutable. 912855714Skris [Ben Laurie] 912955714Skris 913055714Skris *) "linux-sparc64" configuration (ultrapenguin). 913155714Skris [Ray Miller <ray.miller@oucs.ox.ac.uk>] 913255714Skris "linux-sparc" configuration. 913355714Skris [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 913455714Skris 913555714Skris *) config now generates no-xxx options for missing ciphers. 913655714Skris [Ulf M�ller] 913755714Skris 913855714Skris *) Support the EBCDIC character set (work in progress). 913955714Skris File ebcdic.c not yet included because it has a different license. 914055714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 914155714Skris 914255714Skris *) Support BS2000/OSD-POSIX. 914355714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 914455714Skris 914555714Skris *) Make callbacks for key generation use void * instead of char *. 914655714Skris [Ben Laurie] 914755714Skris 914855714Skris *) Make S/MIME samples compile (not yet tested). 914955714Skris [Ben Laurie] 915055714Skris 915155714Skris *) Additional typesafe stacks. 915255714Skris [Ben Laurie] 915355714Skris 915455714Skris *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 915555714Skris [Bodo Moeller] 915655714Skris 915755714Skris 915855714Skris Changes between 0.9.3 and 0.9.3a [29 May 1999] 915955714Skris 916055714Skris *) New configuration variant "sco5-gcc". 916155714Skris 916255714Skris *) Updated some demos. 916355714Skris [Sean O Riordain, Wade Scholine] 916455714Skris 916555714Skris *) Add missing BIO_free at exit of pkcs12 application. 916655714Skris [Wu Zhigang] 916755714Skris 916855714Skris *) Fix memory leak in conf.c. 916955714Skris [Steve Henson] 917055714Skris 917155714Skris *) Updates for Win32 to assembler version of MD5. 917255714Skris [Steve Henson] 917355714Skris 917455714Skris *) Set #! path to perl in apps/der_chop to where we found it 917555714Skris instead of using a fixed path. 917655714Skris [Bodo Moeller] 917755714Skris 917855714Skris *) SHA library changes for irix64-mips4-cc. 917955714Skris [Andy Polyakov] 918055714Skris 918155714Skris *) Improvements for VMS support. 918255714Skris [Richard Levitte] 918355714Skris 918455714Skris 918555714Skris Changes between 0.9.2b and 0.9.3 [24 May 1999] 918655714Skris 918755714Skris *) Bignum library bug fix. IRIX 6 passes "make test" now! 918855714Skris This also avoids the problems with SC4.2 and unpatched SC5. 918955714Skris [Andy Polyakov <appro@fy.chalmers.se>] 919055714Skris 919155714Skris *) New functions sk_num, sk_value and sk_set to replace the previous macros. 919255714Skris These are required because of the typesafe stack would otherwise break 919355714Skris existing code. If old code used a structure member which used to be STACK 919455714Skris and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 919555714Skris sk_num or sk_value it would produce an error because the num, data members 919655714Skris are not present in STACK_OF. Now it just produces a warning. sk_set 919755714Skris replaces the old method of assigning a value to sk_value 919855714Skris (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 919955714Skris that does this will no longer work (and should use sk_set instead) but 920055714Skris this could be regarded as a "questionable" behaviour anyway. 920155714Skris [Steve Henson] 920255714Skris 920355714Skris *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 920455714Skris correctly handle encrypted S/MIME data. 920555714Skris [Steve Henson] 920655714Skris 920755714Skris *) Change type of various DES function arguments from des_cblock 920855714Skris (which means, in function argument declarations, pointer to char) 920955714Skris to des_cblock * (meaning pointer to array with 8 char elements), 921055714Skris which allows the compiler to do more typechecking; it was like 921155714Skris that back in SSLeay, but with lots of ugly casts. 921255714Skris 921355714Skris Introduce new type const_des_cblock. 921455714Skris [Bodo Moeller] 921555714Skris 921655714Skris *) Reorganise the PKCS#7 library and get rid of some of the more obvious 921755714Skris problems: find RecipientInfo structure that matches recipient certificate 921855714Skris and initialise the ASN1 structures properly based on passed cipher. 921955714Skris [Steve Henson] 922055714Skris 922155714Skris *) Belatedly make the BN tests actually check the results. 922255714Skris [Ben Laurie] 922355714Skris 922455714Skris *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 922555714Skris to and from BNs: it was completely broken. New compilation option 922655714Skris NEG_PUBKEY_BUG to allow for some broken certificates that encode public 922755714Skris key elements as negative integers. 922855714Skris [Steve Henson] 922955714Skris 923055714Skris *) Reorganize and speed up MD5. 923155714Skris [Andy Polyakov <appro@fy.chalmers.se>] 923255714Skris 923355714Skris *) VMS support. 923455714Skris [Richard Levitte <richard@levitte.org>] 923555714Skris 923655714Skris *) New option -out to asn1parse to allow the parsed structure to be 923755714Skris output to a file. This is most useful when combined with the -strparse 923855714Skris option to examine the output of things like OCTET STRINGS. 923955714Skris [Steve Henson] 924055714Skris 924155714Skris *) Make SSL library a little more fool-proof by not requiring any longer 924255714Skris that SSL_set_{accept,connect}_state be called before 924355714Skris SSL_{accept,connect} may be used (SSL_set_..._state is omitted 924455714Skris in many applications because usually everything *appeared* to work as 924555714Skris intended anyway -- now it really works as intended). 924655714Skris [Bodo Moeller] 924755714Skris 924855714Skris *) Move openssl.cnf out of lib/. 924955714Skris [Ulf M�ller] 925055714Skris 925155714Skris *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 925255714Skris -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 925355714Skris -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 925455714Skris [Ralf S. Engelschall] 925555714Skris 925655714Skris *) Various fixes to the EVP and PKCS#7 code. It may now be able to 925755714Skris handle PKCS#7 enveloped data properly. 925855714Skris [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 925955714Skris 926055714Skris *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 926155714Skris copying pointers. The cert_st handling is changed by this in 926255714Skris various ways (and thus what used to be known as ctx->default_cert 926355714Skris is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 926455714Skris any longer when s->cert does not give us what we need). 926555714Skris ssl_cert_instantiate becomes obsolete by this change. 926655714Skris As soon as we've got the new code right (possibly it already is?), 926755714Skris we have solved a couple of bugs of the earlier code where s->cert 926855714Skris was used as if it could not have been shared with other SSL structures. 926955714Skris 927055714Skris Note that using the SSL API in certain dirty ways now will result 927155714Skris in different behaviour than observed with earlier library versions: 927255714Skris Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 927355714Skris does not influence s as it used to. 927455714Skris 927555714Skris In order to clean up things more thoroughly, inside SSL_SESSION 927655714Skris we don't use CERT any longer, but a new structure SESS_CERT 927755714Skris that holds per-session data (if available); currently, this is 927855714Skris the peer's certificate chain and, for clients, the server's certificate 927955714Skris and temporary key. CERT holds only those values that can have 928055714Skris meaningful defaults in an SSL_CTX. 928155714Skris [Bodo Moeller] 928255714Skris 928355714Skris *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 928455714Skris from the internal representation. Various PKCS#7 fixes: remove some 928555714Skris evil casts and set the enc_dig_alg field properly based on the signing 928655714Skris key type. 928755714Skris [Steve Henson] 928855714Skris 928955714Skris *) Allow PKCS#12 password to be set from the command line or the 929055714Skris environment. Let 'ca' get its config file name from the environment 929155714Skris variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 929255714Skris and 'x509'). 929355714Skris [Steve Henson] 929455714Skris 929555714Skris *) Allow certificate policies extension to use an IA5STRING for the 929655714Skris organization field. This is contrary to the PKIX definition but 929755714Skris VeriSign uses it and IE5 only recognises this form. Document 'x509' 929855714Skris extension option. 929955714Skris [Steve Henson] 930055714Skris 930155714Skris *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 930255714Skris without disallowing inline assembler and the like for non-pedantic builds. 930355714Skris [Ben Laurie] 930455714Skris 930555714Skris *) Support Borland C++ builder. 930655714Skris [Janez Jere <jj@void.si>, modified by Ulf M�ller] 930755714Skris 930855714Skris *) Support Mingw32. 930955714Skris [Ulf M�ller] 931055714Skris 931155714Skris *) SHA-1 cleanups and performance enhancements. 931255714Skris [Andy Polyakov <appro@fy.chalmers.se>] 931355714Skris 931455714Skris *) Sparc v8plus assembler for the bignum library. 931555714Skris [Andy Polyakov <appro@fy.chalmers.se>] 931655714Skris 931755714Skris *) Accept any -xxx and +xxx compiler options in Configure. 931855714Skris [Ulf M�ller] 931955714Skris 932055714Skris *) Update HPUX configuration. 932155714Skris [Anonymous] 932255714Skris 932355714Skris *) Add missing sk_<type>_unshift() function to safestack.h 932455714Skris [Ralf S. Engelschall] 932555714Skris 932655714Skris *) New function SSL_CTX_use_certificate_chain_file that sets the 932755714Skris "extra_cert"s in addition to the certificate. (This makes sense 932855714Skris only for "PEM" format files, as chains as a whole are not 932955714Skris DER-encoded.) 933055714Skris [Bodo Moeller] 933155714Skris 933255714Skris *) Support verify_depth from the SSL API. 933355714Skris x509_vfy.c had what can be considered an off-by-one-error: 933455714Skris Its depth (which was not part of the external interface) 933555714Skris was actually counting the number of certificates in a chain; 933655714Skris now it really counts the depth. 933755714Skris [Bodo Moeller] 933855714Skris 933955714Skris *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 934055714Skris instead of X509err, which often resulted in confusing error 934155714Skris messages since the error codes are not globally unique 934255714Skris (e.g. an alleged error in ssl3_accept when a certificate 934355714Skris didn't match the private key). 934455714Skris 934555714Skris *) New function SSL_CTX_set_session_id_context that allows to set a default 934655714Skris value (so that you don't need SSL_set_session_id_context for each 934755714Skris connection using the SSL_CTX). 934855714Skris [Bodo Moeller] 934955714Skris 935055714Skris *) OAEP decoding bug fix. 935155714Skris [Ulf M�ller] 935255714Skris 935355714Skris *) Support INSTALL_PREFIX for package builders, as proposed by 935455714Skris David Harris. 935555714Skris [Bodo Moeller] 935655714Skris 935755714Skris *) New Configure options "threads" and "no-threads". For systems 935855714Skris where the proper compiler options are known (currently Solaris 935955714Skris and Linux), "threads" is the default. 936055714Skris [Bodo Moeller] 936155714Skris 936255714Skris *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 936355714Skris [Bodo Moeller] 936455714Skris 936555714Skris *) Install various scripts to $(OPENSSLDIR)/misc, not to 936655714Skris $(INSTALLTOP)/bin -- they shouldn't clutter directories 936755714Skris such as /usr/local/bin. 936855714Skris [Bodo Moeller] 936955714Skris 937055714Skris *) "make linux-shared" to build shared libraries. 937155714Skris [Niels Poppe <niels@netbox.org>] 937255714Skris 937355714Skris *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 937455714Skris [Ulf M�ller] 937555714Skris 937655714Skris *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 937755714Skris extension adding in x509 utility. 937855714Skris [Steve Henson] 937955714Skris 938055714Skris *) Remove NOPROTO sections and error code comments. 938155714Skris [Ulf M�ller] 938255714Skris 938355714Skris *) Partial rewrite of the DEF file generator to now parse the ANSI 938455714Skris prototypes. 938555714Skris [Steve Henson] 938655714Skris 938755714Skris *) New Configure options --prefix=DIR and --openssldir=DIR. 938855714Skris [Ulf M�ller] 938955714Skris 939055714Skris *) Complete rewrite of the error code script(s). It is all now handled 939155714Skris by one script at the top level which handles error code gathering, 939255714Skris header rewriting and C source file generation. It should be much better 939355714Skris than the old method: it now uses a modified version of Ulf's parser to 939455714Skris read the ANSI prototypes in all header files (thus the old K&R definitions 939555714Skris aren't needed for error creation any more) and do a better job of 939655714Skris translating function codes into names. The old 'ASN1 error code imbedded 939755714Skris in a comment' is no longer necessary and it doesn't use .err files which 939855714Skris have now been deleted. Also the error code call doesn't have to appear all 939955714Skris on one line (which resulted in some large lines...). 940055714Skris [Steve Henson] 940155714Skris 940255714Skris *) Change #include filenames from <foo.h> to <openssl/foo.h>. 940355714Skris [Bodo Moeller] 940455714Skris 940555714Skris *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 940655714Skris 0 (which usually indicates a closed connection), but continue reading. 940755714Skris [Bodo Moeller] 940855714Skris 940955714Skris *) Fix some race conditions. 941055714Skris [Bodo Moeller] 941155714Skris 941255714Skris *) Add support for CRL distribution points extension. Add Certificate 941355714Skris Policies and CRL distribution points documentation. 941455714Skris [Steve Henson] 941555714Skris 941655714Skris *) Move the autogenerated header file parts to crypto/opensslconf.h. 941755714Skris [Ulf M�ller] 941855714Skris 941955714Skris *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 942055714Skris 8 of keying material. Merlin has also confirmed interop with this fix 942155714Skris between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 942255714Skris [Merlin Hughes <merlin@baltimore.ie>] 942355714Skris 942455714Skris *) Fix lots of warnings. 942555714Skris [Richard Levitte <levitte@stacken.kth.se>] 942655714Skris 942755714Skris *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 942855714Skris the directory spec didn't end with a LIST_SEPARATOR_CHAR. 942955714Skris [Richard Levitte <levitte@stacken.kth.se>] 943055714Skris 943155714Skris *) Fix problems with sizeof(long) == 8. 943255714Skris [Andy Polyakov <appro@fy.chalmers.se>] 943355714Skris 943455714Skris *) Change functions to ANSI C. 943555714Skris [Ulf M�ller] 943655714Skris 943755714Skris *) Fix typos in error codes. 943855714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 943955714Skris 944055714Skris *) Remove defunct assembler files from Configure. 944155714Skris [Ulf M�ller] 944255714Skris 944355714Skris *) SPARC v8 assembler BIGNUM implementation. 944455714Skris [Andy Polyakov <appro@fy.chalmers.se>] 944555714Skris 944655714Skris *) Support for Certificate Policies extension: both print and set. 944755714Skris Various additions to support the r2i method this uses. 944855714Skris [Steve Henson] 944955714Skris 945055714Skris *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 945155714Skris return a const string when you are expecting an allocated buffer. 945255714Skris [Ben Laurie] 945355714Skris 945455714Skris *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 945555714Skris types DirectoryString and DisplayText. 945655714Skris [Steve Henson] 945755714Skris 945855714Skris *) Add code to allow r2i extensions to access the configuration database, 945955714Skris add an LHASH database driver and add several ctx helper functions. 946055714Skris [Steve Henson] 946155714Skris 946255714Skris *) Fix an evil bug in bn_expand2() which caused various BN functions to 946355714Skris fail when they extended the size of a BIGNUM. 946455714Skris [Steve Henson] 946555714Skris 946655714Skris *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 946755714Skris support typesafe stack. 946855714Skris [Steve Henson] 946955714Skris 947055714Skris *) Fix typo in SSL_[gs]et_options(). 947155714Skris [Nils Frostberg <nils@medcom.se>] 947255714Skris 947355714Skris *) Delete various functions and files that belonged to the (now obsolete) 947455714Skris old X509V3 handling code. 947555714Skris [Steve Henson] 947655714Skris 947755714Skris *) New Configure option "rsaref". 947855714Skris [Ulf M�ller] 947955714Skris 948055714Skris *) Don't auto-generate pem.h. 948155714Skris [Bodo Moeller] 948255714Skris 948355714Skris *) Introduce type-safe ASN.1 SETs. 948455714Skris [Ben Laurie] 948555714Skris 948655714Skris *) Convert various additional casted stacks to type-safe STACK_OF() variants. 948755714Skris [Ben Laurie, Ralf S. Engelschall, Steve Henson] 948855714Skris 948955714Skris *) Introduce type-safe STACKs. This will almost certainly break lots of code 949055714Skris that links with OpenSSL (well at least cause lots of warnings), but fear 949155714Skris not: the conversion is trivial, and it eliminates loads of evil casts. A 949255714Skris few STACKed things have been converted already. Feel free to convert more. 949355714Skris In the fullness of time, I'll do away with the STACK type altogether. 949455714Skris [Ben Laurie] 949555714Skris 949655714Skris *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 949755714Skris specified in <certfile> by updating the entry in the index.txt file. 949855714Skris This way one no longer has to edit the index.txt file manually for 949955714Skris revoking a certificate. The -revoke option does the gory details now. 950055714Skris [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 950155714Skris 950255714Skris *) Fix `openssl crl -noout -text' combination where `-noout' killed the 950355714Skris `-text' option at all and this way the `-noout -text' combination was 950455714Skris inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 950555714Skris [Ralf S. Engelschall] 950655714Skris 950755714Skris *) Make sure a corresponding plain text error message exists for the 950855714Skris X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 950955714Skris verify callback function determined that a certificate was revoked. 951055714Skris [Ralf S. Engelschall] 951155714Skris 951255714Skris *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 951355714Skris ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 951455714Skris all available cipers including rc5, which was forgotten until now. 951555714Skris In order to let the testing shell script know which algorithms 951655714Skris are available, a new (up to now undocumented) command 951755714Skris "openssl list-cipher-commands" is used. 951855714Skris [Bodo Moeller] 951955714Skris 952055714Skris *) Bugfix: s_client occasionally would sleep in select() when 952155714Skris it should have checked SSL_pending() first. 952255714Skris [Bodo Moeller] 952355714Skris 952455714Skris *) New functions DSA_do_sign and DSA_do_verify to provide access to 952555714Skris the raw DSA values prior to ASN.1 encoding. 952655714Skris [Ulf M�ller] 952755714Skris 952855714Skris *) Tweaks to Configure 952955714Skris [Niels Poppe <niels@netbox.org>] 953055714Skris 953155714Skris *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 953255714Skris yet... 953355714Skris [Steve Henson] 953455714Skris 953555714Skris *) New variables $(RANLIB) and $(PERL) in the Makefiles. 953655714Skris [Ulf M�ller] 953755714Skris 953855714Skris *) New config option to avoid instructions that are illegal on the 80386. 953955714Skris The default code is faster, but requires at least a 486. 954055714Skris [Ulf M�ller] 954155714Skris 954255714Skris *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 954355714Skris SSL2_SERVER_VERSION (not used at all) macros, which are now the 954455714Skris same as SSL2_VERSION anyway. 954555714Skris [Bodo Moeller] 954655714Skris 954755714Skris *) New "-showcerts" option for s_client. 954855714Skris [Bodo Moeller] 954955714Skris 955055714Skris *) Still more PKCS#12 integration. Add pkcs12 application to openssl 955155714Skris application. Various cleanups and fixes. 955255714Skris [Steve Henson] 955355714Skris 955455714Skris *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 955555714Skris modify error routines to work internally. Add error codes and PBE init 955655714Skris to library startup routines. 955755714Skris [Steve Henson] 955855714Skris 955955714Skris *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 956055714Skris packing functions to asn1 and evp. Changed function names and error 956155714Skris codes along the way. 956255714Skris [Steve Henson] 956355714Skris 956455714Skris *) PKCS12 integration: and so it begins... First of several patches to 956555714Skris slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 956655714Skris objects to objects.h 956755714Skris [Steve Henson] 956855714Skris 956955714Skris *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 957055714Skris and display support for Thawte strong extranet extension. 957155714Skris [Steve Henson] 957255714Skris 957355714Skris *) Add LinuxPPC support. 957455714Skris [Jeff Dubrule <igor@pobox.org>] 957555714Skris 957655714Skris *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 957755714Skris bn_div_words in alpha.s. 957855714Skris [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 957955714Skris 958055714Skris *) Make sure the RSA OAEP test is skipped under -DRSAref because 958155714Skris OAEP isn't supported when OpenSSL is built with RSAref. 958255714Skris [Ulf Moeller <ulf@fitug.de>] 958355714Skris 958455714Skris *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 958555714Skris so they no longer are missing under -DNOPROTO. 958655714Skris [Soren S. Jorvang <soren@t.dk>] 958755714Skris 958855714Skris 958955714Skris Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 959055714Skris 959155714Skris *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 959255714Skris doesn't work when the session is reused. Coming soon! 959355714Skris [Ben Laurie] 959455714Skris 959555714Skris *) Fix a security hole, that allows sessions to be reused in the wrong 959655714Skris context thus bypassing client cert protection! All software that uses 959755714Skris client certs and session caches in multiple contexts NEEDS PATCHING to 959855714Skris allow session reuse! A fuller solution is in the works. 959955714Skris [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 960055714Skris 960155714Skris *) Some more source tree cleanups (removed obsolete files 960255714Skris crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 960355714Skris permission on "config" script to be executable) and a fix for the INSTALL 960455714Skris document. 960555714Skris [Ulf Moeller <ulf@fitug.de>] 960655714Skris 960755714Skris *) Remove some legacy and erroneous uses of malloc, free instead of 960855714Skris Malloc, Free. 960955714Skris [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 961055714Skris 961155714Skris *) Make rsa_oaep_test return non-zero on error. 961255714Skris [Ulf Moeller <ulf@fitug.de>] 961355714Skris 961455714Skris *) Add support for native Solaris shared libraries. Configure 961555714Skris solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 961655714Skris if someone would make that last step automatic. 961755714Skris [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 961855714Skris 961955714Skris *) ctx_size was not built with the right compiler during "make links". Fixed. 962055714Skris [Ben Laurie] 962155714Skris 962255714Skris *) Change the meaning of 'ALL' in the cipher list. It now means "everything 962355714Skris except NULL ciphers". This means the default cipher list will no longer 962455714Skris enable NULL ciphers. They need to be specifically enabled e.g. with 962555714Skris the string "DEFAULT:eNULL". 962655714Skris [Steve Henson] 962755714Skris 962855714Skris *) Fix to RSA private encryption routines: if p < q then it would 962955714Skris occasionally produce an invalid result. This will only happen with 963055714Skris externally generated keys because OpenSSL (and SSLeay) ensure p > q. 963155714Skris [Steve Henson] 963255714Skris 963355714Skris *) Be less restrictive and allow also `perl util/perlpath.pl 963455714Skris /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 963555714Skris because this way one can also use an interpreter named `perl5' (which is 963655714Skris usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 963755714Skris installed as `perl'). 963855714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 963955714Skris 964055714Skris *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 964155714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 964255714Skris 964355714Skris *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 964455714Skris advapi32.lib to Win32 build and change the pem test comparision 964555714Skris to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 964655714Skris suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 964755714Skris and crypto/des/ede_cbcm_enc.c. 964855714Skris [Steve Henson] 964955714Skris 965055714Skris *) DES quad checksum was broken on big-endian architectures. Fixed. 965155714Skris [Ben Laurie] 965255714Skris 965355714Skris *) Comment out two functions in bio.h that aren't implemented. Fix up the 965455714Skris Win32 test batch file so it (might) work again. The Win32 test batch file 965555714Skris is horrible: I feel ill.... 965655714Skris [Steve Henson] 965755714Skris 965855714Skris *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 965955714Skris in e_os.h. Audit of header files to check ANSI and non ANSI 966055714Skris sections: 10 functions were absent from non ANSI section and not exported 966155714Skris from Windows DLLs. Fixed up libeay.num for new functions. 966255714Skris [Steve Henson] 966355714Skris 966455714Skris *) Make `openssl version' output lines consistent. 966555714Skris [Ralf S. Engelschall] 966655714Skris 966755714Skris *) Fix Win32 symbol export lists for BIO functions: Added 966855714Skris BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 966955714Skris to ms/libeay{16,32}.def. 967055714Skris [Ralf S. Engelschall] 967155714Skris 967255714Skris *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 967355714Skris fine under Unix and passes some trivial tests I've now added. But the 967455714Skris whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 967555714Skris added to make sure no one expects that this stuff really works in the 967655714Skris OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 967755714Skris up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 967855714Skris openssl_bio.xs. 967955714Skris [Ralf S. Engelschall] 968055714Skris 968155714Skris *) Fix the generation of two part addresses in perl. 968255714Skris [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 968355714Skris 968455714Skris *) Add config entry for Linux on MIPS. 968555714Skris [John Tobey <jtobey@channel1.com>] 968655714Skris 968755714Skris *) Make links whenever Configure is run, unless we are on Windoze. 968855714Skris [Ben Laurie] 968955714Skris 969055714Skris *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 969155714Skris Currently only issuerAltName and AuthorityKeyIdentifier make any sense 969255714Skris in CRLs. 969355714Skris [Steve Henson] 969455714Skris 969555714Skris *) Add a useful kludge to allow package maintainers to specify compiler and 969655714Skris other platforms details on the command line without having to patch the 969755714Skris Configure script everytime: One now can use ``perl Configure 969855714Skris <id>:<details>'', i.e. platform ids are allowed to have details appended 969955714Skris to them (seperated by colons). This is treated as there would be a static 970055714Skris pre-configured entry in Configure's %table under key <id> with value 970155714Skris <details> and ``perl Configure <id>'' is called. So, when you want to 970255714Skris perform a quick test-compile under FreeBSD 3.1 with pgcc and without 970355714Skris assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 970455714Skris now, which overrides the FreeBSD-elf entry on-the-fly. 970555714Skris [Ralf S. Engelschall] 970655714Skris 970755714Skris *) Disable new TLS1 ciphersuites by default: they aren't official yet. 970855714Skris [Ben Laurie] 970955714Skris 971055714Skris *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 971155714Skris on the `perl Configure ...' command line. This way one can compile 971255714Skris OpenSSL libraries with Position Independent Code (PIC) which is needed 971355714Skris for linking it into DSOs. 971455714Skris [Ralf S. Engelschall] 971555714Skris 971655714Skris *) Remarkably, export ciphers were totally broken and no-one had noticed! 971755714Skris Fixed. 971855714Skris [Ben Laurie] 971955714Skris 972055714Skris *) Cleaned up the LICENSE document: The official contact for any license 972155714Skris questions now is the OpenSSL core team under openssl-core@openssl.org. 972255714Skris And add a paragraph about the dual-license situation to make sure people 972355714Skris recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 972455714Skris to the OpenSSL toolkit. 972555714Skris [Ralf S. Engelschall] 972655714Skris 972755714Skris *) General source tree makefile cleanups: Made `making xxx in yyy...' 972855714Skris display consistent in the source tree and replaced `/bin/rm' by `rm'. 972955714Skris Additonally cleaned up the `make links' target: Remove unnecessary 973055714Skris semicolons, subsequent redundant removes, inline point.sh into mklink.sh 973155714Skris to speed processing and no longer clutter the display with confusing 973255714Skris stuff. Instead only the actually done links are displayed. 973355714Skris [Ralf S. Engelschall] 973455714Skris 973555714Skris *) Permit null encryption ciphersuites, used for authentication only. It used 973655714Skris to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 973755714Skris It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 973855714Skris encryption. 973955714Skris [Ben Laurie] 974055714Skris 974155714Skris *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 974255714Skris signed attributes when verifying signatures (this would break them), 974355714Skris the detached data encoding was wrong and public keys obtained using 974455714Skris X509_get_pubkey() weren't freed. 974555714Skris [Steve Henson] 974655714Skris 974755714Skris *) Add text documentation for the BUFFER functions. Also added a work around 974855714Skris to a Win95 console bug. This was triggered by the password read stuff: the 974955714Skris last character typed gets carried over to the next fread(). If you were 975055714Skris generating a new cert request using 'req' for example then the last 975155714Skris character of the passphrase would be CR which would then enter the first 975255714Skris field as blank. 975355714Skris [Steve Henson] 975455714Skris 975555714Skris *) Added the new `Includes OpenSSL Cryptography Software' button as 975655714Skris doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 975755714Skris button and can be used by applications based on OpenSSL to show the 975855714Skris relationship to the OpenSSL project. 975955714Skris [Ralf S. Engelschall] 976055714Skris 976155714Skris *) Remove confusing variables in function signatures in files 976255714Skris ssl/ssl_lib.c and ssl/ssl.h. 976355714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 976455714Skris 976555714Skris *) Don't install bss_file.c under PREFIX/include/ 976655714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 976755714Skris 976855714Skris *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 976955714Skris functions that return function pointers and has support for NT specific 977055714Skris stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 977155714Skris #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 977255714Skris unsigned to signed types: this was killing the Win32 compile. 977355714Skris [Steve Henson] 977455714Skris 977555714Skris *) Add new certificate file to stack functions, 977655714Skris SSL_add_dir_cert_subjects_to_stack() and 977755714Skris SSL_add_file_cert_subjects_to_stack(). These largely supplant 977855714Skris SSL_load_client_CA_file(), and can be used to add multiple certs easily 977955714Skris to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 978055714Skris This means that Apache-SSL and similar packages don't have to mess around 978155714Skris to add as many CAs as they want to the preferred list. 978255714Skris [Ben Laurie] 978355714Skris 978455714Skris *) Experiment with doxygen documentation. Currently only partially applied to 978555714Skris ssl/ssl_lib.c. 978655714Skris See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 978755714Skris openssl.doxy as the configuration file. 978855714Skris [Ben Laurie] 978955714Skris 979055714Skris *) Get rid of remaining C++-style comments which strict C compilers hate. 979155714Skris [Ralf S. Engelschall, pointed out by Carlos Amengual] 979255714Skris 979355714Skris *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 979455714Skris compiled in by default: it has problems with large keys. 979555714Skris [Steve Henson] 979655714Skris 979755714Skris *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 979855714Skris DH private keys and/or callback functions which directly correspond to 979955714Skris their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 980055714Skris is needed for applications which have to configure certificates on a 980155714Skris per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 980255714Skris (e.g. s_server). 980355714Skris For the RSA certificate situation is makes no difference, but 980455714Skris for the DSA certificate situation this fixes the "no shared cipher" 980555714Skris problem where the OpenSSL cipher selection procedure failed because the 980655714Skris temporary keys were not overtaken from the context and the API provided 980755714Skris no way to reconfigure them. 980855714Skris The new functions now let applications reconfigure the stuff and they 980955714Skris are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 981055714Skris SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 981155714Skris non-public-API function ssl_cert_instantiate() is used as a helper 981255714Skris function and also to reduce code redundancy inside ssl_rsa.c. 981355714Skris [Ralf S. Engelschall] 981455714Skris 981555714Skris *) Move s_server -dcert and -dkey options out of the undocumented feature 981655714Skris area because they are useful for the DSA situation and should be 981755714Skris recognized by the users. 981855714Skris [Ralf S. Engelschall] 981955714Skris 982055714Skris *) Fix the cipher decision scheme for export ciphers: the export bits are 982155714Skris *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 982255714Skris SSL_EXP_MASK. So, the original variable has to be used instead of the 982355714Skris already masked variable. 982455714Skris [Richard Levitte <levitte@stacken.kth.se>] 982555714Skris 982655714Skris *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 982755714Skris [Richard Levitte <levitte@stacken.kth.se>] 982855714Skris 982955714Skris *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 983055714Skris from `int' to `unsigned int' because it's a length and initialized by 983155714Skris EVP_DigestFinal() which expects an `unsigned int *'. 983255714Skris [Richard Levitte <levitte@stacken.kth.se>] 983355714Skris 983455714Skris *) Don't hard-code path to Perl interpreter on shebang line of Configure 983555714Skris script. Instead use the usual Shell->Perl transition trick. 983655714Skris [Ralf S. Engelschall] 983755714Skris 983855714Skris *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 983955714Skris (in addition to RSA certificates) to match the behaviour of `openssl dsa 984055714Skris -noout -modulus' as it's already the case for `openssl rsa -noout 984155714Skris -modulus'. For RSA the -modulus is the real "modulus" while for DSA 984255714Skris currently the public key is printed (a decision which was already done by 984355714Skris `openssl dsa -modulus' in the past) which serves a similar purpose. 984455714Skris Additionally the NO_RSA no longer completely removes the whole -modulus 984555714Skris option; it now only avoids using the RSA stuff. Same applies to NO_DSA 984655714Skris now, too. 984755714Skris [Ralf S. Engelschall] 984855714Skris 984955714Skris *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 985055714Skris BIO. See the source (crypto/evp/bio_ok.c) for more info. 985155714Skris [Arne Ansper <arne@ats.cyber.ee>] 985255714Skris 985355714Skris *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 985455714Skris to be added. Now both 'req' and 'ca' can use new objects defined in the 985555714Skris config file. 985655714Skris [Steve Henson] 985755714Skris 985855714Skris *) Add cool BIO that does syslog (or event log on NT). 985955714Skris [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 986055714Skris 986155714Skris *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 986255714Skris TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 986355714Skris TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 986455714Skris Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 986555714Skris [Ben Laurie] 986655714Skris 986755714Skris *) Add preliminary config info for new extension code. 986855714Skris [Steve Henson] 986955714Skris 987055714Skris *) Make RSA_NO_PADDING really use no padding. 987155714Skris [Ulf Moeller <ulf@fitug.de>] 987255714Skris 987355714Skris *) Generate errors when private/public key check is done. 987455714Skris [Ben Laurie] 987555714Skris 987655714Skris *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 987755714Skris for some CRL extensions and new objects added. 987855714Skris [Steve Henson] 987955714Skris 988055714Skris *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 988155714Skris key usage extension and fuller support for authority key id. 988255714Skris [Steve Henson] 988355714Skris 988455714Skris *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 988555714Skris padding method for RSA, which is recommended for new applications in PKCS 988655714Skris #1 v2.0 (RFC 2437, October 1998). 988755714Skris OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 988855714Skris foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 988955714Skris against Bleichbacher's attack on RSA. 989055714Skris [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 989155714Skris Ben Laurie] 989255714Skris 989355714Skris *) Updates to the new SSL compression code 989455714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 989555714Skris 989655714Skris *) Fix so that the version number in the master secret, when passed 989755714Skris via RSA, checks that if TLS was proposed, but we roll back to SSLv3 989855714Skris (because the server will not accept higher), that the version number 989955714Skris is 0x03,0x01, not 0x03,0x00 990055714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 990155714Skris 990255714Skris *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 990355714Skris leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 990455714Skris in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 990555714Skris [Steve Henson] 990655714Skris 990755714Skris *) Support for RAW extensions where an arbitrary extension can be 990855714Skris created by including its DER encoding. See apps/openssl.cnf for 990955714Skris an example. 991055714Skris [Steve Henson] 991155714Skris 991255714Skris *) Make sure latest Perl versions don't interpret some generated C array 991355714Skris code as Perl array code in the crypto/err/err_genc.pl script. 991455714Skris [Lars Weber <3weber@informatik.uni-hamburg.de>] 991555714Skris 991655714Skris *) Modify ms/do_ms.bat to not generate assembly language makefiles since 991755714Skris not many people have the assembler. Various Win32 compilation fixes and 991855714Skris update to the INSTALL.W32 file with (hopefully) more accurate Win32 991955714Skris build instructions. 992055714Skris [Steve Henson] 992155714Skris 992255714Skris *) Modify configure script 'Configure' to automatically create crypto/date.h 992355714Skris file under Win32 and also build pem.h from pem.org. New script 992455714Skris util/mkfiles.pl to create the MINFO file on environments that can't do a 992555714Skris 'make files': perl util/mkfiles.pl >MINFO should work. 992655714Skris [Steve Henson] 992755714Skris 992855714Skris *) Major rework of DES function declarations, in the pursuit of correctness 992955714Skris and purity. As a result, many evil casts evaporated, and some weirdness, 993055714Skris too. You may find this causes warnings in your code. Zapping your evil 993155714Skris casts will probably fix them. Mostly. 993255714Skris [Ben Laurie] 993355714Skris 993455714Skris *) Fix for a typo in asn1.h. Bug fix to object creation script 993555714Skris obj_dat.pl. It considered a zero in an object definition to mean 993655714Skris "end of object": none of the objects in objects.h have any zeros 993755714Skris so it wasn't spotted. 993855714Skris [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 993955714Skris 994055714Skris *) Add support for Triple DES Cipher Block Chaining with Output Feedback 994155714Skris Masking (CBCM). In the absence of test vectors, the best I have been able 994255714Skris to do is check that the decrypt undoes the encrypt, so far. Send me test 994355714Skris vectors if you have them. 994455714Skris [Ben Laurie] 994555714Skris 994655714Skris *) Correct calculation of key length for export ciphers (too much space was 994755714Skris allocated for null ciphers). This has not been tested! 994855714Skris [Ben Laurie] 994955714Skris 995055714Skris *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 995155714Skris message is now correct (it understands "crypto" and "ssl" on its 995255714Skris command line). There is also now an "update" option. This will update 995355714Skris the util/ssleay.num and util/libeay.num files with any new functions. 995455714Skris If you do a: 995555714Skris perl util/mkdef.pl crypto ssl update 995655714Skris it will update them. 995755714Skris [Steve Henson] 995855714Skris 995955714Skris *) Overhauled the Perl interface (perl/*): 996055714Skris - ported BN stuff to OpenSSL's different BN library 996155714Skris - made the perl/ source tree CVS-aware 996255714Skris - renamed the package from SSLeay to OpenSSL (the files still contain 996355714Skris their history because I've copied them in the repository) 996455714Skris - removed obsolete files (the test scripts will be replaced 996555714Skris by better Test::Harness variants in the future) 996655714Skris [Ralf S. Engelschall] 996755714Skris 996855714Skris *) First cut for a very conservative source tree cleanup: 996955714Skris 1. merge various obsolete readme texts into doc/ssleay.txt 997055714Skris where we collect the old documents and readme texts. 997155714Skris 2. remove the first part of files where I'm already sure that we no 997255714Skris longer need them because of three reasons: either they are just temporary 997355714Skris files which were left by Eric or they are preserved original files where 997455714Skris I've verified that the diff is also available in the CVS via "cvs diff 997555714Skris -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 997655714Skris the crypto/md/ stuff). 997755714Skris [Ralf S. Engelschall] 997855714Skris 997955714Skris *) More extension code. Incomplete support for subject and issuer alt 998055714Skris name, issuer and authority key id. Change the i2v function parameters 998155714Skris and add an extra 'crl' parameter in the X509V3_CTX structure: guess 998255714Skris what that's for :-) Fix to ASN1 macro which messed up 998355714Skris IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 998455714Skris [Steve Henson] 998555714Skris 998655714Skris *) Preliminary support for ENUMERATED type. This is largely copied from the 998755714Skris INTEGER code. 998855714Skris [Steve Henson] 998955714Skris 999055714Skris *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 999155714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 999255714Skris 999355714Skris *) Make sure `make rehash' target really finds the `openssl' program. 999455714Skris [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 999555714Skris 999655714Skris *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 999755714Skris like to hear about it if this slows down other processors. 999855714Skris [Ben Laurie] 999955714Skris 1000055714Skris *) Add CygWin32 platform information to Configure script. 1000155714Skris [Alan Batie <batie@aahz.jf.intel.com>] 1000255714Skris 1000355714Skris *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 1000455714Skris [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 1000555714Skris 1000655714Skris *) New program nseq to manipulate netscape certificate sequences 1000755714Skris [Steve Henson] 1000855714Skris 1000955714Skris *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 1001055714Skris few typos. 1001155714Skris [Steve Henson] 1001255714Skris 1001355714Skris *) Fixes to BN code. Previously the default was to define BN_RECURSION 1001455714Skris but the BN code had some problems that would cause failures when 1001555714Skris doing certificate verification and some other functions. 1001655714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 1001755714Skris 1001855714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 1001955714Skris [Steve Henson] 1002055714Skris 1002155714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 1002255714Skris [Steve Henson] 1002355714Skris 1002455714Skris *) Add several PKIX and private extended key usage OIDs. 1002555714Skris [Steve Henson] 1002655714Skris 1002755714Skris *) Modify the 'ca' program to handle the new extension code. Modify 1002855714Skris openssl.cnf for new extension format, add comments. 1002955714Skris [Steve Henson] 1003055714Skris 1003155714Skris *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 1003255714Skris and add a sample to openssl.cnf so req -x509 now adds appropriate 1003355714Skris CA extensions. 1003455714Skris [Steve Henson] 1003555714Skris 1003655714Skris *) Continued X509 V3 changes. Add to other makefiles, integrate with the 1003755714Skris error code, add initial support to X509_print() and x509 application. 1003855714Skris [Steve Henson] 1003955714Skris 1004055714Skris *) Takes a deep breath and start addding X509 V3 extension support code. Add 1004155714Skris files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 1004255714Skris stuff is currently isolated and isn't even compiled yet. 1004355714Skris [Steve Henson] 1004455714Skris 1004555714Skris *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 1004655714Skris ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 1004755714Skris Removed the versions check from X509 routines when loading extensions: 1004855714Skris this allows certain broken certificates that don't set the version 1004955714Skris properly to be processed. 1005055714Skris [Steve Henson] 1005155714Skris 1005255714Skris *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 1005355714Skris Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 1005455714Skris can still be regenerated with "make depend". 1005555714Skris [Ben Laurie] 1005655714Skris 1005755714Skris *) Spelling mistake in C version of CAST-128. 1005855714Skris [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 1005955714Skris 1006055714Skris *) Changes to the error generation code. The perl script err-code.pl 1006155714Skris now reads in the old error codes and retains the old numbers, only 1006255714Skris adding new ones if necessary. It also only changes the .err files if new 1006355714Skris codes are added. The makefiles have been modified to only insert errors 1006455714Skris when needed (to avoid needlessly modifying header files). This is done 1006555714Skris by only inserting errors if the .err file is newer than the auto generated 1006655714Skris C file. To rebuild all the error codes from scratch (the old behaviour) 1006755714Skris either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 1006855714Skris or delete all the .err files. 1006955714Skris [Steve Henson] 1007055714Skris 1007155714Skris *) CAST-128 was incorrectly implemented for short keys. The C version has 1007255714Skris been fixed, but is untested. The assembler versions are also fixed, but 1007355714Skris new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 1007455714Skris to regenerate it if needed. 1007555714Skris [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 1007655714Skris Hagino <itojun@kame.net>] 1007755714Skris 1007855714Skris *) File was opened incorrectly in randfile.c. 1007955714Skris [Ulf M�ller <ulf@fitug.de>] 1008055714Skris 1008155714Skris *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 1008255714Skris functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 1008355714Skris GeneralizedTime. ASN1_TIME is the proper type used in certificates et 1008455714Skris al: it's just almost always a UTCTime. Note this patch adds new error 1008555714Skris codes so do a "make errors" if there are problems. 1008655714Skris [Steve Henson] 1008755714Skris 1008855714Skris *) Correct Linux 1 recognition in config. 1008955714Skris [Ulf M�ller <ulf@fitug.de>] 1009055714Skris 1009155714Skris *) Remove pointless MD5 hash when using DSA keys in ca. 1009255714Skris [Anonymous <nobody@replay.com>] 1009355714Skris 1009455714Skris *) Generate an error if given an empty string as a cert directory. Also 1009555714Skris generate an error if handed NULL (previously returned 0 to indicate an 1009655714Skris error, but didn't set one). 1009755714Skris [Ben Laurie, reported by Anonymous <nobody@replay.com>] 1009855714Skris 1009955714Skris *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 1010055714Skris [Ben Laurie] 1010155714Skris 1010255714Skris *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 1010355714Skris parameters. This was causing a warning which killed off the Win32 compile. 1010455714Skris [Steve Henson] 1010555714Skris 1010655714Skris *) Remove C++ style comments from crypto/bn/bn_local.h. 1010755714Skris [Neil Costigan <neil.costigan@celocom.com>] 1010855714Skris 1010955714Skris *) The function OBJ_txt2nid was broken. It was supposed to return a nid 1011055714Skris based on a text string, looking up short and long names and finally 1011155714Skris "dot" format. The "dot" format stuff didn't work. Added new function 1011255714Skris OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 1011355714Skris OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 1011455714Skris OID is not part of the table. 1011555714Skris [Steve Henson] 1011655714Skris 1011755714Skris *) Add prototypes to X509 lookup/verify methods, fixing a bug in 1011855714Skris X509_LOOKUP_by_alias(). 1011955714Skris [Ben Laurie] 1012055714Skris 1012155714Skris *) Sort openssl functions by name. 1012255714Skris [Ben Laurie] 1012355714Skris 1012455714Skris *) Get the gendsa program working (hopefully) and add it to app list. Remove 1012555714Skris encryption from sample DSA keys (in case anyone is interested the password 1012655714Skris was "1234"). 1012755714Skris [Steve Henson] 1012855714Skris 1012955714Skris *) Make _all_ *_free functions accept a NULL pointer. 1013055714Skris [Frans Heymans <fheymans@isaserver.be>] 1013155714Skris 1013255714Skris *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 1013355714Skris NULL pointers. 1013455714Skris [Anonymous <nobody@replay.com>] 1013555714Skris 1013655714Skris *) s_server should send the CAfile as acceptable CAs, not its own cert. 1013755714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 1013855714Skris 1013955714Skris *) Don't blow it for numeric -newkey arguments to apps/req. 1014055714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 1014155714Skris 1014255714Skris *) Temp key "for export" tests were wrong in s3_srvr.c. 1014355714Skris [Anonymous <nobody@replay.com>] 1014455714Skris 1014555714Skris *) Add prototype for temp key callback functions 1014655714Skris SSL_CTX_set_tmp_{rsa,dh}_callback(). 1014755714Skris [Ben Laurie] 1014855714Skris 1014955714Skris *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 1015055714Skris DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 1015155714Skris [Steve Henson] 1015255714Skris 1015355714Skris *) X509_name_add_entry() freed the wrong thing after an error. 1015455714Skris [Arne Ansper <arne@ats.cyber.ee>] 1015555714Skris 1015655714Skris *) rsa_eay.c would attempt to free a NULL context. 1015755714Skris [Arne Ansper <arne@ats.cyber.ee>] 1015855714Skris 1015955714Skris *) BIO_s_socket() had a broken should_retry() on Windoze. 1016055714Skris [Arne Ansper <arne@ats.cyber.ee>] 1016155714Skris 1016255714Skris *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 1016355714Skris [Arne Ansper <arne@ats.cyber.ee>] 1016455714Skris 1016555714Skris *) Make sure the already existing X509_STORE->depth variable is initialized 1016655714Skris in X509_STORE_new(), but document the fact that this variable is still 1016755714Skris unused in the certificate verification process. 1016855714Skris [Ralf S. Engelschall] 1016955714Skris 1017055714Skris *) Fix the various library and apps files to free up pkeys obtained from 1017155714Skris X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 1017255714Skris [Steve Henson] 1017355714Skris 1017455714Skris *) Fix reference counting in X509_PUBKEY_get(). This makes 1017555714Skris demos/maurice/example2.c work, amongst others, probably. 1017655714Skris [Steve Henson and Ben Laurie] 1017755714Skris 1017855714Skris *) First cut of a cleanup for apps/. First the `ssleay' program is now named 1017955714Skris `openssl' and second, the shortcut symlinks for the `openssl <command>' 1018055714Skris are no longer created. This way we have a single and consistent command 1018155714Skris line interface `openssl <command>', similar to `cvs <command>'. 1018255714Skris [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 1018355714Skris 1018455714Skris *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 1018555714Skris BIT STRING wrapper always have zero unused bits. 1018655714Skris [Steve Henson] 1018755714Skris 1018855714Skris *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 1018955714Skris [Steve Henson] 1019055714Skris 1019155714Skris *) Make the top-level INSTALL documentation easier to understand. 1019255714Skris [Paul Sutton] 1019355714Skris 1019455714Skris *) Makefiles updated to exit if an error occurs in a sub-directory 1019555714Skris make (including if user presses ^C) [Paul Sutton] 1019655714Skris 1019755714Skris *) Make Montgomery context stuff explicit in RSA data structure. 1019855714Skris [Ben Laurie] 1019955714Skris 1020055714Skris *) Fix build order of pem and err to allow for generated pem.h. 1020155714Skris [Ben Laurie] 1020255714Skris 1020355714Skris *) Fix renumbering bug in X509_NAME_delete_entry(). 1020455714Skris [Ben Laurie] 1020555714Skris 1020655714Skris *) Enhanced the err-ins.pl script so it makes the error library number 1020755714Skris global and can add a library name. This is needed for external ASN1 and 1020855714Skris other error libraries. 1020955714Skris [Steve Henson] 1021055714Skris 1021155714Skris *) Fixed sk_insert which never worked properly. 1021255714Skris [Steve Henson] 1021355714Skris 1021455714Skris *) Fix ASN1 macros so they can handle indefinite length construted 1021555714Skris EXPLICIT tags. Some non standard certificates use these: they can now 1021655714Skris be read in. 1021755714Skris [Steve Henson] 1021855714Skris 1021955714Skris *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 1022055714Skris into a single doc/ssleay.txt bundle. This way the information is still 1022155714Skris preserved but no longer messes up this directory. Now it's new room for 1022255714Skris the new set of documenation files. 1022355714Skris [Ralf S. Engelschall] 1022455714Skris 1022555714Skris *) SETs were incorrectly DER encoded. This was a major pain, because they 1022655714Skris shared code with SEQUENCEs, which aren't coded the same. This means that 1022755714Skris almost everything to do with SETs or SEQUENCEs has either changed name or 1022855714Skris number of arguments. 1022955714Skris [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 1023055714Skris 1023155714Skris *) Fix test data to work with the above. 1023255714Skris [Ben Laurie] 1023355714Skris 1023455714Skris *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 1023555714Skris was already fixed by Eric for 0.9.1 it seems. 1023655714Skris [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 1023755714Skris 1023855714Skris *) Autodetect FreeBSD3. 1023955714Skris [Ben Laurie] 1024055714Skris 1024155714Skris *) Fix various bugs in Configure. This affects the following platforms: 1024255714Skris nextstep 1024355714Skris ncr-scde 1024455714Skris unixware-2.0 1024555714Skris unixware-2.0-pentium 1024655714Skris sco5-cc. 1024755714Skris [Ben Laurie] 1024855714Skris 1024955714Skris *) Eliminate generated files from CVS. Reorder tests to regenerate files 1025055714Skris before they are needed. 1025155714Skris [Ben Laurie] 1025255714Skris 1025355714Skris *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 1025455714Skris [Ben Laurie] 1025555714Skris 1025655714Skris 1025755714Skris Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 1025855714Skris 1025955714Skris *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 1026055714Skris changed SSLeay to OpenSSL in version strings. 1026155714Skris [Ralf S. Engelschall] 1026255714Skris 1026355714Skris *) Some fixups to the top-level documents. 1026455714Skris [Paul Sutton] 1026555714Skris 1026655714Skris *) Fixed the nasty bug where rsaref.h was not found under compile-time 1026755714Skris because the symlink to include/ was missing. 1026855714Skris [Ralf S. Engelschall] 1026955714Skris 1027055714Skris *) Incorporated the popular no-RSA/DSA-only patches 1027155714Skris which allow to compile a RSA-free SSLeay. 1027255714Skris [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 1027355714Skris 1027455714Skris *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 1027555714Skris when "ssleay" is still not found. 1027655714Skris [Ralf S. Engelschall] 1027755714Skris 1027855714Skris *) Added more platforms to Configure: Cray T3E, HPUX 11, 1027955714Skris [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 1028055714Skris 1028155714Skris *) Updated the README file. 1028255714Skris [Ralf S. Engelschall] 1028355714Skris 1028455714Skris *) Added various .cvsignore files in the CVS repository subdirs 1028555714Skris to make a "cvs update" really silent. 1028655714Skris [Ralf S. Engelschall] 1028755714Skris 1028855714Skris *) Recompiled the error-definition header files and added 1028955714Skris missing symbols to the Win32 linker tables. 1029055714Skris [Ralf S. Engelschall] 1029155714Skris 1029255714Skris *) Cleaned up the top-level documents; 1029355714Skris o new files: CHANGES and LICENSE 1029455714Skris o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 1029555714Skris o merged COPYRIGHT into LICENSE 1029655714Skris o removed obsolete TODO file 1029755714Skris o renamed MICROSOFT to INSTALL.W32 1029855714Skris [Ralf S. Engelschall] 1029955714Skris 1030055714Skris *) Removed dummy files from the 0.9.1b source tree: 1030155714Skris crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 1030255714Skris crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 1030355714Skris crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 1030455714Skris crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 1030555714Skris util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 1030655714Skris [Ralf S. Engelschall] 1030755714Skris 1030855714Skris *) Added various platform portability fixes. 1030955714Skris [Mark J. Cox] 1031055714Skris 1031155714Skris *) The Genesis of the OpenSSL rpject: 1031255714Skris We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 1031355714Skris Young and Tim J. Hudson created while they were working for C2Net until 1031455714Skris summer 1998. 1031555714Skris [The OpenSSL Project] 1031655714Skris 1031755714Skris 1031855714Skris Changes between 0.9.0b and 0.9.1b [not released] 1031955714Skris 1032055714Skris *) Updated a few CA certificates under certs/ 1032155714Skris [Eric A. Young] 1032255714Skris 1032355714Skris *) Changed some BIGNUM api stuff. 1032455714Skris [Eric A. Young] 1032555714Skris 1032655714Skris *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 1032755714Skris DGUX x86, Linux Alpha, etc. 1032855714Skris [Eric A. Young] 1032955714Skris 1033055714Skris *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 1033155714Skris RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 1033255714Skris available). 1033355714Skris [Eric A. Young] 1033455714Skris 1033555714Skris *) Add -strparse option to asn1pars program which parses nested 1033655714Skris binary structures 1033755714Skris [Dr Stephen Henson <shenson@bigfoot.com>] 1033855714Skris 1033955714Skris *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 1034055714Skris [Eric A. Young] 1034155714Skris 1034255714Skris *) DSA fix for "ca" program. 1034355714Skris [Eric A. Young] 1034455714Skris 1034555714Skris *) Added "-genkey" option to "dsaparam" program. 1034655714Skris [Eric A. Young] 1034755714Skris 1034855714Skris *) Added RIPE MD160 (rmd160) message digest. 1034955714Skris [Eric A. Young] 1035055714Skris 1035155714Skris *) Added -a (all) option to "ssleay version" command. 1035255714Skris [Eric A. Young] 1035355714Skris 1035455714Skris *) Added PLATFORM define which is the id given to Configure. 1035555714Skris [Eric A. Young] 1035655714Skris 1035755714Skris *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 1035855714Skris [Eric A. Young] 1035955714Skris 1036055714Skris *) Extended the ASN.1 parser routines. 1036155714Skris [Eric A. Young] 1036255714Skris 1036355714Skris *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 1036455714Skris [Eric A. Young] 1036555714Skris 1036655714Skris *) Added a BN_CTX to the BN library. 1036755714Skris [Eric A. Young] 1036855714Skris 1036955714Skris *) Fixed the weak key values in DES library 1037055714Skris [Eric A. Young] 1037155714Skris 1037255714Skris *) Changed API in EVP library for cipher aliases. 1037355714Skris [Eric A. Young] 1037455714Skris 1037555714Skris *) Added support for RC2/64bit cipher. 1037655714Skris [Eric A. Young] 1037755714Skris 1037855714Skris *) Converted the lhash library to the crypto/mem.c functions. 1037955714Skris [Eric A. Young] 1038055714Skris 1038155714Skris *) Added more recognized ASN.1 object ids. 1038255714Skris [Eric A. Young] 1038355714Skris 1038455714Skris *) Added more RSA padding checks for SSL/TLS. 1038555714Skris [Eric A. Young] 1038655714Skris 1038755714Skris *) Added BIO proxy/filter functionality. 1038855714Skris [Eric A. Young] 1038955714Skris 1039055714Skris *) Added extra_certs to SSL_CTX which can be used 1039155714Skris send extra CA certificates to the client in the CA cert chain sending 1039255714Skris process. It can be configured with SSL_CTX_add_extra_chain_cert(). 1039355714Skris [Eric A. Young] 1039455714Skris 1039555714Skris *) Now Fortezza is denied in the authentication phase because 1039655714Skris this is key exchange mechanism is not supported by SSLeay at all. 1039755714Skris [Eric A. Young] 1039855714Skris 1039955714Skris *) Additional PKCS1 checks. 1040055714Skris [Eric A. Young] 1040155714Skris 1040255714Skris *) Support the string "TLSv1" for all TLS v1 ciphers. 1040355714Skris [Eric A. Young] 1040455714Skris 1040555714Skris *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 1040655714Skris ex_data index of the SSL context in the X509_STORE_CTX ex_data. 1040755714Skris [Eric A. Young] 1040855714Skris 1040955714Skris *) Fixed a few memory leaks. 1041055714Skris [Eric A. Young] 1041155714Skris 1041255714Skris *) Fixed various code and comment typos. 1041355714Skris [Eric A. Young] 1041455714Skris 1041555714Skris *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 1041655714Skris bytes sent in the client random. 1041755714Skris [Edward Bishop <ebishop@spyglass.com>] 1041855714Skris 10419