CHANGES revision 264331
155714Skris 255714Skris OpenSSL CHANGES 355714Skris _______________ 455714Skris 5264331Sjkim Changes between 1.0.1f and 1.0.1g [7 Apr 2014] 6264331Sjkim 7264331Sjkim *) A missing bounds check in the handling of the TLS heartbeat extension 8264331Sjkim can be used to reveal up to 64k of memory to a connected client or 9264331Sjkim server. 10264331Sjkim 11264331Sjkim Thanks for Neel Mehta of Google Security for discovering this bug and to 12264331Sjkim Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for 13264331Sjkim preparing the fix (CVE-2014-0160) 14264331Sjkim [Adam Langley, Bodo Moeller] 15264331Sjkim 16264331Sjkim *) Fix for the attack described in the paper "Recovering OpenSSL 17264331Sjkim ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack" 18264331Sjkim by Yuval Yarom and Naomi Benger. Details can be obtained from: 19264331Sjkim http://eprint.iacr.org/2014/140 20264331Sjkim 21264331Sjkim Thanks to Yuval Yarom and Naomi Benger for discovering this 22264331Sjkim flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076) 23264331Sjkim [Yuval Yarom and Naomi Benger] 24264331Sjkim 25264331Sjkim *) TLS pad extension: draft-agl-tls-padding-03 26264331Sjkim 27264331Sjkim Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the 28264331Sjkim TLS client Hello record length value would otherwise be > 255 and 29264331Sjkim less that 512 pad with a dummy extension containing zeroes so it 30264331Sjkim is at least 512 bytes long. 31264331Sjkim 32264331Sjkim [Adam Langley, Steve Henson] 33264331Sjkim 34264331Sjkim Changes between 1.0.1e and 1.0.1f [6 Jan 2014] 35264331Sjkim 36264331Sjkim *) Fix for TLS record tampering bug. A carefully crafted invalid 37264331Sjkim handshake could crash OpenSSL with a NULL pointer exception. 38264331Sjkim Thanks to Anton Johansson for reporting this issues. 39264331Sjkim (CVE-2013-4353) 40264331Sjkim 41264331Sjkim *) Keep original DTLS digest and encryption contexts in retransmission 42264331Sjkim structures so we can use the previous session parameters if they need 43264331Sjkim to be resent. (CVE-2013-6450) 44264331Sjkim [Steve Henson] 45264331Sjkim 46264331Sjkim *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which 47264331Sjkim avoids preferring ECDHE-ECDSA ciphers when the client appears to be 48264331Sjkim Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for 49264331Sjkim several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug 50264331Sjkim is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing 51264331Sjkim 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer. 52264331Sjkim [Rob Stradling, Adam Langley] 53264331Sjkim 54246772Sjkim Changes between 1.0.1d and 1.0.1e [11 Feb 2013] 55246772Sjkim 56264331Sjkim *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI 57264331Sjkim supporting platforms or when small records were transferred. 58264331Sjkim [Andy Polyakov, Steve Henson] 59246772Sjkim 60246772Sjkim Changes between 1.0.1c and 1.0.1d [5 Feb 2013] 61246772Sjkim 62246772Sjkim *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time. 63246772Sjkim 64246772Sjkim This addresses the flaw in CBC record processing discovered by 65246772Sjkim Nadhem Alfardan and Kenny Paterson. Details of this attack can be found 66246772Sjkim at: http://www.isg.rhul.ac.uk/tls/ 67246772Sjkim 68246772Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 69246772Sjkim Security Group at Royal Holloway, University of London 70246772Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and 71246772Sjkim Emilia K�sper for the initial patch. 72246772Sjkim (CVE-2013-0169) 73246772Sjkim [Emilia K�sper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson] 74246772Sjkim 75246772Sjkim *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode 76246772Sjkim ciphersuites which can be exploited in a denial of service attack. 77246772Sjkim Thanks go to and to Adam Langley <agl@chromium.org> for discovering 78246772Sjkim and detecting this bug and to Wolfgang Ettlinger 79246772Sjkim <wolfgang.ettlinger@gmail.com> for independently discovering this issue. 80246772Sjkim (CVE-2012-2686) 81246772Sjkim [Adam Langley] 82246772Sjkim 83246772Sjkim *) Return an error when checking OCSP signatures when key is NULL. 84246772Sjkim This fixes a DoS attack. (CVE-2013-0166) 85246772Sjkim [Steve Henson] 86246772Sjkim 87246772Sjkim *) Make openssl verify return errors. 88246772Sjkim [Chris Palmer <palmer@google.com> and Ben Laurie] 89246772Sjkim 90246772Sjkim *) Call OCSP Stapling callback after ciphersuite has been chosen, so 91246772Sjkim the right response is stapled. Also change SSL_get_certificate() 92246772Sjkim so it returns the certificate actually sent. 93246772Sjkim See http://rt.openssl.org/Ticket/Display.html?id=2836. 94246772Sjkim [Rob Stradling <rob.stradling@comodo.com>] 95246772Sjkim 96246772Sjkim *) Fix possible deadlock when decoding public keys. 97246772Sjkim [Steve Henson] 98246772Sjkim 99246772Sjkim *) Don't use TLS 1.0 record version number in initial client hello 100246772Sjkim if renegotiating. 101246772Sjkim [Steve Henson] 102246772Sjkim 103238405Sjkim Changes between 1.0.1b and 1.0.1c [10 May 2012] 104237657Sjkim 105238405Sjkim *) Sanity check record length before skipping explicit IV in TLS 106238405Sjkim 1.2, 1.1 and DTLS to fix DoS attack. 107237657Sjkim 108237657Sjkim Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic 109237657Sjkim fuzzing as a service testing platform. 110237657Sjkim (CVE-2012-2333) 111237657Sjkim [Steve Henson] 112237657Sjkim 113237657Sjkim *) Initialise tkeylen properly when encrypting CMS messages. 114237657Sjkim Thanks to Solar Designer of Openwall for reporting this issue. 115237657Sjkim [Steve Henson] 116237657Sjkim 117238405Sjkim *) In FIPS mode don't try to use composite ciphers as they are not 118238405Sjkim approved. 119238405Sjkim [Steve Henson] 120237657Sjkim 121238405Sjkim Changes between 1.0.1a and 1.0.1b [26 Apr 2012] 122237657Sjkim 123238405Sjkim *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and 124238405Sjkim 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately 125238405Sjkim mean any application compiled against OpenSSL 1.0.0 headers setting 126238405Sjkim SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng 127238405Sjkim TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to 128238405Sjkim 0x10000000L Any application which was previously compiled against 129238405Sjkim OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1 130238405Sjkim will need to be recompiled as a result. Letting be results in 131238405Sjkim inability to disable specifically TLS 1.1 and in client context, 132238405Sjkim in unlike event, limit maximum offered version to TLS 1.0 [see below]. 133238405Sjkim [Steve Henson] 134237657Sjkim 135238405Sjkim *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not 136238405Sjkim disable just protocol X, but all protocols above X *if* there are 137238405Sjkim protocols *below* X still enabled. In more practical terms it means 138238405Sjkim that if application wants to disable TLS1.0 in favor of TLS1.1 and 139238405Sjkim above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass 140238405Sjkim SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to 141238405Sjkim client side. 142238405Sjkim [Andy Polyakov] 143238405Sjkim 144238405Sjkim Changes between 1.0.1 and 1.0.1a [19 Apr 2012] 145238405Sjkim 146237657Sjkim *) Check for potentially exploitable overflows in asn1_d2i_read_bio 147237657Sjkim BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer 148237657Sjkim in CRYPTO_realloc_clean. 149237657Sjkim 150237657Sjkim Thanks to Tavis Ormandy, Google Security Team, for discovering this 151237657Sjkim issue and to Adam Langley <agl@chromium.org> for fixing it. 152237657Sjkim (CVE-2012-2110) 153237657Sjkim [Adam Langley (Google), Tavis Ormandy, Google Security Team] 154237657Sjkim 155238405Sjkim *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections. 156238405Sjkim [Adam Langley] 157237657Sjkim 158238405Sjkim *) Workarounds for some broken servers that "hang" if a client hello 159238405Sjkim record length exceeds 255 bytes. 160238405Sjkim 161238405Sjkim 1. Do not use record version number > TLS 1.0 in initial client 162238405Sjkim hello: some (but not all) hanging servers will now work. 163238405Sjkim 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate 164238405Sjkim the number of ciphers sent in the client hello. This should be 165238405Sjkim set to an even number, such as 50, for example by passing: 166238405Sjkim -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure. 167238405Sjkim Most broken servers should now work. 168238405Sjkim 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable 169238405Sjkim TLS 1.2 client support entirely. 170238405Sjkim [Steve Henson] 171238405Sjkim 172238405Sjkim *) Fix SEGV in Vector Permutation AES module observed in OpenSSH. 173238405Sjkim [Andy Polyakov] 174238405Sjkim 175238405Sjkim Changes between 1.0.0h and 1.0.1 [14 Mar 2012] 176238405Sjkim 177238405Sjkim *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET 178238405Sjkim STRING form instead of a DigestInfo. 179238405Sjkim [Steve Henson] 180238405Sjkim 181238405Sjkim *) The format used for MDC2 RSA signatures is inconsistent between EVP 182238405Sjkim and the RSA_sign/RSA_verify functions. This was made more apparent when 183238405Sjkim OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular 184238405Sjkim those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect 185238405Sjkim the correct format in RSA_verify so both forms transparently work. 186238405Sjkim [Steve Henson] 187238405Sjkim 188238405Sjkim *) Some servers which support TLS 1.0 can choke if we initially indicate 189238405Sjkim support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA 190238405Sjkim encrypted premaster secret. As a workaround use the maximum pemitted 191238405Sjkim client version in client hello, this should keep such servers happy 192238405Sjkim and still work with previous versions of OpenSSL. 193238405Sjkim [Steve Henson] 194238405Sjkim 195238405Sjkim *) Add support for TLS/DTLS heartbeats. 196238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>] 197238405Sjkim 198238405Sjkim *) Add support for SCTP. 199238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>] 200238405Sjkim 201238405Sjkim *) Improved PRNG seeding for VOS. 202238405Sjkim [Paul Green <Paul.Green@stratus.com>] 203238405Sjkim 204238405Sjkim *) Extensive assembler packs updates, most notably: 205238405Sjkim 206238405Sjkim - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support; 207238405Sjkim - x86[_64]: SSSE3 support (SHA1, vector-permutation AES); 208238405Sjkim - x86_64: bit-sliced AES implementation; 209238405Sjkim - ARM: NEON support, contemporary platforms optimizations; 210238405Sjkim - s390x: z196 support; 211238405Sjkim - *: GHASH and GF(2^m) multiplication implementations; 212238405Sjkim 213238405Sjkim [Andy Polyakov] 214238405Sjkim 215238405Sjkim *) Make TLS-SRP code conformant with RFC 5054 API cleanup 216238405Sjkim (removal of unnecessary code) 217238405Sjkim [Peter Sylvester <peter.sylvester@edelweb.fr>] 218238405Sjkim 219238405Sjkim *) Add TLS key material exporter from RFC 5705. 220238405Sjkim [Eric Rescorla] 221238405Sjkim 222238405Sjkim *) Add DTLS-SRTP negotiation from RFC 5764. 223238405Sjkim [Eric Rescorla] 224238405Sjkim 225238405Sjkim *) Add Next Protocol Negotiation, 226238405Sjkim http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be 227238405Sjkim disabled with a no-npn flag to config or Configure. Code donated 228238405Sjkim by Google. 229238405Sjkim [Adam Langley <agl@google.com> and Ben Laurie] 230238405Sjkim 231238405Sjkim *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224, 232238405Sjkim NIST-P256, NIST-P521, with constant-time single point multiplication on 233238405Sjkim typical inputs. Compiler support for the nonstandard type __uint128_t is 234238405Sjkim required to use this (present in gcc 4.4 and later, for 64-bit builds). 235238405Sjkim Code made available under Apache License version 2.0. 236238405Sjkim 237238405Sjkim Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command 238238405Sjkim line to include this in your build of OpenSSL, and run "make depend" (or 239238405Sjkim "make update"). This enables the following EC_METHODs: 240238405Sjkim 241238405Sjkim EC_GFp_nistp224_method() 242238405Sjkim EC_GFp_nistp256_method() 243238405Sjkim EC_GFp_nistp521_method() 244238405Sjkim 245238405Sjkim EC_GROUP_new_by_curve_name() will automatically use these (while 246238405Sjkim EC_GROUP_new_curve_GFp() currently prefers the more flexible 247238405Sjkim implementations). 248238405Sjkim [Emilia K�sper, Adam Langley, Bodo Moeller (Google)] 249238405Sjkim 250238405Sjkim *) Use type ossl_ssize_t instad of ssize_t which isn't available on 251238405Sjkim all platforms. Move ssize_t definition from e_os.h to the public 252238405Sjkim header file e_os2.h as it now appears in public header file cms.h 253238405Sjkim [Steve Henson] 254238405Sjkim 255238405Sjkim *) New -sigopt option to the ca, req and x509 utilities. Additional 256238405Sjkim signature parameters can be passed using this option and in 257238405Sjkim particular PSS. 258238405Sjkim [Steve Henson] 259238405Sjkim 260238405Sjkim *) Add RSA PSS signing function. This will generate and set the 261238405Sjkim appropriate AlgorithmIdentifiers for PSS based on those in the 262238405Sjkim corresponding EVP_MD_CTX structure. No application support yet. 263238405Sjkim [Steve Henson] 264238405Sjkim 265238405Sjkim *) Support for companion algorithm specific ASN1 signing routines. 266238405Sjkim New function ASN1_item_sign_ctx() signs a pre-initialised 267238405Sjkim EVP_MD_CTX structure and sets AlgorithmIdentifiers based on 268238405Sjkim the appropriate parameters. 269238405Sjkim [Steve Henson] 270238405Sjkim 271238405Sjkim *) Add new algorithm specific ASN1 verification initialisation function 272238405Sjkim to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1 273238405Sjkim handling will be the same no matter what EVP_PKEY_METHOD is used. 274238405Sjkim Add a PSS handler to support verification of PSS signatures: checked 275238405Sjkim against a number of sample certificates. 276238405Sjkim [Steve Henson] 277238405Sjkim 278238405Sjkim *) Add signature printing for PSS. Add PSS OIDs. 279238405Sjkim [Steve Henson, Martin Kaiser <lists@kaiser.cx>] 280238405Sjkim 281238405Sjkim *) Add algorithm specific signature printing. An individual ASN1 method 282238405Sjkim can now print out signatures instead of the standard hex dump. 283238405Sjkim 284238405Sjkim More complex signatures (e.g. PSS) can print out more meaningful 285238405Sjkim information. Include DSA version that prints out the signature 286238405Sjkim parameters r, s. 287238405Sjkim [Steve Henson] 288238405Sjkim 289238405Sjkim *) Password based recipient info support for CMS library: implementing 290238405Sjkim RFC3211. 291238405Sjkim [Steve Henson] 292238405Sjkim 293238405Sjkim *) Split password based encryption into PBES2 and PBKDF2 functions. This 294238405Sjkim neatly separates the code into cipher and PBE sections and is required 295238405Sjkim for some algorithms that split PBES2 into separate pieces (such as 296238405Sjkim password based CMS). 297238405Sjkim [Steve Henson] 298238405Sjkim 299238405Sjkim *) Session-handling fixes: 300238405Sjkim - Fix handling of connections that are resuming with a session ID, 301238405Sjkim but also support Session Tickets. 302238405Sjkim - Fix a bug that suppressed issuing of a new ticket if the client 303238405Sjkim presented a ticket with an expired session. 304238405Sjkim - Try to set the ticket lifetime hint to something reasonable. 305238405Sjkim - Make tickets shorter by excluding irrelevant information. 306238405Sjkim - On the client side, don't ignore renewed tickets. 307238405Sjkim [Adam Langley, Bodo Moeller (Google)] 308238405Sjkim 309238405Sjkim *) Fix PSK session representation. 310238405Sjkim [Bodo Moeller] 311238405Sjkim 312238405Sjkim *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations. 313238405Sjkim 314238405Sjkim This work was sponsored by Intel. 315238405Sjkim [Andy Polyakov] 316238405Sjkim 317238405Sjkim *) Add GCM support to TLS library. Some custom code is needed to split 318238405Sjkim the IV between the fixed (from PRF) and explicit (from TLS record) 319238405Sjkim portions. This adds all GCM ciphersuites supported by RFC5288 and 320238405Sjkim RFC5289. Generalise some AES* cipherstrings to inlclude GCM and 321238405Sjkim add a special AESGCM string for GCM only. 322238405Sjkim [Steve Henson] 323238405Sjkim 324238405Sjkim *) Expand range of ctrls for AES GCM. Permit setting invocation 325238405Sjkim field on decrypt and retrieval of invocation field only on encrypt. 326238405Sjkim [Steve Henson] 327238405Sjkim 328238405Sjkim *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support. 329238405Sjkim As required by RFC5289 these ciphersuites cannot be used if for 330238405Sjkim versions of TLS earlier than 1.2. 331238405Sjkim [Steve Henson] 332238405Sjkim 333238405Sjkim *) For FIPS capable OpenSSL interpret a NULL default public key method 334238405Sjkim as unset and return the appopriate default but do *not* set the default. 335238405Sjkim This means we can return the appopriate method in applications that 336238405Sjkim swicth between FIPS and non-FIPS modes. 337238405Sjkim [Steve Henson] 338238405Sjkim 339238405Sjkim *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an 340238405Sjkim ENGINE is used then we cannot handle that in the FIPS module so we 341238405Sjkim keep original code iff non-FIPS operations are allowed. 342238405Sjkim [Steve Henson] 343238405Sjkim 344238405Sjkim *) Add -attime option to openssl utilities. 345238405Sjkim [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson] 346238405Sjkim 347238405Sjkim *) Redirect DSA and DH operations to FIPS module in FIPS mode. 348238405Sjkim [Steve Henson] 349238405Sjkim 350238405Sjkim *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use 351238405Sjkim FIPS EC methods unconditionally for now. 352238405Sjkim [Steve Henson] 353238405Sjkim 354238405Sjkim *) New build option no-ec2m to disable characteristic 2 code. 355238405Sjkim [Steve Henson] 356238405Sjkim 357238405Sjkim *) Backport libcrypto audit of return value checking from 1.1.0-dev; not 358238405Sjkim all cases can be covered as some introduce binary incompatibilities. 359238405Sjkim [Steve Henson] 360238405Sjkim 361238405Sjkim *) Redirect RSA operations to FIPS module including keygen, 362238405Sjkim encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods. 363238405Sjkim [Steve Henson] 364238405Sjkim 365238405Sjkim *) Add similar low level API blocking to ciphers. 366238405Sjkim [Steve Henson] 367238405Sjkim 368238405Sjkim *) Low level digest APIs are not approved in FIPS mode: any attempt 369238405Sjkim to use these will cause a fatal error. Applications that *really* want 370238405Sjkim to use them can use the private_* version instead. 371238405Sjkim [Steve Henson] 372238405Sjkim 373238405Sjkim *) Redirect cipher operations to FIPS module for FIPS builds. 374238405Sjkim [Steve Henson] 375238405Sjkim 376238405Sjkim *) Redirect digest operations to FIPS module for FIPS builds. 377238405Sjkim [Steve Henson] 378238405Sjkim 379238405Sjkim *) Update build system to add "fips" flag which will link in fipscanister.o 380238405Sjkim for static and shared library builds embedding a signature if needed. 381238405Sjkim [Steve Henson] 382238405Sjkim 383238405Sjkim *) Output TLS supported curves in preference order instead of numerical 384238405Sjkim order. This is currently hardcoded for the highest order curves first. 385238405Sjkim This should be configurable so applications can judge speed vs strength. 386238405Sjkim [Steve Henson] 387238405Sjkim 388238405Sjkim *) Add TLS v1.2 server support for client authentication. 389238405Sjkim [Steve Henson] 390238405Sjkim 391238405Sjkim *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers 392238405Sjkim and enable MD5. 393238405Sjkim [Steve Henson] 394238405Sjkim 395238405Sjkim *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying 396238405Sjkim FIPS modules versions. 397238405Sjkim [Steve Henson] 398238405Sjkim 399238405Sjkim *) Add TLS v1.2 client side support for client authentication. Keep cache 400238405Sjkim of handshake records longer as we don't know the hash algorithm to use 401238405Sjkim until after the certificate request message is received. 402238405Sjkim [Steve Henson] 403238405Sjkim 404238405Sjkim *) Initial TLS v1.2 client support. Add a default signature algorithms 405238405Sjkim extension including all the algorithms we support. Parse new signature 406238405Sjkim format in client key exchange. Relax some ECC signing restrictions for 407238405Sjkim TLS v1.2 as indicated in RFC5246. 408238405Sjkim [Steve Henson] 409238405Sjkim 410238405Sjkim *) Add server support for TLS v1.2 signature algorithms extension. Switch 411238405Sjkim to new signature format when needed using client digest preference. 412238405Sjkim All server ciphersuites should now work correctly in TLS v1.2. No client 413238405Sjkim support yet and no support for client certificates. 414238405Sjkim [Steve Henson] 415238405Sjkim 416238405Sjkim *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch 417238405Sjkim to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based 418238405Sjkim ciphersuites. At present only RSA key exchange ciphersuites work with 419238405Sjkim TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete 420238405Sjkim SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods 421238405Sjkim and version checking. 422238405Sjkim [Steve Henson] 423238405Sjkim 424238405Sjkim *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled 425238405Sjkim with this defined it will not be affected by any changes to ssl internal 426238405Sjkim structures. Add several utility functions to allow openssl application 427238405Sjkim to work with OPENSSL_NO_SSL_INTERN defined. 428238405Sjkim [Steve Henson] 429238405Sjkim 430238405Sjkim *) Add SRP support. 431238405Sjkim [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie] 432238405Sjkim 433238405Sjkim *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id. 434238405Sjkim [Steve Henson] 435238405Sjkim 436238405Sjkim *) Permit abbreviated handshakes when renegotiating using the function 437238405Sjkim SSL_renegotiate_abbreviated(). 438238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>] 439238405Sjkim 440238405Sjkim *) Add call to ENGINE_register_all_complete() to 441238405Sjkim ENGINE_load_builtin_engines(), so some implementations get used 442238405Sjkim automatically instead of needing explicit application support. 443238405Sjkim [Steve Henson] 444238405Sjkim 445238405Sjkim *) Add support for TLS key exporter as described in RFC5705. 446238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson] 447238405Sjkim 448238405Sjkim *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only 449238405Sjkim a few changes are required: 450238405Sjkim 451238405Sjkim Add SSL_OP_NO_TLSv1_1 flag. 452238405Sjkim Add TLSv1_1 methods. 453238405Sjkim Update version checking logic to handle version 1.1. 454238405Sjkim Add explicit IV handling (ported from DTLS code). 455238405Sjkim Add command line options to s_client/s_server. 456238405Sjkim [Steve Henson] 457238405Sjkim 458264331Sjkim Changes between 1.0.0j and 1.0.0k [5 Feb 2013] 459264331Sjkim 460264331Sjkim *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time. 461264331Sjkim 462264331Sjkim This addresses the flaw in CBC record processing discovered by 463264331Sjkim Nadhem Alfardan and Kenny Paterson. Details of this attack can be found 464264331Sjkim at: http://www.isg.rhul.ac.uk/tls/ 465264331Sjkim 466264331Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 467264331Sjkim Security Group at Royal Holloway, University of London 468264331Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and 469264331Sjkim Emilia K�sper for the initial patch. 470264331Sjkim (CVE-2013-0169) 471264331Sjkim [Emilia K�sper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson] 472264331Sjkim 473264331Sjkim *) Return an error when checking OCSP signatures when key is NULL. 474264331Sjkim This fixes a DoS attack. (CVE-2013-0166) 475264331Sjkim [Steve Henson] 476264331Sjkim 477264331Sjkim *) Call OCSP Stapling callback after ciphersuite has been chosen, so 478264331Sjkim the right response is stapled. Also change SSL_get_certificate() 479264331Sjkim so it returns the certificate actually sent. 480264331Sjkim See http://rt.openssl.org/Ticket/Display.html?id=2836. 481264331Sjkim (This is a backport) 482264331Sjkim [Rob Stradling <rob.stradling@comodo.com>] 483264331Sjkim 484264331Sjkim *) Fix possible deadlock when decoding public keys. 485264331Sjkim [Steve Henson] 486264331Sjkim 487264331Sjkim Changes between 1.0.0i and 1.0.0j [10 May 2012] 488264331Sjkim 489264331Sjkim [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after 490264331Sjkim OpenSSL 1.0.1.] 491264331Sjkim 492264331Sjkim *) Sanity check record length before skipping explicit IV in DTLS 493264331Sjkim to fix DoS attack. 494264331Sjkim 495264331Sjkim Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic 496264331Sjkim fuzzing as a service testing platform. 497264331Sjkim (CVE-2012-2333) 498264331Sjkim [Steve Henson] 499264331Sjkim 500264331Sjkim *) Initialise tkeylen properly when encrypting CMS messages. 501264331Sjkim Thanks to Solar Designer of Openwall for reporting this issue. 502264331Sjkim [Steve Henson] 503264331Sjkim 504264331Sjkim Changes between 1.0.0h and 1.0.0i [19 Apr 2012] 505264331Sjkim 506264331Sjkim *) Check for potentially exploitable overflows in asn1_d2i_read_bio 507264331Sjkim BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer 508264331Sjkim in CRYPTO_realloc_clean. 509264331Sjkim 510264331Sjkim Thanks to Tavis Ormandy, Google Security Team, for discovering this 511264331Sjkim issue and to Adam Langley <agl@chromium.org> for fixing it. 512264331Sjkim (CVE-2012-2110) 513264331Sjkim [Adam Langley (Google), Tavis Ormandy, Google Security Team] 514264331Sjkim 515238405Sjkim Changes between 1.0.0g and 1.0.0h [12 Mar 2012] 516238405Sjkim 517237657Sjkim *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness 518237657Sjkim in CMS and PKCS7 code. When RSA decryption fails use a random key for 519237657Sjkim content decryption and always return the same error. Note: this attack 520237657Sjkim needs on average 2^20 messages so it only affects automated senders. The 521237657Sjkim old behaviour can be reenabled in the CMS code by setting the 522237657Sjkim CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where 523237657Sjkim an MMA defence is not necessary. 524237657Sjkim Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering 525237657Sjkim this issue. (CVE-2012-0884) 526237657Sjkim [Steve Henson] 527237657Sjkim 528237657Sjkim *) Fix CVE-2011-4619: make sure we really are receiving a 529237657Sjkim client hello before rejecting multiple SGC restarts. Thanks to 530237657Sjkim Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug. 531237657Sjkim [Steve Henson] 532237657Sjkim 533238405Sjkim Changes between 1.0.0f and 1.0.0g [18 Jan 2012] 534238405Sjkim 535238405Sjkim *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109. 536238405Sjkim Thanks to Antonio Martin, Enterprise Secure Access Research and 537238405Sjkim Development, Cisco Systems, Inc. for discovering this bug and 538238405Sjkim preparing a fix. (CVE-2012-0050) 539238405Sjkim [Antonio Martin] 540238405Sjkim 541238405Sjkim Changes between 1.0.0e and 1.0.0f [4 Jan 2012] 542238405Sjkim 543238405Sjkim *) Nadhem Alfardan and Kenny Paterson have discovered an extension 544238405Sjkim of the Vaudenay padding oracle attack on CBC mode encryption 545238405Sjkim which enables an efficient plaintext recovery attack against 546238405Sjkim the OpenSSL implementation of DTLS. Their attack exploits timing 547238405Sjkim differences arising during decryption processing. A research 548238405Sjkim paper describing this attack can be found at: 549238405Sjkim http://www.isg.rhul.ac.uk/~kp/dtls.pdf 550238405Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 551238405Sjkim Security Group at Royal Holloway, University of London 552238405Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann 553238405Sjkim <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de> 554238405Sjkim for preparing the fix. (CVE-2011-4108) 555238405Sjkim [Robin Seggelmann, Michael Tuexen] 556238405Sjkim 557238405Sjkim *) Clear bytes used for block padding of SSL 3.0 records. 558238405Sjkim (CVE-2011-4576) 559238405Sjkim [Adam Langley (Google)] 560238405Sjkim 561238405Sjkim *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George 562238405Sjkim Kadianakis <desnacked@gmail.com> for discovering this issue and 563238405Sjkim Adam Langley for preparing the fix. (CVE-2011-4619) 564238405Sjkim [Adam Langley (Google)] 565238405Sjkim 566238405Sjkim *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027) 567238405Sjkim [Andrey Kulikov <amdeich@gmail.com>] 568238405Sjkim 569238405Sjkim *) Prevent malformed RFC3779 data triggering an assertion failure. 570238405Sjkim Thanks to Andrew Chi, BBN Technologies, for discovering the flaw 571238405Sjkim and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577) 572238405Sjkim [Rob Austein <sra@hactrn.net>] 573238405Sjkim 574238405Sjkim *) Improved PRNG seeding for VOS. 575238405Sjkim [Paul Green <Paul.Green@stratus.com>] 576238405Sjkim 577238405Sjkim *) Fix ssl_ciph.c set-up race. 578238405Sjkim [Adam Langley (Google)] 579238405Sjkim 580238405Sjkim *) Fix spurious failures in ecdsatest.c. 581238405Sjkim [Emilia K�sper (Google)] 582238405Sjkim 583238405Sjkim *) Fix the BIO_f_buffer() implementation (which was mixing different 584238405Sjkim interpretations of the '..._len' fields). 585238405Sjkim [Adam Langley (Google)] 586238405Sjkim 587238405Sjkim *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than 588238405Sjkim BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent 589238405Sjkim threads won't reuse the same blinding coefficients. 590238405Sjkim 591238405Sjkim This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING 592238405Sjkim lock to call BN_BLINDING_invert_ex, and avoids one use of 593238405Sjkim BN_BLINDING_update for each BN_BLINDING structure (previously, 594238405Sjkim the last update always remained unused). 595238405Sjkim [Emilia K�sper (Google)] 596238405Sjkim 597238405Sjkim *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf. 598238405Sjkim [Bob Buckholz (Google)] 599238405Sjkim 600238405Sjkim Changes between 1.0.0d and 1.0.0e [6 Sep 2011] 601238405Sjkim 602238405Sjkim *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted 603238405Sjkim by initialising X509_STORE_CTX properly. (CVE-2011-3207) 604238405Sjkim [Kaspar Brand <ossl@velox.ch>] 605238405Sjkim 606238405Sjkim *) Fix SSL memory handling for (EC)DH ciphersuites, in particular 607238405Sjkim for multi-threaded use of ECDH. (CVE-2011-3210) 608238405Sjkim [Adam Langley (Google)] 609238405Sjkim 610238405Sjkim *) Fix x509_name_ex_d2i memory leak on bad inputs. 611238405Sjkim [Bodo Moeller] 612238405Sjkim 613238405Sjkim *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check 614238405Sjkim signature public key algorithm by using OID xref utilities instead. 615238405Sjkim Before this you could only use some ECC ciphersuites with SHA1 only. 616238405Sjkim [Steve Henson] 617238405Sjkim 618238405Sjkim *) Add protection against ECDSA timing attacks as mentioned in the paper 619238405Sjkim by Billy Bob Brumley and Nicola Tuveri, see: 620238405Sjkim 621238405Sjkim http://eprint.iacr.org/2011/232.pdf 622238405Sjkim 623238405Sjkim [Billy Bob Brumley and Nicola Tuveri] 624238405Sjkim 625238405Sjkim Changes between 1.0.0c and 1.0.0d [8 Feb 2011] 626238405Sjkim 627238405Sjkim *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014 628238405Sjkim [Neel Mehta, Adam Langley, Bodo Moeller (Google)] 629238405Sjkim 630238405Sjkim *) Fix bug in string printing code: if *any* escaping is enabled we must 631238405Sjkim escape the escape character (backslash) or the resulting string is 632238405Sjkim ambiguous. 633238405Sjkim [Steve Henson] 634238405Sjkim 635238405Sjkim Changes between 1.0.0b and 1.0.0c [2 Dec 2010] 636238405Sjkim 637238405Sjkim *) Disable code workaround for ancient and obsolete Netscape browsers 638238405Sjkim and servers: an attacker can use it in a ciphersuite downgrade attack. 639238405Sjkim Thanks to Martin Rex for discovering this bug. CVE-2010-4180 640238405Sjkim [Steve Henson] 641238405Sjkim 642238405Sjkim *) Fixed J-PAKE implementation error, originally discovered by 643238405Sjkim Sebastien Martini, further info and confirmation from Stefan 644238405Sjkim Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252 645238405Sjkim [Ben Laurie] 646238405Sjkim 647238405Sjkim Changes between 1.0.0a and 1.0.0b [16 Nov 2010] 648238405Sjkim 649238405Sjkim *) Fix extension code to avoid race conditions which can result in a buffer 650238405Sjkim overrun vulnerability: resumed sessions must not be modified as they can 651238405Sjkim be shared by multiple threads. CVE-2010-3864 652238405Sjkim [Steve Henson] 653238405Sjkim 654238405Sjkim *) Fix WIN32 build system to correctly link an ENGINE directory into 655238405Sjkim a DLL. 656238405Sjkim [Steve Henson] 657238405Sjkim 658238405Sjkim Changes between 1.0.0 and 1.0.0a [01 Jun 2010] 659238405Sjkim 660238405Sjkim *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover 661238405Sjkim (CVE-2010-1633) 662238405Sjkim [Steve Henson, Peter-Michael Hager <hager@dortmund.net>] 663238405Sjkim 664238405Sjkim Changes between 0.9.8n and 1.0.0 [29 Mar 2010] 665238405Sjkim 666238405Sjkim *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher 667238405Sjkim context. The operation can be customised via the ctrl mechanism in 668238405Sjkim case ENGINEs want to include additional functionality. 669238405Sjkim [Steve Henson] 670238405Sjkim 671238405Sjkim *) Tolerate yet another broken PKCS#8 key format: private key value negative. 672238405Sjkim [Steve Henson] 673238405Sjkim 674238405Sjkim *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to 675238405Sjkim output hashes compatible with older versions of OpenSSL. 676238405Sjkim [Willy Weisz <weisz@vcpc.univie.ac.at>] 677238405Sjkim 678238405Sjkim *) Fix compression algorithm handling: if resuming a session use the 679238405Sjkim compression algorithm of the resumed session instead of determining 680238405Sjkim it from client hello again. Don't allow server to change algorithm. 681238405Sjkim [Steve Henson] 682238405Sjkim 683238405Sjkim *) Add load_crls() function to apps tidying load_certs() too. Add option 684238405Sjkim to verify utility to allow additional CRLs to be included. 685238405Sjkim [Steve Henson] 686238405Sjkim 687238405Sjkim *) Update OCSP request code to permit adding custom headers to the request: 688238405Sjkim some responders need this. 689238405Sjkim [Steve Henson] 690238405Sjkim 691238405Sjkim *) The function EVP_PKEY_sign() returns <=0 on error: check return code 692238405Sjkim correctly. 693238405Sjkim [Julia Lawall <julia@diku.dk>] 694238405Sjkim 695238405Sjkim *) Update verify callback code in apps/s_cb.c and apps/verify.c, it 696238405Sjkim needlessly dereferenced structures, used obsolete functions and 697238405Sjkim didn't handle all updated verify codes correctly. 698238405Sjkim [Steve Henson] 699238405Sjkim 700238405Sjkim *) Disable MD2 in the default configuration. 701238405Sjkim [Steve Henson] 702238405Sjkim 703238405Sjkim *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to 704238405Sjkim indicate the initial BIO being pushed or popped. This makes it possible 705238405Sjkim to determine whether the BIO is the one explicitly called or as a result 706238405Sjkim of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so 707238405Sjkim it handles reference counts correctly and doesn't zero out the I/O bio 708238405Sjkim when it is not being explicitly popped. WARNING: applications which 709238405Sjkim included workarounds for the old buggy behaviour will need to be modified 710238405Sjkim or they could free up already freed BIOs. 711238405Sjkim [Steve Henson] 712238405Sjkim 713238405Sjkim *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni 714238405Sjkim renaming to all platforms (within the 0.9.8 branch, this was 715238405Sjkim done conditionally on Netware platforms to avoid a name clash). 716238405Sjkim [Guenter <lists@gknw.net>] 717238405Sjkim 718238405Sjkim *) Add ECDHE and PSK support to DTLS. 719238405Sjkim [Michael Tuexen <tuexen@fh-muenster.de>] 720238405Sjkim 721238405Sjkim *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't 722238405Sjkim be used on C++. 723238405Sjkim [Steve Henson] 724238405Sjkim 725238405Sjkim *) Add "missing" function EVP_MD_flags() (without this the only way to 726238405Sjkim retrieve a digest flags is by accessing the structure directly. Update 727238405Sjkim EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest 728238405Sjkim or cipher is registered as in the "from" argument. Print out all 729238405Sjkim registered digests in the dgst usage message instead of manually 730238405Sjkim attempting to work them out. 731238405Sjkim [Steve Henson] 732238405Sjkim 733238405Sjkim *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello: 734238405Sjkim this allows the use of compression and extensions. Change default cipher 735238405Sjkim string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2 736238405Sjkim by default unless an application cipher string requests it. 737238405Sjkim [Steve Henson] 738238405Sjkim 739238405Sjkim *) Alter match criteria in PKCS12_parse(). It used to try to use local 740238405Sjkim key ids to find matching certificates and keys but some PKCS#12 files 741238405Sjkim don't follow the (somewhat unwritten) rules and this strategy fails. 742238405Sjkim Now just gather all certificates together and the first private key 743238405Sjkim then look for the first certificate that matches the key. 744238405Sjkim [Steve Henson] 745238405Sjkim 746238405Sjkim *) Support use of registered digest and cipher names for dgst and cipher 747238405Sjkim commands instead of having to add each one as a special case. So now 748238405Sjkim you can do: 749238405Sjkim 750238405Sjkim openssl sha256 foo 751238405Sjkim 752238405Sjkim as well as: 753238405Sjkim 754238405Sjkim openssl dgst -sha256 foo 755238405Sjkim 756238405Sjkim and this works for ENGINE based algorithms too. 757238405Sjkim 758238405Sjkim [Steve Henson] 759238405Sjkim 760238405Sjkim *) Update Gost ENGINE to support parameter files. 761238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>] 762238405Sjkim 763238405Sjkim *) Support GeneralizedTime in ca utility. 764238405Sjkim [Oliver Martin <oliver@volatilevoid.net>, Steve Henson] 765238405Sjkim 766238405Sjkim *) Enhance the hash format used for certificate directory links. The new 767238405Sjkim form uses the canonical encoding (meaning equivalent names will work 768238405Sjkim even if they aren't identical) and uses SHA1 instead of MD5. This form 769238405Sjkim is incompatible with the older format and as a result c_rehash should 770238405Sjkim be used to rebuild symbolic links. 771238405Sjkim [Steve Henson] 772238405Sjkim 773238405Sjkim *) Make PKCS#8 the default write format for private keys, replacing the 774238405Sjkim traditional format. This form is standardised, more secure and doesn't 775238405Sjkim include an implicit MD5 dependency. 776238405Sjkim [Steve Henson] 777238405Sjkim 778238405Sjkim *) Add a $gcc_devteam_warn option to Configure. The idea is that any code 779238405Sjkim committed to OpenSSL should pass this lot as a minimum. 780238405Sjkim [Steve Henson] 781238405Sjkim 782238405Sjkim *) Add session ticket override functionality for use by EAP-FAST. 783238405Sjkim [Jouni Malinen <j@w1.fi>] 784238405Sjkim 785238405Sjkim *) Modify HMAC functions to return a value. Since these can be implemented 786238405Sjkim in an ENGINE errors can occur. 787238405Sjkim [Steve Henson] 788238405Sjkim 789238405Sjkim *) Type-checked OBJ_bsearch_ex. 790238405Sjkim [Ben Laurie] 791238405Sjkim 792238405Sjkim *) Type-checked OBJ_bsearch. Also some constification necessitated 793238405Sjkim by type-checking. Still to come: TXT_DB, bsearch(?), 794238405Sjkim OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING, 795238405Sjkim CONF_VALUE. 796238405Sjkim [Ben Laurie] 797238405Sjkim 798238405Sjkim *) New function OPENSSL_gmtime_adj() to add a specific number of days and 799238405Sjkim seconds to a tm structure directly, instead of going through OS 800238405Sjkim specific date routines. This avoids any issues with OS routines such 801238405Sjkim as the year 2038 bug. New *_adj() functions for ASN1 time structures 802238405Sjkim and X509_time_adj_ex() to cover the extended range. The existing 803238405Sjkim X509_time_adj() is still usable and will no longer have any date issues. 804238405Sjkim [Steve Henson] 805238405Sjkim 806238405Sjkim *) Delta CRL support. New use deltas option which will attempt to locate 807238405Sjkim and search any appropriate delta CRLs available. 808238405Sjkim 809238405Sjkim This work was sponsored by Google. 810238405Sjkim [Steve Henson] 811238405Sjkim 812238405Sjkim *) Support for CRLs partitioned by reason code. Reorganise CRL processing 813238405Sjkim code and add additional score elements. Validate alternate CRL paths 814238405Sjkim as part of the CRL checking and indicate a new error "CRL path validation 815238405Sjkim error" in this case. Applications wanting additional details can use 816238405Sjkim the verify callback and check the new "parent" field. If this is not 817238405Sjkim NULL CRL path validation is taking place. Existing applications wont 818238405Sjkim see this because it requires extended CRL support which is off by 819238405Sjkim default. 820238405Sjkim 821238405Sjkim This work was sponsored by Google. 822238405Sjkim [Steve Henson] 823238405Sjkim 824238405Sjkim *) Support for freshest CRL extension. 825238405Sjkim 826238405Sjkim This work was sponsored by Google. 827238405Sjkim [Steve Henson] 828238405Sjkim 829238405Sjkim *) Initial indirect CRL support. Currently only supported in the CRLs 830238405Sjkim passed directly and not via lookup. Process certificate issuer 831238405Sjkim CRL entry extension and lookup CRL entries by bother issuer name 832238405Sjkim and serial number. Check and process CRL issuer entry in IDP extension. 833238405Sjkim 834238405Sjkim This work was sponsored by Google. 835238405Sjkim [Steve Henson] 836238405Sjkim 837238405Sjkim *) Add support for distinct certificate and CRL paths. The CRL issuer 838238405Sjkim certificate is validated separately in this case. Only enabled if 839238405Sjkim an extended CRL support flag is set: this flag will enable additional 840238405Sjkim CRL functionality in future. 841238405Sjkim 842238405Sjkim This work was sponsored by Google. 843238405Sjkim [Steve Henson] 844238405Sjkim 845238405Sjkim *) Add support for policy mappings extension. 846238405Sjkim 847238405Sjkim This work was sponsored by Google. 848238405Sjkim [Steve Henson] 849238405Sjkim 850238405Sjkim *) Fixes to pathlength constraint, self issued certificate handling, 851238405Sjkim policy processing to align with RFC3280 and PKITS tests. 852238405Sjkim 853238405Sjkim This work was sponsored by Google. 854238405Sjkim [Steve Henson] 855238405Sjkim 856238405Sjkim *) Support for name constraints certificate extension. DN, email, DNS 857238405Sjkim and URI types are currently supported. 858238405Sjkim 859238405Sjkim This work was sponsored by Google. 860238405Sjkim [Steve Henson] 861238405Sjkim 862238405Sjkim *) To cater for systems that provide a pointer-based thread ID rather 863238405Sjkim than numeric, deprecate the current numeric thread ID mechanism and 864238405Sjkim replace it with a structure and associated callback type. This 865238405Sjkim mechanism allows a numeric "hash" to be extracted from a thread ID in 866238405Sjkim either case, and on platforms where pointers are larger than 'long', 867238405Sjkim mixing is done to help ensure the numeric 'hash' is usable even if it 868238405Sjkim can't be guaranteed unique. The default mechanism is to use "&errno" 869238405Sjkim as a pointer-based thread ID to distinguish between threads. 870238405Sjkim 871238405Sjkim Applications that want to provide their own thread IDs should now use 872238405Sjkim CRYPTO_THREADID_set_callback() to register a callback that will call 873238405Sjkim either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer(). 874238405Sjkim 875238405Sjkim Note that ERR_remove_state() is now deprecated, because it is tied 876238405Sjkim to the assumption that thread IDs are numeric. ERR_remove_state(0) 877238405Sjkim to free the current thread's error state should be replaced by 878238405Sjkim ERR_remove_thread_state(NULL). 879238405Sjkim 880238405Sjkim (This new approach replaces the functions CRYPTO_set_idptr_callback(), 881238405Sjkim CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in 882238405Sjkim OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an 883238405Sjkim application was previously providing a numeric thread callback that 884238405Sjkim was inappropriate for distinguishing threads, then uniqueness might 885238405Sjkim have been obtained with &errno that happened immediately in the 886238405Sjkim intermediate development versions of OpenSSL; this is no longer the 887238405Sjkim case, the numeric thread callback will now override the automatic use 888238405Sjkim of &errno.) 889238405Sjkim [Geoff Thorpe, with help from Bodo Moeller] 890238405Sjkim 891238405Sjkim *) Initial support for different CRL issuing certificates. This covers a 892238405Sjkim simple case where the self issued certificates in the chain exist and 893238405Sjkim the real CRL issuer is higher in the existing chain. 894238405Sjkim 895238405Sjkim This work was sponsored by Google. 896238405Sjkim [Steve Henson] 897238405Sjkim 898238405Sjkim *) Removed effectively defunct crypto/store from the build. 899238405Sjkim [Ben Laurie] 900238405Sjkim 901238405Sjkim *) Revamp of STACK to provide stronger type-checking. Still to come: 902238405Sjkim TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE, 903238405Sjkim ASN1_STRING, CONF_VALUE. 904238405Sjkim [Ben Laurie] 905238405Sjkim 906238405Sjkim *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer 907238405Sjkim RAM on SSL connections. This option can save about 34k per idle SSL. 908238405Sjkim [Nick Mathewson] 909238405Sjkim 910238405Sjkim *) Revamp of LHASH to provide stronger type-checking. Still to come: 911238405Sjkim STACK, TXT_DB, bsearch, qsort. 912238405Sjkim [Ben Laurie] 913238405Sjkim 914238405Sjkim *) Initial support for Cryptographic Message Syntax (aka CMS) based 915238405Sjkim on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility, 916238405Sjkim support for data, signedData, compressedData, digestedData and 917238405Sjkim encryptedData, envelopedData types included. Scripts to check against 918238405Sjkim RFC4134 examples draft and interop and consistency checks of many 919238405Sjkim content types and variants. 920238405Sjkim [Steve Henson] 921238405Sjkim 922238405Sjkim *) Add options to enc utility to support use of zlib compression BIO. 923238405Sjkim [Steve Henson] 924238405Sjkim 925238405Sjkim *) Extend mk1mf to support importing of options and assembly language 926238405Sjkim files from Configure script, currently only included in VC-WIN32. 927238405Sjkim The assembly language rules can now optionally generate the source 928238405Sjkim files from the associated perl scripts. 929238405Sjkim [Steve Henson] 930238405Sjkim 931238405Sjkim *) Implement remaining functionality needed to support GOST ciphersuites. 932238405Sjkim Interop testing has been performed using CryptoPro implementations. 933238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>] 934238405Sjkim 935238405Sjkim *) s390x assembler pack. 936238405Sjkim [Andy Polyakov] 937238405Sjkim 938238405Sjkim *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU 939238405Sjkim "family." 940238405Sjkim [Andy Polyakov] 941238405Sjkim 942238405Sjkim *) Implement Opaque PRF Input TLS extension as specified in 943238405Sjkim draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an 944238405Sjkim official specification yet and no extension type assignment by 945238405Sjkim IANA exists, this extension (for now) will have to be explicitly 946238405Sjkim enabled when building OpenSSL by providing the extension number 947238405Sjkim to use. For example, specify an option 948238405Sjkim 949238405Sjkim -DTLSEXT_TYPE_opaque_prf_input=0x9527 950238405Sjkim 951238405Sjkim to the "config" or "Configure" script to enable the extension, 952238405Sjkim assuming extension number 0x9527 (which is a completely arbitrary 953238405Sjkim and unofficial assignment based on the MD5 hash of the Internet 954238405Sjkim Draft). Note that by doing so, you potentially lose 955238405Sjkim interoperability with other TLS implementations since these might 956238405Sjkim be using the same extension number for other purposes. 957238405Sjkim 958238405Sjkim SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the 959238405Sjkim opaque PRF input value to use in the handshake. This will create 960238405Sjkim an interal copy of the length-'len' string at 'src', and will 961238405Sjkim return non-zero for success. 962238405Sjkim 963238405Sjkim To get more control and flexibility, provide a callback function 964238405Sjkim by using 965238405Sjkim 966238405Sjkim SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb) 967238405Sjkim SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg) 968238405Sjkim 969238405Sjkim where 970238405Sjkim 971238405Sjkim int (*cb)(SSL *, void *peerinput, size_t len, void *arg); 972238405Sjkim void *arg; 973238405Sjkim 974238405Sjkim Callback function 'cb' will be called in handshakes, and is 975238405Sjkim expected to use SSL_set_tlsext_opaque_prf_input() as appropriate. 976238405Sjkim Argument 'arg' is for application purposes (the value as given to 977238405Sjkim SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly 978238405Sjkim be provided to the callback function). The callback function 979238405Sjkim has to return non-zero to report success: usually 1 to use opaque 980238405Sjkim PRF input just if possible, or 2 to enforce use of the opaque PRF 981238405Sjkim input. In the latter case, the library will abort the handshake 982238405Sjkim if opaque PRF input is not successfully negotiated. 983238405Sjkim 984238405Sjkim Arguments 'peerinput' and 'len' given to the callback function 985238405Sjkim will always be NULL and 0 in the case of a client. A server will 986238405Sjkim see the client's opaque PRF input through these variables if 987238405Sjkim available (NULL and 0 otherwise). Note that if the server 988238405Sjkim provides an opaque PRF input, the length must be the same as the 989238405Sjkim length of the client's opaque PRF input. 990238405Sjkim 991238405Sjkim Note that the callback function will only be called when creating 992238405Sjkim a new session (session resumption can resume whatever was 993238405Sjkim previously negotiated), and will not be called in SSL 2.0 994238405Sjkim handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or 995238405Sjkim SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended 996238405Sjkim for applications that need to enforce opaque PRF input. 997238405Sjkim 998238405Sjkim [Bodo Moeller] 999238405Sjkim 1000238405Sjkim *) Update ssl code to support digests other than SHA1+MD5 for handshake 1001238405Sjkim MAC. 1002238405Sjkim 1003238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>] 1004238405Sjkim 1005238405Sjkim *) Add RFC4507 support to OpenSSL. This includes the corrections in 1006238405Sjkim RFC4507bis. The encrypted ticket format is an encrypted encoded 1007238405Sjkim SSL_SESSION structure, that way new session features are automatically 1008238405Sjkim supported. 1009238405Sjkim 1010238405Sjkim If a client application caches session in an SSL_SESSION structure 1011238405Sjkim support is transparent because tickets are now stored in the encoded 1012238405Sjkim SSL_SESSION. 1013238405Sjkim 1014238405Sjkim The SSL_CTX structure automatically generates keys for ticket 1015238405Sjkim protection in servers so again support should be possible 1016238405Sjkim with no application modification. 1017238405Sjkim 1018238405Sjkim If a client or server wishes to disable RFC4507 support then the option 1019238405Sjkim SSL_OP_NO_TICKET can be set. 1020238405Sjkim 1021238405Sjkim Add a TLS extension debugging callback to allow the contents of any client 1022238405Sjkim or server extensions to be examined. 1023238405Sjkim 1024238405Sjkim This work was sponsored by Google. 1025238405Sjkim [Steve Henson] 1026238405Sjkim 1027238405Sjkim *) Final changes to avoid use of pointer pointer casts in OpenSSL. 1028238405Sjkim OpenSSL should now compile cleanly on gcc 4.2 1029238405Sjkim [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson] 1030238405Sjkim 1031238405Sjkim *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC 1032238405Sjkim support including streaming MAC support: this is required for GOST 1033238405Sjkim ciphersuite support. 1034238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson] 1035238405Sjkim 1036238405Sjkim *) Add option -stream to use PKCS#7 streaming in smime utility. New 1037238405Sjkim function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream() 1038238405Sjkim to output in BER and PEM format. 1039238405Sjkim [Steve Henson] 1040238405Sjkim 1041238405Sjkim *) Experimental support for use of HMAC via EVP_PKEY interface. This 1042238405Sjkim allows HMAC to be handled via the EVP_DigestSign*() interface. The 1043238405Sjkim EVP_PKEY "key" in this case is the HMAC key, potentially allowing 1044238405Sjkim ENGINE support for HMAC keys which are unextractable. New -mac and 1045238405Sjkim -macopt options to dgst utility. 1046238405Sjkim [Steve Henson] 1047238405Sjkim 1048238405Sjkim *) New option -sigopt to dgst utility. Update dgst to use 1049238405Sjkim EVP_Digest{Sign,Verify}*. These two changes make it possible to use 1050238405Sjkim alternative signing paramaters such as X9.31 or PSS in the dgst 1051238405Sjkim utility. 1052238405Sjkim [Steve Henson] 1053238405Sjkim 1054238405Sjkim *) Change ssl_cipher_apply_rule(), the internal function that does 1055238405Sjkim the work each time a ciphersuite string requests enabling 1056238405Sjkim ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or 1057238405Sjkim removing ("!foo+bar") a class of ciphersuites: Now it maintains 1058238405Sjkim the order of disabled ciphersuites such that those ciphersuites 1059238405Sjkim that most recently went from enabled to disabled not only stay 1060238405Sjkim in order with respect to each other, but also have higher priority 1061238405Sjkim than other disabled ciphersuites the next time ciphersuites are 1062238405Sjkim enabled again. 1063238405Sjkim 1064238405Sjkim This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable 1065238405Sjkim the same ciphersuites as with "HIGH" alone, but in a specific 1066238405Sjkim order where the PSK ciphersuites come first (since they are the 1067238405Sjkim most recently disabled ciphersuites when "HIGH" is parsed). 1068238405Sjkim 1069238405Sjkim Also, change ssl_create_cipher_list() (using this new 1070238405Sjkim funcionality) such that between otherwise identical 1071238405Sjkim cihpersuites, ephemeral ECDH is preferred over ephemeral DH in 1072238405Sjkim the default order. 1073238405Sjkim [Bodo Moeller] 1074238405Sjkim 1075238405Sjkim *) Change ssl_create_cipher_list() so that it automatically 1076238405Sjkim arranges the ciphersuites in reasonable order before starting 1077238405Sjkim to process the rule string. Thus, the definition for "DEFAULT" 1078238405Sjkim (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but 1079238405Sjkim remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH". 1080238405Sjkim This makes it much easier to arrive at a reasonable default order 1081238405Sjkim in applications for which anonymous ciphers are OK (meaning 1082238405Sjkim that you can't actually use DEFAULT). 1083238405Sjkim [Bodo Moeller; suggested by Victor Duchovni] 1084238405Sjkim 1085238405Sjkim *) Split the SSL/TLS algorithm mask (as used for ciphersuite string 1086238405Sjkim processing) into multiple integers instead of setting 1087238405Sjkim "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK", 1088238405Sjkim "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer. 1089238405Sjkim (These masks as well as the individual bit definitions are hidden 1090238405Sjkim away into the non-exported interface ssl/ssl_locl.h, so this 1091238405Sjkim change to the definition of the SSL_CIPHER structure shouldn't 1092238405Sjkim affect applications.) This give us more bits for each of these 1093238405Sjkim categories, so there is no longer a need to coagulate AES128 and 1094238405Sjkim AES256 into a single algorithm bit, and to coagulate Camellia128 1095238405Sjkim and Camellia256 into a single algorithm bit, which has led to all 1096238405Sjkim kinds of kludges. 1097238405Sjkim 1098238405Sjkim Thus, among other things, the kludge introduced in 0.9.7m and 1099238405Sjkim 0.9.8e for masking out AES256 independently of AES128 or masking 1100238405Sjkim out Camellia256 independently of AES256 is not needed here in 0.9.9. 1101238405Sjkim 1102238405Sjkim With the change, we also introduce new ciphersuite aliases that 1103238405Sjkim so far were missing: "AES128", "AES256", "CAMELLIA128", and 1104238405Sjkim "CAMELLIA256". 1105238405Sjkim [Bodo Moeller] 1106238405Sjkim 1107238405Sjkim *) Add support for dsa-with-SHA224 and dsa-with-SHA256. 1108238405Sjkim Use the leftmost N bytes of the signature input if the input is 1109238405Sjkim larger than the prime q (with N being the size in bytes of q). 1110238405Sjkim [Nils Larsch] 1111238405Sjkim 1112238405Sjkim *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses 1113238405Sjkim it yet and it is largely untested. 1114238405Sjkim [Steve Henson] 1115238405Sjkim 1116238405Sjkim *) Add support for the ecdsa-with-SHA224/256/384/512 signature types. 1117238405Sjkim [Nils Larsch] 1118238405Sjkim 1119238405Sjkim *) Initial incomplete changes to avoid need for function casts in OpenSSL 1120238405Sjkim some compilers (gcc 4.2 and later) reject their use. Safestack is 1121238405Sjkim reimplemented. Update ASN1 to avoid use of legacy functions. 1122238405Sjkim [Steve Henson] 1123238405Sjkim 1124238405Sjkim *) Win32/64 targets are linked with Winsock2. 1125238405Sjkim [Andy Polyakov] 1126238405Sjkim 1127238405Sjkim *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected 1128238405Sjkim to external functions. This can be used to increase CRL handling 1129238405Sjkim efficiency especially when CRLs are very large by (for example) storing 1130238405Sjkim the CRL revoked certificates in a database. 1131238405Sjkim [Steve Henson] 1132238405Sjkim 1133238405Sjkim *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so 1134238405Sjkim new CRLs added to a directory can be used. New command line option 1135238405Sjkim -verify_return_error to s_client and s_server. This causes real errors 1136238405Sjkim to be returned by the verify callback instead of carrying on no matter 1137238405Sjkim what. This reflects the way a "real world" verify callback would behave. 1138238405Sjkim [Steve Henson] 1139238405Sjkim 1140238405Sjkim *) GOST engine, supporting several GOST algorithms and public key formats. 1141238405Sjkim Kindly donated by Cryptocom. 1142238405Sjkim [Cryptocom] 1143238405Sjkim 1144238405Sjkim *) Partial support for Issuing Distribution Point CRL extension. CRLs 1145238405Sjkim partitioned by DP are handled but no indirect CRL or reason partitioning 1146238405Sjkim (yet). Complete overhaul of CRL handling: now the most suitable CRL is 1147238405Sjkim selected via a scoring technique which handles IDP and AKID in CRLs. 1148238405Sjkim [Steve Henson] 1149238405Sjkim 1150238405Sjkim *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which 1151238405Sjkim will ultimately be used for all verify operations: this will remove the 1152238405Sjkim X509_STORE dependency on certificate verification and allow alternative 1153238405Sjkim lookup methods. X509_STORE based implementations of these two callbacks. 1154238405Sjkim [Steve Henson] 1155238405Sjkim 1156238405Sjkim *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names. 1157238405Sjkim Modify get_crl() to find a valid (unexpired) CRL if possible. 1158238405Sjkim [Steve Henson] 1159238405Sjkim 1160238405Sjkim *) New function X509_CRL_match() to check if two CRLs are identical. Normally 1161238405Sjkim this would be called X509_CRL_cmp() but that name is already used by 1162238405Sjkim a function that just compares CRL issuer names. Cache several CRL 1163238405Sjkim extensions in X509_CRL structure and cache CRLDP in X509. 1164238405Sjkim [Steve Henson] 1165238405Sjkim 1166238405Sjkim *) Store a "canonical" representation of X509_NAME structure (ASN1 Name) 1167238405Sjkim this maps equivalent X509_NAME structures into a consistent structure. 1168238405Sjkim Name comparison can then be performed rapidly using memcmp(). 1169238405Sjkim [Steve Henson] 1170238405Sjkim 1171238405Sjkim *) Non-blocking OCSP request processing. Add -timeout option to ocsp 1172238405Sjkim utility. 1173238405Sjkim [Steve Henson] 1174238405Sjkim 1175238405Sjkim *) Allow digests to supply their own micalg string for S/MIME type using 1176238405Sjkim the ctrl EVP_MD_CTRL_MICALG. 1177238405Sjkim [Steve Henson] 1178238405Sjkim 1179238405Sjkim *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the 1180238405Sjkim EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN 1181238405Sjkim ctrl. It can then customise the structure before and/or after signing 1182238405Sjkim if necessary. 1183238405Sjkim [Steve Henson] 1184238405Sjkim 1185238405Sjkim *) New function OBJ_add_sigid() to allow application defined signature OIDs 1186238405Sjkim to be added to OpenSSLs internal tables. New function OBJ_sigid_free() 1187238405Sjkim to free up any added signature OIDs. 1188238405Sjkim [Steve Henson] 1189238405Sjkim 1190238405Sjkim *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(), 1191238405Sjkim EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal 1192238405Sjkim digest and cipher tables. New options added to openssl utility: 1193238405Sjkim list-message-digest-algorithms and list-cipher-algorithms. 1194238405Sjkim [Steve Henson] 1195238405Sjkim 1196238405Sjkim *) Change the array representation of binary polynomials: the list 1197238405Sjkim of degrees of non-zero coefficients is now terminated with -1. 1198238405Sjkim Previously it was terminated with 0, which was also part of the 1199238405Sjkim value; thus, the array representation was not applicable to 1200238405Sjkim polynomials where t^0 has coefficient zero. This change makes 1201238405Sjkim the array representation useful in a more general context. 1202238405Sjkim [Douglas Stebila] 1203238405Sjkim 1204238405Sjkim *) Various modifications and fixes to SSL/TLS cipher string 1205238405Sjkim handling. For ECC, the code now distinguishes between fixed ECDH 1206238405Sjkim with RSA certificates on the one hand and with ECDSA certificates 1207238405Sjkim on the other hand, since these are separate ciphersuites. The 1208238405Sjkim unused code for Fortezza ciphersuites has been removed. 1209238405Sjkim 1210238405Sjkim For consistency with EDH, ephemeral ECDH is now called "EECDH" 1211238405Sjkim (not "ECDHE"). For consistency with the code for DH 1212238405Sjkim certificates, use of ECDH certificates is now considered ECDH 1213238405Sjkim authentication, not RSA or ECDSA authentication (the latter is 1214238405Sjkim merely the CA's signing algorithm and not actively used in the 1215238405Sjkim protocol). 1216238405Sjkim 1217238405Sjkim The temporary ciphersuite alias "ECCdraft" is no longer 1218238405Sjkim available, and ECC ciphersuites are no longer excluded from "ALL" 1219238405Sjkim and "DEFAULT". The following aliases now exist for RFC 4492 1220238405Sjkim ciphersuites, most of these by analogy with the DH case: 1221238405Sjkim 1222238405Sjkim kECDHr - ECDH cert, signed with RSA 1223238405Sjkim kECDHe - ECDH cert, signed with ECDSA 1224238405Sjkim kECDH - ECDH cert (signed with either RSA or ECDSA) 1225238405Sjkim kEECDH - ephemeral ECDH 1226238405Sjkim ECDH - ECDH cert or ephemeral ECDH 1227238405Sjkim 1228238405Sjkim aECDH - ECDH cert 1229238405Sjkim aECDSA - ECDSA cert 1230238405Sjkim ECDSA - ECDSA cert 1231238405Sjkim 1232238405Sjkim AECDH - anonymous ECDH 1233238405Sjkim EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH") 1234238405Sjkim 1235238405Sjkim [Bodo Moeller] 1236238405Sjkim 1237238405Sjkim *) Add additional S/MIME capabilities for AES and GOST ciphers if supported. 1238238405Sjkim Use correct micalg parameters depending on digest(s) in signed message. 1239238405Sjkim [Steve Henson] 1240238405Sjkim 1241238405Sjkim *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process 1242238405Sjkim an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code. 1243238405Sjkim [Steve Henson] 1244238405Sjkim 1245238405Sjkim *) Initial engine support for EVP_PKEY_METHOD. New functions to permit 1246238405Sjkim an engine to register a method. Add ENGINE lookups for methods and 1247238405Sjkim functional reference processing. 1248238405Sjkim [Steve Henson] 1249238405Sjkim 1250238405Sjkim *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of 1251238405Sjkim EVP_{Sign,Verify}* which allow an application to customise the signature 1252238405Sjkim process. 1253238405Sjkim [Steve Henson] 1254238405Sjkim 1255238405Sjkim *) New -resign option to smime utility. This adds one or more signers 1256238405Sjkim to an existing PKCS#7 signedData structure. Also -md option to use an 1257238405Sjkim alternative message digest algorithm for signing. 1258238405Sjkim [Steve Henson] 1259238405Sjkim 1260238405Sjkim *) Tidy up PKCS#7 routines and add new functions to make it easier to 1261238405Sjkim create PKCS7 structures containing multiple signers. Update smime 1262238405Sjkim application to support multiple signers. 1263238405Sjkim [Steve Henson] 1264238405Sjkim 1265238405Sjkim *) New -macalg option to pkcs12 utility to allow setting of an alternative 1266238405Sjkim digest MAC. 1267238405Sjkim [Steve Henson] 1268238405Sjkim 1269238405Sjkim *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC. 1270238405Sjkim Reorganize PBE internals to lookup from a static table using NIDs, 1271238405Sjkim add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl: 1272238405Sjkim EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative 1273238405Sjkim PRF which will be automatically used with PBES2. 1274238405Sjkim [Steve Henson] 1275238405Sjkim 1276238405Sjkim *) Replace the algorithm specific calls to generate keys in "req" with the 1277238405Sjkim new API. 1278238405Sjkim [Steve Henson] 1279238405Sjkim 1280238405Sjkim *) Update PKCS#7 enveloped data routines to use new API. This is now 1281238405Sjkim supported by any public key method supporting the encrypt operation. A 1282238405Sjkim ctrl is added to allow the public key algorithm to examine or modify 1283238405Sjkim the PKCS#7 RecipientInfo structure if it needs to: for RSA this is 1284238405Sjkim a no op. 1285238405Sjkim [Steve Henson] 1286238405Sjkim 1287238405Sjkim *) Add a ctrl to asn1 method to allow a public key algorithm to express 1288238405Sjkim a default digest type to use. In most cases this will be SHA1 but some 1289238405Sjkim algorithms (such as GOST) need to specify an alternative digest. The 1290238405Sjkim return value indicates how strong the prefernce is 1 means optional and 1291238405Sjkim 2 is mandatory (that is it is the only supported type). Modify 1292238405Sjkim ASN1_item_sign() to accept a NULL digest argument to indicate it should 1293238405Sjkim use the default md. Update openssl utilities to use the default digest 1294238405Sjkim type for signing if it is not explicitly indicated. 1295238405Sjkim [Steve Henson] 1296238405Sjkim 1297238405Sjkim *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New 1298238405Sjkim EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant 1299238405Sjkim signing method from the key type. This effectively removes the link 1300238405Sjkim between digests and public key types. 1301238405Sjkim [Steve Henson] 1302238405Sjkim 1303238405Sjkim *) Add an OID cross reference table and utility functions. Its purpose is to 1304238405Sjkim translate between signature OIDs such as SHA1WithrsaEncryption and SHA1, 1305238405Sjkim rsaEncryption. This will allow some of the algorithm specific hackery 1306238405Sjkim needed to use the correct OID to be removed. 1307238405Sjkim [Steve Henson] 1308238405Sjkim 1309238405Sjkim *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO 1310238405Sjkim structures for PKCS7_sign(). They are now set up by the relevant public 1311238405Sjkim key ASN1 method. 1312238405Sjkim [Steve Henson] 1313238405Sjkim 1314238405Sjkim *) Add provisional EC pkey method with support for ECDSA and ECDH. 1315238405Sjkim [Steve Henson] 1316238405Sjkim 1317238405Sjkim *) Add support for key derivation (agreement) in the API, DH method and 1318238405Sjkim pkeyutl. 1319238405Sjkim [Steve Henson] 1320238405Sjkim 1321238405Sjkim *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support 1322238405Sjkim public and private key formats. As a side effect these add additional 1323238405Sjkim command line functionality not previously available: DSA signatures can be 1324238405Sjkim generated and verified using pkeyutl and DH key support and generation in 1325238405Sjkim pkey, genpkey. 1326238405Sjkim [Steve Henson] 1327238405Sjkim 1328238405Sjkim *) BeOS support. 1329238405Sjkim [Oliver Tappe <zooey@hirschkaefer.de>] 1330238405Sjkim 1331238405Sjkim *) New make target "install_html_docs" installs HTML renditions of the 1332238405Sjkim manual pages. 1333238405Sjkim [Oliver Tappe <zooey@hirschkaefer.de>] 1334238405Sjkim 1335238405Sjkim *) New utility "genpkey" this is analagous to "genrsa" etc except it can 1336238405Sjkim generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to 1337238405Sjkim support key and parameter generation and add initial key generation 1338238405Sjkim functionality for RSA. 1339238405Sjkim [Steve Henson] 1340238405Sjkim 1341238405Sjkim *) Add functions for main EVP_PKEY_method operations. The undocumented 1342238405Sjkim functions EVP_PKEY_{encrypt,decrypt} have been renamed to 1343238405Sjkim EVP_PKEY_{encrypt,decrypt}_old. 1344238405Sjkim [Steve Henson] 1345238405Sjkim 1346238405Sjkim *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public 1347238405Sjkim key API, doesn't do much yet. 1348238405Sjkim [Steve Henson] 1349238405Sjkim 1350238405Sjkim *) New function EVP_PKEY_asn1_get0_info() to retrieve information about 1351238405Sjkim public key algorithms. New option to openssl utility: 1352238405Sjkim "list-public-key-algorithms" to print out info. 1353238405Sjkim [Steve Henson] 1354238405Sjkim 1355238405Sjkim *) Implement the Supported Elliptic Curves Extension for 1356238405Sjkim ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 1357238405Sjkim [Douglas Stebila] 1358238405Sjkim 1359238405Sjkim *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or 1360238405Sjkim EVP_CIPHER structures to avoid later problems in EVP_cleanup(). 1361238405Sjkim [Steve Henson] 1362238405Sjkim 1363238405Sjkim *) New utilities pkey and pkeyparam. These are similar to algorithm specific 1364238405Sjkim utilities such as rsa, dsa, dsaparam etc except they process any key 1365238405Sjkim type. 1366238405Sjkim [Steve Henson] 1367238405Sjkim 1368238405Sjkim *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New 1369238405Sjkim functions EVP_PKEY_print_public(), EVP_PKEY_print_private(), 1370238405Sjkim EVP_PKEY_print_param() to print public key data from an EVP_PKEY 1371238405Sjkim structure. 1372238405Sjkim [Steve Henson] 1373238405Sjkim 1374238405Sjkim *) Initial support for pluggable public key ASN1. 1375238405Sjkim De-spaghettify the public key ASN1 handling. Move public and private 1376238405Sjkim key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate 1377238405Sjkim algorithm specific handling to a single module within the relevant 1378238405Sjkim algorithm directory. Add functions to allow (near) opaque processing 1379238405Sjkim of public and private key structures. 1380238405Sjkim [Steve Henson] 1381238405Sjkim 1382238405Sjkim *) Implement the Supported Point Formats Extension for 1383238405Sjkim ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 1384238405Sjkim [Douglas Stebila] 1385238405Sjkim 1386238405Sjkim *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members 1387238405Sjkim for the psk identity [hint] and the psk callback functions to the 1388238405Sjkim SSL_SESSION, SSL and SSL_CTX structure. 1389238405Sjkim 1390238405Sjkim New ciphersuites: 1391238405Sjkim PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA, 1392238405Sjkim PSK-AES256-CBC-SHA 1393238405Sjkim 1394238405Sjkim New functions: 1395238405Sjkim SSL_CTX_use_psk_identity_hint 1396238405Sjkim SSL_get_psk_identity_hint 1397238405Sjkim SSL_get_psk_identity 1398238405Sjkim SSL_use_psk_identity_hint 1399238405Sjkim 1400238405Sjkim [Mika Kousa and Pasi Eronen of Nokia Corporation] 1401238405Sjkim 1402238405Sjkim *) Add RFC 3161 compliant time stamp request creation, response generation 1403238405Sjkim and response verification functionality. 1404238405Sjkim [Zolt�n Gl�zik <zglozik@opentsa.org>, The OpenTSA Project] 1405238405Sjkim 1406238405Sjkim *) Add initial support for TLS extensions, specifically for the server_name 1407238405Sjkim extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 1408238405Sjkim have new members for a host name. The SSL data structure has an 1409238405Sjkim additional member SSL_CTX *initial_ctx so that new sessions can be 1410238405Sjkim stored in that context to allow for session resumption, even after the 1411238405Sjkim SSL has been switched to a new SSL_CTX in reaction to a client's 1412238405Sjkim server_name extension. 1413238405Sjkim 1414238405Sjkim New functions (subject to change): 1415238405Sjkim 1416238405Sjkim SSL_get_servername() 1417238405Sjkim SSL_get_servername_type() 1418238405Sjkim SSL_set_SSL_CTX() 1419238405Sjkim 1420238405Sjkim New CTRL codes and macros (subject to change): 1421238405Sjkim 1422238405Sjkim SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 1423238405Sjkim - SSL_CTX_set_tlsext_servername_callback() 1424238405Sjkim SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 1425238405Sjkim - SSL_CTX_set_tlsext_servername_arg() 1426238405Sjkim SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 1427238405Sjkim 1428238405Sjkim openssl s_client has a new '-servername ...' option. 1429238405Sjkim 1430238405Sjkim openssl s_server has new options '-servername_host ...', '-cert2 ...', 1431238405Sjkim '-key2 ...', '-servername_fatal' (subject to change). This allows 1432238405Sjkim testing the HostName extension for a specific single host name ('-cert' 1433238405Sjkim and '-key' remain fallbacks for handshakes without HostName 1434238405Sjkim negotiation). If the unrecogninzed_name alert has to be sent, this by 1435238405Sjkim default is a warning; it becomes fatal with the '-servername_fatal' 1436238405Sjkim option. 1437238405Sjkim 1438238405Sjkim [Peter Sylvester, Remy Allais, Christophe Renou] 1439238405Sjkim 1440238405Sjkim *) Whirlpool hash implementation is added. 1441238405Sjkim [Andy Polyakov] 1442238405Sjkim 1443238405Sjkim *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to 1444238405Sjkim bn(64,32). Because of instruction set limitations it doesn't have 1445238405Sjkim any negative impact on performance. This was done mostly in order 1446238405Sjkim to make it possible to share assembler modules, such as bn_mul_mont 1447238405Sjkim implementations, between 32- and 64-bit builds without hassle. 1448238405Sjkim [Andy Polyakov] 1449238405Sjkim 1450238405Sjkim *) Move code previously exiled into file crypto/ec/ec2_smpt.c 1451238405Sjkim to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP 1452238405Sjkim macro. 1453238405Sjkim [Bodo Moeller] 1454238405Sjkim 1455238405Sjkim *) New candidate for BIGNUM assembler implementation, bn_mul_mont, 1456238405Sjkim dedicated Montgomery multiplication procedure, is introduced. 1457238405Sjkim BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher 1458238405Sjkim "64-bit" performance on certain 32-bit targets. 1459238405Sjkim [Andy Polyakov] 1460238405Sjkim 1461238405Sjkim *) New option SSL_OP_NO_COMP to disable use of compression selectively 1462238405Sjkim in SSL structures. New SSL ctrl to set maximum send fragment size. 1463238405Sjkim Save memory by seeting the I/O buffer sizes dynamically instead of 1464238405Sjkim using the maximum available value. 1465238405Sjkim [Steve Henson] 1466238405Sjkim 1467238405Sjkim *) New option -V for 'openssl ciphers'. This prints the ciphersuite code 1468238405Sjkim in addition to the text details. 1469238405Sjkim [Bodo Moeller] 1470238405Sjkim 1471238405Sjkim *) Very, very preliminary EXPERIMENTAL support for printing of general 1472238405Sjkim ASN1 structures. This currently produces rather ugly output and doesn't 1473238405Sjkim handle several customised structures at all. 1474238405Sjkim [Steve Henson] 1475238405Sjkim 1476238405Sjkim *) Integrated support for PVK file format and some related formats such 1477238405Sjkim as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support 1478238405Sjkim these in the 'rsa' and 'dsa' utilities. 1479238405Sjkim [Steve Henson] 1480238405Sjkim 1481238405Sjkim *) Support for PKCS#1 RSAPublicKey format on rsa utility command line. 1482238405Sjkim [Steve Henson] 1483238405Sjkim 1484238405Sjkim *) Remove the ancient ASN1_METHOD code. This was only ever used in one 1485238405Sjkim place for the (very old) "NETSCAPE" format certificates which are now 1486238405Sjkim handled using new ASN1 code equivalents. 1487238405Sjkim [Steve Henson] 1488238405Sjkim 1489238405Sjkim *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD 1490238405Sjkim pointer and make the SSL_METHOD parameter in SSL_CTX_new, 1491238405Sjkim SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'. 1492238405Sjkim [Nils Larsch] 1493238405Sjkim 1494238405Sjkim *) Modify CRL distribution points extension code to print out previously 1495238405Sjkim unsupported fields. Enhance extension setting code to allow setting of 1496238405Sjkim all fields. 1497238405Sjkim [Steve Henson] 1498238405Sjkim 1499238405Sjkim *) Add print and set support for Issuing Distribution Point CRL extension. 1500238405Sjkim [Steve Henson] 1501238405Sjkim 1502238405Sjkim *) Change 'Configure' script to enable Camellia by default. 1503238405Sjkim [NTT] 1504238405Sjkim 1505264331Sjkim Changes between 0.9.8x and 0.9.8y [5 Feb 2013] 1506264331Sjkim 1507264331Sjkim *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time. 1508264331Sjkim 1509264331Sjkim This addresses the flaw in CBC record processing discovered by 1510264331Sjkim Nadhem Alfardan and Kenny Paterson. Details of this attack can be found 1511264331Sjkim at: http://www.isg.rhul.ac.uk/tls/ 1512264331Sjkim 1513264331Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 1514264331Sjkim Security Group at Royal Holloway, University of London 1515264331Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and 1516264331Sjkim Emilia K�sper for the initial patch. 1517264331Sjkim (CVE-2013-0169) 1518264331Sjkim [Emilia K�sper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson] 1519264331Sjkim 1520264331Sjkim *) Return an error when checking OCSP signatures when key is NULL. 1521264331Sjkim This fixes a DoS attack. (CVE-2013-0166) 1522264331Sjkim [Steve Henson] 1523264331Sjkim 1524264331Sjkim *) Call OCSP Stapling callback after ciphersuite has been chosen, so 1525264331Sjkim the right response is stapled. Also change SSL_get_certificate() 1526264331Sjkim so it returns the certificate actually sent. 1527264331Sjkim See http://rt.openssl.org/Ticket/Display.html?id=2836. 1528264331Sjkim (This is a backport) 1529264331Sjkim [Rob Stradling <rob.stradling@comodo.com>] 1530264331Sjkim 1531264331Sjkim *) Fix possible deadlock when decoding public keys. 1532264331Sjkim [Steve Henson] 1533264331Sjkim 1534264331Sjkim Changes between 0.9.8w and 0.9.8x [10 May 2012] 1535264331Sjkim 1536264331Sjkim *) Sanity check record length before skipping explicit IV in DTLS 1537264331Sjkim to fix DoS attack. 1538264331Sjkim 1539264331Sjkim Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic 1540264331Sjkim fuzzing as a service testing platform. 1541264331Sjkim (CVE-2012-2333) 1542264331Sjkim [Steve Henson] 1543264331Sjkim 1544264331Sjkim *) Initialise tkeylen properly when encrypting CMS messages. 1545264331Sjkim Thanks to Solar Designer of Openwall for reporting this issue. 1546264331Sjkim [Steve Henson] 1547264331Sjkim 1548264331Sjkim Changes between 0.9.8v and 0.9.8w [23 Apr 2012] 1549264331Sjkim 1550264331Sjkim *) The fix for CVE-2012-2110 did not take into account that the 1551264331Sjkim 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an 1552264331Sjkim int in OpenSSL 0.9.8, making it still vulnerable. Fix by 1553264331Sjkim rejecting negative len parameter. (CVE-2012-2131) 1554264331Sjkim [Tomas Hoger <thoger@redhat.com>] 1555264331Sjkim 1556264331Sjkim Changes between 0.9.8u and 0.9.8v [19 Apr 2012] 1557264331Sjkim 1558264331Sjkim *) Check for potentially exploitable overflows in asn1_d2i_read_bio 1559264331Sjkim BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer 1560264331Sjkim in CRYPTO_realloc_clean. 1561264331Sjkim 1562264331Sjkim Thanks to Tavis Ormandy, Google Security Team, for discovering this 1563264331Sjkim issue and to Adam Langley <agl@chromium.org> for fixing it. 1564264331Sjkim (CVE-2012-2110) 1565264331Sjkim [Adam Langley (Google), Tavis Ormandy, Google Security Team] 1566264331Sjkim 1567264331Sjkim Changes between 0.9.8t and 0.9.8u [12 Mar 2012] 1568264331Sjkim 1569264331Sjkim *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness 1570264331Sjkim in CMS and PKCS7 code. When RSA decryption fails use a random key for 1571264331Sjkim content decryption and always return the same error. Note: this attack 1572264331Sjkim needs on average 2^20 messages so it only affects automated senders. The 1573264331Sjkim old behaviour can be reenabled in the CMS code by setting the 1574264331Sjkim CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where 1575264331Sjkim an MMA defence is not necessary. 1576264331Sjkim Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering 1577264331Sjkim this issue. (CVE-2012-0884) 1578264331Sjkim [Steve Henson] 1579264331Sjkim 1580264331Sjkim *) Fix CVE-2011-4619: make sure we really are receiving a 1581264331Sjkim client hello before rejecting multiple SGC restarts. Thanks to 1582264331Sjkim Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug. 1583264331Sjkim [Steve Henson] 1584264331Sjkim 1585237657Sjkim Changes between 0.9.8s and 0.9.8t [18 Jan 2012] 1586237657Sjkim 1587237657Sjkim *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109. 1588237657Sjkim Thanks to Antonio Martin, Enterprise Secure Access Research and 1589237657Sjkim Development, Cisco Systems, Inc. for discovering this bug and 1590237657Sjkim preparing a fix. (CVE-2012-0050) 1591237657Sjkim [Antonio Martin] 1592264331Sjkim 1593237657Sjkim Changes between 0.9.8r and 0.9.8s [4 Jan 2012] 1594237657Sjkim 1595237657Sjkim *) Nadhem Alfardan and Kenny Paterson have discovered an extension 1596237657Sjkim of the Vaudenay padding oracle attack on CBC mode encryption 1597237657Sjkim which enables an efficient plaintext recovery attack against 1598237657Sjkim the OpenSSL implementation of DTLS. Their attack exploits timing 1599237657Sjkim differences arising during decryption processing. A research 1600237657Sjkim paper describing this attack can be found at: 1601237657Sjkim http://www.isg.rhul.ac.uk/~kp/dtls.pdf 1602237657Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 1603237657Sjkim Security Group at Royal Holloway, University of London 1604237657Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann 1605237657Sjkim <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de> 1606237657Sjkim for preparing the fix. (CVE-2011-4108) 1607237657Sjkim [Robin Seggelmann, Michael Tuexen] 1608237657Sjkim 1609237657Sjkim *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109) 1610237657Sjkim [Ben Laurie, Kasper <ekasper@google.com>] 1611237657Sjkim 1612237657Sjkim *) Clear bytes used for block padding of SSL 3.0 records. 1613237657Sjkim (CVE-2011-4576) 1614237657Sjkim [Adam Langley (Google)] 1615237657Sjkim 1616237657Sjkim *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George 1617237657Sjkim Kadianakis <desnacked@gmail.com> for discovering this issue and 1618237657Sjkim Adam Langley for preparing the fix. (CVE-2011-4619) 1619237657Sjkim [Adam Langley (Google)] 1620237657Sjkim 1621237657Sjkim *) Prevent malformed RFC3779 data triggering an assertion failure. 1622237657Sjkim Thanks to Andrew Chi, BBN Technologies, for discovering the flaw 1623237657Sjkim and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577) 1624237657Sjkim [Rob Austein <sra@hactrn.net>] 1625237657Sjkim 1626237657Sjkim *) Fix ssl_ciph.c set-up race. 1627237657Sjkim [Adam Langley (Google)] 1628237657Sjkim 1629237657Sjkim *) Fix spurious failures in ecdsatest.c. 1630237657Sjkim [Emilia K�sper (Google)] 1631237657Sjkim 1632237657Sjkim *) Fix the BIO_f_buffer() implementation (which was mixing different 1633237657Sjkim interpretations of the '..._len' fields). 1634237657Sjkim [Adam Langley (Google)] 1635237657Sjkim 1636237657Sjkim *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than 1637237657Sjkim BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent 1638237657Sjkim threads won't reuse the same blinding coefficients. 1639237657Sjkim 1640237657Sjkim This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING 1641237657Sjkim lock to call BN_BLINDING_invert_ex, and avoids one use of 1642237657Sjkim BN_BLINDING_update for each BN_BLINDING structure (previously, 1643237657Sjkim the last update always remained unused). 1644237657Sjkim [Emilia K�sper (Google)] 1645237657Sjkim 1646237657Sjkim *) Fix SSL memory handling for (EC)DH ciphersuites, in particular 1647237657Sjkim for multi-threaded use of ECDH. 1648237657Sjkim [Adam Langley (Google)] 1649237657Sjkim 1650237657Sjkim *) Fix x509_name_ex_d2i memory leak on bad inputs. 1651237657Sjkim [Bodo Moeller] 1652237657Sjkim 1653237657Sjkim *) Add protection against ECDSA timing attacks as mentioned in the paper 1654237657Sjkim by Billy Bob Brumley and Nicola Tuveri, see: 1655237657Sjkim 1656237657Sjkim http://eprint.iacr.org/2011/232.pdf 1657237657Sjkim 1658237657Sjkim [Billy Bob Brumley and Nicola Tuveri] 1659237657Sjkim 1660237657Sjkim Changes between 0.9.8q and 0.9.8r [8 Feb 2011] 1661237657Sjkim 1662237657Sjkim *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014 1663237657Sjkim [Neel Mehta, Adam Langley, Bodo Moeller (Google)] 1664237657Sjkim 1665237657Sjkim *) Fix bug in string printing code: if *any* escaping is enabled we must 1666237657Sjkim escape the escape character (backslash) or the resulting string is 1667237657Sjkim ambiguous. 1668237657Sjkim [Steve Henson] 1669237657Sjkim 1670216166Ssimon Changes between 0.9.8p and 0.9.8q [2 Dec 2010] 1671216166Ssimon 1672216166Ssimon *) Disable code workaround for ancient and obsolete Netscape browsers 1673216166Ssimon and servers: an attacker can use it in a ciphersuite downgrade attack. 1674216166Ssimon Thanks to Martin Rex for discovering this bug. CVE-2010-4180 1675216166Ssimon [Steve Henson] 1676216166Ssimon 1677216166Ssimon *) Fixed J-PAKE implementation error, originally discovered by 1678216166Ssimon Sebastien Martini, further info and confirmation from Stefan 1679216166Ssimon Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252 1680216166Ssimon [Ben Laurie] 1681216166Ssimon 1682215697Ssimon Changes between 0.9.8o and 0.9.8p [16 Nov 2010] 1683215697Ssimon 1684215697Ssimon *) Fix extension code to avoid race conditions which can result in a buffer 1685215697Ssimon overrun vulnerability: resumed sessions must not be modified as they can 1686215697Ssimon be shared by multiple threads. CVE-2010-3864 1687215697Ssimon [Steve Henson] 1688215697Ssimon 1689215697Ssimon *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939 1690215697Ssimon [Steve Henson] 1691215697Ssimon 1692215697Ssimon *) Don't reencode certificate when calculating signature: cache and use 1693215697Ssimon the original encoding instead. This makes signature verification of 1694215697Ssimon some broken encodings work correctly. 1695215697Ssimon [Steve Henson] 1696215697Ssimon 1697215697Ssimon *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT 1698215697Ssimon is also one of the inputs. 1699215697Ssimon [Emilia K�sper <emilia.kasper@esat.kuleuven.be> (Google)] 1700215697Ssimon 1701215697Ssimon *) Don't repeatedly append PBE algorithms to table if they already exist. 1702215697Ssimon Sort table on each new add. This effectively makes the table read only 1703215697Ssimon after all algorithms are added and subsequent calls to PKCS12_pbe_add 1704215697Ssimon etc are non-op. 1705215697Ssimon [Steve Henson] 1706215697Ssimon 1707215697Ssimon Changes between 0.9.8n and 0.9.8o [01 Jun 2010] 1708215697Ssimon 1709215697Ssimon [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after 1710215697Ssimon OpenSSL 1.0.0.] 1711215697Ssimon 1712215697Ssimon *) Correct a typo in the CMS ASN1 module which can result in invalid memory 1713215697Ssimon access or freeing data twice (CVE-2010-0742) 1714215697Ssimon [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>] 1715215697Ssimon 1716215697Ssimon *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more 1717215697Ssimon common in certificates and some applications which only call 1718215697Ssimon SSL_library_init and not OpenSSL_add_all_algorithms() will fail. 1719215697Ssimon [Steve Henson] 1720215697Ssimon 1721215697Ssimon *) VMS fixes: 1722215697Ssimon Reduce copying into .apps and .test in makevms.com 1723215697Ssimon Don't try to use blank CA certificate in CA.com 1724215697Ssimon Allow use of C files from original directories in maketests.com 1725215697Ssimon [Steven M. Schweda" <sms@antinode.info>] 1726215697Ssimon 1727206046Ssimon Changes between 0.9.8m and 0.9.8n [24 Mar 2010] 1728206046Ssimon 1729206046Ssimon *) When rejecting SSL/TLS records due to an incorrect version number, never 1730206046Ssimon update s->server with a new major version number. As of 1731206046Ssimon - OpenSSL 0.9.8m if 'short' is a 16-bit type, 1732206046Ssimon - OpenSSL 0.9.8f if 'short' is longer than 16 bits, 1733206046Ssimon the previous behavior could result in a read attempt at NULL when 1734206046Ssimon receiving specific incorrect SSL/TLS records once record payload 1735206046Ssimon protection is active. (CVE-2010-0740) 1736206046Ssimon [Bodo Moeller, Adam Langley <agl@chromium.org>] 1737206046Ssimon 1738206046Ssimon *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL 1739206046Ssimon could be crashed if the relevant tables were not present (e.g. chrooted). 1740206046Ssimon [Tomas Hoger <thoger@redhat.com>] 1741206046Ssimon 1742205128Ssimon Changes between 0.9.8l and 0.9.8m [25 Feb 2010] 1743205128Ssimon 1744205128Ssimon *) Always check bn_wexpend() return values for failure. (CVE-2009-3245) 1745205128Ssimon [Martin Olsson, Neel Mehta] 1746205128Ssimon 1747205128Ssimon *) Fix X509_STORE locking: Every 'objs' access requires a lock (to 1748205128Ssimon accommodate for stack sorting, always a write lock!). 1749205128Ssimon [Bodo Moeller] 1750205128Ssimon 1751205128Ssimon *) On some versions of WIN32 Heap32Next is very slow. This can cause 1752205128Ssimon excessive delays in the RAND_poll(): over a minute. As a workaround 1753205128Ssimon include a time check in the inner Heap32Next loop too. 1754205128Ssimon [Steve Henson] 1755205128Ssimon 1756205128Ssimon *) The code that handled flushing of data in SSL/TLS originally used the 1757205128Ssimon BIO_CTRL_INFO ctrl to see if any data was pending first. This caused 1758205128Ssimon the problem outlined in PR#1949. The fix suggested there however can 1759205128Ssimon trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions 1760205128Ssimon of Apache). So instead simplify the code to flush unconditionally. 1761205128Ssimon This should be fine since flushing with no data to flush is a no op. 1762205128Ssimon [Steve Henson] 1763205128Ssimon 1764205128Ssimon *) Handle TLS versions 2.0 and later properly and correctly use the 1765205128Ssimon highest version of TLS/SSL supported. Although TLS >= 2.0 is some way 1766205128Ssimon off ancient servers have a habit of sticking around for a while... 1767205128Ssimon [Steve Henson] 1768205128Ssimon 1769205128Ssimon *) Modify compression code so it frees up structures without using the 1770205128Ssimon ex_data callbacks. This works around a problem where some applications 1771205128Ssimon call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when 1772205128Ssimon restarting) then use compression (e.g. SSL with compression) later. 1773205128Ssimon This results in significant per-connection memory leaks and 1774205128Ssimon has caused some security issues including CVE-2008-1678 and 1775205128Ssimon CVE-2009-4355. 1776205128Ssimon [Steve Henson] 1777205128Ssimon 1778205128Ssimon *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't 1779205128Ssimon change when encrypting or decrypting. 1780205128Ssimon [Bodo Moeller] 1781205128Ssimon 1782205128Ssimon *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to 1783205128Ssimon connect and renegotiate with servers which do not support RI. 1784205128Ssimon Until RI is more widely deployed this option is enabled by default. 1785205128Ssimon [Steve Henson] 1786205128Ssimon 1787205128Ssimon *) Add "missing" ssl ctrls to clear options and mode. 1788205128Ssimon [Steve Henson] 1789205128Ssimon 1790205128Ssimon *) If client attempts to renegotiate and doesn't support RI respond with 1791205128Ssimon a no_renegotiation alert as required by RFC5746. Some renegotiating 1792205128Ssimon TLS clients will continue a connection gracefully when they receive 1793205128Ssimon the alert. Unfortunately OpenSSL mishandled this alert and would hang 1794205128Ssimon waiting for a server hello which it will never receive. Now we treat a 1795205128Ssimon received no_renegotiation alert as a fatal error. This is because 1796205128Ssimon applications requesting a renegotiation might well expect it to succeed 1797205128Ssimon and would have no code in place to handle the server denying it so the 1798205128Ssimon only safe thing to do is to terminate the connection. 1799205128Ssimon [Steve Henson] 1800205128Ssimon 1801205128Ssimon *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if 1802205128Ssimon peer supports secure renegotiation and 0 otherwise. Print out peer 1803205128Ssimon renegotiation support in s_client/s_server. 1804205128Ssimon [Steve Henson] 1805205128Ssimon 1806205128Ssimon *) Replace the highly broken and deprecated SPKAC certification method with 1807205128Ssimon the updated NID creation version. This should correctly handle UTF8. 1808205128Ssimon [Steve Henson] 1809205128Ssimon 1810205128Ssimon *) Implement RFC5746. Re-enable renegotiation but require the extension 1811205128Ssimon as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION 1812205128Ssimon turns out to be a bad idea. It has been replaced by 1813205128Ssimon SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with 1814205128Ssimon SSL_CTX_set_options(). This is really not recommended unless you 1815205128Ssimon know what you are doing. 1816205128Ssimon [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson] 1817205128Ssimon 1818205128Ssimon *) Fixes to stateless session resumption handling. Use initial_ctx when 1819205128Ssimon issuing and attempting to decrypt tickets in case it has changed during 1820205128Ssimon servername handling. Use a non-zero length session ID when attempting 1821205128Ssimon stateless session resumption: this makes it possible to determine if 1822205128Ssimon a resumption has occurred immediately after receiving server hello 1823205128Ssimon (several places in OpenSSL subtly assume this) instead of later in 1824205128Ssimon the handshake. 1825205128Ssimon [Steve Henson] 1826205128Ssimon 1827205128Ssimon *) The functions ENGINE_ctrl(), OPENSSL_isservice(), 1828205128Ssimon CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error 1829205128Ssimon fixes for a few places where the return code is not checked 1830205128Ssimon correctly. 1831205128Ssimon [Julia Lawall <julia@diku.dk>] 1832205128Ssimon 1833205128Ssimon *) Add --strict-warnings option to Configure script to include devteam 1834205128Ssimon warnings in other configurations. 1835205128Ssimon [Steve Henson] 1836205128Ssimon 1837205128Ssimon *) Add support for --libdir option and LIBDIR variable in makefiles. This 1838205128Ssimon makes it possible to install openssl libraries in locations which 1839205128Ssimon have names other than "lib", for example "/usr/lib64" which some 1840205128Ssimon systems need. 1841205128Ssimon [Steve Henson, based on patch from Jeremy Utley] 1842205128Ssimon 1843205128Ssimon *) Don't allow the use of leading 0x80 in OIDs. This is a violation of 1844205128Ssimon X690 8.9.12 and can produce some misleading textual output of OIDs. 1845205128Ssimon [Steve Henson, reported by Dan Kaminsky] 1846205128Ssimon 1847205128Ssimon *) Delete MD2 from algorithm tables. This follows the recommendation in 1848205128Ssimon several standards that it is not used in new applications due to 1849205128Ssimon several cryptographic weaknesses. For binary compatibility reasons 1850205128Ssimon the MD2 API is still compiled in by default. 1851205128Ssimon [Steve Henson] 1852205128Ssimon 1853205128Ssimon *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved 1854205128Ssimon and restored. 1855205128Ssimon [Steve Henson] 1856205128Ssimon 1857205128Ssimon *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and 1858205128Ssimon OPENSSL_asc2uni conditionally on Netware platforms to avoid a name 1859205128Ssimon clash. 1860205128Ssimon [Guenter <lists@gknw.net>] 1861205128Ssimon 1862205128Ssimon *) Fix the server certificate chain building code to use X509_verify_cert(), 1863205128Ssimon it used to have an ad-hoc builder which was unable to cope with anything 1864205128Ssimon other than a simple chain. 1865205128Ssimon [David Woodhouse <dwmw2@infradead.org>, Steve Henson] 1866205128Ssimon 1867205128Ssimon *) Don't check self signed certificate signatures in X509_verify_cert() 1868205128Ssimon by default (a flag can override this): it just wastes time without 1869205128Ssimon adding any security. As a useful side effect self signed root CAs 1870205128Ssimon with non-FIPS digests are now usable in FIPS mode. 1871205128Ssimon [Steve Henson] 1872205128Ssimon 1873205128Ssimon *) In dtls1_process_out_of_seq_message() the check if the current message 1874205128Ssimon is already buffered was missing. For every new message was memory 1875205128Ssimon allocated, allowing an attacker to perform an denial of service attack 1876205128Ssimon with sending out of seq handshake messages until there is no memory 1877205128Ssimon left. Additionally every future messege was buffered, even if the 1878205128Ssimon sequence number made no sense and would be part of another handshake. 1879205128Ssimon So only messages with sequence numbers less than 10 in advance will be 1880205128Ssimon buffered. (CVE-2009-1378) 1881205128Ssimon [Robin Seggelmann, discovered by Daniel Mentz] 1882205128Ssimon 1883205128Ssimon *) Records are buffered if they arrive with a future epoch to be 1884205128Ssimon processed after finishing the corresponding handshake. There is 1885205128Ssimon currently no limitation to this buffer allowing an attacker to perform 1886205128Ssimon a DOS attack with sending records with future epochs until there is no 1887205128Ssimon memory left. This patch adds the pqueue_size() function to detemine 1888205128Ssimon the size of a buffer and limits the record buffer to 100 entries. 1889205128Ssimon (CVE-2009-1377) 1890205128Ssimon [Robin Seggelmann, discovered by Daniel Mentz] 1891205128Ssimon 1892205128Ssimon *) Keep a copy of frag->msg_header.frag_len so it can be used after the 1893205128Ssimon parent structure is freed. (CVE-2009-1379) 1894205128Ssimon [Daniel Mentz] 1895205128Ssimon 1896205128Ssimon *) Handle non-blocking I/O properly in SSL_shutdown() call. 1897205128Ssimon [Darryl Miles <darryl-mailinglists@netbauds.net>] 1898205128Ssimon 1899205128Ssimon *) Add 2.5.4.* OIDs 1900205128Ssimon [Ilya O. <vrghost@gmail.com>] 1901205128Ssimon 1902205128Ssimon Changes between 0.9.8k and 0.9.8l [5 Nov 2009] 1903205128Ssimon 1904205128Ssimon *) Disable renegotiation completely - this fixes a severe security 1905205128Ssimon problem (CVE-2009-3555) at the cost of breaking all 1906205128Ssimon renegotiation. Renegotiation can be re-enabled by setting 1907205128Ssimon SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at 1908205128Ssimon run-time. This is really not recommended unless you know what 1909205128Ssimon you're doing. 1910205128Ssimon [Ben Laurie] 1911205128Ssimon 1912194206Ssimon Changes between 0.9.8j and 0.9.8k [25 Mar 2009] 1913167612Ssimon 1914194206Ssimon *) Don't set val to NULL when freeing up structures, it is freed up by 1915194206Ssimon underlying code. If sizeof(void *) > sizeof(long) this can result in 1916194206Ssimon zeroing past the valid field. (CVE-2009-0789) 1917194206Ssimon [Paolo Ganci <Paolo.Ganci@AdNovum.CH>] 1918194206Ssimon 1919194206Ssimon *) Fix bug where return value of CMS_SignerInfo_verify_content() was not 1920194206Ssimon checked correctly. This would allow some invalid signed attributes to 1921194206Ssimon appear to verify correctly. (CVE-2009-0591) 1922194206Ssimon [Ivan Nestlerode <inestlerode@us.ibm.com>] 1923194206Ssimon 1924194206Ssimon *) Reject UniversalString and BMPString types with invalid lengths. This 1925194206Ssimon prevents a crash in ASN1_STRING_print_ex() which assumes the strings have 1926194206Ssimon a legal length. (CVE-2009-0590) 1927194206Ssimon [Steve Henson] 1928194206Ssimon 1929194206Ssimon *) Set S/MIME signing as the default purpose rather than setting it 1930194206Ssimon unconditionally. This allows applications to override it at the store 1931194206Ssimon level. 1932194206Ssimon [Steve Henson] 1933194206Ssimon 1934194206Ssimon *) Permit restricted recursion of ASN1 strings. This is needed in practice 1935194206Ssimon to handle some structures. 1936194206Ssimon [Steve Henson] 1937194206Ssimon 1938194206Ssimon *) Improve efficiency of mem_gets: don't search whole buffer each time 1939194206Ssimon for a '\n' 1940194206Ssimon [Jeremy Shapiro <jnshapir@us.ibm.com>] 1941194206Ssimon 1942194206Ssimon *) New -hex option for openssl rand. 1943194206Ssimon [Matthieu Herrb] 1944194206Ssimon 1945194206Ssimon *) Print out UTF8String and NumericString when parsing ASN1. 1946194206Ssimon [Steve Henson] 1947194206Ssimon 1948194206Ssimon *) Support NumericString type for name components. 1949194206Ssimon [Steve Henson] 1950194206Ssimon 1951194206Ssimon *) Allow CC in the environment to override the automatically chosen 1952194206Ssimon compiler. Note that nothing is done to ensure flags work with the 1953194206Ssimon chosen compiler. 1954194206Ssimon [Ben Laurie] 1955194206Ssimon 1956194206Ssimon Changes between 0.9.8i and 0.9.8j [07 Jan 2009] 1957194206Ssimon 1958194206Ssimon *) Properly check EVP_VerifyFinal() and similar return values 1959194206Ssimon (CVE-2008-5077). 1960194206Ssimon [Ben Laurie, Bodo Moeller, Google Security Team] 1961194206Ssimon 1962194206Ssimon *) Enable TLS extensions by default. 1963194206Ssimon [Ben Laurie] 1964194206Ssimon 1965194206Ssimon *) Allow the CHIL engine to be loaded, whether the application is 1966194206Ssimon multithreaded or not. (This does not release the developer from the 1967194206Ssimon obligation to set up the dynamic locking callbacks.) 1968194206Ssimon [Sander Temme <sander@temme.net>] 1969194206Ssimon 1970194206Ssimon *) Use correct exit code if there is an error in dgst command. 1971194206Ssimon [Steve Henson; problem pointed out by Roland Dirlewanger] 1972194206Ssimon 1973194206Ssimon *) Tweak Configure so that you need to say "experimental-jpake" to enable 1974194206Ssimon JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications. 1975194206Ssimon [Bodo Moeller] 1976194206Ssimon 1977194206Ssimon *) Add experimental JPAKE support, including demo authentication in 1978194206Ssimon s_client and s_server. 1979194206Ssimon [Ben Laurie] 1980194206Ssimon 1981194206Ssimon *) Set the comparison function in v3_addr_canonize(). 1982194206Ssimon [Rob Austein <sra@hactrn.net>] 1983194206Ssimon 1984194206Ssimon *) Add support for XMPP STARTTLS in s_client. 1985194206Ssimon [Philip Paeps <philip@freebsd.org>] 1986194206Ssimon 1987194206Ssimon *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior 1988194206Ssimon to ensure that even with this option, only ciphersuites in the 1989194206Ssimon server's preference list will be accepted. (Note that the option 1990194206Ssimon applies only when resuming a session, so the earlier behavior was 1991194206Ssimon just about the algorithm choice for symmetric cryptography.) 1992194206Ssimon [Bodo Moeller] 1993194206Ssimon 1994194206Ssimon Changes between 0.9.8h and 0.9.8i [15 Sep 2008] 1995194206Ssimon 1996205128Ssimon *) Fix NULL pointer dereference if a DTLS server received 1997205128Ssimon ChangeCipherSpec as first record (CVE-2009-1386). 1998205128Ssimon [PR #1679] 1999205128Ssimon 2000194206Ssimon *) Fix a state transitition in s3_srvr.c and d1_srvr.c 2001194206Ssimon (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...). 2002194206Ssimon [Nagendra Modadugu] 2003194206Ssimon 2004194206Ssimon *) The fix in 0.9.8c that supposedly got rid of unsafe 2005194206Ssimon double-checked locking was incomplete for RSA blinding, 2006194206Ssimon addressing just one layer of what turns out to have been 2007194206Ssimon doubly unsafe triple-checked locking. 2008194206Ssimon 2009194206Ssimon So now fix this for real by retiring the MONT_HELPER macro 2010194206Ssimon in crypto/rsa/rsa_eay.c. 2011194206Ssimon 2012194206Ssimon [Bodo Moeller; problem pointed out by Marius Schilder] 2013194206Ssimon 2014194206Ssimon *) Various precautionary measures: 2015194206Ssimon 2016194206Ssimon - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h). 2017194206Ssimon 2018194206Ssimon - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c). 2019194206Ssimon (NB: This would require knowledge of the secret session ticket key 2020194206Ssimon to exploit, in which case you'd be SOL either way.) 2021194206Ssimon 2022194206Ssimon - Change bn_nist.c so that it will properly handle input BIGNUMs 2023194206Ssimon outside the expected range. 2024194206Ssimon 2025194206Ssimon - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG 2026194206Ssimon builds. 2027194206Ssimon 2028194206Ssimon [Neel Mehta, Bodo Moeller] 2029194206Ssimon 2030194206Ssimon *) Allow engines to be "soft loaded" - i.e. optionally don't die if 2031194206Ssimon the load fails. Useful for distros. 2032194206Ssimon [Ben Laurie and the FreeBSD team] 2033194206Ssimon 2034194206Ssimon *) Add support for Local Machine Keyset attribute in PKCS#12 files. 2035194206Ssimon [Steve Henson] 2036194206Ssimon 2037194206Ssimon *) Fix BN_GF2m_mod_arr() top-bit cleanup code. 2038194206Ssimon [Huang Ying] 2039194206Ssimon 2040194206Ssimon *) Expand ENGINE to support engine supplied SSL client certificate functions. 2041194206Ssimon 2042194206Ssimon This work was sponsored by Logica. 2043194206Ssimon [Steve Henson] 2044194206Ssimon 2045194206Ssimon *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows 2046194206Ssimon keystores. Support for SSL/TLS client authentication too. 2047194206Ssimon Not compiled unless enable-capieng specified to Configure. 2048194206Ssimon 2049194206Ssimon This work was sponsored by Logica. 2050194206Ssimon [Steve Henson] 2051194206Ssimon 2052194206Ssimon *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using 2053194206Ssimon ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain 2054194206Ssimon attribute creation routines such as certifcate requests and PKCS#12 2055194206Ssimon files. 2056194206Ssimon [Steve Henson] 2057194206Ssimon 2058194206Ssimon Changes between 0.9.8g and 0.9.8h [28 May 2008] 2059194206Ssimon 2060194206Ssimon *) Fix flaw if 'Server Key exchange message' is omitted from a TLS 2061194206Ssimon handshake which could lead to a cilent crash as found using the 2062194206Ssimon Codenomicon TLS test suite (CVE-2008-1672) 2063194206Ssimon [Steve Henson, Mark Cox] 2064194206Ssimon 2065194206Ssimon *) Fix double free in TLS server name extensions which could lead to 2066194206Ssimon a remote crash found by Codenomicon TLS test suite (CVE-2008-0891) 2067194206Ssimon [Joe Orton] 2068194206Ssimon 2069194206Ssimon *) Clear error queue in SSL_CTX_use_certificate_chain_file() 2070194206Ssimon 2071194206Ssimon Clear the error queue to ensure that error entries left from 2072194206Ssimon older function calls do not interfere with the correct operation. 2073194206Ssimon [Lutz Jaenicke, Erik de Castro Lopo] 2074194206Ssimon 2075194206Ssimon *) Remove root CA certificates of commercial CAs: 2076194206Ssimon 2077194206Ssimon The OpenSSL project does not recommend any specific CA and does not 2078194206Ssimon have any policy with respect to including or excluding any CA. 2079194206Ssimon Therefore it does not make any sense to ship an arbitrary selection 2080194206Ssimon of root CA certificates with the OpenSSL software. 2081194206Ssimon [Lutz Jaenicke] 2082194206Ssimon 2083194206Ssimon *) RSA OAEP patches to fix two separate invalid memory reads. 2084194206Ssimon The first one involves inputs when 'lzero' is greater than 2085194206Ssimon 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes 2086194206Ssimon before the beginning of from). The second one involves inputs where 2087194206Ssimon the 'db' section contains nothing but zeroes (there is a one-byte 2088194206Ssimon invalid read after the end of 'db'). 2089194206Ssimon [Ivan Nestlerode <inestlerode@us.ibm.com>] 2090194206Ssimon 2091194206Ssimon *) Partial backport from 0.9.9-dev: 2092194206Ssimon 2093194206Ssimon Introduce bn_mul_mont (dedicated Montgomery multiplication 2094194206Ssimon procedure) as a candidate for BIGNUM assembler implementation. 2095194206Ssimon While 0.9.9-dev uses assembler for various architectures, only 2096194206Ssimon x86_64 is available by default here in the 0.9.8 branch, and 2097194206Ssimon 32-bit x86 is available through a compile-time setting. 2098194206Ssimon 2099194206Ssimon To try the 32-bit x86 assembler implementation, use Configure 2100194206Ssimon option "enable-montasm" (which exists only for this backport). 2101194206Ssimon 2102194206Ssimon As "enable-montasm" for 32-bit x86 disclaims code stability 2103194206Ssimon anyway, in this constellation we activate additional code 2104194206Ssimon backported from 0.9.9-dev for further performance improvements, 2105194206Ssimon namely BN_from_montgomery_word. (To enable this otherwise, 2106194206Ssimon e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".) 2107194206Ssimon 2108194206Ssimon [Andy Polyakov (backport partially by Bodo Moeller)] 2109194206Ssimon 2110194206Ssimon *) Add TLS session ticket callback. This allows an application to set 2111194206Ssimon TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed 2112194206Ssimon values. This is useful for key rollover for example where several key 2113194206Ssimon sets may exist with different names. 2114194206Ssimon [Steve Henson] 2115194206Ssimon 2116194206Ssimon *) Reverse ENGINE-internal logic for caching default ENGINE handles. 2117194206Ssimon This was broken until now in 0.9.8 releases, such that the only way 2118194206Ssimon a registered ENGINE could be used (assuming it initialises 2119194206Ssimon successfully on the host) was to explicitly set it as the default 2120194206Ssimon for the relevant algorithms. This is in contradiction with 0.9.7 2121194206Ssimon behaviour and the documentation. With this fix, when an ENGINE is 2122194206Ssimon registered into a given algorithm's table of implementations, the 2123194206Ssimon 'uptodate' flag is reset so that auto-discovery will be used next 2124194206Ssimon time a new context for that algorithm attempts to select an 2125194206Ssimon implementation. 2126194206Ssimon [Ian Lister (tweaked by Geoff Thorpe)] 2127194206Ssimon 2128194206Ssimon *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9 2129194206Ssimon implemention in the following ways: 2130194206Ssimon 2131194206Ssimon Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be 2132194206Ssimon hard coded. 2133194206Ssimon 2134194206Ssimon Lack of BER streaming support means one pass streaming processing is 2135194206Ssimon only supported if data is detached: setting the streaming flag is 2136194206Ssimon ignored for embedded content. 2137194206Ssimon 2138194206Ssimon CMS support is disabled by default and must be explicitly enabled 2139194206Ssimon with the enable-cms configuration option. 2140194206Ssimon [Steve Henson] 2141194206Ssimon 2142194206Ssimon *) Update the GMP engine glue to do direct copies between BIGNUM and 2143194206Ssimon mpz_t when openssl and GMP use the same limb size. Otherwise the 2144194206Ssimon existing "conversion via a text string export" trick is still used. 2145194206Ssimon [Paul Sheer <paulsheer@gmail.com>] 2146194206Ssimon 2147194206Ssimon *) Zlib compression BIO. This is a filter BIO which compressed and 2148194206Ssimon uncompresses any data passed through it. 2149194206Ssimon [Steve Henson] 2150194206Ssimon 2151194206Ssimon *) Add AES_wrap_key() and AES_unwrap_key() functions to implement 2152194206Ssimon RFC3394 compatible AES key wrapping. 2153194206Ssimon [Steve Henson] 2154194206Ssimon 2155194206Ssimon *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0(): 2156194206Ssimon sets string data without copying. X509_ALGOR_set0() and 2157194206Ssimon X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier) 2158194206Ssimon data. Attribute function X509at_get0_data_by_OBJ(): retrieves data 2159194206Ssimon from an X509_ATTRIBUTE structure optionally checking it occurs only 2160194206Ssimon once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied 2161194206Ssimon data. 2162194206Ssimon [Steve Henson] 2163194206Ssimon 2164194206Ssimon *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set() 2165194206Ssimon to get the expected BN_FLG_CONSTTIME behavior. 2166194206Ssimon [Bodo Moeller (Google)] 2167194206Ssimon 2168194206Ssimon *) Netware support: 2169194206Ssimon 2170194206Ssimon - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets 2171194206Ssimon - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT) 2172194206Ssimon - added some more tests to do_tests.pl 2173194206Ssimon - fixed RunningProcess usage so that it works with newer LIBC NDKs too 2174194206Ssimon - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency 2175194206Ssimon - added new Configure targets netware-clib-bsdsock, netware-clib-gcc, 2176194206Ssimon netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc 2177194206Ssimon - various changes to netware.pl to enable gcc-cross builds on Win32 2178194206Ssimon platform 2179194206Ssimon - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD) 2180194206Ssimon - various changes to fix missing prototype warnings 2181194206Ssimon - fixed x86nasm.pl to create correct asm files for NASM COFF output 2182194206Ssimon - added AES, WHIRLPOOL and CPUID assembler code to build files 2183194206Ssimon - added missing AES assembler make rules to mk1mf.pl 2184194206Ssimon - fixed order of includes in apps/ocsp.c so that e_os.h settings apply 2185194206Ssimon [Guenter Knauf <eflash@gmx.net>] 2186194206Ssimon 2187194206Ssimon *) Implement certificate status request TLS extension defined in RFC3546. 2188194206Ssimon A client can set the appropriate parameters and receive the encoded 2189194206Ssimon OCSP response via a callback. A server can query the supplied parameters 2190194206Ssimon and set the encoded OCSP response in the callback. Add simplified examples 2191194206Ssimon to s_client and s_server. 2192194206Ssimon [Steve Henson] 2193194206Ssimon 2194194206Ssimon Changes between 0.9.8f and 0.9.8g [19 Oct 2007] 2195194206Ssimon 2196194206Ssimon *) Fix various bugs: 2197194206Ssimon + Binary incompatibility of ssl_ctx_st structure 2198194206Ssimon + DTLS interoperation with non-compliant servers 2199194206Ssimon + Don't call get_session_cb() without proposed session 2200194206Ssimon + Fix ia64 assembler code 2201194206Ssimon [Andy Polyakov, Steve Henson] 2202194206Ssimon 2203194206Ssimon Changes between 0.9.8e and 0.9.8f [11 Oct 2007] 2204194206Ssimon 2205194206Ssimon *) DTLS Handshake overhaul. There were longstanding issues with 2206194206Ssimon OpenSSL DTLS implementation, which were making it impossible for 2207194206Ssimon RFC 4347 compliant client to communicate with OpenSSL server. 2208194206Ssimon Unfortunately just fixing these incompatibilities would "cut off" 2209194206Ssimon pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e 2210194206Ssimon server keeps tolerating non RFC compliant syntax. The opposite is 2211194206Ssimon not true, 0.9.8f client can not communicate with earlier server. 2212194206Ssimon This update even addresses CVE-2007-4995. 2213194206Ssimon [Andy Polyakov] 2214194206Ssimon 2215194206Ssimon *) Changes to avoid need for function casts in OpenSSL: some compilers 2216194206Ssimon (gcc 4.2 and later) reject their use. 2217194206Ssimon [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>, 2218194206Ssimon Steve Henson] 2219194206Ssimon 2220194206Ssimon *) Add RFC4507 support to OpenSSL. This includes the corrections in 2221194206Ssimon RFC4507bis. The encrypted ticket format is an encrypted encoded 2222194206Ssimon SSL_SESSION structure, that way new session features are automatically 2223194206Ssimon supported. 2224194206Ssimon 2225194206Ssimon If a client application caches session in an SSL_SESSION structure 2226194206Ssimon support is transparent because tickets are now stored in the encoded 2227194206Ssimon SSL_SESSION. 2228194206Ssimon 2229194206Ssimon The SSL_CTX structure automatically generates keys for ticket 2230194206Ssimon protection in servers so again support should be possible 2231194206Ssimon with no application modification. 2232194206Ssimon 2233194206Ssimon If a client or server wishes to disable RFC4507 support then the option 2234194206Ssimon SSL_OP_NO_TICKET can be set. 2235194206Ssimon 2236194206Ssimon Add a TLS extension debugging callback to allow the contents of any client 2237194206Ssimon or server extensions to be examined. 2238194206Ssimon 2239194206Ssimon This work was sponsored by Google. 2240194206Ssimon [Steve Henson] 2241194206Ssimon 2242194206Ssimon *) Add initial support for TLS extensions, specifically for the server_name 2243194206Ssimon extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 2244194206Ssimon have new members for a host name. The SSL data structure has an 2245194206Ssimon additional member SSL_CTX *initial_ctx so that new sessions can be 2246194206Ssimon stored in that context to allow for session resumption, even after the 2247194206Ssimon SSL has been switched to a new SSL_CTX in reaction to a client's 2248194206Ssimon server_name extension. 2249194206Ssimon 2250194206Ssimon New functions (subject to change): 2251194206Ssimon 2252194206Ssimon SSL_get_servername() 2253194206Ssimon SSL_get_servername_type() 2254194206Ssimon SSL_set_SSL_CTX() 2255194206Ssimon 2256194206Ssimon New CTRL codes and macros (subject to change): 2257194206Ssimon 2258194206Ssimon SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 2259194206Ssimon - SSL_CTX_set_tlsext_servername_callback() 2260194206Ssimon SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 2261194206Ssimon - SSL_CTX_set_tlsext_servername_arg() 2262194206Ssimon SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 2263194206Ssimon 2264194206Ssimon openssl s_client has a new '-servername ...' option. 2265194206Ssimon 2266194206Ssimon openssl s_server has new options '-servername_host ...', '-cert2 ...', 2267194206Ssimon '-key2 ...', '-servername_fatal' (subject to change). This allows 2268194206Ssimon testing the HostName extension for a specific single host name ('-cert' 2269194206Ssimon and '-key' remain fallbacks for handshakes without HostName 2270194206Ssimon negotiation). If the unrecogninzed_name alert has to be sent, this by 2271194206Ssimon default is a warning; it becomes fatal with the '-servername_fatal' 2272194206Ssimon option. 2273194206Ssimon 2274194206Ssimon [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson] 2275194206Ssimon 2276194206Ssimon *) Add AES and SSE2 assembly language support to VC++ build. 2277194206Ssimon [Steve Henson] 2278194206Ssimon 2279194206Ssimon *) Mitigate attack on final subtraction in Montgomery reduction. 2280194206Ssimon [Andy Polyakov] 2281194206Ssimon 2282194206Ssimon *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0 2283194206Ssimon (which previously caused an internal error). 2284194206Ssimon [Bodo Moeller] 2285194206Ssimon 2286194206Ssimon *) Squeeze another 10% out of IGE mode when in != out. 2287194206Ssimon [Ben Laurie] 2288194206Ssimon 2289194206Ssimon *) AES IGE mode speedup. 2290194206Ssimon [Dean Gaudet (Google)] 2291194206Ssimon 2292194206Ssimon *) Add the Korean symmetric 128-bit cipher SEED (see 2293194206Ssimon http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and 2294194206Ssimon add SEED ciphersuites from RFC 4162: 2295194206Ssimon 2296194206Ssimon TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA" 2297194206Ssimon TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA" 2298194206Ssimon TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA" 2299194206Ssimon TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA" 2300194206Ssimon 2301194206Ssimon To minimize changes between patchlevels in the OpenSSL 0.9.8 2302194206Ssimon series, SEED remains excluded from compilation unless OpenSSL 2303194206Ssimon is configured with 'enable-seed'. 2304194206Ssimon [KISA, Bodo Moeller] 2305194206Ssimon 2306194206Ssimon *) Mitigate branch prediction attacks, which can be practical if a 2307194206Ssimon single processor is shared, allowing a spy process to extract 2308194206Ssimon information. For detailed background information, see 2309194206Ssimon http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron, 2310194206Ssimon J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL 2311194206Ssimon and Necessary Software Countermeasures"). The core of the change 2312194206Ssimon are new versions BN_div_no_branch() and 2313194206Ssimon BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(), 2314194206Ssimon respectively, which are slower, but avoid the security-relevant 2315194206Ssimon conditional branches. These are automatically called by BN_div() 2316194206Ssimon and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one 2317194206Ssimon of the input BIGNUMs. Also, BN_is_bit_set() has been changed to 2318194206Ssimon remove a conditional branch. 2319194206Ssimon 2320194206Ssimon BN_FLG_CONSTTIME is the new name for the previous 2321194206Ssimon BN_FLG_EXP_CONSTTIME flag, since it now affects more than just 2322194206Ssimon modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag 2323194206Ssimon in the exponent causes BN_mod_exp_mont() to use the alternative 2324194206Ssimon implementation in BN_mod_exp_mont_consttime().) The old name 2325194206Ssimon remains as a deprecated alias. 2326194206Ssimon 2327194206Ssimon Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general 2328194206Ssimon RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses 2329194206Ssimon constant-time implementations for more than just exponentiation. 2330194206Ssimon Here too the old name is kept as a deprecated alias. 2331194206Ssimon 2332194206Ssimon BN_BLINDING_new() will now use BN_dup() for the modulus so that 2333194206Ssimon the BN_BLINDING structure gets an independent copy of the 2334194206Ssimon modulus. This means that the previous "BIGNUM *m" argument to 2335194206Ssimon BN_BLINDING_new() and to BN_BLINDING_create_param() now 2336194206Ssimon essentially becomes "const BIGNUM *m", although we can't actually 2337194206Ssimon change this in the header file before 0.9.9. It allows 2338194206Ssimon RSA_setup_blinding() to use BN_with_flags() on the modulus to 2339194206Ssimon enable BN_FLG_CONSTTIME. 2340194206Ssimon 2341194206Ssimon [Matthew D Wood (Intel Corp)] 2342194206Ssimon 2343194206Ssimon *) In the SSL/TLS server implementation, be strict about session ID 2344194206Ssimon context matching (which matters if an application uses a single 2345194206Ssimon external cache for different purposes). Previously, 2346194206Ssimon out-of-context reuse was forbidden only if SSL_VERIFY_PEER was 2347194206Ssimon set. This did ensure strict client verification, but meant that, 2348194206Ssimon with applications using a single external cache for quite 2349194206Ssimon different requirements, clients could circumvent ciphersuite 2350194206Ssimon restrictions for a given session ID context by starting a session 2351194206Ssimon in a different context. 2352194206Ssimon [Bodo Moeller] 2353194206Ssimon 2354167612Ssimon *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 2355167612Ssimon a ciphersuite string such as "DEFAULT:RSA" cannot enable 2356167612Ssimon authentication-only ciphersuites. 2357167612Ssimon [Bodo Moeller] 2358167612Ssimon 2359194206Ssimon *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was 2360194206Ssimon not complete and could lead to a possible single byte overflow 2361194206Ssimon (CVE-2007-5135) [Ben Laurie] 2362194206Ssimon 2363194206Ssimon Changes between 0.9.8d and 0.9.8e [23 Feb 2007] 2364194206Ssimon 2365167612Ssimon *) Since AES128 and AES256 (and similarly Camellia128 and 2366167612Ssimon Camellia256) share a single mask bit in the logic of 2367167612Ssimon ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 2368167612Ssimon kludge to work properly if AES128 is available and AES256 isn't 2369167612Ssimon (or if Camellia128 is available and Camellia256 isn't). 2370167612Ssimon [Victor Duchovni] 2371167612Ssimon 2372167612Ssimon *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c 2373167612Ssimon (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters): 2374167612Ssimon When a point or a seed is encoded in a BIT STRING, we need to 2375167612Ssimon prevent the removal of trailing zero bits to get the proper DER 2376167612Ssimon encoding. (By default, crypto/asn1/a_bitstr.c assumes the case 2377167612Ssimon of a NamedBitList, for which trailing 0 bits need to be removed.) 2378167612Ssimon [Bodo Moeller] 2379167612Ssimon 2380167612Ssimon *) Have SSL/TLS server implementation tolerate "mismatched" record 2381167612Ssimon protocol version while receiving ClientHello even if the 2382167612Ssimon ClientHello is fragmented. (The server can't insist on the 2383167612Ssimon particular protocol version it has chosen before the ServerHello 2384167612Ssimon message has informed the client about his choice.) 2385167612Ssimon [Bodo Moeller] 2386167612Ssimon 2387167612Ssimon *) Add RFC 3779 support. 2388167612Ssimon [Rob Austein for ARIN, Ben Laurie] 2389167612Ssimon 2390167612Ssimon *) Load error codes if they are not already present instead of using a 2391167612Ssimon static variable. This allows them to be cleanly unloaded and reloaded. 2392167612Ssimon Improve header file function name parsing. 2393167612Ssimon [Steve Henson] 2394167612Ssimon 2395167612Ssimon *) extend SMTP and IMAP protocol emulation in s_client to use EHLO 2396167612Ssimon or CAPABILITY handshake as required by RFCs. 2397167612Ssimon [Goetz Babin-Ebell] 2398167612Ssimon 2399162911Ssimon Changes between 0.9.8c and 0.9.8d [28 Sep 2006] 2400162911Ssimon 2401162911Ssimon *) Introduce limits to prevent malicious keys being able to 2402162911Ssimon cause a denial of service. (CVE-2006-2940) 2403162911Ssimon [Steve Henson, Bodo Moeller] 2404162911Ssimon 2405162911Ssimon *) Fix ASN.1 parsing of certain invalid structures that can result 2406162911Ssimon in a denial of service. (CVE-2006-2937) [Steve Henson] 2407162911Ssimon 2408162911Ssimon *) Fix buffer overflow in SSL_get_shared_ciphers() function. 2409162911Ssimon (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 2410162911Ssimon 2411162911Ssimon *) Fix SSL client code which could crash if connecting to a 2412162911Ssimon malicious SSLv2 server. (CVE-2006-4343) 2413162911Ssimon [Tavis Ormandy and Will Drewry, Google Security Team] 2414162911Ssimon 2415162911Ssimon *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites 2416162911Ssimon match only those. Before that, "AES256-SHA" would be interpreted 2417162911Ssimon as a pattern and match "AES128-SHA" too (since AES128-SHA got 2418162911Ssimon the same strength classification in 0.9.7h) as we currently only 2419162911Ssimon have a single AES bit in the ciphersuite description bitmap. 2420162911Ssimon That change, however, also applied to ciphersuite strings such as 2421162911Ssimon "RC4-MD5" that intentionally matched multiple ciphersuites -- 2422162911Ssimon namely, SSL 2.0 ciphersuites in addition to the more common ones 2423162911Ssimon from SSL 3.0/TLS 1.0. 2424162911Ssimon 2425162911Ssimon So we change the selection algorithm again: Naming an explicit 2426162911Ssimon ciphersuite selects this one ciphersuite, and any other similar 2427162911Ssimon ciphersuite (same bitmap) from *other* protocol versions. 2428162911Ssimon Thus, "RC4-MD5" again will properly select both the SSL 2.0 2429162911Ssimon ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite. 2430162911Ssimon 2431162911Ssimon Since SSL 2.0 does not have any ciphersuites for which the 2432162911Ssimon 128/256 bit distinction would be relevant, this works for now. 2433162911Ssimon The proper fix will be to use different bits for AES128 and 2434162911Ssimon AES256, which would have avoided the problems from the beginning; 2435162911Ssimon however, bits are scarce, so we can only do this in a new release 2436162911Ssimon (not just a patchlevel) when we can change the SSL_CIPHER 2437162911Ssimon definition to split the single 'unsigned long mask' bitmap into 2438162911Ssimon multiple values to extend the available space. 2439162911Ssimon 2440162911Ssimon [Bodo Moeller] 2441162911Ssimon 2442162911Ssimon Changes between 0.9.8b and 0.9.8c [05 Sep 2006] 2443162911Ssimon 2444162911Ssimon *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 2445162911Ssimon (CVE-2006-4339) [Ben Laurie and Google Security Team] 2446162911Ssimon 2447162911Ssimon *) Add AES IGE and biIGE modes. 2448162911Ssimon [Ben Laurie] 2449162911Ssimon 2450162911Ssimon *) Change the Unix randomness entropy gathering to use poll() when 2451162911Ssimon possible instead of select(), since the latter has some 2452162911Ssimon undesirable limitations. 2453162911Ssimon [Darryl Miles via Richard Levitte and Bodo Moeller] 2454162911Ssimon 2455162911Ssimon *) Disable "ECCdraft" ciphersuites more thoroughly. Now special 2456162911Ssimon treatment in ssl/ssl_ciph.s makes sure that these ciphersuites 2457162911Ssimon cannot be implicitly activated as part of, e.g., the "AES" alias. 2458162911Ssimon However, please upgrade to OpenSSL 0.9.9[-dev] for 2459162911Ssimon non-experimental use of the ECC ciphersuites to get TLS extension 2460162911Ssimon support, which is required for curve and point format negotiation 2461162911Ssimon to avoid potential handshake problems. 2462162911Ssimon [Bodo Moeller] 2463162911Ssimon 2464162911Ssimon *) Disable rogue ciphersuites: 2465162911Ssimon 2466162911Ssimon - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 2467162911Ssimon - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 2468162911Ssimon - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 2469162911Ssimon 2470162911Ssimon The latter two were purportedly from 2471162911Ssimon draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 2472162911Ssimon appear there. 2473162911Ssimon 2474167612Ssimon Also deactivate the remaining ciphersuites from 2475162911Ssimon draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 2476162911Ssimon unofficial, and the ID has long expired. 2477162911Ssimon [Bodo Moeller] 2478162911Ssimon 2479162911Ssimon *) Fix RSA blinding Heisenbug (problems sometimes occured on 2480162911Ssimon dual-core machines) and other potential thread-safety issues. 2481162911Ssimon [Bodo Moeller] 2482162911Ssimon 2483162911Ssimon *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key 2484162911Ssimon versions), which is now available for royalty-free use 2485162911Ssimon (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html). 2486162911Ssimon Also, add Camellia TLS ciphersuites from RFC 4132. 2487162911Ssimon 2488162911Ssimon To minimize changes between patchlevels in the OpenSSL 0.9.8 2489162911Ssimon series, Camellia remains excluded from compilation unless OpenSSL 2490162911Ssimon is configured with 'enable-camellia'. 2491162911Ssimon [NTT] 2492162911Ssimon 2493162911Ssimon *) Disable the padding bug check when compression is in use. The padding 2494162911Ssimon bug check assumes the first packet is of even length, this is not 2495162911Ssimon necessarily true if compresssion is enabled and can result in false 2496162911Ssimon positives causing handshake failure. The actual bug test is ancient 2497162911Ssimon code so it is hoped that implementations will either have fixed it by 2498162911Ssimon now or any which still have the bug do not support compression. 2499162911Ssimon [Steve Henson] 2500162911Ssimon 2501160814Ssimon Changes between 0.9.8a and 0.9.8b [04 May 2006] 2502160814Ssimon 2503160814Ssimon *) When applying a cipher rule check to see if string match is an explicit 2504160814Ssimon cipher suite and only match that one cipher suite if it is. 2505160814Ssimon [Steve Henson] 2506160814Ssimon 2507160814Ssimon *) Link in manifests for VC++ if needed. 2508160814Ssimon [Austin Ziegler <halostatue@gmail.com>] 2509160814Ssimon 2510160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 2511160814Ssimon draft-ietf-tls-ecc-12.txt with proposed changes (but without 2512160814Ssimon TLS extensions, which are supported starting with the 0.9.9 2513160814Ssimon branch, not in the OpenSSL 0.9.8 branch). 2514160814Ssimon [Douglas Stebila] 2515160814Ssimon 2516160814Ssimon *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support 2517160814Ssimon opaque EVP_CIPHER_CTX handling. 2518160814Ssimon [Steve Henson] 2519160814Ssimon 2520160814Ssimon *) Fixes and enhancements to zlib compression code. We now only use 2521160814Ssimon "zlib1.dll" and use the default __cdecl calling convention on Win32 2522160814Ssimon to conform with the standards mentioned here: 2523160814Ssimon http://www.zlib.net/DLL_FAQ.txt 2524160814Ssimon Static zlib linking now works on Windows and the new --with-zlib-include 2525160814Ssimon --with-zlib-lib options to Configure can be used to supply the location 2526160814Ssimon of the headers and library. Gracefully handle case where zlib library 2527160814Ssimon can't be loaded. 2528160814Ssimon [Steve Henson] 2529160814Ssimon 2530160814Ssimon *) Several fixes and enhancements to the OID generation code. The old code 2531160814Ssimon sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't 2532160814Ssimon handle numbers larger than ULONG_MAX, truncated printing and had a 2533160814Ssimon non standard OBJ_obj2txt() behaviour. 2534160814Ssimon [Steve Henson] 2535160814Ssimon 2536160814Ssimon *) Add support for building of engines under engine/ as shared libraries 2537160814Ssimon under VC++ build system. 2538160814Ssimon [Steve Henson] 2539160814Ssimon 2540160814Ssimon *) Corrected the numerous bugs in the Win32 path splitter in DSO. 2541160814Ssimon Hopefully, we will not see any false combination of paths any more. 2542160814Ssimon [Richard Levitte] 2543160814Ssimon 2544160814Ssimon Changes between 0.9.8 and 0.9.8a [11 Oct 2005] 2545160814Ssimon 2546160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 2547160814Ssimon (part of SSL_OP_ALL). This option used to disable the 2548160814Ssimon countermeasure against man-in-the-middle protocol-version 2549160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 2550160814Ssimon idea. (CVE-2005-2969) 2551160814Ssimon 2552160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 2553160814Ssimon for Information Security, National Institute of Advanced Industrial 2554160814Ssimon Science and Technology [AIST], Japan)] 2555160814Ssimon 2556160814Ssimon *) Add two function to clear and return the verify parameter flags. 2557160814Ssimon [Steve Henson] 2558160814Ssimon 2559160814Ssimon *) Keep cipherlists sorted in the source instead of sorting them at 2560160814Ssimon runtime, thus removing the need for a lock. 2561160814Ssimon [Nils Larsch] 2562160814Ssimon 2563160814Ssimon *) Avoid some small subgroup attacks in Diffie-Hellman. 2564160814Ssimon [Nick Mathewson and Ben Laurie] 2565160814Ssimon 2566160814Ssimon *) Add functions for well-known primes. 2567160814Ssimon [Nick Mathewson] 2568160814Ssimon 2569160814Ssimon *) Extended Windows CE support. 2570160814Ssimon [Satoshi Nakamura and Andy Polyakov] 2571160814Ssimon 2572160814Ssimon *) Initialize SSL_METHOD structures at compile time instead of during 2573160814Ssimon runtime, thus removing the need for a lock. 2574160814Ssimon [Steve Henson] 2575160814Ssimon 2576160814Ssimon *) Make PKCS7_decrypt() work even if no certificate is supplied by 2577160814Ssimon attempting to decrypt each encrypted key in turn. Add support to 2578160814Ssimon smime utility. 2579160814Ssimon [Steve Henson] 2580160814Ssimon 2581160814Ssimon Changes between 0.9.7h and 0.9.8 [05 Jul 2005] 2582160814Ssimon 2583162911Ssimon [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after 2584162911Ssimon OpenSSL 0.9.8.] 2585162911Ssimon 2586160814Ssimon *) Add libcrypto.pc and libssl.pc for those who feel they need them. 2587160814Ssimon [Richard Levitte] 2588160814Ssimon 2589160814Ssimon *) Change CA.sh and CA.pl so they don't bundle the CSR and the private 2590160814Ssimon key into the same file any more. 2591160814Ssimon [Richard Levitte] 2592160814Ssimon 2593160814Ssimon *) Add initial support for Win64, both IA64 and AMD64/x64 flavors. 2594160814Ssimon [Andy Polyakov] 2595160814Ssimon 2596160814Ssimon *) Add -utf8 command line and config file option to 'ca'. 2597160814Ssimon [Stefan <stf@udoma.org] 2598160814Ssimon 2599160814Ssimon *) Removed the macro des_crypt(), as it seems to conflict with some 2600160814Ssimon libraries. Use DES_crypt(). 2601160814Ssimon [Richard Levitte] 2602160814Ssimon 2603160814Ssimon *) Correct naming of the 'chil' and '4758cca' ENGINEs. This 2604160814Ssimon involves renaming the source and generated shared-libs for 2605160814Ssimon both. The engines will accept the corrected or legacy ids 2606160814Ssimon ('ncipher' and '4758_cca' respectively) when binding. NB, 2607160814Ssimon this only applies when building 'shared'. 2608160814Ssimon [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe] 2609160814Ssimon 2610160814Ssimon *) Add attribute functions to EVP_PKEY structure. Modify 2611160814Ssimon PKCS12_create() to recognize a CSP name attribute and 2612160814Ssimon use it. Make -CSP option work again in pkcs12 utility. 2613160814Ssimon [Steve Henson] 2614160814Ssimon 2615160814Ssimon *) Add new functionality to the bn blinding code: 2616160814Ssimon - automatic re-creation of the BN_BLINDING parameters after 2617160814Ssimon a fixed number of uses (currently 32) 2618160814Ssimon - add new function for parameter creation 2619160814Ssimon - introduce flags to control the update behaviour of the 2620160814Ssimon BN_BLINDING parameters 2621160814Ssimon - hide BN_BLINDING structure 2622160814Ssimon Add a second BN_BLINDING slot to the RSA structure to improve 2623160814Ssimon performance when a single RSA object is shared among several 2624160814Ssimon threads. 2625160814Ssimon [Nils Larsch] 2626160814Ssimon 2627160814Ssimon *) Add support for DTLS. 2628160814Ssimon [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie] 2629160814Ssimon 2630160814Ssimon *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1) 2631160814Ssimon to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file() 2632160814Ssimon [Walter Goulet] 2633160814Ssimon 2634160814Ssimon *) Remove buggy and incompletet DH cert support from 2635160814Ssimon ssl/ssl_rsa.c and ssl/s3_both.c 2636160814Ssimon [Nils Larsch] 2637160814Ssimon 2638160814Ssimon *) Use SHA-1 instead of MD5 as the default digest algorithm for 2639160814Ssimon the apps/openssl applications. 2640160814Ssimon [Nils Larsch] 2641160814Ssimon 2642160814Ssimon *) Compile clean with "-Wall -Wmissing-prototypes 2643160814Ssimon -Wstrict-prototypes -Wmissing-declarations -Werror". Currently 2644160814Ssimon DEBUG_SAFESTACK must also be set. 2645160814Ssimon [Ben Laurie] 2646160814Ssimon 2647160814Ssimon *) Change ./Configure so that certain algorithms can be disabled by default. 2648160814Ssimon The new counterpiece to "no-xxx" is "enable-xxx". 2649160814Ssimon 2650160814Ssimon The patented RC5 and MDC2 algorithms will now be disabled unless 2651160814Ssimon "enable-rc5" and "enable-mdc2", respectively, are specified. 2652160814Ssimon 2653160814Ssimon (IDEA remains enabled despite being patented. This is because IDEA 2654160814Ssimon is frequently required for interoperability, and there is no license 2655160814Ssimon fee for non-commercial use. As before, "no-idea" can be used to 2656160814Ssimon avoid this algorithm.) 2657160814Ssimon 2658160814Ssimon [Bodo Moeller] 2659160814Ssimon 2660160814Ssimon *) Add processing of proxy certificates (see RFC 3820). This work was 2661160814Ssimon sponsored by KTH (The Royal Institute of Technology in Stockholm) and 2662160814Ssimon EGEE (Enabling Grids for E-science in Europe). 2663160814Ssimon [Richard Levitte] 2664160814Ssimon 2665160814Ssimon *) RC4 performance overhaul on modern architectures/implementations, such 2666160814Ssimon as Intel P4, IA-64 and AMD64. 2667160814Ssimon [Andy Polyakov] 2668160814Ssimon 2669160814Ssimon *) New utility extract-section.pl. This can be used specify an alternative 2670160814Ssimon section number in a pod file instead of having to treat each file as 2671160814Ssimon a separate case in Makefile. This can be done by adding two lines to the 2672160814Ssimon pod file: 2673160814Ssimon 2674160814Ssimon =for comment openssl_section:XXX 2675160814Ssimon 2676160814Ssimon The blank line is mandatory. 2677160814Ssimon 2678160814Ssimon [Steve Henson] 2679160814Ssimon 2680160814Ssimon *) New arguments -certform, -keyform and -pass for s_client and s_server 2681160814Ssimon to allow alternative format key and certificate files and passphrase 2682160814Ssimon sources. 2683160814Ssimon [Steve Henson] 2684160814Ssimon 2685160814Ssimon *) New structure X509_VERIFY_PARAM which combines current verify parameters, 2686160814Ssimon update associated structures and add various utility functions. 2687160814Ssimon 2688160814Ssimon Add new policy related verify parameters, include policy checking in 2689160814Ssimon standard verify code. Enhance 'smime' application with extra parameters 2690160814Ssimon to support policy checking and print out. 2691160814Ssimon [Steve Henson] 2692160814Ssimon 2693160814Ssimon *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3 2694160814Ssimon Nehemiah processors. These extensions support AES encryption in hardware 2695160814Ssimon as well as RNG (though RNG support is currently disabled). 2696160814Ssimon [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov] 2697160814Ssimon 2698160814Ssimon *) Deprecate BN_[get|set]_params() functions (they were ignored internally). 2699160814Ssimon [Geoff Thorpe] 2700160814Ssimon 2701160814Ssimon *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented. 2702160814Ssimon [Andy Polyakov and a number of other people] 2703160814Ssimon 2704160814Ssimon *) Improved PowerPC platform support. Most notably BIGNUM assembler 2705160814Ssimon implementation contributed by IBM. 2706160814Ssimon [Suresh Chari, Peter Waltenberg, Andy Polyakov] 2707160814Ssimon 2708160814Ssimon *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public 2709160814Ssimon exponent rather than 'unsigned long'. There is a corresponding change to 2710160814Ssimon the new 'rsa_keygen' element of the RSA_METHOD structure. 2711160814Ssimon [Jelte Jansen, Geoff Thorpe] 2712160814Ssimon 2713160814Ssimon *) Functionality for creating the initial serial number file is now 2714160814Ssimon moved from CA.pl to the 'ca' utility with a new option -create_serial. 2715160814Ssimon 2716160814Ssimon (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial 2717160814Ssimon number file to 1, which is bound to cause problems. To avoid 2718160814Ssimon the problems while respecting compatibility between different 0.9.7 2719160814Ssimon patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in 2720160814Ssimon CA.pl for serial number initialization. With the new release 0.9.8, 2721160814Ssimon we can fix the problem directly in the 'ca' utility.) 2722160814Ssimon [Steve Henson] 2723160814Ssimon 2724160814Ssimon *) Reduced header interdepencies by declaring more opaque objects in 2725160814Ssimon ossl_typ.h. As a consequence, including some headers (eg. engine.h) will 2726160814Ssimon give fewer recursive includes, which could break lazy source code - so 2727160814Ssimon this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always, 2728160814Ssimon developers should define this symbol when building and using openssl to 2729160814Ssimon ensure they track the recommended behaviour, interfaces, [etc], but 2730160814Ssimon backwards-compatible behaviour prevails when this isn't defined. 2731160814Ssimon [Geoff Thorpe] 2732160814Ssimon 2733160814Ssimon *) New function X509_POLICY_NODE_print() which prints out policy nodes. 2734160814Ssimon [Steve Henson] 2735160814Ssimon 2736160814Ssimon *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality. 2737160814Ssimon This will generate a random key of the appropriate length based on the 2738160814Ssimon cipher context. The EVP_CIPHER can provide its own random key generation 2739160814Ssimon routine to support keys of a specific form. This is used in the des and 2740160814Ssimon 3des routines to generate a key of the correct parity. Update S/MIME 2741160814Ssimon code to use new functions and hence generate correct parity DES keys. 2742160814Ssimon Add EVP_CHECK_DES_KEY #define to return an error if the key is not 2743160814Ssimon valid (weak or incorrect parity). 2744160814Ssimon [Steve Henson] 2745160814Ssimon 2746160814Ssimon *) Add a local set of CRLs that can be used by X509_verify_cert() as well 2747160814Ssimon as looking them up. This is useful when the verified structure may contain 2748160814Ssimon CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs 2749160814Ssimon present unless the new PKCS7_NO_CRL flag is asserted. 2750160814Ssimon [Steve Henson] 2751160814Ssimon 2752160814Ssimon *) Extend ASN1 oid configuration module. It now additionally accepts the 2753160814Ssimon syntax: 2754160814Ssimon 2755160814Ssimon shortName = some long name, 1.2.3.4 2756160814Ssimon [Steve Henson] 2757160814Ssimon 2758160814Ssimon *) Reimplemented the BN_CTX implementation. There is now no more static 2759160814Ssimon limitation on the number of variables it can handle nor the depth of the 2760160814Ssimon "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack 2761160814Ssimon information can now expand as required, and rather than having a single 2762160814Ssimon static array of bignums, BN_CTX now uses a linked-list of such arrays 2763160814Ssimon allowing it to expand on demand whilst maintaining the usefulness of 2764160814Ssimon BN_CTX's "bundling". 2765160814Ssimon [Geoff Thorpe] 2766160814Ssimon 2767160814Ssimon *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD 2768160814Ssimon to allow all RSA operations to function using a single BN_CTX. 2769160814Ssimon [Geoff Thorpe] 2770160814Ssimon 2771160814Ssimon *) Preliminary support for certificate policy evaluation and checking. This 2772160814Ssimon is initially intended to pass the tests outlined in "Conformance Testing 2773160814Ssimon of Relying Party Client Certificate Path Processing Logic" v1.07. 2774160814Ssimon [Steve Henson] 2775160814Ssimon 2776160814Ssimon *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and 2777160814Ssimon remained unused and not that useful. A variety of other little bignum 2778160814Ssimon tweaks and fixes have also been made continuing on from the audit (see 2779160814Ssimon below). 2780160814Ssimon [Geoff Thorpe] 2781160814Ssimon 2782160814Ssimon *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with 2783160814Ssimon associated ASN1, EVP and SSL functions and old ASN1 macros. 2784160814Ssimon [Richard Levitte] 2785160814Ssimon 2786160814Ssimon *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results, 2787160814Ssimon and this should never fail. So the return value from the use of 2788160814Ssimon BN_set_word() (which can fail due to needless expansion) is now deprecated; 2789160814Ssimon if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro. 2790160814Ssimon [Geoff Thorpe] 2791160814Ssimon 2792160814Ssimon *) BN_CTX_get() should return zero-valued bignums, providing the same 2793160814Ssimon initialised value as BN_new(). 2794160814Ssimon [Geoff Thorpe, suggested by Ulf M�ller] 2795160814Ssimon 2796160814Ssimon *) Support for inhibitAnyPolicy certificate extension. 2797160814Ssimon [Steve Henson] 2798160814Ssimon 2799160814Ssimon *) An audit of the BIGNUM code is underway, for which debugging code is 2800160814Ssimon enabled when BN_DEBUG is defined. This makes stricter enforcements on what 2801160814Ssimon is considered valid when processing BIGNUMs, and causes execution to 2802160814Ssimon assert() when a problem is discovered. If BN_DEBUG_RAND is defined, 2803160814Ssimon further steps are taken to deliberately pollute unused data in BIGNUM 2804160814Ssimon structures to try and expose faulty code further on. For now, openssl will 2805160814Ssimon (in its default mode of operation) continue to tolerate the inconsistent 2806160814Ssimon forms that it has tolerated in the past, but authors and packagers should 2807160814Ssimon consider trying openssl and their own applications when compiled with 2808160814Ssimon these debugging symbols defined. It will help highlight potential bugs in 2809160814Ssimon their own code, and will improve the test coverage for OpenSSL itself. At 2810160814Ssimon some point, these tighter rules will become openssl's default to improve 2811160814Ssimon maintainability, though the assert()s and other overheads will remain only 2812160814Ssimon in debugging configurations. See bn.h for more details. 2813160814Ssimon [Geoff Thorpe, Nils Larsch, Ulf M�ller] 2814160814Ssimon 2815160814Ssimon *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure 2816160814Ssimon that can only be obtained through BN_CTX_new() (which implicitly 2817160814Ssimon initialises it). The presence of this function only made it possible 2818160814Ssimon to overwrite an existing structure (and cause memory leaks). 2819160814Ssimon [Geoff Thorpe] 2820160814Ssimon 2821160814Ssimon *) Because of the callback-based approach for implementing LHASH as a 2822160814Ssimon template type, lh_insert() adds opaque objects to hash-tables and 2823160814Ssimon lh_doall() or lh_doall_arg() are typically used with a destructor callback 2824160814Ssimon to clean up those corresponding objects before destroying the hash table 2825160814Ssimon (and losing the object pointers). So some over-zealous constifications in 2826160814Ssimon LHASH have been relaxed so that lh_insert() does not take (nor store) the 2827160814Ssimon objects as "const" and the lh_doall[_arg] callback wrappers are not 2828160814Ssimon prototyped to have "const" restrictions on the object pointers they are 2829160814Ssimon given (and so aren't required to cast them away any more). 2830160814Ssimon [Geoff Thorpe] 2831160814Ssimon 2832160814Ssimon *) The tmdiff.h API was so ugly and minimal that our own timing utility 2833160814Ssimon (speed) prefers to use its own implementation. The two implementations 2834160814Ssimon haven't been consolidated as yet (volunteers?) but the tmdiff API has had 2835160814Ssimon its object type properly exposed (MS_TM) instead of casting to/from "char 2836160814Ssimon *". This may still change yet if someone realises MS_TM and "ms_time_***" 2837160814Ssimon aren't necessarily the greatest nomenclatures - but this is what was used 2838160814Ssimon internally to the implementation so I've used that for now. 2839160814Ssimon [Geoff Thorpe] 2840160814Ssimon 2841160814Ssimon *) Ensure that deprecated functions do not get compiled when 2842160814Ssimon OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of 2843160814Ssimon the self-tests were still using deprecated key-generation functions so 2844160814Ssimon these have been updated also. 2845160814Ssimon [Geoff Thorpe] 2846160814Ssimon 2847160814Ssimon *) Reorganise PKCS#7 code to separate the digest location functionality 2848160814Ssimon into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest(). 2849160814Ssimon New function PKCS7_set_digest() to set the digest type for PKCS#7 2850160814Ssimon digestedData type. Add additional code to correctly generate the 2851160814Ssimon digestedData type and add support for this type in PKCS7 initialization 2852160814Ssimon functions. 2853160814Ssimon [Steve Henson] 2854160814Ssimon 2855160814Ssimon *) New function PKCS7_set0_type_other() this initializes a PKCS7 2856160814Ssimon structure of type "other". 2857160814Ssimon [Steve Henson] 2858160814Ssimon 2859160814Ssimon *) Fix prime generation loop in crypto/bn/bn_prime.pl by making 2860160814Ssimon sure the loop does correctly stop and breaking ("division by zero") 2861160814Ssimon modulus operations are not performed. The (pre-generated) prime 2862160814Ssimon table crypto/bn/bn_prime.h was already correct, but it could not be 2863160814Ssimon re-generated on some platforms because of the "division by zero" 2864160814Ssimon situation in the script. 2865160814Ssimon [Ralf S. Engelschall] 2866160814Ssimon 2867160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 2868160814Ssimon draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with 2869160814Ssimon SHA-1 now is only used for "small" curves (where the 2870160814Ssimon representation of a field element takes up to 24 bytes); for 2871160814Ssimon larger curves, the field element resulting from ECDH is directly 2872160814Ssimon used as premaster secret. 2873160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 2874160814Ssimon 2875160814Ssimon *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2 2876160814Ssimon curve secp160r1 to the tests. 2877160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 2878160814Ssimon 2879160814Ssimon *) Add the possibility to load symbols globally with DSO. 2880160814Ssimon [G�tz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte] 2881160814Ssimon 2882160814Ssimon *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better 2883160814Ssimon control of the error stack. 2884160814Ssimon [Richard Levitte] 2885160814Ssimon 2886160814Ssimon *) Add support for STORE in ENGINE. 2887160814Ssimon [Richard Levitte] 2888160814Ssimon 2889160814Ssimon *) Add the STORE type. The intention is to provide a common interface 2890160814Ssimon to certificate and key stores, be they simple file-based stores, or 2891160814Ssimon HSM-type store, or LDAP stores, or... 2892160814Ssimon NOTE: The code is currently UNTESTED and isn't really used anywhere. 2893160814Ssimon [Richard Levitte] 2894160814Ssimon 2895160814Ssimon *) Add a generic structure called OPENSSL_ITEM. This can be used to 2896160814Ssimon pass a list of arguments to any function as well as provide a way 2897160814Ssimon for a function to pass data back to the caller. 2898160814Ssimon [Richard Levitte] 2899160814Ssimon 2900160814Ssimon *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup() 2901160814Ssimon works like BUF_strdup() but can be used to duplicate a portion of 2902160814Ssimon a string. The copy gets NUL-terminated. BUF_memdup() duplicates 2903160814Ssimon a memory area. 2904160814Ssimon [Richard Levitte] 2905160814Ssimon 2906160814Ssimon *) Add the function sk_find_ex() which works like sk_find(), but will 2907160814Ssimon return an index to an element even if an exact match couldn't be 2908160814Ssimon found. The index is guaranteed to point at the element where the 2909160814Ssimon searched-for key would be inserted to preserve sorting order. 2910160814Ssimon [Richard Levitte] 2911160814Ssimon 2912160814Ssimon *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but 2913160814Ssimon takes an extra flags argument for optional functionality. Currently, 2914160814Ssimon the following flags are defined: 2915160814Ssimon 2916160814Ssimon OBJ_BSEARCH_VALUE_ON_NOMATCH 2917160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 2918160814Ssimon element where the comparing function returns a negative or zero 2919160814Ssimon number. 2920160814Ssimon 2921160814Ssimon OBJ_BSEARCH_FIRST_VALUE_ON_MATCH 2922160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 2923160814Ssimon element where the comparing function returns zero. This is useful 2924160814Ssimon if there are more than one element where the comparing function 2925160814Ssimon returns zero. 2926160814Ssimon [Richard Levitte] 2927160814Ssimon 2928160814Ssimon *) Make it possible to create self-signed certificates with 'openssl ca' 2929160814Ssimon in such a way that the self-signed certificate becomes part of the 2930160814Ssimon CA database and uses the same mechanisms for serial number generation 2931160814Ssimon as all other certificate signing. The new flag '-selfsign' enables 2932160814Ssimon this functionality. Adapt CA.sh and CA.pl.in. 2933160814Ssimon [Richard Levitte] 2934160814Ssimon 2935160814Ssimon *) Add functionality to check the public key of a certificate request 2936160814Ssimon against a given private. This is useful to check that a certificate 2937160814Ssimon request can be signed by that key (self-signing). 2938160814Ssimon [Richard Levitte] 2939160814Ssimon 2940160814Ssimon *) Make it possible to have multiple active certificates with the same 2941160814Ssimon subject in the CA index file. This is done only if the keyword 2942160814Ssimon 'unique_subject' is set to 'no' in the main CA section (default 2943160814Ssimon if 'CA_default') of the configuration file. The value is saved 2944160814Ssimon with the database itself in a separate index attribute file, 2945160814Ssimon named like the index file with '.attr' appended to the name. 2946160814Ssimon [Richard Levitte] 2947160814Ssimon 2948160814Ssimon *) Generate muti valued AVAs using '+' notation in config files for 2949160814Ssimon req and dirName. 2950160814Ssimon [Steve Henson] 2951160814Ssimon 2952160814Ssimon *) Support for nameConstraints certificate extension. 2953160814Ssimon [Steve Henson] 2954160814Ssimon 2955160814Ssimon *) Support for policyConstraints certificate extension. 2956160814Ssimon [Steve Henson] 2957160814Ssimon 2958160814Ssimon *) Support for policyMappings certificate extension. 2959160814Ssimon [Steve Henson] 2960160814Ssimon 2961160814Ssimon *) Make sure the default DSA_METHOD implementation only uses its 2962160814Ssimon dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL, 2963160814Ssimon and change its own handlers to be NULL so as to remove unnecessary 2964160814Ssimon indirection. This lets alternative implementations fallback to the 2965160814Ssimon default implementation more easily. 2966160814Ssimon [Geoff Thorpe] 2967160814Ssimon 2968160814Ssimon *) Support for directoryName in GeneralName related extensions 2969160814Ssimon in config files. 2970160814Ssimon [Steve Henson] 2971160814Ssimon 2972160814Ssimon *) Make it possible to link applications using Makefile.shared. 2973160814Ssimon Make that possible even when linking against static libraries! 2974160814Ssimon [Richard Levitte] 2975160814Ssimon 2976160814Ssimon *) Support for single pass processing for S/MIME signing. This now 2977160814Ssimon means that S/MIME signing can be done from a pipe, in addition 2978160814Ssimon cleartext signing (multipart/signed type) is effectively streaming 2979160814Ssimon and the signed data does not need to be all held in memory. 2980160814Ssimon 2981160814Ssimon This is done with a new flag PKCS7_STREAM. When this flag is set 2982160814Ssimon PKCS7_sign() only initializes the PKCS7 structure and the actual signing 2983160814Ssimon is done after the data is output (and digests calculated) in 2984160814Ssimon SMIME_write_PKCS7(). 2985160814Ssimon [Steve Henson] 2986160814Ssimon 2987160814Ssimon *) Add full support for -rpath/-R, both in shared libraries and 2988160814Ssimon applications, at least on the platforms where it's known how 2989160814Ssimon to do it. 2990160814Ssimon [Richard Levitte] 2991160814Ssimon 2992160814Ssimon *) In crypto/ec/ec_mult.c, implement fast point multiplication with 2993160814Ssimon precomputation, based on wNAF splitting: EC_GROUP_precompute_mult() 2994160814Ssimon will now compute a table of multiples of the generator that 2995160814Ssimon makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul() 2996160814Ssimon faster (notably in the case of a single point multiplication, 2997160814Ssimon scalar * generator). 2998160814Ssimon [Nils Larsch, Bodo Moeller] 2999160814Ssimon 3000160814Ssimon *) IPv6 support for certificate extensions. The various extensions 3001160814Ssimon which use the IP:a.b.c.d can now take IPv6 addresses using the 3002160814Ssimon formats of RFC1884 2.2 . IPv6 addresses are now also displayed 3003160814Ssimon correctly. 3004160814Ssimon [Steve Henson] 3005160814Ssimon 3006160814Ssimon *) Added an ENGINE that implements RSA by performing private key 3007160814Ssimon exponentiations with the GMP library. The conversions to and from 3008160814Ssimon GMP's mpz_t format aren't optimised nor are any montgomery forms 3009160814Ssimon cached, and on x86 it appears OpenSSL's own performance has caught up. 3010160814Ssimon However there are likely to be other architectures where GMP could 3011160814Ssimon provide a boost. This ENGINE is not built in by default, but it can be 3012160814Ssimon specified at Configure time and should be accompanied by the necessary 3013160814Ssimon linker additions, eg; 3014160814Ssimon ./config -DOPENSSL_USE_GMP -lgmp 3015160814Ssimon [Geoff Thorpe] 3016160814Ssimon 3017160814Ssimon *) "openssl engine" will not display ENGINE/DSO load failure errors when 3018160814Ssimon testing availability of engines with "-t" - the old behaviour is 3019160814Ssimon produced by increasing the feature's verbosity with "-tt". 3020160814Ssimon [Geoff Thorpe] 3021160814Ssimon 3022160814Ssimon *) ECDSA routines: under certain error conditions uninitialized BN objects 3023160814Ssimon could be freed. Solution: make sure initialization is performed early 3024160814Ssimon enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de> 3025160814Ssimon via PR#459) 3026160814Ssimon [Lutz Jaenicke] 3027160814Ssimon 3028160814Ssimon *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD 3029160814Ssimon and DH_METHOD (eg. by ENGINE implementations) to override the normal 3030160814Ssimon software implementations. For DSA and DH, parameter generation can 3031160814Ssimon also be overriden by providing the appropriate method callbacks. 3032160814Ssimon [Geoff Thorpe] 3033160814Ssimon 3034160814Ssimon *) Change the "progress" mechanism used in key-generation and 3035160814Ssimon primality testing to functions that take a new BN_GENCB pointer in 3036160814Ssimon place of callback/argument pairs. The new API functions have "_ex" 3037160814Ssimon postfixes and the older functions are reimplemented as wrappers for 3038160814Ssimon the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide 3039160814Ssimon declarations of the old functions to help (graceful) attempts to 3040160814Ssimon migrate to the new functions. Also, the new key-generation API 3041160814Ssimon functions operate on a caller-supplied key-structure and return 3042160814Ssimon success/failure rather than returning a key or NULL - this is to 3043160814Ssimon help make "keygen" another member function of RSA_METHOD etc. 3044160814Ssimon 3045160814Ssimon Example for using the new callback interface: 3046160814Ssimon 3047160814Ssimon int (*my_callback)(int a, int b, BN_GENCB *cb) = ...; 3048160814Ssimon void *my_arg = ...; 3049160814Ssimon BN_GENCB my_cb; 3050160814Ssimon 3051160814Ssimon BN_GENCB_set(&my_cb, my_callback, my_arg); 3052160814Ssimon 3053160814Ssimon return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb); 3054160814Ssimon /* For the meaning of a, b in calls to my_callback(), see the 3055160814Ssimon * documentation of the function that calls the callback. 3056160814Ssimon * cb will point to my_cb; my_arg can be retrieved as cb->arg. 3057160814Ssimon * my_callback should return 1 if it wants BN_is_prime_ex() 3058160814Ssimon * to continue, or 0 to stop. 3059160814Ssimon */ 3060160814Ssimon 3061160814Ssimon [Geoff Thorpe] 3062160814Ssimon 3063160814Ssimon *) Change the ZLIB compression method to be stateful, and make it 3064160814Ssimon available to TLS with the number defined in 3065160814Ssimon draft-ietf-tls-compression-04.txt. 3066160814Ssimon [Richard Levitte] 3067160814Ssimon 3068160814Ssimon *) Add the ASN.1 structures and functions for CertificatePair, which 3069160814Ssimon is defined as follows (according to X.509_4thEditionDraftV6.pdf): 3070160814Ssimon 3071160814Ssimon CertificatePair ::= SEQUENCE { 3072160814Ssimon forward [0] Certificate OPTIONAL, 3073160814Ssimon reverse [1] Certificate OPTIONAL, 3074160814Ssimon -- at least one of the pair shall be present -- } 3075160814Ssimon 3076160814Ssimon Also implement the PEM functions to read and write certificate 3077160814Ssimon pairs, and defined the PEM tag as "CERTIFICATE PAIR". 3078160814Ssimon 3079160814Ssimon This needed to be defined, mostly for the sake of the LDAP 3080160814Ssimon attribute crossCertificatePair, but may prove useful elsewhere as 3081160814Ssimon well. 3082160814Ssimon [Richard Levitte] 3083160814Ssimon 3084160814Ssimon *) Make it possible to inhibit symlinking of shared libraries in 3085160814Ssimon Makefile.shared, for Cygwin's sake. 3086160814Ssimon [Richard Levitte] 3087160814Ssimon 3088160814Ssimon *) Extend the BIGNUM API by creating a function 3089160814Ssimon void BN_set_negative(BIGNUM *a, int neg); 3090160814Ssimon and a macro that behave like 3091160814Ssimon int BN_is_negative(const BIGNUM *a); 3092160814Ssimon 3093160814Ssimon to avoid the need to access 'a->neg' directly in applications. 3094160814Ssimon [Nils Larsch] 3095160814Ssimon 3096160814Ssimon *) Implement fast modular reduction for pseudo-Mersenne primes 3097160814Ssimon used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c). 3098160814Ssimon EC_GROUP_new_curve_GFp() will now automatically use this 3099160814Ssimon if applicable. 3100160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3101160814Ssimon 3102160814Ssimon *) Add new lock type (CRYPTO_LOCK_BN). 3103160814Ssimon [Bodo Moeller] 3104160814Ssimon 3105160814Ssimon *) Change the ENGINE framework to automatically load engines 3106160814Ssimon dynamically from specific directories unless they could be 3107160814Ssimon found to already be built in or loaded. Move all the 3108160814Ssimon current engines except for the cryptodev one to a new 3109160814Ssimon directory engines/. 3110160814Ssimon The engines in engines/ are built as shared libraries if 3111160814Ssimon the "shared" options was given to ./Configure or ./config. 3112160814Ssimon Otherwise, they are inserted in libcrypto.a. 3113160814Ssimon /usr/local/ssl/engines is the default directory for dynamic 3114160814Ssimon engines, but that can be overriden at configure time through 3115160814Ssimon the usual use of --prefix and/or --openssldir, and at run 3116160814Ssimon time with the environment variable OPENSSL_ENGINES. 3117160814Ssimon [Geoff Thorpe and Richard Levitte] 3118160814Ssimon 3119160814Ssimon *) Add Makefile.shared, a helper makefile to build shared 3120160814Ssimon libraries. Addapt Makefile.org. 3121160814Ssimon [Richard Levitte] 3122160814Ssimon 3123160814Ssimon *) Add version info to Win32 DLLs. 3124160814Ssimon [Peter 'Luna' Runestig" <peter@runestig.com>] 3125160814Ssimon 3126160814Ssimon *) Add new 'medium level' PKCS#12 API. Certificates and keys 3127160814Ssimon can be added using this API to created arbitrary PKCS#12 3128160814Ssimon files while avoiding the low level API. 3129160814Ssimon 3130160814Ssimon New options to PKCS12_create(), key or cert can be NULL and 3131160814Ssimon will then be omitted from the output file. The encryption 3132160814Ssimon algorithm NIDs can be set to -1 for no encryption, the mac 3133160814Ssimon iteration count can be set to 0 to omit the mac. 3134160814Ssimon 3135160814Ssimon Enhance pkcs12 utility by making the -nokeys and -nocerts 3136160814Ssimon options work when creating a PKCS#12 file. New option -nomac 3137160814Ssimon to omit the mac, NONE can be set for an encryption algorithm. 3138160814Ssimon New code is modified to use the enhanced PKCS12_create() 3139160814Ssimon instead of the low level API. 3140160814Ssimon [Steve Henson] 3141160814Ssimon 3142160814Ssimon *) Extend ASN1 encoder to support indefinite length constructed 3143160814Ssimon encoding. This can output sequences tags and octet strings in 3144160814Ssimon this form. Modify pk7_asn1.c to support indefinite length 3145160814Ssimon encoding. This is experimental and needs additional code to 3146160814Ssimon be useful, such as an ASN1 bio and some enhanced streaming 3147160814Ssimon PKCS#7 code. 3148160814Ssimon 3149160814Ssimon Extend template encode functionality so that tagging is passed 3150160814Ssimon down to the template encoder. 3151160814Ssimon [Steve Henson] 3152160814Ssimon 3153160814Ssimon *) Let 'openssl req' fail if an argument to '-newkey' is not 3154160814Ssimon recognized instead of using RSA as a default. 3155160814Ssimon [Bodo Moeller] 3156160814Ssimon 3157160814Ssimon *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt. 3158160814Ssimon As these are not official, they are not included in "ALL"; 3159160814Ssimon the "ECCdraft" ciphersuite group alias can be used to select them. 3160160814Ssimon [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)] 3161160814Ssimon 3162160814Ssimon *) Add ECDH engine support. 3163160814Ssimon [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)] 3164160814Ssimon 3165160814Ssimon *) Add ECDH in new directory crypto/ecdh/. 3166160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 3167160814Ssimon 3168160814Ssimon *) Let BN_rand_range() abort with an error after 100 iterations 3169160814Ssimon without success (which indicates a broken PRNG). 3170160814Ssimon [Bodo Moeller] 3171160814Ssimon 3172160814Ssimon *) Change BN_mod_sqrt() so that it verifies that the input value 3173160814Ssimon is really the square of the return value. (Previously, 3174160814Ssimon BN_mod_sqrt would show GIGO behaviour.) 3175160814Ssimon [Bodo Moeller] 3176160814Ssimon 3177160814Ssimon *) Add named elliptic curves over binary fields from X9.62, SECG, 3178160814Ssimon and WAP/WTLS; add OIDs that were still missing. 3179160814Ssimon 3180160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3181160814Ssimon (Sun Microsystems Laboratories)] 3182160814Ssimon 3183160814Ssimon *) Extend the EC library for elliptic curves over binary fields 3184160814Ssimon (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/). 3185160814Ssimon New EC_METHOD: 3186160814Ssimon 3187160814Ssimon EC_GF2m_simple_method 3188160814Ssimon 3189160814Ssimon New API functions: 3190160814Ssimon 3191160814Ssimon EC_GROUP_new_curve_GF2m 3192160814Ssimon EC_GROUP_set_curve_GF2m 3193160814Ssimon EC_GROUP_get_curve_GF2m 3194160814Ssimon EC_POINT_set_affine_coordinates_GF2m 3195160814Ssimon EC_POINT_get_affine_coordinates_GF2m 3196160814Ssimon EC_POINT_set_compressed_coordinates_GF2m 3197160814Ssimon 3198160814Ssimon Point compression for binary fields is disabled by default for 3199160814Ssimon patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to 3200160814Ssimon enable it). 3201160814Ssimon 3202160814Ssimon As binary polynomials are represented as BIGNUMs, various members 3203160814Ssimon of the EC_GROUP and EC_POINT data structures can be shared 3204160814Ssimon between the implementations for prime fields and binary fields; 3205160814Ssimon the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m) 3206160814Ssimon are essentially identical to their ..._GFp counterparts. 3207160814Ssimon (For simplicity, the '..._GFp' prefix has been dropped from 3208160814Ssimon various internal method names.) 3209160814Ssimon 3210160814Ssimon An internal 'field_div' method (similar to 'field_mul' and 3211160814Ssimon 'field_sqr') has been added; this is used only for binary fields. 3212160814Ssimon 3213160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3214160814Ssimon (Sun Microsystems Laboratories)] 3215160814Ssimon 3216160814Ssimon *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult() 3217160814Ssimon through methods ('mul', 'precompute_mult'). 3218160814Ssimon 3219160814Ssimon The generic implementations (now internally called 'ec_wNAF_mul' 3220160814Ssimon and 'ec_wNAF_precomputed_mult') remain the default if these 3221160814Ssimon methods are undefined. 3222160814Ssimon 3223160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3224160814Ssimon (Sun Microsystems Laboratories)] 3225160814Ssimon 3226160814Ssimon *) New function EC_GROUP_get_degree, which is defined through 3227160814Ssimon EC_METHOD. For curves over prime fields, this returns the bit 3228160814Ssimon length of the modulus. 3229160814Ssimon 3230160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3231160814Ssimon (Sun Microsystems Laboratories)] 3232160814Ssimon 3233160814Ssimon *) New functions EC_GROUP_dup, EC_POINT_dup. 3234160814Ssimon (These simply call ..._new and ..._copy). 3235160814Ssimon 3236160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3237160814Ssimon (Sun Microsystems Laboratories)] 3238160814Ssimon 3239160814Ssimon *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c. 3240160814Ssimon Polynomials are represented as BIGNUMs (where the sign bit is not 3241160814Ssimon used) in the following functions [macros]: 3242160814Ssimon 3243160814Ssimon BN_GF2m_add 3244160814Ssimon BN_GF2m_sub [= BN_GF2m_add] 3245160814Ssimon BN_GF2m_mod [wrapper for BN_GF2m_mod_arr] 3246160814Ssimon BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr] 3247160814Ssimon BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr] 3248160814Ssimon BN_GF2m_mod_inv 3249160814Ssimon BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr] 3250160814Ssimon BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr] 3251160814Ssimon BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr] 3252160814Ssimon BN_GF2m_cmp [= BN_ucmp] 3253160814Ssimon 3254160814Ssimon (Note that only the 'mod' functions are actually for fields GF(2^m). 3255160814Ssimon BN_GF2m_add() is misnomer, but this is for the sake of consistency.) 3256160814Ssimon 3257160814Ssimon For some functions, an the irreducible polynomial defining a 3258160814Ssimon field can be given as an 'unsigned int[]' with strictly 3259160814Ssimon decreasing elements giving the indices of those bits that are set; 3260160814Ssimon i.e., p[] represents the polynomial 3261160814Ssimon f(t) = t^p[0] + t^p[1] + ... + t^p[k] 3262160814Ssimon where 3263160814Ssimon p[0] > p[1] > ... > p[k] = 0. 3264160814Ssimon This applies to the following functions: 3265160814Ssimon 3266160814Ssimon BN_GF2m_mod_arr 3267160814Ssimon BN_GF2m_mod_mul_arr 3268160814Ssimon BN_GF2m_mod_sqr_arr 3269160814Ssimon BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv] 3270160814Ssimon BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div] 3271160814Ssimon BN_GF2m_mod_exp_arr 3272160814Ssimon BN_GF2m_mod_sqrt_arr 3273160814Ssimon BN_GF2m_mod_solve_quad_arr 3274160814Ssimon BN_GF2m_poly2arr 3275160814Ssimon BN_GF2m_arr2poly 3276160814Ssimon 3277160814Ssimon Conversion can be performed by the following functions: 3278160814Ssimon 3279160814Ssimon BN_GF2m_poly2arr 3280160814Ssimon BN_GF2m_arr2poly 3281160814Ssimon 3282160814Ssimon bntest.c has additional tests for binary polynomial arithmetic. 3283160814Ssimon 3284160814Ssimon Two implementations for BN_GF2m_mod_div() are available. 3285160814Ssimon The default algorithm simply uses BN_GF2m_mod_inv() and 3286160814Ssimon BN_GF2m_mod_mul(). The alternative algorithm is compiled in only 3287160814Ssimon if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the 3288160814Ssimon copyright notice in crypto/bn/bn_gf2m.c before enabling it). 3289160814Ssimon 3290160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3291160814Ssimon (Sun Microsystems Laboratories)] 3292160814Ssimon 3293160814Ssimon *) Add new error code 'ERR_R_DISABLED' that can be used when some 3294160814Ssimon functionality is disabled at compile-time. 3295160814Ssimon [Douglas Stebila <douglas.stebila@sun.com>] 3296160814Ssimon 3297160814Ssimon *) Change default behaviour of 'openssl asn1parse' so that more 3298160814Ssimon information is visible when viewing, e.g., a certificate: 3299160814Ssimon 3300160814Ssimon Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump' 3301160814Ssimon mode the content of non-printable OCTET STRINGs is output in a 3302160814Ssimon style similar to INTEGERs, but with '[HEX DUMP]' prepended to 3303160814Ssimon avoid the appearance of a printable string. 3304160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3305160814Ssimon 3306160814Ssimon *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access 3307160814Ssimon functions 3308160814Ssimon EC_GROUP_set_asn1_flag() 3309160814Ssimon EC_GROUP_get_asn1_flag() 3310160814Ssimon EC_GROUP_set_point_conversion_form() 3311160814Ssimon EC_GROUP_get_point_conversion_form() 3312160814Ssimon These control ASN1 encoding details: 3313160814Ssimon - Curves (i.e., groups) are encoded explicitly unless asn1_flag 3314160814Ssimon has been set to OPENSSL_EC_NAMED_CURVE. 3315160814Ssimon - Points are encoded in uncompressed form by default; options for 3316160814Ssimon asn1_for are as for point2oct, namely 3317160814Ssimon POINT_CONVERSION_COMPRESSED 3318160814Ssimon POINT_CONVERSION_UNCOMPRESSED 3319160814Ssimon POINT_CONVERSION_HYBRID 3320160814Ssimon 3321160814Ssimon Also add 'seed' and 'seed_len' members to EC_GROUP with access 3322160814Ssimon functions 3323160814Ssimon EC_GROUP_set_seed() 3324160814Ssimon EC_GROUP_get0_seed() 3325160814Ssimon EC_GROUP_get_seed_len() 3326160814Ssimon This is used only for ASN1 purposes (so far). 3327160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3328160814Ssimon 3329160814Ssimon *) Add 'field_type' member to EC_METHOD, which holds the NID 3330160814Ssimon of the appropriate field type OID. The new function 3331160814Ssimon EC_METHOD_get_field_type() returns this value. 3332160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3333160814Ssimon 3334160814Ssimon *) Add functions 3335160814Ssimon EC_POINT_point2bn() 3336160814Ssimon EC_POINT_bn2point() 3337160814Ssimon EC_POINT_point2hex() 3338160814Ssimon EC_POINT_hex2point() 3339160814Ssimon providing useful interfaces to EC_POINT_point2oct() and 3340160814Ssimon EC_POINT_oct2point(). 3341160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3342160814Ssimon 3343160814Ssimon *) Change internals of the EC library so that the functions 3344160814Ssimon EC_GROUP_set_generator() 3345160814Ssimon EC_GROUP_get_generator() 3346160814Ssimon EC_GROUP_get_order() 3347160814Ssimon EC_GROUP_get_cofactor() 3348160814Ssimon are implemented directly in crypto/ec/ec_lib.c and not dispatched 3349160814Ssimon to methods, which would lead to unnecessary code duplication when 3350160814Ssimon adding different types of curves. 3351160814Ssimon [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller] 3352160814Ssimon 3353160814Ssimon *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM 3354160814Ssimon arithmetic, and such that modified wNAFs are generated 3355160814Ssimon (which avoid length expansion in many cases). 3356160814Ssimon [Bodo Moeller] 3357160814Ssimon 3358160814Ssimon *) Add a function EC_GROUP_check_discriminant() (defined via 3359160814Ssimon EC_METHOD) that verifies that the curve discriminant is non-zero. 3360160814Ssimon 3361160814Ssimon Add a function EC_GROUP_check() that makes some sanity tests 3362160814Ssimon on a EC_GROUP, its generator and order. This includes 3363160814Ssimon EC_GROUP_check_discriminant(). 3364160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3365160814Ssimon 3366160814Ssimon *) Add ECDSA in new directory crypto/ecdsa/. 3367160814Ssimon 3368160814Ssimon Add applications 'openssl ecparam' and 'openssl ecdsa' 3369160814Ssimon (these are based on 'openssl dsaparam' and 'openssl dsa'). 3370160814Ssimon 3371160814Ssimon ECDSA support is also included in various other files across the 3372160814Ssimon library. Most notably, 3373160814Ssimon - 'openssl req' now has a '-newkey ecdsa:file' option; 3374160814Ssimon - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA; 3375160814Ssimon - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and 3376160814Ssimon d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make 3377160814Ssimon them suitable for ECDSA where domain parameters must be 3378160814Ssimon extracted before the specific public key; 3379160814Ssimon - ECDSA engine support has been added. 3380160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3381160814Ssimon 3382160814Ssimon *) Include some named elliptic curves, and add OIDs from X9.62, 3383160814Ssimon SECG, and WAP/WTLS. Each curve can be obtained from the new 3384160814Ssimon function 3385160814Ssimon EC_GROUP_new_by_curve_name(), 3386160814Ssimon and the list of available named curves can be obtained with 3387160814Ssimon EC_get_builtin_curves(). 3388160814Ssimon Also add a 'curve_name' member to EC_GROUP objects, which can be 3389160814Ssimon accessed via 3390160814Ssimon EC_GROUP_set_curve_name() 3391160814Ssimon EC_GROUP_get_curve_name() 3392160814Ssimon [Nils Larsch <larsch@trustcenter.de, Bodo Moeller] 3393160814Ssimon 3394160814Ssimon *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 3395160814Ssimon was actually never needed) and in BN_mul(). The removal in BN_mul() 3396160814Ssimon required a small change in bn_mul_part_recursive() and the addition 3397160814Ssimon of the functions bn_cmp_part_words(), bn_sub_part_words() and 3398160814Ssimon bn_add_part_words(), which do the same thing as bn_cmp_words(), 3399160814Ssimon bn_sub_words() and bn_add_words() except they take arrays with 3400160814Ssimon differing sizes. 3401160814Ssimon [Richard Levitte] 3402160814Ssimon 3403194206Ssimon Changes between 0.9.7l and 0.9.7m [23 Feb 2007] 3404194206Ssimon 3405167612Ssimon *) Cleanse PEM buffers before freeing them since they may contain 3406167612Ssimon sensitive data. 3407167612Ssimon [Benjamin Bennett <ben@psc.edu>] 3408167612Ssimon 3409167612Ssimon *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 3410167612Ssimon a ciphersuite string such as "DEFAULT:RSA" cannot enable 3411167612Ssimon authentication-only ciphersuites. 3412167612Ssimon [Bodo Moeller] 3413167612Ssimon 3414167612Ssimon *) Since AES128 and AES256 share a single mask bit in the logic of 3415167612Ssimon ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 3416167612Ssimon kludge to work properly if AES128 is available and AES256 isn't. 3417167612Ssimon [Victor Duchovni] 3418167612Ssimon 3419194206Ssimon *) Expand security boundary to match 1.1.1 module. 3420194206Ssimon [Steve Henson] 3421194206Ssimon 3422194206Ssimon *) Remove redundant features: hash file source, editing of test vectors 3423194206Ssimon modify fipsld to use external fips_premain.c signature. 3424194206Ssimon [Steve Henson] 3425194206Ssimon 3426194206Ssimon *) New perl script mkfipsscr.pl to create shell scripts or batch files to 3427194206Ssimon run algorithm test programs. 3428194206Ssimon [Steve Henson] 3429194206Ssimon 3430194206Ssimon *) Make algorithm test programs more tolerant of whitespace. 3431194206Ssimon [Steve Henson] 3432194206Ssimon 3433167612Ssimon *) Have SSL/TLS server implementation tolerate "mismatched" record 3434167612Ssimon protocol version while receiving ClientHello even if the 3435167612Ssimon ClientHello is fragmented. (The server can't insist on the 3436167612Ssimon particular protocol version it has chosen before the ServerHello 3437167612Ssimon message has informed the client about his choice.) 3438167612Ssimon [Bodo Moeller] 3439167612Ssimon 3440167612Ssimon *) Load error codes if they are not already present instead of using a 3441167612Ssimon static variable. This allows them to be cleanly unloaded and reloaded. 3442167612Ssimon [Steve Henson] 3443167612Ssimon 3444167612Ssimon Changes between 0.9.7k and 0.9.7l [28 Sep 2006] 3445167612Ssimon 3446167612Ssimon *) Introduce limits to prevent malicious keys being able to 3447167612Ssimon cause a denial of service. (CVE-2006-2940) 3448167612Ssimon [Steve Henson, Bodo Moeller] 3449167612Ssimon 3450167612Ssimon *) Fix ASN.1 parsing of certain invalid structures that can result 3451167612Ssimon in a denial of service. (CVE-2006-2937) [Steve Henson] 3452167612Ssimon 3453167612Ssimon *) Fix buffer overflow in SSL_get_shared_ciphers() function. 3454167612Ssimon (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 3455167612Ssimon 3456167612Ssimon *) Fix SSL client code which could crash if connecting to a 3457167612Ssimon malicious SSLv2 server. (CVE-2006-4343) 3458167612Ssimon [Tavis Ormandy and Will Drewry, Google Security Team] 3459167612Ssimon 3460162911Ssimon *) Change ciphersuite string processing so that an explicit 3461162911Ssimon ciphersuite selects this one ciphersuite (so that "AES256-SHA" 3462162911Ssimon will no longer include "AES128-SHA"), and any other similar 3463162911Ssimon ciphersuite (same bitmap) from *other* protocol versions (so that 3464162911Ssimon "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the 3465162911Ssimon SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining 3466162911Ssimon changes from 0.9.8b and 0.9.8d. 3467162911Ssimon [Bodo Moeller] 3468162911Ssimon 3469162911Ssimon Changes between 0.9.7j and 0.9.7k [05 Sep 2006] 3470162911Ssimon 3471162911Ssimon *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 3472162911Ssimon (CVE-2006-4339) [Ben Laurie and Google Security Team] 3473162911Ssimon 3474162911Ssimon *) Change the Unix randomness entropy gathering to use poll() when 3475162911Ssimon possible instead of select(), since the latter has some 3476162911Ssimon undesirable limitations. 3477162911Ssimon [Darryl Miles via Richard Levitte and Bodo Moeller] 3478162911Ssimon 3479162911Ssimon *) Disable rogue ciphersuites: 3480162911Ssimon 3481162911Ssimon - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 3482162911Ssimon - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 3483162911Ssimon - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 3484162911Ssimon 3485162911Ssimon The latter two were purportedly from 3486162911Ssimon draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 3487162911Ssimon appear there. 3488162911Ssimon 3489162911Ssimon Also deactive the remaining ciphersuites from 3490162911Ssimon draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 3491162911Ssimon unofficial, and the ID has long expired. 3492162911Ssimon [Bodo Moeller] 3493162911Ssimon 3494162911Ssimon *) Fix RSA blinding Heisenbug (problems sometimes occured on 3495162911Ssimon dual-core machines) and other potential thread-safety issues. 3496162911Ssimon [Bodo Moeller] 3497162911Ssimon 3498162911Ssimon Changes between 0.9.7i and 0.9.7j [04 May 2006] 3499162911Ssimon 3500162911Ssimon *) Adapt fipsld and the build system to link against the validated FIPS 3501162911Ssimon module in FIPS mode. 3502162911Ssimon [Steve Henson] 3503162911Ssimon 3504162911Ssimon *) Fixes for VC++ 2005 build under Windows. 3505162911Ssimon [Steve Henson] 3506162911Ssimon 3507162911Ssimon *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make 3508162911Ssimon from a Windows bash shell such as MSYS. It is autodetected from the 3509162911Ssimon "config" script when run from a VC++ environment. Modify standard VC++ 3510162911Ssimon build to use fipscanister.o from the GNU make build. 3511162911Ssimon [Steve Henson] 3512162911Ssimon 3513160814Ssimon Changes between 0.9.7h and 0.9.7i [14 Oct 2005] 3514160814Ssimon 3515160814Ssimon *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS. 3516160814Ssimon The value now differs depending on if you build for FIPS or not. 3517160814Ssimon BEWARE! A program linked with a shared FIPSed libcrypto can't be 3518160814Ssimon safely run with a non-FIPSed libcrypto, as it may crash because of 3519160814Ssimon the difference induced by this change. 3520160814Ssimon [Andy Polyakov] 3521160814Ssimon 3522160814Ssimon Changes between 0.9.7g and 0.9.7h [11 Oct 2005] 3523160814Ssimon 3524160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 3525160814Ssimon (part of SSL_OP_ALL). This option used to disable the 3526160814Ssimon countermeasure against man-in-the-middle protocol-version 3527160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 3528160814Ssimon idea. (CVE-2005-2969) 3529160814Ssimon 3530160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 3531160814Ssimon for Information Security, National Institute of Advanced Industrial 3532160814Ssimon Science and Technology [AIST], Japan)] 3533160814Ssimon 3534160814Ssimon *) Minimal support for X9.31 signatures and PSS padding modes. This is 3535160814Ssimon mainly for FIPS compliance and not fully integrated at this stage. 3536160814Ssimon [Steve Henson] 3537160814Ssimon 3538160814Ssimon *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform 3539160814Ssimon the exponentiation using a fixed-length exponent. (Otherwise, 3540160814Ssimon the information leaked through timing could expose the secret key 3541160814Ssimon after many signatures; cf. Bleichenbacher's attack on DSA with 3542160814Ssimon biased k.) 3543160814Ssimon [Bodo Moeller] 3544160814Ssimon 3545160814Ssimon *) Make a new fixed-window mod_exp implementation the default for 3546160814Ssimon RSA, DSA, and DH private-key operations so that the sequence of 3547160814Ssimon squares and multiplies and the memory access pattern are 3548160814Ssimon independent of the particular secret key. This will mitigate 3549160814Ssimon cache-timing and potential related attacks. 3550160814Ssimon 3551160814Ssimon BN_mod_exp_mont_consttime() is the new exponentiation implementation, 3552160814Ssimon and this is automatically used by BN_mod_exp_mont() if the new flag 3553160814Ssimon BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH 3554160814Ssimon will use this BN flag for private exponents unless the flag 3555160814Ssimon RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or 3556160814Ssimon DH_FLAG_NO_EXP_CONSTTIME, respectively, is set. 3557160814Ssimon 3558160814Ssimon [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller] 3559160814Ssimon 3560160814Ssimon *) Change the client implementation for SSLv23_method() and 3561160814Ssimon SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0 3562160814Ssimon Client Hello message format if the SSL_OP_NO_SSLv2 option is set. 3563160814Ssimon (Previously, the SSL 2.0 backwards compatible Client Hello 3564160814Ssimon message format would be used even with SSL_OP_NO_SSLv2.) 3565160814Ssimon [Bodo Moeller] 3566160814Ssimon 3567160814Ssimon *) Add support for smime-type MIME parameter in S/MIME messages which some 3568160814Ssimon clients need. 3569160814Ssimon [Steve Henson] 3570160814Ssimon 3571160814Ssimon *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in 3572160814Ssimon a threadsafe manner. Modify rsa code to use new function and add calls 3573160814Ssimon to dsa and dh code (which had race conditions before). 3574160814Ssimon [Steve Henson] 3575160814Ssimon 3576160814Ssimon *) Include the fixed error library code in the C error file definitions 3577160814Ssimon instead of fixing them up at runtime. This keeps the error code 3578160814Ssimon structures constant. 3579160814Ssimon [Steve Henson] 3580160814Ssimon 3581160814Ssimon Changes between 0.9.7f and 0.9.7g [11 Apr 2005] 3582160814Ssimon 3583160814Ssimon [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after 3584160814Ssimon OpenSSL 0.9.8.] 3585160814Ssimon 3586160814Ssimon *) Fixes for newer kerberos headers. NB: the casts are needed because 3587160814Ssimon the 'length' field is signed on one version and unsigned on another 3588160814Ssimon with no (?) obvious way to tell the difference, without these VC++ 3589160814Ssimon complains. Also the "definition" of FAR (blank) is no longer included 3590160814Ssimon nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up 3591160814Ssimon some needed definitions. 3592160814Ssimon [Steve Henson] 3593160814Ssimon 3594160814Ssimon *) Undo Cygwin change. 3595160814Ssimon [Ulf M�ller] 3596160814Ssimon 3597160814Ssimon *) Added support for proxy certificates according to RFC 3820. 3598160814Ssimon Because they may be a security thread to unaware applications, 3599160814Ssimon they must be explicitely allowed in run-time. See 3600160814Ssimon docs/HOWTO/proxy_certificates.txt for further information. 3601160814Ssimon [Richard Levitte] 3602160814Ssimon 3603160814Ssimon Changes between 0.9.7e and 0.9.7f [22 Mar 2005] 3604160814Ssimon 3605160814Ssimon *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating 3606160814Ssimon server and client random values. Previously 3607160814Ssimon (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in 3608160814Ssimon less random data when sizeof(time_t) > 4 (some 64 bit platforms). 3609160814Ssimon 3610160814Ssimon This change has negligible security impact because: 3611160814Ssimon 3612160814Ssimon 1. Server and client random values still have 24 bytes of pseudo random 3613160814Ssimon data. 3614160814Ssimon 3615160814Ssimon 2. Server and client random values are sent in the clear in the initial 3616160814Ssimon handshake. 3617160814Ssimon 3618160814Ssimon 3. The master secret is derived using the premaster secret (48 bytes in 3619160814Ssimon size for static RSA ciphersuites) as well as client server and random 3620160814Ssimon values. 3621160814Ssimon 3622160814Ssimon The OpenSSL team would like to thank the UK NISCC for bringing this issue 3623160814Ssimon to our attention. 3624160814Ssimon 3625160814Ssimon [Stephen Henson, reported by UK NISCC] 3626160814Ssimon 3627160814Ssimon *) Use Windows randomness collection on Cygwin. 3628160814Ssimon [Ulf M�ller] 3629160814Ssimon 3630160814Ssimon *) Fix hang in EGD/PRNGD query when communication socket is closed 3631160814Ssimon prematurely by EGD/PRNGD. 3632160814Ssimon [Darren Tucker <dtucker@zip.com.au> via Lutz J�nicke, resolves #1014] 3633160814Ssimon 3634160814Ssimon *) Prompt for pass phrases when appropriate for PKCS12 input format. 3635160814Ssimon [Steve Henson] 3636160814Ssimon 3637160814Ssimon *) Back-port of selected performance improvements from development 3638160814Ssimon branch, as well as improved support for PowerPC platforms. 3639160814Ssimon [Andy Polyakov] 3640160814Ssimon 3641160814Ssimon *) Add lots of checks for memory allocation failure, error codes to indicate 3642160814Ssimon failure and freeing up memory if a failure occurs. 3643160814Ssimon [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson] 3644160814Ssimon 3645160814Ssimon *) Add new -passin argument to dgst. 3646160814Ssimon [Steve Henson] 3647160814Ssimon 3648160814Ssimon *) Perform some character comparisons of different types in X509_NAME_cmp: 3649160814Ssimon this is needed for some certificates that reencode DNs into UTF8Strings 3650160814Ssimon (in violation of RFC3280) and can't or wont issue name rollover 3651160814Ssimon certificates. 3652160814Ssimon [Steve Henson] 3653160814Ssimon 3654160814Ssimon *) Make an explicit check during certificate validation to see that 3655160814Ssimon the CA setting in each certificate on the chain is correct. As a 3656160814Ssimon side effect always do the following basic checks on extensions, 3657160814Ssimon not just when there's an associated purpose to the check: 3658160814Ssimon 3659160814Ssimon - if there is an unhandled critical extension (unless the user 3660160814Ssimon has chosen to ignore this fault) 3661160814Ssimon - if the path length has been exceeded (if one is set at all) 3662160814Ssimon - that certain extensions fit the associated purpose (if one has 3663160814Ssimon been given) 3664160814Ssimon [Richard Levitte] 3665160814Ssimon 3666142425Snectar Changes between 0.9.7d and 0.9.7e [25 Oct 2004] 3667142425Snectar 3668142425Snectar *) Avoid a race condition when CRLs are checked in a multi threaded 3669142425Snectar environment. This would happen due to the reordering of the revoked 3670142425Snectar entries during signature checking and serial number lookup. Now the 3671142425Snectar encoding is cached and the serial number sort performed under a lock. 3672142425Snectar Add new STACK function sk_is_sorted(). 3673142425Snectar [Steve Henson] 3674142425Snectar 3675142425Snectar *) Add Delta CRL to the extension code. 3676142425Snectar [Steve Henson] 3677142425Snectar 3678142425Snectar *) Various fixes to s3_pkt.c so alerts are sent properly. 3679142425Snectar [David Holmes <d.holmes@f5.com>] 3680142425Snectar 3681142425Snectar *) Reduce the chances of duplicate issuer name and serial numbers (in 3682142425Snectar violation of RFC3280) using the OpenSSL certificate creation utilities. 3683142425Snectar This is done by creating a random 64 bit value for the initial serial 3684142425Snectar number when a serial number file is created or when a self signed 3685142425Snectar certificate is created using 'openssl req -x509'. The initial serial 3686142425Snectar number file is created using 'openssl x509 -next_serial' in CA.pl 3687142425Snectar rather than being initialized to 1. 3688142425Snectar [Steve Henson] 3689142425Snectar 3690127128Snectar Changes between 0.9.7c and 0.9.7d [17 Mar 2004] 3691127128Snectar 3692127128Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 3693160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 3694127128Snectar [Joe Orton, Steve Henson] 3695127128Snectar 3696127128Snectar *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites 3697160814Ssimon (CVE-2004-0112) 3698127128Snectar [Joe Orton, Steve Henson] 3699127128Snectar 3700127128Snectar *) Make it possible to have multiple active certificates with the same 3701127128Snectar subject in the CA index file. This is done only if the keyword 3702127128Snectar 'unique_subject' is set to 'no' in the main CA section (default 3703127128Snectar if 'CA_default') of the configuration file. The value is saved 3704127128Snectar with the database itself in a separate index attribute file, 3705127128Snectar named like the index file with '.attr' appended to the name. 3706127128Snectar [Richard Levitte] 3707127128Snectar 3708127128Snectar *) X509 verify fixes. Disable broken certificate workarounds when 3709127128Snectar X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if 3710127128Snectar keyUsage extension present. Don't accept CRLs with unhandled critical 3711127128Snectar extensions: since verify currently doesn't process CRL extensions this 3712127128Snectar rejects a CRL with *any* critical extensions. Add new verify error codes 3713127128Snectar for these cases. 3714127128Snectar [Steve Henson] 3715127128Snectar 3716127128Snectar *) When creating an OCSP nonce use an OCTET STRING inside the extnValue. 3717127128Snectar A clarification of RFC2560 will require the use of OCTET STRINGs and 3718127128Snectar some implementations cannot handle the current raw format. Since OpenSSL 3719127128Snectar copies and compares OCSP nonces as opaque blobs without any attempt at 3720127128Snectar parsing them this should not create any compatibility issues. 3721127128Snectar [Steve Henson] 3722127128Snectar 3723127128Snectar *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when 3724127128Snectar calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without 3725127128Snectar this HMAC (and other) operations are several times slower than OpenSSL 3726127128Snectar < 0.9.7. 3727127128Snectar [Steve Henson] 3728127128Snectar 3729127128Snectar *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex(). 3730127128Snectar [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 3731127128Snectar 3732127128Snectar *) Use the correct content when signing type "other". 3733127128Snectar [Steve Henson] 3734127128Snectar 3735120631Snectar Changes between 0.9.7b and 0.9.7c [30 Sep 2003] 3736120631Snectar 3737120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 3738120631Snectar 3739120631Snectar Stop out of bounds reads in the ASN1 code when presented with 3740160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 3741120631Snectar 3742160814Ssimon Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545). 3743120631Snectar 3744120631Snectar If verify callback ignores invalid public key errors don't try to check 3745120631Snectar certificate signature with the NULL public key. 3746120631Snectar 3747120631Snectar [Steve Henson] 3748120631Snectar 3749120631Snectar *) New -ignore_err option in ocsp application to stop the server 3750120631Snectar exiting on the first error in a request. 3751120631Snectar [Steve Henson] 3752120631Snectar 3753120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 3754120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 3755120631Snectar specifications. 3756120631Snectar [Steve Henson] 3757120631Snectar 3758120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 3759120631Snectar extra data after the compression methods not only for TLS 1.0 3760120631Snectar but also for SSL 3.0 (as required by the specification). 3761120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 3762120631Snectar 3763120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 3764120631Snectar when it's 512 *bits* long, not 512 bytes. 3765120631Snectar [Richard Levitte] 3766120631Snectar 3767120631Snectar *) Change AES_cbc_encrypt() so it outputs exact multiple of 3768120631Snectar blocks during encryption. 3769120631Snectar [Richard Levitte] 3770120631Snectar 3771120631Snectar *) Various fixes to base64 BIO and non blocking I/O. On write 3772120631Snectar flushes were not handled properly if the BIO retried. On read 3773120631Snectar data was not being buffered properly and had various logic bugs. 3774120631Snectar This also affects blocking I/O when the data being decoded is a 3775120631Snectar certain size. 3776120631Snectar [Steve Henson] 3777120631Snectar 3778120631Snectar *) Various S/MIME bugfixes and compatibility changes: 3779120631Snectar output correct application/pkcs7 MIME type if 3780120631Snectar PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures. 3781120631Snectar Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening 3782120631Snectar of files as .eml work). Correctly handle very long lines in MIME 3783120631Snectar parser. 3784120631Snectar [Steve Henson] 3785120631Snectar 3786120631Snectar Changes between 0.9.7a and 0.9.7b [10 Apr 2003] 3787120631Snectar 3788120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 3789120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 3790120631Snectar a protocol version number mismatch like a decryption error 3791120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 3792120631Snectar [Bodo Moeller] 3793120631Snectar 3794120631Snectar *) Turn on RSA blinding by default in the default implementation 3795120631Snectar to avoid a timing attack. Applications that don't want it can call 3796120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 3797120631Snectar They would be ill-advised to do so in most cases. 3798120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 3799120631Snectar 3800120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 3801120631Snectar seeded (in this case, the secret RSA exponent is abused as 3802120631Snectar an unpredictable seed -- if it is not unpredictable, there 3803120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 3804120631Snectar by remembering the creator's thread ID in rsa->blinding and 3805120631Snectar having all other threads use local one-time blinding factors 3806120631Snectar (this requires more computation than sharing rsa->blinding, but 3807120631Snectar avoids excessive locking; and if an RSA object is not shared 3808120631Snectar between threads, blinding will still be very fast). 3809120631Snectar [Bodo Moeller] 3810120631Snectar 3811120631Snectar *) Fixed a typo bug that would cause ENGINE_set_default() to set an 3812120631Snectar ENGINE as defaults for all supported algorithms irrespective of 3813120631Snectar the 'flags' parameter. 'flags' is now honoured, so applications 3814120631Snectar should make sure they are passing it correctly. 3815120631Snectar [Geoff Thorpe] 3816120631Snectar 3817120631Snectar *) Target "mingw" now allows native Windows code to be generated in 3818120631Snectar the Cygwin environment as well as with the MinGW compiler. 3819120631Snectar [Ulf Moeller] 3820120631Snectar 3821111147Snectar Changes between 0.9.7 and 0.9.7a [19 Feb 2003] 3822111147Snectar 3823111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 3824111147Snectar via timing by performing a MAC computation even if incorrrect 3825111147Snectar block cipher padding has been found. This is a countermeasure 3826111147Snectar against active attacks where the attacker has to distinguish 3827160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 3828111147Snectar 3829111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 3830111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 3831111147Snectar Martin Vuagnoux (EPFL, Ilion)] 3832111147Snectar 3833111147Snectar *) Make the no-err option work as intended. The intention with no-err 3834111147Snectar is not to have the whole error stack handling routines removed from 3835111147Snectar libcrypto, it's only intended to remove all the function name and 3836111147Snectar reason texts, thereby removing some of the footprint that may not 3837111147Snectar be interesting if those errors aren't displayed anyway. 3838111147Snectar 3839111147Snectar NOTE: it's still possible for any application or module to have it's 3840111147Snectar own set of error texts inserted. The routines are there, just not 3841111147Snectar used by default when no-err is given. 3842111147Snectar [Richard Levitte] 3843111147Snectar 3844111147Snectar *) Add support for FreeBSD on IA64. 3845111147Snectar [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454] 3846111147Snectar 3847111147Snectar *) Adjust DES_cbc_cksum() so it returns the same value as the MIT 3848111147Snectar Kerberos function mit_des_cbc_cksum(). Before this change, 3849111147Snectar the value returned by DES_cbc_cksum() was like the one from 3850111147Snectar mit_des_cbc_cksum(), except the bytes were swapped. 3851111147Snectar [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte] 3852111147Snectar 3853111147Snectar *) Allow an application to disable the automatic SSL chain building. 3854111147Snectar Before this a rather primitive chain build was always performed in 3855111147Snectar ssl3_output_cert_chain(): an application had no way to send the 3856111147Snectar correct chain if the automatic operation produced an incorrect result. 3857111147Snectar 3858111147Snectar Now the chain builder is disabled if either: 3859111147Snectar 3860111147Snectar 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert(). 3861111147Snectar 3862111147Snectar 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set. 3863111147Snectar 3864111147Snectar The reasoning behind this is that an application would not want the 3865111147Snectar auto chain building to take place if extra chain certificates are 3866111147Snectar present and it might also want a means of sending no additional 3867111147Snectar certificates (for example the chain has two certificates and the 3868111147Snectar root is omitted). 3869111147Snectar [Steve Henson] 3870111147Snectar 3871111147Snectar *) Add the possibility to build without the ENGINE framework. 3872111147Snectar [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 3873111147Snectar 3874111147Snectar *) Under Win32 gmtime() can return NULL: check return value in 3875111147Snectar OPENSSL_gmtime(). Add error code for case where gmtime() fails. 3876111147Snectar [Steve Henson] 3877111147Snectar 3878111147Snectar *) DSA routines: under certain error conditions uninitialized BN objects 3879111147Snectar could be freed. Solution: make sure initialization is performed early 3880111147Snectar enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>, 3881111147Snectar Nils Larsch <nla@trustcenter.de> via PR#459) 3882111147Snectar [Lutz Jaenicke] 3883111147Snectar 3884111147Snectar *) Another fix for SSLv2 session ID handling: the session ID was incorrectly 3885111147Snectar checked on reconnect on the client side, therefore session resumption 3886111147Snectar could still fail with a "ssl session id is different" error. This 3887111147Snectar behaviour is masked when SSL_OP_ALL is used due to 3888111147Snectar SSL_OP_MICROSOFT_SESS_ID_BUG being set. 3889111147Snectar Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 3890111147Snectar followup to PR #377. 3891111147Snectar [Lutz Jaenicke] 3892111147Snectar 3893111147Snectar *) IA-32 assembler support enhancements: unified ELF targets, support 3894111147Snectar for SCO/Caldera platforms, fix for Cygwin shared build. 3895111147Snectar [Andy Polyakov] 3896111147Snectar 3897111147Snectar *) Add support for FreeBSD on sparc64. As a consequence, support for 3898111147Snectar FreeBSD on non-x86 processors is separate from x86 processors on 3899111147Snectar the config script, much like the NetBSD support. 3900111147Snectar [Richard Levitte & Kris Kennaway <kris@obsecurity.org>] 3901111147Snectar 3902109998Smarkm Changes between 0.9.6h and 0.9.7 [31 Dec 2002] 3903109998Smarkm 3904120631Snectar [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after 3905120631Snectar OpenSSL 0.9.7.] 3906120631Snectar 3907109998Smarkm *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED 3908109998Smarkm code (06) was taken as the first octet of the session ID and the last 3909109998Smarkm octet was ignored consequently. As a result SSLv2 client side session 3910109998Smarkm caching could not have worked due to the session ID mismatch between 3911109998Smarkm client and server. 3912109998Smarkm Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 3913109998Smarkm PR #377. 3914109998Smarkm [Lutz Jaenicke] 3915109998Smarkm 3916109998Smarkm *) Change the declaration of needed Kerberos libraries to use EX_LIBS 3917109998Smarkm instead of the special (and badly supported) LIBKRB5. LIBKRB5 is 3918109998Smarkm removed entirely. 3919109998Smarkm [Richard Levitte] 3920109998Smarkm 3921109998Smarkm *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it 3922109998Smarkm seems that in spite of existing for more than a year, many application 3923109998Smarkm author have done nothing to provide the necessary callbacks, which 3924109998Smarkm means that this particular engine will not work properly anywhere. 3925109998Smarkm This is a very unfortunate situation which forces us, in the name 3926109998Smarkm of usability, to give the hw_ncipher.c a static lock, which is part 3927109998Smarkm of libcrypto. 3928109998Smarkm NOTE: This is for the 0.9.7 series ONLY. This hack will never 3929109998Smarkm appear in 0.9.8 or later. We EXPECT application authors to have 3930109998Smarkm dealt properly with this when 0.9.8 is released (unless we actually 3931109998Smarkm make such changes in the libcrypto locking code that changes will 3932109998Smarkm have to be made anyway). 3933109998Smarkm [Richard Levitte] 3934109998Smarkm 3935109998Smarkm *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content 3936109998Smarkm octets have been read, EOF or an error occurs. Without this change 3937109998Smarkm some truncated ASN1 structures will not produce an error. 3938109998Smarkm [Steve Henson] 3939109998Smarkm 3940109998Smarkm *) Disable Heimdal support, since it hasn't been fully implemented. 3941109998Smarkm Still give the possibility to force the use of Heimdal, but with 3942109998Smarkm warnings and a request that patches get sent to openssl-dev. 3943109998Smarkm [Richard Levitte] 3944109998Smarkm 3945109998Smarkm *) Add the VC-CE target, introduce the WINCE sysname, and add 3946109998Smarkm INSTALL.WCE and appropriate conditionals to make it build. 3947109998Smarkm [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 3948109998Smarkm 3949109998Smarkm *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and 3950109998Smarkm cygssl-x.y.z.dll, where x, y and z are the major, minor and 3951109998Smarkm edit numbers of the version. 3952109998Smarkm [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 3953109998Smarkm 3954109998Smarkm *) Introduce safe string copy and catenation functions 3955109998Smarkm (BUF_strlcpy() and BUF_strlcat()). 3956109998Smarkm [Ben Laurie (CHATS) and Richard Levitte] 3957109998Smarkm 3958109998Smarkm *) Avoid using fixed-size buffers for one-line DNs. 3959109998Smarkm [Ben Laurie (CHATS)] 3960109998Smarkm 3961109998Smarkm *) Add BUF_MEM_grow_clean() to avoid information leakage when 3962109998Smarkm resizing buffers containing secrets, and use where appropriate. 3963109998Smarkm [Ben Laurie (CHATS)] 3964109998Smarkm 3965109998Smarkm *) Avoid using fixed size buffers for configuration file location. 3966109998Smarkm [Ben Laurie (CHATS)] 3967109998Smarkm 3968109998Smarkm *) Avoid filename truncation for various CA files. 3969109998Smarkm [Ben Laurie (CHATS)] 3970109998Smarkm 3971109998Smarkm *) Use sizeof in preference to magic numbers. 3972109998Smarkm [Ben Laurie (CHATS)] 3973109998Smarkm 3974109998Smarkm *) Avoid filename truncation in cert requests. 3975109998Smarkm [Ben Laurie (CHATS)] 3976109998Smarkm 3977109998Smarkm *) Add assertions to check for (supposedly impossible) buffer 3978109998Smarkm overflows. 3979109998Smarkm [Ben Laurie (CHATS)] 3980109998Smarkm 3981109998Smarkm *) Don't cache truncated DNS entries in the local cache (this could 3982109998Smarkm potentially lead to a spoofing attack). 3983109998Smarkm [Ben Laurie (CHATS)] 3984109998Smarkm 3985109998Smarkm *) Fix various buffers to be large enough for hex/decimal 3986109998Smarkm representations in a platform independent manner. 3987109998Smarkm [Ben Laurie (CHATS)] 3988109998Smarkm 3989109998Smarkm *) Add CRYPTO_realloc_clean() to avoid information leakage when 3990109998Smarkm resizing buffers containing secrets, and use where appropriate. 3991109998Smarkm [Ben Laurie (CHATS)] 3992109998Smarkm 3993109998Smarkm *) Add BIO_indent() to avoid much slightly worrying code to do 3994109998Smarkm indents. 3995109998Smarkm [Ben Laurie (CHATS)] 3996109998Smarkm 3997109998Smarkm *) Convert sprintf()/BIO_puts() to BIO_printf(). 3998109998Smarkm [Ben Laurie (CHATS)] 3999109998Smarkm 4000109998Smarkm *) buffer_gets() could terminate with the buffer only half 4001109998Smarkm full. Fixed. 4002109998Smarkm [Ben Laurie (CHATS)] 4003109998Smarkm 4004109998Smarkm *) Add assertions to prevent user-supplied crypto functions from 4005109998Smarkm overflowing internal buffers by having large block sizes, etc. 4006109998Smarkm [Ben Laurie (CHATS)] 4007109998Smarkm 4008109998Smarkm *) New OPENSSL_assert() macro (similar to assert(), but enabled 4009109998Smarkm unconditionally). 4010109998Smarkm [Ben Laurie (CHATS)] 4011109998Smarkm 4012109998Smarkm *) Eliminate unused copy of key in RC4. 4013109998Smarkm [Ben Laurie (CHATS)] 4014109998Smarkm 4015109998Smarkm *) Eliminate unused and incorrectly sized buffers for IV in pem.h. 4016109998Smarkm [Ben Laurie (CHATS)] 4017109998Smarkm 4018109998Smarkm *) Fix off-by-one error in EGD path. 4019109998Smarkm [Ben Laurie (CHATS)] 4020109998Smarkm 4021109998Smarkm *) If RANDFILE path is too long, ignore instead of truncating. 4022109998Smarkm [Ben Laurie (CHATS)] 4023109998Smarkm 4024109998Smarkm *) Eliminate unused and incorrectly sized X.509 structure 4025109998Smarkm CBCParameter. 4026109998Smarkm [Ben Laurie (CHATS)] 4027109998Smarkm 4028109998Smarkm *) Eliminate unused and dangerous function knumber(). 4029109998Smarkm [Ben Laurie (CHATS)] 4030109998Smarkm 4031109998Smarkm *) Eliminate unused and dangerous structure, KSSL_ERR. 4032109998Smarkm [Ben Laurie (CHATS)] 4033109998Smarkm 4034109998Smarkm *) Protect against overlong session ID context length in an encoded 4035109998Smarkm session object. Since these are local, this does not appear to be 4036109998Smarkm exploitable. 4037109998Smarkm [Ben Laurie (CHATS)] 4038109998Smarkm 4039109998Smarkm *) Change from security patch (see 0.9.6e below) that did not affect 4040109998Smarkm the 0.9.6 release series: 4041109998Smarkm 4042109998Smarkm Remote buffer overflow in SSL3 protocol - an attacker could 4043109998Smarkm supply an oversized master key in Kerberos-enabled versions. 4044160814Ssimon (CVE-2002-0657) 4045109998Smarkm [Ben Laurie (CHATS)] 4046109998Smarkm 4047109998Smarkm *) Change the SSL kerb5 codes to match RFC 2712. 4048109998Smarkm [Richard Levitte] 4049109998Smarkm 4050109998Smarkm *) Make -nameopt work fully for req and add -reqopt switch. 4051109998Smarkm [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson] 4052109998Smarkm 4053109998Smarkm *) The "block size" for block ciphers in CFB and OFB mode should be 1. 4054109998Smarkm [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>] 4055109998Smarkm 4056109998Smarkm *) Make sure tests can be performed even if the corresponding algorithms 4057109998Smarkm have been removed entirely. This was also the last step to make 4058109998Smarkm OpenSSL compilable with DJGPP under all reasonable conditions. 4059109998Smarkm [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>] 4060109998Smarkm 4061109998Smarkm *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT 4062109998Smarkm to allow version independent disabling of normally unselected ciphers, 4063109998Smarkm which may be activated as a side-effect of selecting a single cipher. 4064109998Smarkm 4065109998Smarkm (E.g., cipher list string "RSA" enables ciphersuites that are left 4066109998Smarkm out of "ALL" because they do not provide symmetric encryption. 4067109998Smarkm "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) 4068109998Smarkm [Lutz Jaenicke, Bodo Moeller] 4069109998Smarkm 4070109998Smarkm *) Add appropriate support for separate platform-dependent build 4071109998Smarkm directories. The recommended way to make a platform-dependent 4072109998Smarkm build directory is the following (tested on Linux), maybe with 4073109998Smarkm some local tweaks: 4074109998Smarkm 4075109998Smarkm # Place yourself outside of the OpenSSL source tree. In 4076109998Smarkm # this example, the environment variable OPENSSL_SOURCE 4077109998Smarkm # is assumed to contain the absolute OpenSSL source directory. 4078109998Smarkm mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" 4079109998Smarkm cd objtree/"`uname -s`-`uname -r`-`uname -m`" 4080111147Snectar (cd $OPENSSL_SOURCE; find . -type f) | while read F; do 4081109998Smarkm mkdir -p `dirname $F` 4082109998Smarkm ln -s $OPENSSL_SOURCE/$F $F 4083109998Smarkm done 4084109998Smarkm 4085109998Smarkm To be absolutely sure not to disturb the source tree, a "make clean" 4086109998Smarkm is a good thing. If it isn't successfull, don't worry about it, 4087109998Smarkm it probably means the source directory is very clean. 4088109998Smarkm [Richard Levitte] 4089109998Smarkm 4090109998Smarkm *) Make sure any ENGINE control commands make local copies of string 4091109998Smarkm pointers passed to them whenever necessary. Otherwise it is possible 4092109998Smarkm the caller may have overwritten (or deallocated) the original string 4093109998Smarkm data when a later ENGINE operation tries to use the stored values. 4094109998Smarkm [G�tz Babin-Ebell <babinebell@trustcenter.de>] 4095109998Smarkm 4096109998Smarkm *) Improve diagnostics in file reading and command-line digests. 4097109998Smarkm [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>] 4098109998Smarkm 4099109998Smarkm *) Add AES modes CFB and OFB to the object database. Correct an 4100109998Smarkm error in AES-CFB decryption. 4101109998Smarkm [Richard Levitte] 4102109998Smarkm 4103109998Smarkm *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this 4104109998Smarkm allows existing EVP_CIPHER_CTX structures to be reused after 4105109998Smarkm calling EVP_*Final(). This behaviour is used by encryption 4106109998Smarkm BIOs and some applications. This has the side effect that 4107109998Smarkm applications must explicitly clean up cipher contexts with 4108109998Smarkm EVP_CIPHER_CTX_cleanup() or they will leak memory. 4109109998Smarkm [Steve Henson] 4110109998Smarkm 4111109998Smarkm *) Check the values of dna and dnb in bn_mul_recursive before calling 4112109998Smarkm bn_mul_comba (a non zero value means the a or b arrays do not contain 4113109998Smarkm n2 elements) and fallback to bn_mul_normal if either is not zero. 4114109998Smarkm [Steve Henson] 4115109998Smarkm 4116109998Smarkm *) Fix escaping of non-ASCII characters when using the -subj option 4117109998Smarkm of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) 4118109998Smarkm [Lutz Jaenicke] 4119109998Smarkm 4120109998Smarkm *) Make object definitions compliant to LDAP (RFC2256): SN is the short 4121109998Smarkm form for "surname", serialNumber has no short form. 4122109998Smarkm Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; 4123109998Smarkm therefore remove "mail" short name for "internet 7". 4124109998Smarkm The OID for unique identifiers in X509 certificates is 4125109998Smarkm x500UniqueIdentifier, not uniqueIdentifier. 4126109998Smarkm Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) 4127109998Smarkm [Lutz Jaenicke] 4128109998Smarkm 4129109998Smarkm *) Add an "init" command to the ENGINE config module and auto initialize 4130109998Smarkm ENGINEs. Without any "init" command the ENGINE will be initialized 4131109998Smarkm after all ctrl commands have been executed on it. If init=1 the 4132109998Smarkm ENGINE is initailized at that point (ctrls before that point are run 4133109998Smarkm on the uninitialized ENGINE and after on the initialized one). If 4134109998Smarkm init=0 then the ENGINE will not be iniatialized at all. 4135109998Smarkm [Steve Henson] 4136109998Smarkm 4137109998Smarkm *) Fix the 'app_verify_callback' interface so that the user-defined 4138109998Smarkm argument is actually passed to the callback: In the 4139109998Smarkm SSL_CTX_set_cert_verify_callback() prototype, the callback 4140109998Smarkm declaration has been changed from 4141109998Smarkm int (*cb)() 4142109998Smarkm into 4143109998Smarkm int (*cb)(X509_STORE_CTX *,void *); 4144109998Smarkm in ssl_verify_cert_chain (ssl/ssl_cert.c), the call 4145109998Smarkm i=s->ctx->app_verify_callback(&ctx) 4146109998Smarkm has been changed into 4147109998Smarkm i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). 4148109998Smarkm 4149109998Smarkm To update applications using SSL_CTX_set_cert_verify_callback(), 4150109998Smarkm a dummy argument can be added to their callback functions. 4151109998Smarkm [D. K. Smetters <smetters@parc.xerox.com>] 4152109998Smarkm 4153109998Smarkm *) Added the '4758cca' ENGINE to support IBM 4758 cards. 4154109998Smarkm [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe] 4155109998Smarkm 4156109998Smarkm *) Add and OPENSSL_LOAD_CONF define which will cause 4157109998Smarkm OpenSSL_add_all_algorithms() to load the openssl.cnf config file. 4158109998Smarkm This allows older applications to transparently support certain 4159109998Smarkm OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. 4160109998Smarkm Two new functions OPENSSL_add_all_algorithms_noconf() which will never 4161109998Smarkm load the config file and OPENSSL_add_all_algorithms_conf() which will 4162109998Smarkm always load it have also been added. 4163109998Smarkm [Steve Henson] 4164109998Smarkm 4165109998Smarkm *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. 4166109998Smarkm Adjust NIDs and EVP layer. 4167109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 4168109998Smarkm 4169109998Smarkm *) Config modules support in openssl utility. 4170109998Smarkm 4171109998Smarkm Most commands now load modules from the config file, 4172109998Smarkm though in a few (such as version) this isn't done 4173109998Smarkm because it couldn't be used for anything. 4174109998Smarkm 4175109998Smarkm In the case of ca and req the config file used is 4176109998Smarkm the same as the utility itself: that is the -config 4177109998Smarkm command line option can be used to specify an 4178109998Smarkm alternative file. 4179109998Smarkm [Steve Henson] 4180109998Smarkm 4181109998Smarkm *) Move default behaviour from OPENSSL_config(). If appname is NULL 4182109998Smarkm use "openssl_conf" if filename is NULL use default openssl config file. 4183109998Smarkm [Steve Henson] 4184109998Smarkm 4185109998Smarkm *) Add an argument to OPENSSL_config() to allow the use of an alternative 4186109998Smarkm config section name. Add a new flag to tolerate a missing config file 4187109998Smarkm and move code to CONF_modules_load_file(). 4188109998Smarkm [Steve Henson] 4189109998Smarkm 4190109998Smarkm *) Support for crypto accelerator cards from Accelerated Encryption 4191109998Smarkm Processing, www.aep.ie. (Use engine 'aep') 4192109998Smarkm The support was copied from 0.9.6c [engine] and adapted/corrected 4193109998Smarkm to work with the new engine framework. 4194109998Smarkm [AEP Inc. and Richard Levitte] 4195109998Smarkm 4196109998Smarkm *) Support for SureWare crypto accelerator cards from Baltimore 4197109998Smarkm Technologies. (Use engine 'sureware') 4198109998Smarkm The support was copied from 0.9.6c [engine] and adapted 4199109998Smarkm to work with the new engine framework. 4200109998Smarkm [Richard Levitte] 4201109998Smarkm 4202109998Smarkm *) Have the CHIL engine fork-safe (as defined by nCipher) and actually 4203109998Smarkm make the newer ENGINE framework commands for the CHIL engine work. 4204109998Smarkm [Toomas Kiisk <vix@cyber.ee> and Richard Levitte] 4205109998Smarkm 4206109998Smarkm *) Make it possible to produce shared libraries on ReliantUNIX. 4207109998Smarkm [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte] 4208109998Smarkm 4209109998Smarkm *) Add the configuration target debug-linux-ppro. 4210109998Smarkm Make 'openssl rsa' use the general key loading routines 4211109998Smarkm implemented in apps.c, and make those routines able to 4212109998Smarkm handle the key format FORMAT_NETSCAPE and the variant 4213109998Smarkm FORMAT_IISSGC. 4214109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 4215109998Smarkm 4216109998Smarkm *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 4217109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 4218109998Smarkm 4219109998Smarkm *) Add -keyform to rsautl, and document -engine. 4220109998Smarkm [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>] 4221109998Smarkm 4222109998Smarkm *) Change BIO_new_file (crypto/bio/bss_file.c) to use new 4223109998Smarkm BIO_R_NO_SUCH_FILE error code rather than the generic 4224109998Smarkm ERR_R_SYS_LIB error code if fopen() fails with ENOENT. 4225109998Smarkm [Ben Laurie] 4226109998Smarkm 4227109998Smarkm *) Add new functions 4228109998Smarkm ERR_peek_last_error 4229109998Smarkm ERR_peek_last_error_line 4230109998Smarkm ERR_peek_last_error_line_data. 4231109998Smarkm These are similar to 4232109998Smarkm ERR_peek_error 4233109998Smarkm ERR_peek_error_line 4234109998Smarkm ERR_peek_error_line_data, 4235109998Smarkm but report on the latest error recorded rather than the first one 4236109998Smarkm still in the error queue. 4237109998Smarkm [Ben Laurie, Bodo Moeller] 4238109998Smarkm 4239109998Smarkm *) default_algorithms option in ENGINE config module. This allows things 4240109998Smarkm like: 4241109998Smarkm default_algorithms = ALL 4242109998Smarkm default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS 4243109998Smarkm [Steve Henson] 4244109998Smarkm 4245109998Smarkm *) Prelminary ENGINE config module. 4246109998Smarkm [Steve Henson] 4247109998Smarkm 4248109998Smarkm *) New experimental application configuration code. 4249109998Smarkm [Steve Henson] 4250109998Smarkm 4251109998Smarkm *) Change the AES code to follow the same name structure as all other 4252109998Smarkm symmetric ciphers, and behave the same way. Move everything to 4253109998Smarkm the directory crypto/aes, thereby obsoleting crypto/rijndael. 4254109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 4255109998Smarkm 4256109998Smarkm *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. 4257109998Smarkm [Ben Laurie and Theo de Raadt] 4258109998Smarkm 4259109998Smarkm *) Add option to output public keys in req command. 4260109998Smarkm [Massimiliano Pala madwolf@openca.org] 4261109998Smarkm 4262109998Smarkm *) Use wNAFs in EC_POINTs_mul() for improved efficiency 4263109998Smarkm (up to about 10% better than before for P-192 and P-224). 4264109998Smarkm [Bodo Moeller] 4265109998Smarkm 4266109998Smarkm *) New functions/macros 4267109998Smarkm 4268109998Smarkm SSL_CTX_set_msg_callback(ctx, cb) 4269109998Smarkm SSL_CTX_set_msg_callback_arg(ctx, arg) 4270109998Smarkm SSL_set_msg_callback(ssl, cb) 4271109998Smarkm SSL_set_msg_callback_arg(ssl, arg) 4272109998Smarkm 4273109998Smarkm to request calling a callback function 4274109998Smarkm 4275109998Smarkm void cb(int write_p, int version, int content_type, 4276109998Smarkm const void *buf, size_t len, SSL *ssl, void *arg) 4277109998Smarkm 4278109998Smarkm whenever a protocol message has been completely received 4279109998Smarkm (write_p == 0) or sent (write_p == 1). Here 'version' is the 4280109998Smarkm protocol version according to which the SSL library interprets 4281109998Smarkm the current protocol message (SSL2_VERSION, SSL3_VERSION, or 4282109998Smarkm TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or 4283109998Smarkm the content type as defined in the SSL 3.0/TLS 1.0 protocol 4284109998Smarkm specification (change_cipher_spec(20), alert(21), handshake(22)). 4285109998Smarkm 'buf' and 'len' point to the actual message, 'ssl' to the 4286109998Smarkm SSL object, and 'arg' is the application-defined value set by 4287109998Smarkm SSL[_CTX]_set_msg_callback_arg(). 4288109998Smarkm 4289109998Smarkm 'openssl s_client' and 'openssl s_server' have new '-msg' options 4290109998Smarkm to enable a callback that displays all protocol messages. 4291109998Smarkm [Bodo Moeller] 4292109998Smarkm 4293109998Smarkm *) Change the shared library support so shared libraries are built as 4294109998Smarkm soon as the corresponding static library is finished, and thereby get 4295109998Smarkm openssl and the test programs linked against the shared library. 4296109998Smarkm This still only happens when the keyword "shard" has been given to 4297109998Smarkm the configuration scripts. 4298109998Smarkm 4299109998Smarkm NOTE: shared library support is still an experimental thing, and 4300109998Smarkm backward binary compatibility is still not guaranteed. 4301109998Smarkm ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte] 4302109998Smarkm 4303109998Smarkm *) Add support for Subject Information Access extension. 4304109998Smarkm [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 4305109998Smarkm 4306109998Smarkm *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero 4307109998Smarkm additional bytes when new memory had to be allocated, not just 4308109998Smarkm when reusing an existing buffer. 4309109998Smarkm [Bodo Moeller] 4310109998Smarkm 4311109998Smarkm *) New command line and configuration option 'utf8' for the req command. 4312109998Smarkm This allows field values to be specified as UTF8 strings. 4313109998Smarkm [Steve Henson] 4314109998Smarkm 4315109998Smarkm *) Add -multi and -mr options to "openssl speed" - giving multiple parallel 4316109998Smarkm runs for the former and machine-readable output for the latter. 4317109998Smarkm [Ben Laurie] 4318109998Smarkm 4319109998Smarkm *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion 4320109998Smarkm of the e-mail address in the DN (i.e., it will go into a certificate 4321109998Smarkm extension only). The new configuration file option 'email_in_dn = no' 4322109998Smarkm has the same effect. 4323109998Smarkm [Massimiliano Pala madwolf@openca.org] 4324109998Smarkm 4325109998Smarkm *) Change all functions with names starting with des_ to be starting 4326109998Smarkm with DES_ instead. Add wrappers that are compatible with libdes, 4327109998Smarkm but are named _ossl_old_des_*. Finally, add macros that map the 4328109998Smarkm des_* symbols to the corresponding _ossl_old_des_* if libdes 4329109998Smarkm compatibility is desired. If OpenSSL 0.9.6c compatibility is 4330109998Smarkm desired, the des_* symbols will be mapped to DES_*, with one 4331109998Smarkm exception. 4332109998Smarkm 4333109998Smarkm Since we provide two compatibility mappings, the user needs to 4334109998Smarkm define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes 4335109998Smarkm compatibility is desired. The default (i.e., when that macro 4336109998Smarkm isn't defined) is OpenSSL 0.9.6c compatibility. 4337109998Smarkm 4338109998Smarkm There are also macros that enable and disable the support of old 4339109998Smarkm des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT 4340109998Smarkm and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those 4341109998Smarkm are defined, the default will apply: to support the old des routines. 4342109998Smarkm 4343109998Smarkm In either case, one must include openssl/des.h to get the correct 4344109998Smarkm definitions. Do not try to just include openssl/des_old.h, that 4345109998Smarkm won't work. 4346109998Smarkm 4347109998Smarkm NOTE: This is a major break of an old API into a new one. Software 4348109998Smarkm authors are encouraged to switch to the DES_ style functions. Some 4349109998Smarkm time in the future, des_old.h and the libdes compatibility functions 4350109998Smarkm will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the 4351109998Smarkm default), and then completely removed. 4352109998Smarkm [Richard Levitte] 4353109998Smarkm 4354109998Smarkm *) Test for certificates which contain unsupported critical extensions. 4355109998Smarkm If such a certificate is found during a verify operation it is 4356109998Smarkm rejected by default: this behaviour can be overridden by either 4357109998Smarkm handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or 4358109998Smarkm by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function 4359109998Smarkm X509_supported_extension() has also been added which returns 1 if a 4360109998Smarkm particular extension is supported. 4361109998Smarkm [Steve Henson] 4362109998Smarkm 4363109998Smarkm *) Modify the behaviour of EVP cipher functions in similar way to digests 4364109998Smarkm to retain compatibility with existing code. 4365109998Smarkm [Steve Henson] 4366109998Smarkm 4367109998Smarkm *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain 4368109998Smarkm compatibility with existing code. In particular the 'ctx' parameter does 4369109998Smarkm not have to be to be initialized before the call to EVP_DigestInit() and 4370109998Smarkm it is tidied up after a call to EVP_DigestFinal(). New function 4371109998Smarkm EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function 4372109998Smarkm EVP_MD_CTX_copy() changed to not require the destination to be 4373109998Smarkm initialized valid and new function EVP_MD_CTX_copy_ex() added which 4374109998Smarkm requires the destination to be valid. 4375109998Smarkm 4376109998Smarkm Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), 4377109998Smarkm EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). 4378109998Smarkm [Steve Henson] 4379109998Smarkm 4380109998Smarkm *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it 4381109998Smarkm so that complete 'Handshake' protocol structures are kept in memory 4382109998Smarkm instead of overwriting 'msg_type' and 'length' with 'body' data. 4383109998Smarkm [Bodo Moeller] 4384109998Smarkm 4385109998Smarkm *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. 4386109998Smarkm [Massimo Santin via Richard Levitte] 4387109998Smarkm 4388109998Smarkm *) Major restructuring to the underlying ENGINE code. This includes 4389109998Smarkm reduction of linker bloat, separation of pure "ENGINE" manipulation 4390109998Smarkm (initialisation, etc) from functionality dealing with implementations 4391109998Smarkm of specific crypto iterfaces. This change also introduces integrated 4392109998Smarkm support for symmetric ciphers and digest implementations - so ENGINEs 4393109998Smarkm can now accelerate these by providing EVP_CIPHER and EVP_MD 4394109998Smarkm implementations of their own. This is detailed in crypto/engine/README 4395109998Smarkm as it couldn't be adequately described here. However, there are a few 4396109998Smarkm API changes worth noting - some RSA, DSA, DH, and RAND functions that 4397109998Smarkm were changed in the original introduction of ENGINE code have now 4398109998Smarkm reverted back - the hooking from this code to ENGINE is now a good 4399109998Smarkm deal more passive and at run-time, operations deal directly with 4400109998Smarkm RSA_METHODs, DSA_METHODs (etc) as they did before, rather than 4401109998Smarkm dereferencing through an ENGINE pointer any more. Also, the ENGINE 4402109998Smarkm functions dealing with BN_MOD_EXP[_CRT] handlers have been removed - 4403109998Smarkm they were not being used by the framework as there is no concept of a 4404109998Smarkm BIGNUM_METHOD and they could not be generalised to the new 4405109998Smarkm 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, 4406109998Smarkm ENGINE_cpy() has been removed as it cannot be consistently defined in 4407109998Smarkm the new code. 4408109998Smarkm [Geoff Thorpe] 4409109998Smarkm 4410109998Smarkm *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. 4411109998Smarkm [Steve Henson] 4412109998Smarkm 4413109998Smarkm *) Change mkdef.pl to sort symbols that get the same entry number, 4414109998Smarkm and make sure the automatically generated functions ERR_load_* 4415109998Smarkm become part of libeay.num as well. 4416109998Smarkm [Richard Levitte] 4417109998Smarkm 4418109998Smarkm *) New function SSL_renegotiate_pending(). This returns true once 4419109998Smarkm renegotiation has been requested (either SSL_renegotiate() call 4420109998Smarkm or HelloRequest/ClientHello receveived from the peer) and becomes 4421109998Smarkm false once a handshake has been completed. 4422109998Smarkm (For servers, SSL_renegotiate() followed by SSL_do_handshake() 4423109998Smarkm sends a HelloRequest, but does not ensure that a handshake takes 4424109998Smarkm place. SSL_renegotiate_pending() is useful for checking if the 4425109998Smarkm client has followed the request.) 4426109998Smarkm [Bodo Moeller] 4427109998Smarkm 4428109998Smarkm *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. 4429109998Smarkm By default, clients may request session resumption even during 4430109998Smarkm renegotiation (if session ID contexts permit); with this option, 4431109998Smarkm session resumption is possible only in the first handshake. 4432109998Smarkm 4433109998Smarkm SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes 4434109998Smarkm more bits available for options that should not be part of 4435109998Smarkm SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). 4436109998Smarkm [Bodo Moeller] 4437109998Smarkm 4438109998Smarkm *) Add some demos for certificate and certificate request creation. 4439109998Smarkm [Steve Henson] 4440109998Smarkm 4441109998Smarkm *) Make maximum certificate chain size accepted from the peer application 4442109998Smarkm settable (SSL*_get/set_max_cert_list()), as proposed by 4443109998Smarkm "Douglas E. Engert" <deengert@anl.gov>. 4444109998Smarkm [Lutz Jaenicke] 4445109998Smarkm 4446109998Smarkm *) Add support for shared libraries for Unixware-7 4447109998Smarkm (Boyd Lynn Gerber <gerberb@zenez.com>). 4448109998Smarkm [Lutz Jaenicke] 4449109998Smarkm 4450109998Smarkm *) Add a "destroy" handler to ENGINEs that allows structural cleanup to 4451109998Smarkm be done prior to destruction. Use this to unload error strings from 4452109998Smarkm ENGINEs that load their own error strings. NB: This adds two new API 4453109998Smarkm functions to "get" and "set" this destroy handler in an ENGINE. 4454109998Smarkm [Geoff Thorpe] 4455109998Smarkm 4456109998Smarkm *) Alter all existing ENGINE implementations (except "openssl" and 4457109998Smarkm "openbsd") to dynamically instantiate their own error strings. This 4458109998Smarkm makes them more flexible to be built both as statically-linked ENGINEs 4459109998Smarkm and self-contained shared-libraries loadable via the "dynamic" ENGINE. 4460109998Smarkm Also, add stub code to each that makes building them as self-contained 4461109998Smarkm shared-libraries easier (see README.ENGINE). 4462109998Smarkm [Geoff Thorpe] 4463109998Smarkm 4464109998Smarkm *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE 4465109998Smarkm implementations into applications that are completely implemented in 4466109998Smarkm self-contained shared-libraries. The "dynamic" ENGINE exposes control 4467109998Smarkm commands that can be used to configure what shared-library to load and 4468109998Smarkm to control aspects of the way it is handled. Also, made an update to 4469109998Smarkm the README.ENGINE file that brings its information up-to-date and 4470109998Smarkm provides some information and instructions on the "dynamic" ENGINE 4471109998Smarkm (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). 4472109998Smarkm [Geoff Thorpe] 4473109998Smarkm 4474109998Smarkm *) Make it possible to unload ranges of ERR strings with a new 4475109998Smarkm "ERR_unload_strings" function. 4476109998Smarkm [Geoff Thorpe] 4477109998Smarkm 4478109998Smarkm *) Add a copy() function to EVP_MD. 4479109998Smarkm [Ben Laurie] 4480109998Smarkm 4481109998Smarkm *) Make EVP_MD routines take a context pointer instead of just the 4482109998Smarkm md_data void pointer. 4483109998Smarkm [Ben Laurie] 4484109998Smarkm 4485109998Smarkm *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates 4486109998Smarkm that the digest can only process a single chunk of data 4487109998Smarkm (typically because it is provided by a piece of 4488109998Smarkm hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application 4489109998Smarkm is only going to provide a single chunk of data, and hence the 4490109998Smarkm framework needn't accumulate the data for oneshot drivers. 4491109998Smarkm [Ben Laurie] 4492109998Smarkm 4493109998Smarkm *) As with "ERR", make it possible to replace the underlying "ex_data" 4494109998Smarkm functions. This change also alters the storage and management of global 4495109998Smarkm ex_data state - it's now all inside ex_data.c and all "class" code (eg. 4496109998Smarkm RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class 4497109998Smarkm index counters. The API functions that use this state have been changed 4498109998Smarkm to take a "class_index" rather than pointers to the class's local STACK 4499109998Smarkm and counter, and there is now an API function to dynamically create new 4500109998Smarkm classes. This centralisation allows us to (a) plug a lot of the 4501109998Smarkm thread-safety problems that existed, and (b) makes it possible to clean 4502109998Smarkm up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) 4503109998Smarkm such data would previously have always leaked in application code and 4504109998Smarkm workarounds were in place to make the memory debugging turn a blind eye 4505109998Smarkm to it. Application code that doesn't use this new function will still 4506109998Smarkm leak as before, but their memory debugging output will announce it now 4507109998Smarkm rather than letting it slide. 4508109998Smarkm 4509109998Smarkm Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change 4510109998Smarkm induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now 4511109998Smarkm has a return value to indicate success or failure. 4512109998Smarkm [Geoff Thorpe] 4513109998Smarkm 4514109998Smarkm *) Make it possible to replace the underlying "ERR" functions such that the 4515109998Smarkm global state (2 LHASH tables and 2 locks) is only used by the "default" 4516109998Smarkm implementation. This change also adds two functions to "get" and "set" 4517109998Smarkm the implementation prior to it being automatically set the first time 4518109998Smarkm any other ERR function takes place. Ie. an application can call "get", 4519109998Smarkm pass the return value to a module it has just loaded, and that module 4520109998Smarkm can call its own "set" function using that value. This means the 4521109998Smarkm module's "ERR" operations will use (and modify) the error state in the 4522109998Smarkm application and not in its own statically linked copy of OpenSSL code. 4523109998Smarkm [Geoff Thorpe] 4524109998Smarkm 4525109998Smarkm *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment 4526109998Smarkm reference counts. This performs normal REF_PRINT/REF_CHECK macros on 4527109998Smarkm the operation, and provides a more encapsulated way for external code 4528109998Smarkm (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code 4529109998Smarkm to use these functions rather than manually incrementing the counts. 4530109998Smarkm 4531109998Smarkm Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". 4532109998Smarkm [Geoff Thorpe] 4533109998Smarkm 4534109998Smarkm *) Add EVP test program. 4535109998Smarkm [Ben Laurie] 4536109998Smarkm 4537109998Smarkm *) Add symmetric cipher support to ENGINE. Expect the API to change! 4538109998Smarkm [Ben Laurie] 4539109998Smarkm 4540109998Smarkm *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() 4541109998Smarkm X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), 4542109998Smarkm X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). 4543109998Smarkm These allow a CRL to be built without having to access X509_CRL fields 4544109998Smarkm directly. Modify 'ca' application to use new functions. 4545109998Smarkm [Steve Henson] 4546109998Smarkm 4547109998Smarkm *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended 4548109998Smarkm bug workarounds. Rollback attack detection is a security feature. 4549109998Smarkm The problem will only arise on OpenSSL servers when TLSv1 is not 4550109998Smarkm available (sslv3_server_method() or SSL_OP_NO_TLSv1). 4551109998Smarkm Software authors not wanting to support TLSv1 will have special reasons 4552109998Smarkm for their choice and can explicitly enable this option. 4553109998Smarkm [Bodo Moeller, Lutz Jaenicke] 4554109998Smarkm 4555109998Smarkm *) Rationalise EVP so it can be extended: don't include a union of 4556109998Smarkm cipher/digest structures, add init/cleanup functions for EVP_MD_CTX 4557109998Smarkm (similar to those existing for EVP_CIPHER_CTX). 4558109998Smarkm Usage example: 4559109998Smarkm 4560109998Smarkm EVP_MD_CTX md; 4561109998Smarkm 4562109998Smarkm EVP_MD_CTX_init(&md); /* new function call */ 4563109998Smarkm EVP_DigestInit(&md, EVP_sha1()); 4564109998Smarkm EVP_DigestUpdate(&md, in, len); 4565109998Smarkm EVP_DigestFinal(&md, out, NULL); 4566109998Smarkm EVP_MD_CTX_cleanup(&md); /* new function call */ 4567109998Smarkm 4568109998Smarkm [Ben Laurie] 4569109998Smarkm 4570109998Smarkm *) Make DES key schedule conform to the usual scheme, as well as 4571109998Smarkm correcting its structure. This means that calls to DES functions 4572109998Smarkm now have to pass a pointer to a des_key_schedule instead of a 4573109998Smarkm plain des_key_schedule (which was actually always a pointer 4574109998Smarkm anyway): E.g., 4575109998Smarkm 4576109998Smarkm des_key_schedule ks; 4577109998Smarkm 4578109998Smarkm des_set_key_checked(..., &ks); 4579109998Smarkm des_ncbc_encrypt(..., &ks, ...); 4580109998Smarkm 4581109998Smarkm (Note that a later change renames 'des_...' into 'DES_...'.) 4582109998Smarkm [Ben Laurie] 4583109998Smarkm 4584109998Smarkm *) Initial reduction of linker bloat: the use of some functions, such as 4585109998Smarkm PEM causes large amounts of unused functions to be linked in due to 4586109998Smarkm poor organisation. For example pem_all.c contains every PEM function 4587109998Smarkm which has a knock on effect of linking in large amounts of (unused) 4588109998Smarkm ASN1 code. Grouping together similar functions and splitting unrelated 4589109998Smarkm functions prevents this. 4590109998Smarkm [Steve Henson] 4591109998Smarkm 4592109998Smarkm *) Cleanup of EVP macros. 4593109998Smarkm [Ben Laurie] 4594109998Smarkm 4595109998Smarkm *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the 4596109998Smarkm correct _ecb suffix. 4597109998Smarkm [Ben Laurie] 4598109998Smarkm 4599109998Smarkm *) Add initial OCSP responder support to ocsp application. The 4600109998Smarkm revocation information is handled using the text based index 4601109998Smarkm use by the ca application. The responder can either handle 4602109998Smarkm requests generated internally, supplied in files (for example 4603109998Smarkm via a CGI script) or using an internal minimal server. 4604109998Smarkm [Steve Henson] 4605109998Smarkm 4606109998Smarkm *) Add configuration choices to get zlib compression for TLS. 4607109998Smarkm [Richard Levitte] 4608109998Smarkm 4609109998Smarkm *) Changes to Kerberos SSL for RFC 2712 compliance: 4610109998Smarkm 1. Implemented real KerberosWrapper, instead of just using 4611109998Smarkm KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] 4612109998Smarkm 2. Implemented optional authenticator field of KerberosWrapper. 4613109998Smarkm 4614109998Smarkm Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, 4615109998Smarkm and authenticator structs; see crypto/krb5/. 4616109998Smarkm 4617109998Smarkm Generalized Kerberos calls to support multiple Kerberos libraries. 4618109998Smarkm [Vern Staats <staatsvr@asc.hpc.mil>, 4619109998Smarkm Jeffrey Altman <jaltman@columbia.edu> 4620109998Smarkm via Richard Levitte] 4621109998Smarkm 4622109998Smarkm *) Cause 'openssl speed' to use fully hard-coded DSA keys as it 4623109998Smarkm already does with RSA. testdsa.h now has 'priv_key/pub_key' 4624109998Smarkm values for each of the key sizes rather than having just 4625109998Smarkm parameters (and 'speed' generating keys each time). 4626109998Smarkm [Geoff Thorpe] 4627109998Smarkm 4628109998Smarkm *) Speed up EVP routines. 4629109998Smarkm Before: 4630109998Smarkmencrypt 4631109998Smarkmtype 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 4632109998Smarkmdes-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k 4633109998Smarkmdes-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k 4634109998Smarkmdes-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k 4635109998Smarkmdecrypt 4636109998Smarkmdes-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k 4637109998Smarkmdes-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k 4638109998Smarkmdes-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k 4639109998Smarkm After: 4640109998Smarkmencrypt 4641109998Smarkmdes-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k 4642109998Smarkmdecrypt 4643109998Smarkmdes-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k 4644109998Smarkm [Ben Laurie] 4645109998Smarkm 4646109998Smarkm *) Added the OS2-EMX target. 4647109998Smarkm ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte] 4648109998Smarkm 4649109998Smarkm *) Rewrite apps to use NCONF routines instead of the old CONF. New functions 4650109998Smarkm to support NCONF routines in extension code. New function CONF_set_nconf() 4651109998Smarkm to allow functions which take an NCONF to also handle the old LHASH 4652109998Smarkm structure: this means that the old CONF compatible routines can be 4653109998Smarkm retained (in particular wrt extensions) without having to duplicate the 4654109998Smarkm code. New function X509V3_add_ext_nconf_sk to add extensions to a stack. 4655109998Smarkm [Steve Henson] 4656109998Smarkm 4657109998Smarkm *) Enhance the general user interface with mechanisms for inner control 4658109998Smarkm and with possibilities to have yes/no kind of prompts. 4659109998Smarkm [Richard Levitte] 4660109998Smarkm 4661109998Smarkm *) Change all calls to low level digest routines in the library and 4662109998Smarkm applications to use EVP. Add missing calls to HMAC_cleanup() and 4663109998Smarkm don't assume HMAC_CTX can be copied using memcpy(). 4664109998Smarkm [Verdon Walker <VWalker@novell.com>, Steve Henson] 4665109998Smarkm 4666109998Smarkm *) Add the possibility to control engines through control names but with 4667109998Smarkm arbitrary arguments instead of just a string. 4668109998Smarkm Change the key loaders to take a UI_METHOD instead of a callback 4669109998Smarkm function pointer. NOTE: this breaks binary compatibility with earlier 4670109998Smarkm versions of OpenSSL [engine]. 4671109998Smarkm Adapt the nCipher code for these new conditions and add a card insertion 4672109998Smarkm callback. 4673109998Smarkm [Richard Levitte] 4674109998Smarkm 4675109998Smarkm *) Enhance the general user interface with mechanisms to better support 4676109998Smarkm dialog box interfaces, application-defined prompts, the possibility 4677109998Smarkm to use defaults (for example default passwords from somewhere else) 4678109998Smarkm and interrupts/cancellations. 4679109998Smarkm [Richard Levitte] 4680109998Smarkm 4681109998Smarkm *) Tidy up PKCS#12 attribute handling. Add support for the CSP name 4682109998Smarkm attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. 4683109998Smarkm [Steve Henson] 4684109998Smarkm 4685109998Smarkm *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also 4686109998Smarkm tidy up some unnecessarily weird code in 'sk_new()'). 4687109998Smarkm [Geoff, reported by Diego Tartara <dtartara@novamens.com>] 4688109998Smarkm 4689109998Smarkm *) Change the key loading routines for ENGINEs to use the same kind 4690109998Smarkm callback (pem_password_cb) as all other routines that need this 4691109998Smarkm kind of callback. 4692109998Smarkm [Richard Levitte] 4693109998Smarkm 4694109998Smarkm *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with 4695109998Smarkm 256 bit (=32 byte) keys. Of course seeding with more entropy bytes 4696109998Smarkm than this minimum value is recommended. 4697109998Smarkm [Lutz Jaenicke] 4698109998Smarkm 4699109998Smarkm *) New random seeder for OpenVMS, using the system process statistics 4700109998Smarkm that are easily reachable. 4701109998Smarkm [Richard Levitte] 4702109998Smarkm 4703109998Smarkm *) Windows apparently can't transparently handle global 4704109998Smarkm variables defined in DLLs. Initialisations such as: 4705109998Smarkm 4706109998Smarkm const ASN1_ITEM *it = &ASN1_INTEGER_it; 4707109998Smarkm 4708109998Smarkm wont compile. This is used by the any applications that need to 4709109998Smarkm declare their own ASN1 modules. This was fixed by adding the option 4710109998Smarkm EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly 4711109998Smarkm needed for static libraries under Win32. 4712109998Smarkm [Steve Henson] 4713109998Smarkm 4714109998Smarkm *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle 4715109998Smarkm setting of purpose and trust fields. New X509_STORE trust and 4716109998Smarkm purpose functions and tidy up setting in other SSL functions. 4717109998Smarkm [Steve Henson] 4718109998Smarkm 4719109998Smarkm *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE 4720109998Smarkm structure. These are inherited by X509_STORE_CTX when it is 4721109998Smarkm initialised. This allows various defaults to be set in the 4722109998Smarkm X509_STORE structure (such as flags for CRL checking and custom 4723109998Smarkm purpose or trust settings) for functions which only use X509_STORE_CTX 4724109998Smarkm internally such as S/MIME. 4725109998Smarkm 4726109998Smarkm Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and 4727109998Smarkm trust settings if they are not set in X509_STORE. This allows X509_STORE 4728109998Smarkm purposes and trust (in S/MIME for example) to override any set by default. 4729109998Smarkm 4730109998Smarkm Add command line options for CRL checking to smime, s_client and s_server 4731109998Smarkm applications. 4732109998Smarkm [Steve Henson] 4733109998Smarkm 4734109998Smarkm *) Initial CRL based revocation checking. If the CRL checking flag(s) 4735109998Smarkm are set then the CRL is looked up in the X509_STORE structure and 4736109998Smarkm its validity and signature checked, then if the certificate is found 4737109998Smarkm in the CRL the verify fails with a revoked error. 4738109998Smarkm 4739109998Smarkm Various new CRL related callbacks added to X509_STORE_CTX structure. 4740109998Smarkm 4741109998Smarkm Command line options added to 'verify' application to support this. 4742109998Smarkm 4743109998Smarkm This needs some additional work, such as being able to handle multiple 4744109998Smarkm CRLs with different times, extension based lookup (rather than just 4745109998Smarkm by subject name) and ultimately more complete V2 CRL extension 4746109998Smarkm handling. 4747109998Smarkm [Steve Henson] 4748109998Smarkm 4749109998Smarkm *) Add a general user interface API (crypto/ui/). This is designed 4750109998Smarkm to replace things like des_read_password and friends (backward 4751109998Smarkm compatibility functions using this new API are provided). 4752109998Smarkm The purpose is to remove prompting functions from the DES code 4753109998Smarkm section as well as provide for prompting through dialog boxes in 4754109998Smarkm a window system and the like. 4755109998Smarkm [Richard Levitte] 4756109998Smarkm 4757109998Smarkm *) Add "ex_data" support to ENGINE so implementations can add state at a 4758109998Smarkm per-structure level rather than having to store it globally. 4759109998Smarkm [Geoff] 4760109998Smarkm 4761109998Smarkm *) Make it possible for ENGINE structures to be copied when retrieved by 4762109998Smarkm ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. 4763109998Smarkm This causes the "original" ENGINE structure to act like a template, 4764109998Smarkm analogous to the RSA vs. RSA_METHOD type of separation. Because of this 4765109998Smarkm operational state can be localised to each ENGINE structure, despite the 4766109998Smarkm fact they all share the same "methods". New ENGINE structures returned in 4767109998Smarkm this case have no functional references and the return value is the single 4768109998Smarkm structural reference. This matches the single structural reference returned 4769109998Smarkm by ENGINE_by_id() normally, when it is incremented on the pre-existing 4770109998Smarkm ENGINE structure. 4771109998Smarkm [Geoff] 4772109998Smarkm 4773109998Smarkm *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this 4774109998Smarkm needs to match any other type at all we need to manually clear the 4775109998Smarkm tag cache. 4776109998Smarkm [Steve Henson] 4777109998Smarkm 4778109998Smarkm *) Changes to the "openssl engine" utility to include; 4779109998Smarkm - verbosity levels ('-v', '-vv', and '-vvv') that provide information 4780109998Smarkm about an ENGINE's available control commands. 4781109998Smarkm - executing control commands from command line arguments using the 4782109998Smarkm '-pre' and '-post' switches. '-post' is only used if '-t' is 4783109998Smarkm specified and the ENGINE is successfully initialised. The syntax for 4784109998Smarkm the individual commands are colon-separated, for example; 4785109998Smarkm openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so 4786109998Smarkm [Geoff] 4787109998Smarkm 4788109998Smarkm *) New dynamic control command support for ENGINEs. ENGINEs can now 4789109998Smarkm declare their own commands (numbers), names (strings), descriptions, 4790109998Smarkm and input types for run-time discovery by calling applications. A 4791109998Smarkm subset of these commands are implicitly classed as "executable" 4792109998Smarkm depending on their input type, and only these can be invoked through 4793109998Smarkm the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this 4794109998Smarkm can be based on user input, config files, etc). The distinction is 4795109998Smarkm that "executable" commands cannot return anything other than a boolean 4796109998Smarkm result and can only support numeric or string input, whereas some 4797109998Smarkm discoverable commands may only be for direct use through 4798109998Smarkm ENGINE_ctrl(), eg. supporting the exchange of binary data, function 4799109998Smarkm pointers, or other custom uses. The "executable" commands are to 4800109998Smarkm support parameterisations of ENGINE behaviour that can be 4801109998Smarkm unambiguously defined by ENGINEs and used consistently across any 4802109998Smarkm OpenSSL-based application. Commands have been added to all the 4803109998Smarkm existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow 4804109998Smarkm control over shared-library paths without source code alterations. 4805109998Smarkm [Geoff] 4806109998Smarkm 4807109998Smarkm *) Changed all ENGINE implementations to dynamically allocate their 4808109998Smarkm ENGINEs rather than declaring them statically. Apart from this being 4809109998Smarkm necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, 4810109998Smarkm this also allows the implementations to compile without using the 4811109998Smarkm internal engine_int.h header. 4812109998Smarkm [Geoff] 4813109998Smarkm 4814109998Smarkm *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a 4815109998Smarkm 'const' value. Any code that should be able to modify a RAND_METHOD 4816109998Smarkm should already have non-const pointers to it (ie. they should only 4817109998Smarkm modify their own ones). 4818109998Smarkm [Geoff] 4819109998Smarkm 4820109998Smarkm *) Made a variety of little tweaks to the ENGINE code. 4821109998Smarkm - "atalla" and "ubsec" string definitions were moved from header files 4822109998Smarkm to C code. "nuron" string definitions were placed in variables 4823109998Smarkm rather than hard-coded - allowing parameterisation of these values 4824109998Smarkm later on via ctrl() commands. 4825109998Smarkm - Removed unused "#if 0"'d code. 4826109998Smarkm - Fixed engine list iteration code so it uses ENGINE_free() to release 4827109998Smarkm structural references. 4828109998Smarkm - Constified the RAND_METHOD element of ENGINE structures. 4829109998Smarkm - Constified various get/set functions as appropriate and added 4830109998Smarkm missing functions (including a catch-all ENGINE_cpy that duplicates 4831109998Smarkm all ENGINE values onto a new ENGINE except reference counts/state). 4832109998Smarkm - Removed NULL parameter checks in get/set functions. Setting a method 4833109998Smarkm or function to NULL is a way of cancelling out a previously set 4834109998Smarkm value. Passing a NULL ENGINE parameter is just plain stupid anyway 4835109998Smarkm and doesn't justify the extra error symbols and code. 4836109998Smarkm - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for 4837109998Smarkm flags from engine_int.h to engine.h. 4838109998Smarkm - Changed prototypes for ENGINE handler functions (init(), finish(), 4839109998Smarkm ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. 4840109998Smarkm [Geoff] 4841109998Smarkm 4842109998Smarkm *) Implement binary inversion algorithm for BN_mod_inverse in addition 4843109998Smarkm to the algorithm using long division. The binary algorithm can be 4844109998Smarkm used only if the modulus is odd. On 32-bit systems, it is faster 4845109998Smarkm only for relatively small moduli (roughly 20-30% for 128-bit moduli, 4846109998Smarkm roughly 5-15% for 256-bit moduli), so we use it only for moduli 4847109998Smarkm up to 450 bits. In 64-bit environments, the binary algorithm 4848109998Smarkm appears to be advantageous for much longer moduli; here we use it 4849109998Smarkm for moduli up to 2048 bits. 4850109998Smarkm [Bodo Moeller] 4851109998Smarkm 4852109998Smarkm *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code 4853109998Smarkm could not support the combine flag in choice fields. 4854109998Smarkm [Steve Henson] 4855109998Smarkm 4856109998Smarkm *) Add a 'copy_extensions' option to the 'ca' utility. This copies 4857109998Smarkm extensions from a certificate request to the certificate. 4858109998Smarkm [Steve Henson] 4859109998Smarkm 4860109998Smarkm *) Allow multiple 'certopt' and 'nameopt' options to be separated 4861109998Smarkm by commas. Add 'namopt' and 'certopt' options to the 'ca' config 4862109998Smarkm file: this allows the display of the certificate about to be 4863109998Smarkm signed to be customised, to allow certain fields to be included 4864109998Smarkm or excluded and extension details. The old system didn't display 4865109998Smarkm multicharacter strings properly, omitted fields not in the policy 4866109998Smarkm and couldn't display additional details such as extensions. 4867109998Smarkm [Steve Henson] 4868109998Smarkm 4869109998Smarkm *) Function EC_POINTs_mul for multiple scalar multiplication 4870109998Smarkm of an arbitrary number of elliptic curve points 4871109998Smarkm \sum scalars[i]*points[i], 4872109998Smarkm optionally including the generator defined for the EC_GROUP: 4873109998Smarkm scalar*generator + \sum scalars[i]*points[i]. 4874109998Smarkm 4875109998Smarkm EC_POINT_mul is a simple wrapper function for the typical case 4876109998Smarkm that the point list has just one item (besides the optional 4877109998Smarkm generator). 4878109998Smarkm [Bodo Moeller] 4879109998Smarkm 4880109998Smarkm *) First EC_METHODs for curves over GF(p): 4881109998Smarkm 4882109998Smarkm EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr 4883109998Smarkm operations and provides various method functions that can also 4884109998Smarkm operate with faster implementations of modular arithmetic. 4885109998Smarkm 4886109998Smarkm EC_GFp_mont_method() reuses most functions that are part of 4887109998Smarkm EC_GFp_simple_method, but uses Montgomery arithmetic. 4888109998Smarkm 4889109998Smarkm [Bodo Moeller; point addition and point doubling 4890109998Smarkm implementation directly derived from source code provided by 4891109998Smarkm Lenka Fibikova <fibikova@exp-math.uni-essen.de>] 4892109998Smarkm 4893109998Smarkm *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, 4894109998Smarkm crypto/ec/ec_lib.c): 4895109998Smarkm 4896109998Smarkm Curves are EC_GROUP objects (with an optional group generator) 4897109998Smarkm based on EC_METHODs that are built into the library. 4898109998Smarkm 4899109998Smarkm Points are EC_POINT objects based on EC_GROUP objects. 4900109998Smarkm 4901109998Smarkm Most of the framework would be able to handle curves over arbitrary 4902109998Smarkm finite fields, but as there are no obvious types for fields other 4903109998Smarkm than GF(p), some functions are limited to that for now. 4904109998Smarkm [Bodo Moeller] 4905109998Smarkm 4906109998Smarkm *) Add the -HTTP option to s_server. It is similar to -WWW, but requires 4907109998Smarkm that the file contains a complete HTTP response. 4908109998Smarkm [Richard Levitte] 4909109998Smarkm 4910109998Smarkm *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl 4911109998Smarkm change the def and num file printf format specifier from "%-40sXXX" 4912109998Smarkm to "%-39s XXX". The latter will always guarantee a space after the 4913109998Smarkm field while the former will cause them to run together if the field 4914109998Smarkm is 40 of more characters long. 4915109998Smarkm [Steve Henson] 4916109998Smarkm 4917109998Smarkm *) Constify the cipher and digest 'method' functions and structures 4918109998Smarkm and modify related functions to take constant EVP_MD and EVP_CIPHER 4919109998Smarkm pointers. 4920109998Smarkm [Steve Henson] 4921109998Smarkm 4922109998Smarkm *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them 4923109998Smarkm in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. 4924109998Smarkm [Bodo Moeller] 4925109998Smarkm 4926109998Smarkm *) Modify EVP_Digest*() routines so they now return values. Although the 4927109998Smarkm internal software routines can never fail additional hardware versions 4928109998Smarkm might. 4929109998Smarkm [Steve Henson] 4930109998Smarkm 4931109998Smarkm *) Clean up crypto/err/err.h and change some error codes to avoid conflicts: 4932109998Smarkm 4933109998Smarkm Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 4934109998Smarkm (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. 4935109998Smarkm 4936109998Smarkm ASN1 error codes 4937109998Smarkm ERR_R_NESTED_ASN1_ERROR 4938109998Smarkm ... 4939109998Smarkm ERR_R_MISSING_ASN1_EOS 4940109998Smarkm were 4 .. 9, conflicting with 4941109998Smarkm ERR_LIB_RSA (= ERR_R_RSA_LIB) 4942109998Smarkm ... 4943109998Smarkm ERR_LIB_PEM (= ERR_R_PEM_LIB). 4944109998Smarkm They are now 58 .. 63 (i.e., just below ERR_R_FATAL). 4945109998Smarkm 4946109998Smarkm Add new error code 'ERR_R_INTERNAL_ERROR'. 4947109998Smarkm [Bodo Moeller] 4948109998Smarkm 4949109998Smarkm *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock 4950109998Smarkm suffices. 4951109998Smarkm [Bodo Moeller] 4952109998Smarkm 4953109998Smarkm *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This 4954109998Smarkm sets the subject name for a new request or supersedes the 4955109998Smarkm subject name in a given request. Formats that can be parsed are 4956109998Smarkm 'CN=Some Name, OU=myOU, C=IT' 4957109998Smarkm and 4958109998Smarkm 'CN=Some Name/OU=myOU/C=IT'. 4959109998Smarkm 4960109998Smarkm Add options '-batch' and '-verbose' to 'openssl req'. 4961109998Smarkm [Massimiliano Pala <madwolf@hackmasters.net>] 4962109998Smarkm 4963109998Smarkm *) Introduce the possibility to access global variables through 4964109998Smarkm functions on platform were that's the best way to handle exporting 4965109998Smarkm global variables in shared libraries. To enable this functionality, 4966109998Smarkm one must configure with "EXPORT_VAR_AS_FN" or defined the C macro 4967109998Smarkm "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter 4968109998Smarkm is normally done by Configure or something similar). 4969109998Smarkm 4970109998Smarkm To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL 4971109998Smarkm in the source file (foo.c) like this: 4972109998Smarkm 4973109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; 4974109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(double,bar); 4975109998Smarkm 4976109998Smarkm To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL 4977109998Smarkm and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: 4978109998Smarkm 4979109998Smarkm OPENSSL_DECLARE_GLOBAL(int,foo); 4980109998Smarkm #define foo OPENSSL_GLOBAL_REF(foo) 4981109998Smarkm OPENSSL_DECLARE_GLOBAL(double,bar); 4982109998Smarkm #define bar OPENSSL_GLOBAL_REF(bar) 4983109998Smarkm 4984109998Smarkm The #defines are very important, and therefore so is including the 4985109998Smarkm header file everywhere where the defined globals are used. 4986109998Smarkm 4987109998Smarkm The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition 4988109998Smarkm of ASN.1 items, but that structure is a bit different. 4989109998Smarkm 4990109998Smarkm The largest change is in util/mkdef.pl which has been enhanced with 4991109998Smarkm better and easier to understand logic to choose which symbols should 4992109998Smarkm go into the Windows .def files as well as a number of fixes and code 4993109998Smarkm cleanup (among others, algorithm keywords are now sorted 4994109998Smarkm lexicographically to avoid constant rewrites). 4995109998Smarkm [Richard Levitte] 4996109998Smarkm 4997109998Smarkm *) In BN_div() keep a copy of the sign of 'num' before writing the 4998109998Smarkm result to 'rm' because if rm==num the value will be overwritten 4999109998Smarkm and produce the wrong result if 'num' is negative: this caused 5000109998Smarkm problems with BN_mod() and BN_nnmod(). 5001109998Smarkm [Steve Henson] 5002109998Smarkm 5003109998Smarkm *) Function OCSP_request_verify(). This checks the signature on an 5004109998Smarkm OCSP request and verifies the signer certificate. The signer 5005109998Smarkm certificate is just checked for a generic purpose and OCSP request 5006109998Smarkm trust settings. 5007109998Smarkm [Steve Henson] 5008109998Smarkm 5009109998Smarkm *) Add OCSP_check_validity() function to check the validity of OCSP 5010109998Smarkm responses. OCSP responses are prepared in real time and may only 5011109998Smarkm be a few seconds old. Simply checking that the current time lies 5012109998Smarkm between thisUpdate and nextUpdate max reject otherwise valid responses 5013109998Smarkm caused by either OCSP responder or client clock inaccuracy. Instead 5014109998Smarkm we allow thisUpdate and nextUpdate to fall within a certain period of 5015109998Smarkm the current time. The age of the response can also optionally be 5016109998Smarkm checked. Two new options -validity_period and -status_age added to 5017109998Smarkm ocsp utility. 5018109998Smarkm [Steve Henson] 5019109998Smarkm 5020109998Smarkm *) If signature or public key algorithm is unrecognized print out its 5021109998Smarkm OID rather that just UNKNOWN. 5022109998Smarkm [Steve Henson] 5023109998Smarkm 5024109998Smarkm *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and 5025109998Smarkm OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate 5026109998Smarkm ID to be generated from the issuer certificate alone which can then be 5027109998Smarkm passed to OCSP_id_issuer_cmp(). 5028109998Smarkm [Steve Henson] 5029109998Smarkm 5030109998Smarkm *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new 5031109998Smarkm ASN1 modules to export functions returning ASN1_ITEM pointers 5032109998Smarkm instead of the ASN1_ITEM structures themselves. This adds several 5033109998Smarkm new macros which allow the underlying ASN1 function/structure to 5034109998Smarkm be accessed transparently. As a result code should not use ASN1_ITEM 5035109998Smarkm references directly (such as &X509_it) but instead use the relevant 5036109998Smarkm macros (such as ASN1_ITEM_rptr(X509)). This option is to allow 5037109998Smarkm use of the new ASN1 code on platforms where exporting structures 5038109998Smarkm is problematical (for example in shared libraries) but exporting 5039109998Smarkm functions returning pointers to structures is not. 5040109998Smarkm [Steve Henson] 5041109998Smarkm 5042109998Smarkm *) Add support for overriding the generation of SSL/TLS session IDs. 5043109998Smarkm These callbacks can be registered either in an SSL_CTX or per SSL. 5044109998Smarkm The purpose of this is to allow applications to control, if they wish, 5045109998Smarkm the arbitrary values chosen for use as session IDs, particularly as it 5046109998Smarkm can be useful for session caching in multiple-server environments. A 5047109998Smarkm command-line switch for testing this (and any client code that wishes 5048109998Smarkm to use such a feature) has been added to "s_server". 5049109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 5050109998Smarkm 5051109998Smarkm *) Modify mkdef.pl to recognise and parse preprocessor conditionals 5052109998Smarkm of the form '#if defined(...) || defined(...) || ...' and 5053109998Smarkm '#if !defined(...) && !defined(...) && ...'. This also avoids 5054109998Smarkm the growing number of special cases it was previously handling. 5055109998Smarkm [Richard Levitte] 5056109998Smarkm 5057109998Smarkm *) Make all configuration macros available for application by making 5058109998Smarkm sure they are available in opensslconf.h, by giving them names starting 5059109998Smarkm with "OPENSSL_" to avoid conflicts with other packages and by making 5060109998Smarkm sure e_os2.h will cover all platform-specific cases together with 5061109998Smarkm opensslconf.h. 5062109998Smarkm Additionally, it is now possible to define configuration/platform- 5063109998Smarkm specific names (called "system identities"). In the C code, these 5064109998Smarkm are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another 5065109998Smarkm macro with the name beginning with "OPENSSL_SYS_", which is determined 5066109998Smarkm from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on 5067109998Smarkm what is available. 5068109998Smarkm [Richard Levitte] 5069109998Smarkm 5070109998Smarkm *) New option -set_serial to 'req' and 'x509' this allows the serial 5071109998Smarkm number to use to be specified on the command line. Previously self 5072109998Smarkm signed certificates were hard coded with serial number 0 and the 5073109998Smarkm CA options of 'x509' had to use a serial number in a file which was 5074109998Smarkm auto incremented. 5075109998Smarkm [Steve Henson] 5076109998Smarkm 5077109998Smarkm *) New options to 'ca' utility to support V2 CRL entry extensions. 5078109998Smarkm Currently CRL reason, invalidity date and hold instruction are 5079109998Smarkm supported. Add new CRL extensions to V3 code and some new objects. 5080109998Smarkm [Steve Henson] 5081109998Smarkm 5082109998Smarkm *) New function EVP_CIPHER_CTX_set_padding() this is used to 5083109998Smarkm disable standard block padding (aka PKCS#5 padding) in the EVP 5084109998Smarkm API, which was previously mandatory. This means that the data is 5085109998Smarkm not padded in any way and so the total length much be a multiple 5086109998Smarkm of the block size, otherwise an error occurs. 5087109998Smarkm [Steve Henson] 5088109998Smarkm 5089109998Smarkm *) Initial (incomplete) OCSP SSL support. 5090109998Smarkm [Steve Henson] 5091109998Smarkm 5092109998Smarkm *) New function OCSP_parse_url(). This splits up a URL into its host, 5093109998Smarkm port and path components: primarily to parse OCSP URLs. New -url 5094109998Smarkm option to ocsp utility. 5095109998Smarkm [Steve Henson] 5096109998Smarkm 5097109998Smarkm *) New nonce behavior. The return value of OCSP_check_nonce() now 5098109998Smarkm reflects the various checks performed. Applications can decide 5099109998Smarkm whether to tolerate certain situations such as an absent nonce 5100109998Smarkm in a response when one was present in a request: the ocsp application 5101109998Smarkm just prints out a warning. New function OCSP_add1_basic_nonce() 5102109998Smarkm this is to allow responders to include a nonce in a response even if 5103109998Smarkm the request is nonce-less. 5104109998Smarkm [Steve Henson] 5105109998Smarkm 5106109998Smarkm *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are 5107109998Smarkm skipped when using openssl x509 multiple times on a single input file, 5108109998Smarkm e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs". 5109109998Smarkm [Bodo Moeller] 5110109998Smarkm 5111109998Smarkm *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() 5112109998Smarkm set string type: to handle setting ASN1_TIME structures. Fix ca 5113109998Smarkm utility to correctly initialize revocation date of CRLs. 5114109998Smarkm [Steve Henson] 5115109998Smarkm 5116109998Smarkm *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override 5117109998Smarkm the clients preferred ciphersuites and rather use its own preferences. 5118109998Smarkm Should help to work around M$ SGC (Server Gated Cryptography) bug in 5119109998Smarkm Internet Explorer by ensuring unchanged hash method during stepup. 5120109998Smarkm (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) 5121109998Smarkm [Lutz Jaenicke] 5122109998Smarkm 5123109998Smarkm *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael 5124109998Smarkm to aes and add a new 'exist' option to print out symbols that don't 5125109998Smarkm appear to exist. 5126109998Smarkm [Steve Henson] 5127109998Smarkm 5128109998Smarkm *) Additional options to ocsp utility to allow flags to be set and 5129109998Smarkm additional certificates supplied. 5130109998Smarkm [Steve Henson] 5131109998Smarkm 5132109998Smarkm *) Add the option -VAfile to 'openssl ocsp', so the user can give the 5133109998Smarkm OCSP client a number of certificate to only verify the response 5134109998Smarkm signature against. 5135109998Smarkm [Richard Levitte] 5136109998Smarkm 5137109998Smarkm *) Update Rijndael code to version 3.0 and change EVP AES ciphers to 5138109998Smarkm handle the new API. Currently only ECB, CBC modes supported. Add new 5139109998Smarkm AES OIDs. 5140109998Smarkm 5141109998Smarkm Add TLS AES ciphersuites as described in RFC3268, "Advanced 5142109998Smarkm Encryption Standard (AES) Ciphersuites for Transport Layer 5143109998Smarkm Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were 5144109998Smarkm not enabled by default and were not part of the "ALL" ciphersuite 5145109998Smarkm alias because they were not yet official; they could be 5146109998Smarkm explicitly requested by specifying the "AESdraft" ciphersuite 5147109998Smarkm group alias. In the final release of OpenSSL 0.9.7, the group 5148109998Smarkm alias is called "AES" and is part of "ALL".) 5149109998Smarkm [Ben Laurie, Steve Henson, Bodo Moeller] 5150109998Smarkm 5151109998Smarkm *) New function OCSP_copy_nonce() to copy nonce value (if present) from 5152109998Smarkm request to response. 5153109998Smarkm [Steve Henson] 5154109998Smarkm 5155109998Smarkm *) Functions for OCSP responders. OCSP_request_onereq_count(), 5156109998Smarkm OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() 5157109998Smarkm extract information from a certificate request. OCSP_response_create() 5158109998Smarkm creates a response and optionally adds a basic response structure. 5159109998Smarkm OCSP_basic_add1_status() adds a complete single response to a basic 5160109998Smarkm response and returns the OCSP_SINGLERESP structure just added (to allow 5161109998Smarkm extensions to be included for example). OCSP_basic_add1_cert() adds a 5162109998Smarkm certificate to a basic response and OCSP_basic_sign() signs a basic 5163109998Smarkm response with various flags. New helper functions ASN1_TIME_check() 5164109998Smarkm (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() 5165109998Smarkm (converts ASN1_TIME to GeneralizedTime). 5166109998Smarkm [Steve Henson] 5167109998Smarkm 5168109998Smarkm *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() 5169109998Smarkm in a single operation. X509_get0_pubkey_bitstr() extracts the public_key 5170109998Smarkm structure from a certificate. X509_pubkey_digest() digests the public_key 5171109998Smarkm contents: this is used in various key identifiers. 5172109998Smarkm [Steve Henson] 5173109998Smarkm 5174109998Smarkm *) Make sk_sort() tolerate a NULL argument. 5175109998Smarkm [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>] 5176109998Smarkm 5177109998Smarkm *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates 5178109998Smarkm passed by the function are trusted implicitly. If any of them signed the 5179109998Smarkm response then it is assumed to be valid and is not verified. 5180109998Smarkm [Steve Henson] 5181109998Smarkm 5182109998Smarkm *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT 5183109998Smarkm to data. This was previously part of the PKCS7 ASN1 code. This 5184109998Smarkm was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. 5185109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 5186109998Smarkm <support@securenetterm.com>] 5187109998Smarkm 5188109998Smarkm *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 5189109998Smarkm routines: without these tracing memory leaks is very painful. 5190109998Smarkm Fix leaks in PKCS12 and PKCS7 routines. 5191109998Smarkm [Steve Henson] 5192109998Smarkm 5193109998Smarkm *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). 5194109998Smarkm Previously it initialised the 'type' argument to V_ASN1_UTCTIME which 5195109998Smarkm effectively meant GeneralizedTime would never be used. Now it 5196109998Smarkm is initialised to -1 but X509_time_adj() now has to check the value 5197109998Smarkm and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or 5198109998Smarkm V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. 5199109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 5200109998Smarkm <support@securenetterm.com>] 5201109998Smarkm 5202109998Smarkm *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously 5203109998Smarkm result in a zero length in the ASN1_INTEGER structure which was 5204109998Smarkm not consistent with the structure when d2i_ASN1_INTEGER() was used 5205109998Smarkm and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() 5206109998Smarkm to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() 5207109998Smarkm where it did not print out a minus for negative ASN1_INTEGER. 5208109998Smarkm [Steve Henson] 5209109998Smarkm 5210109998Smarkm *) Add summary printout to ocsp utility. The various functions which 5211109998Smarkm convert status values to strings have been renamed to: 5212109998Smarkm OCSP_response_status_str(), OCSP_cert_status_str() and 5213109998Smarkm OCSP_crl_reason_str() and are no longer static. New options 5214109998Smarkm to verify nonce values and to disable verification. OCSP response 5215109998Smarkm printout format cleaned up. 5216109998Smarkm [Steve Henson] 5217109998Smarkm 5218109998Smarkm *) Add additional OCSP certificate checks. These are those specified 5219109998Smarkm in RFC2560. This consists of two separate checks: the CA of the 5220109998Smarkm certificate being checked must either be the OCSP signer certificate 5221109998Smarkm or the issuer of the OCSP signer certificate. In the latter case the 5222109998Smarkm OCSP signer certificate must contain the OCSP signing extended key 5223109998Smarkm usage. This check is performed by attempting to match the OCSP 5224109998Smarkm signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash 5225109998Smarkm in the OCSP_CERTID structures of the response. 5226109998Smarkm [Steve Henson] 5227109998Smarkm 5228109998Smarkm *) Initial OCSP certificate verification added to OCSP_basic_verify() 5229109998Smarkm and related routines. This uses the standard OpenSSL certificate 5230109998Smarkm verify routines to perform initial checks (just CA validity) and 5231109998Smarkm to obtain the certificate chain. Then additional checks will be 5232109998Smarkm performed on the chain. Currently the root CA is checked to see 5233109998Smarkm if it is explicitly trusted for OCSP signing. This is used to set 5234109998Smarkm a root CA as a global signing root: that is any certificate that 5235109998Smarkm chains to that CA is an acceptable OCSP signing certificate. 5236109998Smarkm [Steve Henson] 5237109998Smarkm 5238109998Smarkm *) New '-extfile ...' option to 'openssl ca' for reading X.509v3 5239109998Smarkm extensions from a separate configuration file. 5240109998Smarkm As when reading extensions from the main configuration file, 5241109998Smarkm the '-extensions ...' option may be used for specifying the 5242109998Smarkm section to use. 5243109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 5244109998Smarkm 5245109998Smarkm *) New OCSP utility. Allows OCSP requests to be generated or 5246109998Smarkm read. The request can be sent to a responder and the output 5247109998Smarkm parsed, outputed or printed in text form. Not complete yet: 5248109998Smarkm still needs to check the OCSP response validity. 5249109998Smarkm [Steve Henson] 5250109998Smarkm 5251109998Smarkm *) New subcommands for 'openssl ca': 5252109998Smarkm 'openssl ca -status <serial>' prints the status of the cert with 5253109998Smarkm the given serial number (according to the index file). 5254109998Smarkm 'openssl ca -updatedb' updates the expiry status of certificates 5255109998Smarkm in the index file. 5256109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 5257109998Smarkm 5258109998Smarkm *) New '-newreq-nodes' command option to CA.pl. This is like 5259109998Smarkm '-newreq', but calls 'openssl req' with the '-nodes' option 5260109998Smarkm so that the resulting key is not encrypted. 5261109998Smarkm [Damien Miller <djm@mindrot.org>] 5262109998Smarkm 5263109998Smarkm *) New configuration for the GNU Hurd. 5264109998Smarkm [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte] 5265109998Smarkm 5266109998Smarkm *) Initial code to implement OCSP basic response verify. This 5267109998Smarkm is currently incomplete. Currently just finds the signer's 5268109998Smarkm certificate and verifies the signature on the response. 5269109998Smarkm [Steve Henson] 5270109998Smarkm 5271109998Smarkm *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in 5272109998Smarkm value of OPENSSLDIR. This is available via the new '-d' option 5273109998Smarkm to 'openssl version', and is also included in 'openssl version -a'. 5274109998Smarkm [Bodo Moeller] 5275109998Smarkm 5276109998Smarkm *) Allowing defining memory allocation callbacks that will be given 5277109998Smarkm file name and line number information in additional arguments 5278109998Smarkm (a const char* and an int). The basic functionality remains, as 5279109998Smarkm well as the original possibility to just replace malloc(), 5280109998Smarkm realloc() and free() by functions that do not know about these 5281109998Smarkm additional arguments. To register and find out the current 5282109998Smarkm settings for extended allocation functions, the following 5283109998Smarkm functions are provided: 5284109998Smarkm 5285109998Smarkm CRYPTO_set_mem_ex_functions 5286109998Smarkm CRYPTO_set_locked_mem_ex_functions 5287109998Smarkm CRYPTO_get_mem_ex_functions 5288109998Smarkm CRYPTO_get_locked_mem_ex_functions 5289109998Smarkm 5290109998Smarkm These work the same way as CRYPTO_set_mem_functions and friends. 5291109998Smarkm CRYPTO_get_[locked_]mem_functions now writes 0 where such an 5292109998Smarkm extended allocation function is enabled. 5293109998Smarkm Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where 5294109998Smarkm a conventional allocation function is enabled. 5295109998Smarkm [Richard Levitte, Bodo Moeller] 5296109998Smarkm 5297109998Smarkm *) Finish off removing the remaining LHASH function pointer casts. 5298109998Smarkm There should no longer be any prototype-casting required when using 5299109998Smarkm the LHASH abstraction, and any casts that remain are "bugs". See 5300109998Smarkm the callback types and macros at the head of lhash.h for details 5301109998Smarkm (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). 5302109998Smarkm [Geoff Thorpe] 5303109998Smarkm 5304109998Smarkm *) Add automatic query of EGD sockets in RAND_poll() for the unix variant. 5305109998Smarkm If /dev/[u]random devices are not available or do not return enough 5306109998Smarkm entropy, EGD style sockets (served by EGD or PRNGD) will automatically 5307109998Smarkm be queried. 5308109998Smarkm The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and 5309109998Smarkm /etc/entropy will be queried once each in this sequence, quering stops 5310109998Smarkm when enough entropy was collected without querying more sockets. 5311109998Smarkm [Lutz Jaenicke] 5312109998Smarkm 5313109998Smarkm *) Change the Unix RAND_poll() variant to be able to poll several 5314109998Smarkm random devices, as specified by DEVRANDOM, until a sufficient amount 5315109998Smarkm of data has been collected. We spend at most 10 ms on each file 5316109998Smarkm (select timeout) and read in non-blocking mode. DEVRANDOM now 5317109998Smarkm defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" 5318109998Smarkm (previously it was just the string "/dev/urandom"), so on typical 5319109998Smarkm platforms the 10 ms delay will never occur. 5320109998Smarkm Also separate out the Unix variant to its own file, rand_unix.c. 5321109998Smarkm For VMS, there's a currently-empty rand_vms.c. 5322109998Smarkm [Richard Levitte] 5323109998Smarkm 5324109998Smarkm *) Move OCSP client related routines to ocsp_cl.c. These 5325109998Smarkm provide utility functions which an application needing 5326109998Smarkm to issue a request to an OCSP responder and analyse the 5327109998Smarkm response will typically need: as opposed to those which an 5328109998Smarkm OCSP responder itself would need which will be added later. 5329109998Smarkm 5330109998Smarkm OCSP_request_sign() signs an OCSP request with an API similar 5331109998Smarkm to PKCS7_sign(). OCSP_response_status() returns status of OCSP 5332109998Smarkm response. OCSP_response_get1_basic() extracts basic response 5333109998Smarkm from response. OCSP_resp_find_status(): finds and extracts status 5334109998Smarkm information from an OCSP_CERTID structure (which will be created 5335109998Smarkm when the request structure is built). These are built from lower 5336109998Smarkm level functions which work on OCSP_SINGLERESP structures but 5337109998Smarkm wont normally be used unless the application wishes to examine 5338109998Smarkm extensions in the OCSP response for example. 5339109998Smarkm 5340109998Smarkm Replace nonce routines with a pair of functions. 5341109998Smarkm OCSP_request_add1_nonce() adds a nonce value and optionally 5342109998Smarkm generates a random value. OCSP_check_nonce() checks the 5343109998Smarkm validity of the nonce in an OCSP response. 5344109998Smarkm [Steve Henson] 5345109998Smarkm 5346109998Smarkm *) Change function OCSP_request_add() to OCSP_request_add0_id(). 5347109998Smarkm This doesn't copy the supplied OCSP_CERTID and avoids the 5348109998Smarkm need to free up the newly created id. Change return type 5349109998Smarkm to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. 5350109998Smarkm This can then be used to add extensions to the request. 5351109998Smarkm Deleted OCSP_request_new(), since most of its functionality 5352109998Smarkm is now in OCSP_REQUEST_new() (and the case insensitive name 5353109998Smarkm clash) apart from the ability to set the request name which 5354109998Smarkm will be added elsewhere. 5355109998Smarkm [Steve Henson] 5356109998Smarkm 5357109998Smarkm *) Update OCSP API. Remove obsolete extensions argument from 5358109998Smarkm various functions. Extensions are now handled using the new 5359109998Smarkm OCSP extension code. New simple OCSP HTTP function which 5360109998Smarkm can be used to send requests and parse the response. 5361109998Smarkm [Steve Henson] 5362109998Smarkm 5363109998Smarkm *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new 5364109998Smarkm ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN 5365109998Smarkm uses the special reorder version of SET OF to sort the attributes 5366109998Smarkm and reorder them to match the encoded order. This resolves a long 5367109998Smarkm standing problem: a verify on a PKCS7 structure just after signing 5368109998Smarkm it used to fail because the attribute order did not match the 5369109998Smarkm encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: 5370109998Smarkm it uses the received order. This is necessary to tolerate some broken 5371109998Smarkm software that does not order SET OF. This is handled by encoding 5372109998Smarkm as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) 5373109998Smarkm to produce the required SET OF. 5374109998Smarkm [Steve Henson] 5375109998Smarkm 5376109998Smarkm *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and 5377109998Smarkm OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header 5378109998Smarkm files to get correct declarations of the ASN.1 item variables. 5379109998Smarkm [Richard Levitte] 5380109998Smarkm 5381109998Smarkm *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many 5382109998Smarkm PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: 5383109998Smarkm asn1_check_tlen() would sometimes attempt to use 'ctx' when it was 5384109998Smarkm NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). 5385109998Smarkm New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant 5386109998Smarkm ASN1_ITEM and no wrapper functions. 5387109998Smarkm [Steve Henson] 5388109998Smarkm 5389109998Smarkm *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These 5390109998Smarkm replace the old function pointer based I/O routines. Change most of 5391109998Smarkm the *_d2i_bio() and *_d2i_fp() functions to use these. 5392109998Smarkm [Steve Henson] 5393109998Smarkm 5394109998Smarkm *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor 5395109998Smarkm lines, recognice more "algorithms" that can be deselected, and make 5396109998Smarkm it complain about algorithm deselection that isn't recognised. 5397109998Smarkm [Richard Levitte] 5398109998Smarkm 5399109998Smarkm *) New ASN1 functions to handle dup, sign, verify, digest, pack and 5400109998Smarkm unpack operations in terms of ASN1_ITEM. Modify existing wrappers 5401109998Smarkm to use new functions. Add NO_ASN1_OLD which can be set to remove 5402109998Smarkm some old style ASN1 functions: this can be used to determine if old 5403109998Smarkm code will still work when these eventually go away. 5404109998Smarkm [Steve Henson] 5405109998Smarkm 5406109998Smarkm *) New extension functions for OCSP structures, these follow the 5407109998Smarkm same conventions as certificates and CRLs. 5408109998Smarkm [Steve Henson] 5409109998Smarkm 5410109998Smarkm *) New function X509V3_add1_i2d(). This automatically encodes and 5411109998Smarkm adds an extension. Its behaviour can be customised with various 5412109998Smarkm flags to append, replace or delete. Various wrappers added for 5413109998Smarkm certifcates and CRLs. 5414109998Smarkm [Steve Henson] 5415109998Smarkm 5416109998Smarkm *) Fix to avoid calling the underlying ASN1 print routine when 5417109998Smarkm an extension cannot be parsed. Correct a typo in the 5418109998Smarkm OCSP_SERVICELOC extension. Tidy up print OCSP format. 5419109998Smarkm [Steve Henson] 5420109998Smarkm 5421109998Smarkm *) Make mkdef.pl parse some of the ASN1 macros and add apropriate 5422109998Smarkm entries for variables. 5423109998Smarkm [Steve Henson] 5424109998Smarkm 5425109998Smarkm *) Add functionality to apps/openssl.c for detecting locking 5426109998Smarkm problems: As the program is single-threaded, all we have 5427109998Smarkm to do is register a locking callback using an array for 5428109998Smarkm storing which locks are currently held by the program. 5429109998Smarkm [Bodo Moeller] 5430109998Smarkm 5431109998Smarkm *) Use a lock around the call to CRYPTO_get_ex_new_index() in 5432109998Smarkm SSL_get_ex_data_X509_STORE_idx(), which is used in 5433109998Smarkm ssl_verify_cert_chain() and thus can be called at any time 5434109998Smarkm during TLS/SSL handshakes so that thread-safety is essential. 5435109998Smarkm Unfortunately, the ex_data design is not at all suited 5436109998Smarkm for multi-threaded use, so it probably should be abolished. 5437109998Smarkm [Bodo Moeller] 5438109998Smarkm 5439109998Smarkm *) Added Broadcom "ubsec" ENGINE to OpenSSL. 5440109998Smarkm [Broadcom, tweaked and integrated by Geoff Thorpe] 5441109998Smarkm 5442109998Smarkm *) Move common extension printing code to new function 5443109998Smarkm X509V3_print_extensions(). Reorganise OCSP print routines and 5444109998Smarkm implement some needed OCSP ASN1 functions. Add OCSP extensions. 5445109998Smarkm [Steve Henson] 5446109998Smarkm 5447109998Smarkm *) New function X509_signature_print() to remove duplication in some 5448109998Smarkm print routines. 5449109998Smarkm [Steve Henson] 5450109998Smarkm 5451109998Smarkm *) Add a special meaning when SET OF and SEQUENCE OF flags are both 5452109998Smarkm set (this was treated exactly the same as SET OF previously). This 5453109998Smarkm is used to reorder the STACK representing the structure to match the 5454109998Smarkm encoding. This will be used to get round a problem where a PKCS7 5455109998Smarkm structure which was signed could not be verified because the STACK 5456109998Smarkm order did not reflect the encoded order. 5457109998Smarkm [Steve Henson] 5458109998Smarkm 5459109998Smarkm *) Reimplement the OCSP ASN1 module using the new code. 5460109998Smarkm [Steve Henson] 5461109998Smarkm 5462109998Smarkm *) Update the X509V3 code to permit the use of an ASN1_ITEM structure 5463109998Smarkm for its ASN1 operations. The old style function pointers still exist 5464109998Smarkm for now but they will eventually go away. 5465109998Smarkm [Steve Henson] 5466109998Smarkm 5467109998Smarkm *) Merge in replacement ASN1 code from the ASN1 branch. This almost 5468109998Smarkm completely replaces the old ASN1 functionality with a table driven 5469109998Smarkm encoder and decoder which interprets an ASN1_ITEM structure describing 5470109998Smarkm the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is 5471109998Smarkm largely maintained. Almost all of the old asn1_mac.h macro based ASN1 5472109998Smarkm has also been converted to the new form. 5473109998Smarkm [Steve Henson] 5474109998Smarkm 5475109998Smarkm *) Change BN_mod_exp_recp so that negative moduli are tolerated 5476109998Smarkm (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set 5477109998Smarkm so that BN_mod_exp_mont and BN_mod_exp_mont_word work 5478109998Smarkm for negative moduli. 5479109998Smarkm [Bodo Moeller] 5480109998Smarkm 5481109998Smarkm *) Fix BN_uadd and BN_usub: Always return non-negative results instead 5482109998Smarkm of not touching the result's sign bit. 5483109998Smarkm [Bodo Moeller] 5484109998Smarkm 5485109998Smarkm *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be 5486109998Smarkm set. 5487109998Smarkm [Bodo Moeller] 5488109998Smarkm 5489109998Smarkm *) Changed the LHASH code to use prototypes for callbacks, and created 5490109998Smarkm macros to declare and implement thin (optionally static) functions 5491109998Smarkm that provide type-safety and avoid function pointer casting for the 5492109998Smarkm type-specific callbacks. 5493109998Smarkm [Geoff Thorpe] 5494109998Smarkm 5495109998Smarkm *) Added Kerberos Cipher Suites to be used with TLS, as written in 5496109998Smarkm RFC 2712. 5497109998Smarkm [Veers Staats <staatsvr@asc.hpc.mil>, 5498109998Smarkm Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte] 5499109998Smarkm 5500109998Smarkm *) Reformat the FAQ so the different questions and answers can be divided 5501109998Smarkm in sections depending on the subject. 5502109998Smarkm [Richard Levitte] 5503109998Smarkm 5504109998Smarkm *) Have the zlib compression code load ZLIB.DLL dynamically under 5505109998Smarkm Windows. 5506109998Smarkm [Richard Levitte] 5507109998Smarkm 5508109998Smarkm *) New function BN_mod_sqrt for computing square roots modulo a prime 5509109998Smarkm (using the probabilistic Tonelli-Shanks algorithm unless 5510109998Smarkm p == 3 (mod 4) or p == 5 (mod 8), which are cases that can 5511109998Smarkm be handled deterministically). 5512109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 5513109998Smarkm 5514109998Smarkm *) Make BN_mod_inverse faster by explicitly handling small quotients 5515109998Smarkm in the Euclid loop. (Speed gain about 20% for small moduli [256 or 5516109998Smarkm 512 bits], about 30% for larger ones [1024 or 2048 bits].) 5517109998Smarkm [Bodo Moeller] 5518109998Smarkm 5519109998Smarkm *) New function BN_kronecker. 5520109998Smarkm [Bodo Moeller] 5521109998Smarkm 5522109998Smarkm *) Fix BN_gcd so that it works on negative inputs; the result is 5523109998Smarkm positive unless both parameters are zero. 5524109998Smarkm Previously something reasonably close to an infinite loop was 5525109998Smarkm possible because numbers could be growing instead of shrinking 5526109998Smarkm in the implementation of Euclid's algorithm. 5527109998Smarkm [Bodo Moeller] 5528109998Smarkm 5529109998Smarkm *) Fix BN_is_word() and BN_is_one() macros to take into account the 5530109998Smarkm sign of the number in question. 5531109998Smarkm 5532109998Smarkm Fix BN_is_word(a,w) to work correctly for w == 0. 5533109998Smarkm 5534109998Smarkm The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) 5535109998Smarkm because its test if the absolute value of 'a' equals 'w'. 5536109998Smarkm Note that BN_abs_is_word does *not* handle w == 0 reliably; 5537109998Smarkm it exists mostly for use in the implementations of BN_is_zero(), 5538109998Smarkm BN_is_one(), and BN_is_word(). 5539109998Smarkm [Bodo Moeller] 5540109998Smarkm 5541109998Smarkm *) New function BN_swap. 5542109998Smarkm [Bodo Moeller] 5543109998Smarkm 5544109998Smarkm *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that 5545109998Smarkm the exponentiation functions are more likely to produce reasonable 5546109998Smarkm results on negative inputs. 5547109998Smarkm [Bodo Moeller] 5548109998Smarkm 5549109998Smarkm *) Change BN_mod_mul so that the result is always non-negative. 5550109998Smarkm Previously, it could be negative if one of the factors was negative; 5551109998Smarkm I don't think anyone really wanted that behaviour. 5552109998Smarkm [Bodo Moeller] 5553109998Smarkm 5554109998Smarkm *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c 5555109998Smarkm (except for exponentiation, which stays in crypto/bn/bn_exp.c, 5556109998Smarkm and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c) 5557109998Smarkm and add new functions: 5558109998Smarkm 5559109998Smarkm BN_nnmod 5560109998Smarkm BN_mod_sqr 5561109998Smarkm BN_mod_add 5562109998Smarkm BN_mod_add_quick 5563109998Smarkm BN_mod_sub 5564109998Smarkm BN_mod_sub_quick 5565109998Smarkm BN_mod_lshift1 5566109998Smarkm BN_mod_lshift1_quick 5567109998Smarkm BN_mod_lshift 5568109998Smarkm BN_mod_lshift_quick 5569109998Smarkm 5570109998Smarkm These functions always generate non-negative results. 5571109998Smarkm 5572109998Smarkm BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r 5573109998Smarkm such that |m| < r < 0, BN_nnmod will output rem + |m| instead). 5574109998Smarkm 5575109998Smarkm BN_mod_XXX_quick(r, a, [b,] m) generates the same result as 5576109998Smarkm BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b] 5577109998Smarkm be reduced modulo m. 5578109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 5579109998Smarkm 5580111147Snectar#if 0 5581111147Snectar The following entry accidentily appeared in the CHANGES file 5582111147Snectar distributed with OpenSSL 0.9.7. The modifications described in 5583111147Snectar it do *not* apply to OpenSSL 0.9.7. 5584111147Snectar 5585109998Smarkm *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 5586109998Smarkm was actually never needed) and in BN_mul(). The removal in BN_mul() 5587109998Smarkm required a small change in bn_mul_part_recursive() and the addition 5588109998Smarkm of the functions bn_cmp_part_words(), bn_sub_part_words() and 5589109998Smarkm bn_add_part_words(), which do the same thing as bn_cmp_words(), 5590109998Smarkm bn_sub_words() and bn_add_words() except they take arrays with 5591109998Smarkm differing sizes. 5592109998Smarkm [Richard Levitte] 5593111147Snectar#endif 5594109998Smarkm 5595109998Smarkm *) In 'openssl passwd', verify passwords read from the terminal 5596109998Smarkm unless the '-salt' option is used (which usually means that 5597109998Smarkm verification would just waste user's time since the resulting 5598109998Smarkm hash is going to be compared with some given password hash) 5599109998Smarkm or the new '-noverify' option is used. 5600109998Smarkm 5601109998Smarkm This is an incompatible change, but it does not affect 5602109998Smarkm non-interactive use of 'openssl passwd' (passwords on the command 5603109998Smarkm line, '-stdin' option, '-in ...' option) and thus should not 5604109998Smarkm cause any problems. 5605109998Smarkm [Bodo Moeller] 5606109998Smarkm 5607109998Smarkm *) Remove all references to RSAref, since there's no more need for it. 5608109998Smarkm [Richard Levitte] 5609109998Smarkm 5610109998Smarkm *) Make DSO load along a path given through an environment variable 5611109998Smarkm (SHLIB_PATH) with shl_load(). 5612109998Smarkm [Richard Levitte] 5613109998Smarkm 5614109998Smarkm *) Constify the ENGINE code as a result of BIGNUM constification. 5615109998Smarkm Also constify the RSA code and most things related to it. In a 5616109998Smarkm few places, most notable in the depth of the ASN.1 code, ugly 5617109998Smarkm casts back to non-const were required (to be solved at a later 5618109998Smarkm time) 5619109998Smarkm [Richard Levitte] 5620109998Smarkm 5621109998Smarkm *) Make it so the openssl application has all engines loaded by default. 5622109998Smarkm [Richard Levitte] 5623109998Smarkm 5624109998Smarkm *) Constify the BIGNUM routines a little more. 5625109998Smarkm [Richard Levitte] 5626109998Smarkm 5627109998Smarkm *) Add the following functions: 5628109998Smarkm 5629109998Smarkm ENGINE_load_cswift() 5630109998Smarkm ENGINE_load_chil() 5631109998Smarkm ENGINE_load_atalla() 5632109998Smarkm ENGINE_load_nuron() 5633109998Smarkm ENGINE_load_builtin_engines() 5634109998Smarkm 5635109998Smarkm That way, an application can itself choose if external engines that 5636109998Smarkm are built-in in OpenSSL shall ever be used or not. The benefit is 5637109998Smarkm that applications won't have to be linked with libdl or other dso 5638109998Smarkm libraries unless it's really needed. 5639109998Smarkm 5640109998Smarkm Changed 'openssl engine' to load all engines on demand. 5641109998Smarkm Changed the engine header files to avoid the duplication of some 5642109998Smarkm declarations (they differed!). 5643109998Smarkm [Richard Levitte] 5644109998Smarkm 5645109998Smarkm *) 'openssl engine' can now list capabilities. 5646109998Smarkm [Richard Levitte] 5647109998Smarkm 5648109998Smarkm *) Better error reporting in 'openssl engine'. 5649109998Smarkm [Richard Levitte] 5650109998Smarkm 5651109998Smarkm *) Never call load_dh_param(NULL) in s_server. 5652109998Smarkm [Bodo Moeller] 5653109998Smarkm 5654109998Smarkm *) Add engine application. It can currently list engines by name and 5655109998Smarkm identity, and test if they are actually available. 5656109998Smarkm [Richard Levitte] 5657109998Smarkm 5658109998Smarkm *) Improve RPM specification file by forcing symbolic linking and making 5659109998Smarkm sure the installed documentation is also owned by root.root. 5660109998Smarkm [Damien Miller <djm@mindrot.org>] 5661109998Smarkm 5662109998Smarkm *) Give the OpenSSL applications more possibilities to make use of 5663109998Smarkm keys (public as well as private) handled by engines. 5664109998Smarkm [Richard Levitte] 5665109998Smarkm 5666109998Smarkm *) Add OCSP code that comes from CertCo. 5667109998Smarkm [Richard Levitte] 5668109998Smarkm 5669109998Smarkm *) Add VMS support for the Rijndael code. 5670109998Smarkm [Richard Levitte] 5671109998Smarkm 5672109998Smarkm *) Added untested support for Nuron crypto accelerator. 5673109998Smarkm [Ben Laurie] 5674109998Smarkm 5675109998Smarkm *) Add support for external cryptographic devices. This code was 5676109998Smarkm previously distributed separately as the "engine" branch. 5677109998Smarkm [Geoff Thorpe, Richard Levitte] 5678109998Smarkm 5679109998Smarkm *) Rework the filename-translation in the DSO code. It is now possible to 5680109998Smarkm have far greater control over how a "name" is turned into a filename 5681109998Smarkm depending on the operating environment and any oddities about the 5682109998Smarkm different shared library filenames on each system. 5683109998Smarkm [Geoff Thorpe] 5684109998Smarkm 5685109998Smarkm *) Support threads on FreeBSD-elf in Configure. 5686109998Smarkm [Richard Levitte] 5687109998Smarkm 5688109998Smarkm *) Fix for SHA1 assembly problem with MASM: it produces 5689109998Smarkm warnings about corrupt line number information when assembling 5690109998Smarkm with debugging information. This is caused by the overlapping 5691109998Smarkm of two sections. 5692109998Smarkm [Bernd Matthes <mainbug@celocom.de>, Steve Henson] 5693109998Smarkm 5694109998Smarkm *) NCONF changes. 5695109998Smarkm NCONF_get_number() has no error checking at all. As a replacement, 5696109998Smarkm NCONF_get_number_e() is defined (_e for "error checking") and is 5697109998Smarkm promoted strongly. The old NCONF_get_number is kept around for 5698109998Smarkm binary backward compatibility. 5699109998Smarkm Make it possible for methods to load from something other than a BIO, 5700109998Smarkm by providing a function pointer that is given a name instead of a BIO. 5701109998Smarkm For example, this could be used to load configuration data from an 5702109998Smarkm LDAP server. 5703109998Smarkm [Richard Levitte] 5704109998Smarkm 5705109998Smarkm *) Fix for non blocking accept BIOs. Added new I/O special reason 5706109998Smarkm BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs 5707109998Smarkm with non blocking I/O was not possible because no retry code was 5708109998Smarkm implemented. Also added new SSL code SSL_WANT_ACCEPT to cover 5709109998Smarkm this case. 5710109998Smarkm [Steve Henson] 5711109998Smarkm 5712109998Smarkm *) Added the beginnings of Rijndael support. 5713109998Smarkm [Ben Laurie] 5714109998Smarkm 5715109998Smarkm *) Fix for bug in DirectoryString mask setting. Add support for 5716109998Smarkm X509_NAME_print_ex() in 'req' and X509_print_ex() function 5717109998Smarkm to allow certificate printing to more controllable, additional 5718109998Smarkm 'certopt' option to 'x509' to allow new printing options to be 5719109998Smarkm set. 5720109998Smarkm [Steve Henson] 5721109998Smarkm 5722109998Smarkm *) Clean old EAY MD5 hack from e_os.h. 5723109998Smarkm [Richard Levitte] 5724109998Smarkm 5725142425Snectar Changes between 0.9.6l and 0.9.6m [17 Mar 2004] 5726142425Snectar 5727142425Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 5728160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 5729142425Snectar [Joe Orton, Steve Henson] 5730142425Snectar 5731142425Snectar Changes between 0.9.6k and 0.9.6l [04 Nov 2003] 5732142425Snectar 5733142425Snectar *) Fix additional bug revealed by the NISCC test suite: 5734142425Snectar 5735142425Snectar Stop bug triggering large recursion when presented with 5736160814Ssimon certain ASN.1 tags (CVE-2003-0851) 5737142425Snectar [Steve Henson] 5738142425Snectar 5739120631Snectar Changes between 0.9.6j and 0.9.6k [30 Sep 2003] 5740120631Snectar 5741120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 5742120631Snectar 5743120631Snectar Stop out of bounds reads in the ASN1 code when presented with 5744160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 5745120631Snectar 5746120631Snectar If verify callback ignores invalid public key errors don't try to check 5747120631Snectar certificate signature with the NULL public key. 5748120631Snectar 5749120631Snectar [Steve Henson] 5750120631Snectar 5751120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 5752120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 5753120631Snectar specifications. 5754120631Snectar [Steve Henson] 5755120631Snectar 5756120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 5757120631Snectar extra data after the compression methods not only for TLS 1.0 5758120631Snectar but also for SSL 3.0 (as required by the specification). 5759120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 5760120631Snectar 5761120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 5762120631Snectar when it's 512 *bits* long, not 512 bytes. 5763120631Snectar [Richard Levitte] 5764120631Snectar 5765120631Snectar Changes between 0.9.6i and 0.9.6j [10 Apr 2003] 5766120631Snectar 5767120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 5768120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 5769120631Snectar a protocol version number mismatch like a decryption error 5770120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 5771120631Snectar [Bodo Moeller] 5772120631Snectar 5773120631Snectar *) Turn on RSA blinding by default in the default implementation 5774120631Snectar to avoid a timing attack. Applications that don't want it can call 5775120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 5776120631Snectar They would be ill-advised to do so in most cases. 5777120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 5778120631Snectar 5779120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 5780120631Snectar seeded (in this case, the secret RSA exponent is abused as 5781120631Snectar an unpredictable seed -- if it is not unpredictable, there 5782120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 5783120631Snectar by remembering the creator's thread ID in rsa->blinding and 5784120631Snectar having all other threads use local one-time blinding factors 5785120631Snectar (this requires more computation than sharing rsa->blinding, but 5786120631Snectar avoids excessive locking; and if an RSA object is not shared 5787120631Snectar between threads, blinding will still be very fast). 5788120631Snectar [Bodo Moeller] 5789120631Snectar 5790111147Snectar Changes between 0.9.6h and 0.9.6i [19 Feb 2003] 5791111147Snectar 5792111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 5793111147Snectar via timing by performing a MAC computation even if incorrrect 5794111147Snectar block cipher padding has been found. This is a countermeasure 5795111147Snectar against active attacks where the attacker has to distinguish 5796160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 5797111147Snectar 5798111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 5799111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 5800111147Snectar Martin Vuagnoux (EPFL, Ilion)] 5801111147Snectar 5802109998Smarkm Changes between 0.9.6g and 0.9.6h [5 Dec 2002] 5803109998Smarkm 5804109998Smarkm *) New function OPENSSL_cleanse(), which is used to cleanse a section of 5805109998Smarkm memory from it's contents. This is done with a counter that will 5806109998Smarkm place alternating values in each byte. This can be used to solve 5807109998Smarkm two issues: 1) the removal of calls to memset() by highly optimizing 5808109998Smarkm compilers, and 2) cleansing with other values than 0, since those can 5809109998Smarkm be read through on certain media, for example a swap space on disk. 5810109998Smarkm [Geoff Thorpe] 5811109998Smarkm 5812109998Smarkm *) Bugfix: client side session caching did not work with external caching, 5813109998Smarkm because the session->cipher setting was not restored when reloading 5814109998Smarkm from the external cache. This problem was masked, when 5815109998Smarkm SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. 5816109998Smarkm (Found by Steve Haslam <steve@araqnid.ddts.net>.) 5817109998Smarkm [Lutz Jaenicke] 5818109998Smarkm 5819109998Smarkm *) Fix client_certificate (ssl/s2_clnt.c): The permissible total 5820109998Smarkm length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. 5821109998Smarkm [Zeev Lieber <zeev-l@yahoo.com>] 5822109998Smarkm 5823109998Smarkm *) Undo an undocumented change introduced in 0.9.6e which caused 5824109998Smarkm repeated calls to OpenSSL_add_all_ciphers() and 5825109998Smarkm OpenSSL_add_all_digests() to be ignored, even after calling 5826109998Smarkm EVP_cleanup(). 5827109998Smarkm [Richard Levitte] 5828109998Smarkm 5829109998Smarkm *) Change the default configuration reader to deal with last line not 5830109998Smarkm being properly terminated. 5831109998Smarkm [Richard Levitte] 5832109998Smarkm 5833109998Smarkm *) Change X509_NAME_cmp() so it applies the special rules on handling 5834109998Smarkm DN values that are of type PrintableString, as well as RDNs of type 5835109998Smarkm emailAddress where the value has the type ia5String. 5836109998Smarkm [stefank@valicert.com via Richard Levitte] 5837109998Smarkm 5838109998Smarkm *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half 5839109998Smarkm the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently 5840109998Smarkm doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be 5841109998Smarkm the bitwise-OR of the two for use by the majority of applications 5842109998Smarkm wanting this behaviour, and update the docs. The documented 5843109998Smarkm behaviour and actual behaviour were inconsistent and had been 5844109998Smarkm changing anyway, so this is more a bug-fix than a behavioural 5845109998Smarkm change. 5846109998Smarkm [Geoff Thorpe, diagnosed by Nadav Har'El] 5847109998Smarkm 5848109998Smarkm *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c 5849109998Smarkm (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). 5850109998Smarkm [Bodo Moeller] 5851109998Smarkm 5852109998Smarkm *) Fix initialization code race conditions in 5853109998Smarkm SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), 5854109998Smarkm SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), 5855109998Smarkm SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), 5856109998Smarkm TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), 5857109998Smarkm ssl2_get_cipher_by_char(), 5858109998Smarkm ssl3_get_cipher_by_char(). 5859109998Smarkm [Patrick McCormick <patrick@tellme.com>, Bodo Moeller] 5860109998Smarkm 5861109998Smarkm *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after 5862109998Smarkm the cached sessions are flushed, as the remove_cb() might use ex_data 5863109998Smarkm contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> 5864109998Smarkm (see [openssl.org #212]). 5865109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 5866109998Smarkm 5867109998Smarkm *) Fix typo in OBJ_txt2obj which incorrectly passed the content 5868109998Smarkm length, instead of the encoding length to d2i_ASN1_OBJECT. 5869109998Smarkm [Steve Henson] 5870109998Smarkm 5871101618Snectar Changes between 0.9.6f and 0.9.6g [9 Aug 2002] 5872101618Snectar 5873101618Snectar *) [In 0.9.6g-engine release:] 5874101618Snectar Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall'). 5875101618Snectar [Lynn Gazis <lgazis@rainbow.com>] 5876101618Snectar 5877101613Snectar Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 5878101613Snectar 5879101613Snectar *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 5880101613Snectar and get fix the header length calculation. 5881101613Snectar [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 5882101613Snectar Alon Kantor <alonk@checkpoint.com> (and others), 5883101613Snectar Steve Henson] 5884101613Snectar 5885101613Snectar *) Use proper error handling instead of 'assertions' in buffer 5886101613Snectar overflow checks added in 0.9.6e. This prevents DoS (the 5887101613Snectar assertions could call abort()). 5888101613Snectar [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] 5889101613Snectar 5890100936Snectar Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 5891100936Snectar 5892109998Smarkm *) Add various sanity checks to asn1_get_length() to reject 5893109998Smarkm the ASN1 length bytes if they exceed sizeof(long), will appear 5894109998Smarkm negative or the content length exceeds the length of the 5895109998Smarkm supplied buffer. 5896109998Smarkm [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 5897109998Smarkm 5898100936Snectar *) Fix cipher selection routines: ciphers without encryption had no flags 5899100936Snectar for the cipher strength set and where therefore not handled correctly 5900100936Snectar by the selection routines (PR #130). 5901100936Snectar [Lutz Jaenicke] 5902100936Snectar 5903100936Snectar *) Fix EVP_dsa_sha macro. 5904100936Snectar [Nils Larsch] 5905100936Snectar 5906100936Snectar *) New option 5907100936Snectar SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 5908100936Snectar for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 5909100936Snectar that was added in OpenSSL 0.9.6d. 5910100936Snectar 5911100936Snectar As the countermeasure turned out to be incompatible with some 5912100936Snectar broken SSL implementations, the new option is part of SSL_OP_ALL. 5913100936Snectar SSL_OP_ALL is usually employed when compatibility with weird SSL 5914100936Snectar implementations is desired (e.g. '-bugs' option to 's_client' and 5915100936Snectar 's_server'), so the new option is automatically set in many 5916100936Snectar applications. 5917100936Snectar [Bodo Moeller] 5918100936Snectar 5919100936Snectar *) Changes in security patch: 5920100936Snectar 5921100936Snectar Changes marked "(CHATS)" were sponsored by the Defense Advanced 5922100936Snectar Research Projects Agency (DARPA) and Air Force Research Laboratory, 5923100936Snectar Air Force Materiel Command, USAF, under agreement number 5924100936Snectar F30602-01-2-0537. 5925100936Snectar 5926100936Snectar *) Add various sanity checks to asn1_get_length() to reject 5927100936Snectar the ASN1 length bytes if they exceed sizeof(long), will appear 5928100936Snectar negative or the content length exceeds the length of the 5929160814Ssimon supplied buffer. (CVE-2002-0659) 5930100936Snectar [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 5931100936Snectar 5932100936Snectar *) Assertions for various potential buffer overflows, not known to 5933100936Snectar happen in practice. 5934100936Snectar [Ben Laurie (CHATS)] 5935100936Snectar 5936100936Snectar *) Various temporary buffers to hold ASCII versions of integers were 5937160814Ssimon too small for 64 bit platforms. (CVE-2002-0655) 5938100936Snectar [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 5939100936Snectar 5940100936Snectar *) Remote buffer overflow in SSL3 protocol - an attacker could 5941160814Ssimon supply an oversized session ID to a client. (CVE-2002-0656) 5942100936Snectar [Ben Laurie (CHATS)] 5943100936Snectar 5944100936Snectar *) Remote buffer overflow in SSL2 protocol - an attacker could 5945160814Ssimon supply an oversized client master key. (CVE-2002-0656) 5946100936Snectar [Ben Laurie (CHATS)] 5947100936Snectar 5948100928Snectar Changes between 0.9.6c and 0.9.6d [9 May 2002] 5949100928Snectar 5950100928Snectar *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 5951100928Snectar encoded as NULL) with id-dsa-with-sha1. 5952100928Snectar [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 5953100928Snectar 5954100928Snectar *) Check various X509_...() return values in apps/req.c. 5955100928Snectar [Nils Larsch <nla@trustcenter.de>] 5956100928Snectar 5957100928Snectar *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 5958100928Snectar an end-of-file condition would erronously be flagged, when the CRLF 5959100928Snectar was just at the end of a processed block. The bug was discovered when 5960100928Snectar processing data through a buffering memory BIO handing the data to a 5961100928Snectar BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 5962100928Snectar <ptsekov@syntrex.com> and Nedelcho Stanev. 5963100928Snectar [Lutz Jaenicke] 5964100928Snectar 5965100928Snectar *) Implement a countermeasure against a vulnerability recently found 5966100928Snectar in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 5967100928Snectar before application data chunks to avoid the use of known IVs 5968100928Snectar with data potentially chosen by the attacker. 5969100928Snectar [Bodo Moeller] 5970100928Snectar 5971100928Snectar *) Fix length checks in ssl3_get_client_hello(). 5972100928Snectar [Bodo Moeller] 5973100928Snectar 5974100928Snectar *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 5975100928Snectar to prevent ssl3_read_internal() from incorrectly assuming that 5976100928Snectar ssl3_read_bytes() found application data while handshake 5977100928Snectar processing was enabled when in fact s->s3->in_read_app_data was 5978100928Snectar merely automatically cleared during the initial handshake. 5979100928Snectar [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 5980100928Snectar 5981100928Snectar *) Fix object definitions for Private and Enterprise: they were not 5982100928Snectar recognized in their shortname (=lowercase) representation. Extend 5983100928Snectar obj_dat.pl to issue an error when using undefined keywords instead 5984100928Snectar of silently ignoring the problem (Svenning Sorensen 5985100928Snectar <sss@sss.dnsalias.net>). 5986100928Snectar [Lutz Jaenicke] 5987100928Snectar 5988100928Snectar *) Fix DH_generate_parameters() so that it works for 'non-standard' 5989100928Snectar generators, i.e. generators other than 2 and 5. (Previously, the 5990100928Snectar code did not properly initialise the 'add' and 'rem' values to 5991100928Snectar BN_generate_prime().) 5992100928Snectar 5993100928Snectar In the new general case, we do not insist that 'generator' is 5994100928Snectar actually a primitive root: This requirement is rather pointless; 5995100928Snectar a generator of the order-q subgroup is just as good, if not 5996100928Snectar better. 5997100928Snectar [Bodo Moeller] 5998100928Snectar 5999100928Snectar *) Map new X509 verification errors to alerts. Discovered and submitted by 6000100928Snectar Tom Wu <tom@arcot.com>. 6001100928Snectar [Lutz Jaenicke] 6002100928Snectar 6003100928Snectar *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 6004100928Snectar returning non-zero before the data has been completely received 6005100928Snectar when using non-blocking I/O. 6006100928Snectar [Bodo Moeller; problem pointed out by John Hughes] 6007100928Snectar 6008100928Snectar *) Some of the ciphers missed the strength entry (SSL_LOW etc). 6009100928Snectar [Ben Laurie, Lutz Jaenicke] 6010100928Snectar 6011100928Snectar *) Fix bug in SSL_clear(): bad sessions were not removed (found by 6012100928Snectar Yoram Zahavi <YoramZ@gilian.com>). 6013100928Snectar [Lutz Jaenicke] 6014100928Snectar 6015100928Snectar *) Add information about CygWin 1.3 and on, and preserve proper 6016100928Snectar configuration for the versions before that. 6017100928Snectar [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 6018100928Snectar 6019100928Snectar *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 6020100928Snectar check whether we deal with a copy of a session and do not delete from 6021100928Snectar the cache in this case. Problem reported by "Izhar Shoshani Levi" 6022100928Snectar <izhar@checkpoint.com>. 6023100928Snectar [Lutz Jaenicke] 6024100928Snectar 6025100928Snectar *) Do not store session data into the internal session cache, if it 6026100928Snectar is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 6027100928Snectar flag is set). Proposed by Aslam <aslam@funk.com>. 6028100928Snectar [Lutz Jaenicke] 6029100928Snectar 6030100928Snectar *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 6031100928Snectar value is 0. 6032100928Snectar [Richard Levitte] 6033100928Snectar 6034109998Smarkm *) [In 0.9.6d-engine release:] 6035109998Smarkm Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 6036100928Snectar [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 6037100928Snectar 6038100928Snectar *) Add the configuration target linux-s390x. 6039100928Snectar [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 6040100928Snectar 6041100928Snectar *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 6042100928Snectar ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 6043100928Snectar variable as an indication that a ClientHello message has been 6044100928Snectar received. As the flag value will be lost between multiple 6045100928Snectar invocations of ssl3_accept when using non-blocking I/O, the 6046100928Snectar function may not be aware that a handshake has actually taken 6047100928Snectar place, thus preventing a new session from being added to the 6048100928Snectar session cache. 6049100928Snectar 6050100928Snectar To avoid this problem, we now set s->new_session to 2 instead of 6051100928Snectar using a local variable. 6052100928Snectar [Lutz Jaenicke, Bodo Moeller] 6053100928Snectar 6054100928Snectar *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 6055100928Snectar if the SSL_R_LENGTH_MISMATCH error is detected. 6056100928Snectar [Geoff Thorpe, Bodo Moeller] 6057100928Snectar 6058100928Snectar *) New 'shared_ldflag' column in Configure platform table. 6059100928Snectar [Richard Levitte] 6060100928Snectar 6061100928Snectar *) Fix EVP_CIPHER_mode macro. 6062100928Snectar ["Dan S. Camper" <dan@bti.net>] 6063100928Snectar 6064100928Snectar *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 6065100928Snectar type, we must throw them away by setting rr->length to 0. 6066100928Snectar [D P Chang <dpc@qualys.com>] 6067100928Snectar 606889837Skris Changes between 0.9.6b and 0.9.6c [21 dec 2001] 606989837Skris 607089837Skris *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 607189837Skris <Dominikus.Scherkl@biodata.com>. (The previous implementation 607289837Skris worked incorrectly for those cases where range = 10..._2 and 607389837Skris 3*range is two bits longer than range.) 607489837Skris [Bodo Moeller] 607589837Skris 607689837Skris *) Only add signing time to PKCS7 structures if it is not already 607789837Skris present. 607889837Skris [Steve Henson] 607989837Skris 608089837Skris *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 608189837Skris OBJ_ld_ce should be OBJ_id_ce. 608289837Skris Also some ip-pda OIDs in crypto/objects/objects.txt were 608389837Skris incorrect (cf. RFC 3039). 608489837Skris [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 608589837Skris 608689837Skris *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 608789837Skris returns early because it has nothing to do. 608889837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 608989837Skris 609089837Skris *) [In 0.9.6c-engine release:] 609189837Skris Fix mutex callback return values in crypto/engine/hw_ncipher.c. 609289837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 609389837Skris 609489837Skris *) [In 0.9.6c-engine release:] 609589837Skris Add support for Cryptographic Appliance's keyserver technology. 609689837Skris (Use engine 'keyclient') 609789837Skris [Cryptographic Appliances and Geoff Thorpe] 609889837Skris 609989837Skris *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 610089837Skris is called via tools/c89.sh because arguments have to be 610189837Skris rearranged (all '-L' options must appear before the first object 610289837Skris modules). 610389837Skris [Richard Shapiro <rshapiro@abinitio.com>] 610489837Skris 610589837Skris *) [In 0.9.6c-engine release:] 610689837Skris Add support for Broadcom crypto accelerator cards, backported 610789837Skris from 0.9.7. 610889837Skris [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 610989837Skris 611089837Skris *) [In 0.9.6c-engine release:] 611189837Skris Add support for SureWare crypto accelerator cards from 611289837Skris Baltimore Technologies. (Use engine 'sureware') 611389837Skris [Baltimore Technologies and Mark Cox] 611489837Skris 611589837Skris *) [In 0.9.6c-engine release:] 611689837Skris Add support for crypto accelerator cards from Accelerated 611789837Skris Encryption Processing, www.aep.ie. (Use engine 'aep') 611889837Skris [AEP Inc. and Mark Cox] 611989837Skris 612089837Skris *) Add a configuration entry for gcc on UnixWare. 612189837Skris [Gary Benson <gbenson@redhat.com>] 612289837Skris 612389837Skris *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 612489837Skris messages are stored in a single piece (fixed-length part and 612589837Skris variable-length part combined) and fix various bugs found on the way. 612689837Skris [Bodo Moeller] 612789837Skris 612889837Skris *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 612989837Skris instead. BIO_gethostbyname() does not know what timeouts are 613089837Skris appropriate, so entries would stay in cache even when they have 613189837Skris become invalid. 613289837Skris [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 613389837Skris 613489837Skris *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 613589837Skris faced with a pathologically small ClientHello fragment that does 613689837Skris not contain client_version: Instead of aborting with an error, 613789837Skris simply choose the highest available protocol version (i.e., 613889837Skris TLS 1.0 unless it is disabled). In practice, ClientHello 613989837Skris messages are never sent like this, but this change gives us 614089837Skris strictly correct behaviour at least for TLS. 614189837Skris [Bodo Moeller] 614289837Skris 614389837Skris *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 614489837Skris never resets s->method to s->ctx->method when called from within 614589837Skris one of the SSL handshake functions. 614689837Skris [Bodo Moeller; problem pointed out by Niko Baric] 614789837Skris 614889837Skris *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 614989837Skris (sent using the client's version number) if client_version is 615089837Skris smaller than the protocol version in use. Also change 615189837Skris ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 615289837Skris the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 615389837Skris the client will at least see that alert. 615489837Skris [Bodo Moeller] 615589837Skris 615689837Skris *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 615789837Skris correctly. 615889837Skris [Bodo Moeller] 615989837Skris 616089837Skris *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 616189837Skris client receives HelloRequest while in a handshake. 616289837Skris [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 616389837Skris 616489837Skris *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 616589837Skris should end in 'break', not 'goto end' which circuments various 616689837Skris cleanups done in state SSL_ST_OK. But session related stuff 616789837Skris must be disabled for SSL_ST_OK in the case that we just sent a 616889837Skris HelloRequest. 616989837Skris 617089837Skris Also avoid some overhead by not calling ssl_init_wbio_buffer() 617189837Skris before just sending a HelloRequest. 617289837Skris [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 617389837Skris 617489837Skris *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 617589837Skris reveal whether illegal block cipher padding was found or a MAC 617689837Skris verification error occured. (Neither SSLerr() codes nor alerts 617789837Skris are directly visible to potential attackers, but the information 617889837Skris may leak via logfiles.) 617989837Skris 618089837Skris Similar changes are not required for the SSL 2.0 implementation 618189837Skris because the number of padding bytes is sent in clear for SSL 2.0, 618289837Skris and the extra bytes are just ignored. However ssl/s2_pkt.c 618389837Skris failed to verify that the purported number of padding bytes is in 618489837Skris the legal range. 618589837Skris [Bodo Moeller] 618689837Skris 618789837Skris *) Add OpenUNIX-8 support including shared libraries 618889837Skris (Boyd Lynn Gerber <gerberb@zenez.com>). 618989837Skris [Lutz Jaenicke] 619089837Skris 619189837Skris *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 619289837Skris 'wristwatch attack' using huge encoding parameters (cf. 619389837Skris James H. Manger's CRYPTO 2001 paper). Note that the 619489837Skris RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 619589837Skris encoding parameters and hence was not vulnerable. 619689837Skris [Bodo Moeller] 619789837Skris 619889837Skris *) BN_sqr() bug fix. 619989837Skris [Ulf M�ller, reported by Jim Ellis <jim.ellis@cavium.com>] 620089837Skris 620189837Skris *) Rabin-Miller test analyses assume uniformly distributed witnesses, 620289837Skris so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 620389837Skris followed by modular reduction. 620489837Skris [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 620589837Skris 620689837Skris *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 620789837Skris equivalent based on BN_pseudo_rand() instead of BN_rand(). 620889837Skris [Bodo Moeller] 620989837Skris 621089837Skris *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 621189837Skris This function was broken, as the check for a new client hello message 621289837Skris to handle SGC did not allow these large messages. 621389837Skris (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 621489837Skris [Lutz Jaenicke] 621589837Skris 621689837Skris *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 621789837Skris [Lutz Jaenicke] 621889837Skris 621989837Skris *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 622089837Skris for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 622189837Skris [Lutz Jaenicke] 622289837Skris 622389837Skris *) Rework the configuration and shared library support for Tru64 Unix. 622489837Skris The configuration part makes use of modern compiler features and 622589837Skris still retains old compiler behavior for those that run older versions 622689837Skris of the OS. The shared library support part includes a variant that 622789837Skris uses the RPATH feature, and is available through the special 622889837Skris configuration target "alpha-cc-rpath", which will never be selected 622989837Skris automatically. 623089837Skris [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 623189837Skris 623289837Skris *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 623389837Skris with the same message size as in ssl3_get_certificate_request(). 623489837Skris Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 623589837Skris messages might inadvertently be reject as too long. 623689837Skris [Petr Lampa <lampa@fee.vutbr.cz>] 623789837Skris 623889837Skris *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 623989837Skris [Andy Polyakov] 624089837Skris 624189837Skris *) Modified SSL library such that the verify_callback that has been set 624289837Skris specificly for an SSL object with SSL_set_verify() is actually being 624389837Skris used. Before the change, a verify_callback set with this function was 624489837Skris ignored and the verify_callback() set in the SSL_CTX at the time of 624589837Skris the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 624689837Skris to allow the necessary settings. 624789837Skris [Lutz Jaenicke] 624889837Skris 624989837Skris *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 625089837Skris explicitly to NULL, as at least on Solaris 8 this seems not always to be 625189837Skris done automatically (in contradiction to the requirements of the C 625289837Skris standard). This made problems when used from OpenSSH. 625389837Skris [Lutz Jaenicke] 625489837Skris 625589837Skris *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 625689837Skris dh->length and always used 625789837Skris 625889837Skris BN_rand_range(priv_key, dh->p). 625989837Skris 626089837Skris BN_rand_range() is not necessary for Diffie-Hellman, and this 626189837Skris specific range makes Diffie-Hellman unnecessarily inefficient if 626289837Skris dh->length (recommended exponent length) is much smaller than the 626389837Skris length of dh->p. We could use BN_rand_range() if the order of 626489837Skris the subgroup was stored in the DH structure, but we only have 626589837Skris dh->length. 626689837Skris 626789837Skris So switch back to 626889837Skris 626989837Skris BN_rand(priv_key, l, ...) 627089837Skris 627189837Skris where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 627289837Skris otherwise. 627389837Skris [Bodo Moeller] 627489837Skris 627589837Skris *) In 627689837Skris 627789837Skris RSA_eay_public_encrypt 627889837Skris RSA_eay_private_decrypt 627989837Skris RSA_eay_private_encrypt (signing) 628089837Skris RSA_eay_public_decrypt (signature verification) 628189837Skris 628289837Skris (default implementations for RSA_public_encrypt, 628389837Skris RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 628489837Skris always reject numbers >= n. 628589837Skris [Bodo Moeller] 628689837Skris 628789837Skris *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 628889837Skris to synchronize access to 'locking_thread'. This is necessary on 628989837Skris systems where access to 'locking_thread' (an 'unsigned long' 629089837Skris variable) is not atomic. 629189837Skris [Bodo Moeller] 629289837Skris 629389837Skris *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 629489837Skris *before* setting the 'crypto_lock_rand' flag. The previous code had 629589837Skris a race condition if 0 is a valid thread ID. 629689837Skris [Travis Vitek <vitek@roguewave.com>] 629789837Skris 629889837Skris *) Add support for shared libraries under Irix. 629989837Skris [Albert Chin-A-Young <china@thewrittenword.com>] 630089837Skris 630189837Skris *) Add configuration option to build on Linux on both big-endian and 630289837Skris little-endian MIPS. 630389837Skris [Ralf Baechle <ralf@uni-koblenz.de>] 630489837Skris 630589837Skris *) Add the possibility to create shared libraries on HP-UX. 630689837Skris [Richard Levitte] 630789837Skris 630879998Skris Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 630979998Skris 631079998Skris *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 631179998Skris to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 631279998Skris Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 631379998Skris PRNG state recovery was possible based on the output of 631479998Skris one PRNG request appropriately sized to gain knowledge on 631579998Skris 'md' followed by enough consecutive 1-byte PRNG requests 631679998Skris to traverse all of 'state'. 631779998Skris 631879998Skris 1. When updating 'md_local' (the current thread's copy of 'md') 631979998Skris during PRNG output generation, hash all of the previous 632079998Skris 'md_local' value, not just the half used for PRNG output. 632179998Skris 632279998Skris 2. Make the number of bytes from 'state' included into the hash 632379998Skris independent from the number of PRNG bytes requested. 632479998Skris 632579998Skris The first measure alone would be sufficient to avoid 632679998Skris Markku-Juhani's attack. (Actually it had never occurred 632779998Skris to me that the half of 'md_local' used for chaining was the 632879998Skris half from which PRNG output bytes were taken -- I had always 632979998Skris assumed that the secret half would be used.) The second 633079998Skris measure makes sure that additional data from 'state' is never 633179998Skris mixed into 'md_local' in small portions; this heuristically 633279998Skris further strengthens the PRNG. 633379998Skris [Bodo Moeller] 633479998Skris 633579998Skris *) Fix crypto/bn/asm/mips3.s. 633679998Skris [Andy Polyakov] 633779998Skris 633879998Skris *) When only the key is given to "enc", the IV is undefined. Print out 633979998Skris an error message in this case. 634079998Skris [Lutz Jaenicke] 634179998Skris 634279998Skris *) Handle special case when X509_NAME is empty in X509 printing routines. 634379998Skris [Steve Henson] 634479998Skris 634579998Skris *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 634679998Skris positive and less than q. 634779998Skris [Bodo Moeller] 634879998Skris 634979998Skris *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 635079998Skris used: it isn't thread safe and the add_lock_callback should handle 635179998Skris that itself. 635279998Skris [Paul Rose <Paul.Rose@bridge.com>] 635379998Skris 635479998Skris *) Verify that incoming data obeys the block size in 635579998Skris ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 635679998Skris [Bodo Moeller] 635779998Skris 635879998Skris *) Fix OAEP check. 635979998Skris [Ulf M�ller, Bodo M�ller] 636079998Skris 636179998Skris *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 636289837Skris RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 636379998Skris when fixing the server behaviour for backwards-compatible 'client 636479998Skris hello' messages. (Note that the attack is impractical against 636579998Skris SSL 3.0 and TLS 1.0 anyway because length and version checking 636679998Skris means that the probability of guessing a valid ciphertext is 636779998Skris around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 636879998Skris paper.) 636979998Skris 637079998Skris Before 0.9.5, the countermeasure (hide the error by generating a 637179998Skris random 'decryption result') did not work properly because 637279998Skris ERR_clear_error() was missing, meaning that SSL_get_error() would 637379998Skris detect the supposedly ignored error. 637479998Skris 637579998Skris Both problems are now fixed. 637679998Skris [Bodo Moeller] 637779998Skris 637879998Skris *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 637979998Skris (previously it was 1024). 638079998Skris [Bodo Moeller] 638179998Skris 638279998Skris *) Fix for compatibility mode trust settings: ignore trust settings 638379998Skris unless some valid trust or reject settings are present. 638479998Skris [Steve Henson] 638579998Skris 638679998Skris *) Fix for blowfish EVP: its a variable length cipher. 638779998Skris [Steve Henson] 638879998Skris 638979998Skris *) Fix various bugs related to DSA S/MIME verification. Handle missing 639079998Skris parameters in DSA public key structures and return an error in the 639179998Skris DSA routines if parameters are absent. 639279998Skris [Steve Henson] 639379998Skris 639479998Skris *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 639579998Skris in the current directory if neither $RANDFILE nor $HOME was set. 639679998Skris RAND_file_name() in 0.9.6a returned NULL in this case. This has 639779998Skris caused some confusion to Windows users who haven't defined $HOME. 639879998Skris Thus RAND_file_name() is changed again: e_os.h can define a 639979998Skris DEFAULT_HOME, which will be used if $HOME is not set. 640079998Skris For Windows, we use "C:"; on other platforms, we still require 640179998Skris environment variables. 640279998Skris 640379998Skris *) Move 'if (!initialized) RAND_poll()' into regions protected by 640479998Skris CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 640579998Skris having multiple threads call RAND_poll() concurrently. 640679998Skris [Bodo Moeller] 640779998Skris 640879998Skris *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 640979998Skris combination of a flag and a thread ID variable. 641079998Skris Otherwise while one thread is in ssleay_rand_bytes (which sets the 641179998Skris flag), *other* threads can enter ssleay_add_bytes without obeying 641289837Skris the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 641379998Skris that they do not hold after the first thread unsets add_do_not_lock). 641479998Skris [Bodo Moeller] 641579998Skris 641679998Skris *) Change bctest again: '-x' expressions are not available in all 641779998Skris versions of 'test'. 641879998Skris [Bodo Moeller] 641979998Skris 642076866Skris Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 642172613Skris 642276866Skris *) Fix a couple of memory leaks in PKCS7_dataDecode() 642376866Skris [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 642476866Skris 642576866Skris *) Change Configure and Makefiles to provide EXE_EXT, which will contain 642676866Skris the default extension for executables, if any. Also, make the perl 642776866Skris scripts that use symlink() to test if it really exists and use "cp" 642876866Skris if it doesn't. All this made OpenSSL compilable and installable in 642976866Skris CygWin. 643076866Skris [Richard Levitte] 643176866Skris 643276866Skris *) Fix for asn1_GetSequence() for indefinite length constructed data. 643376866Skris If SEQUENCE is length is indefinite just set c->slen to the total 643476866Skris amount of data available. 643576866Skris [Steve Henson, reported by shige@FreeBSD.org] 643676866Skris [This change does not apply to 0.9.7.] 643776866Skris 643876866Skris *) Change bctest to avoid here-documents inside command substitution 643976866Skris (workaround for FreeBSD /bin/sh bug). 644076866Skris For compatibility with Ultrix, avoid shell functions (introduced 644176866Skris in the bctest version that searches along $PATH). 644276866Skris [Bodo Moeller] 644376866Skris 644476866Skris *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 644576866Skris with des_encrypt() defined on some operating systems, like Solaris 644676866Skris and UnixWare. 644776866Skris [Richard Levitte] 644876866Skris 644976866Skris *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 645076866Skris On the Importance of Eliminating Errors in Cryptographic 645176866Skris Computations, J. Cryptology 14 (2001) 2, 101-119, 645276866Skris http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 645376866Skris [Ulf Moeller] 645476866Skris 645576866Skris *) MIPS assembler BIGNUM division bug fix. 645676866Skris [Andy Polyakov] 645776866Skris 645876866Skris *) Disabled incorrect Alpha assembler code. 645976866Skris [Richard Levitte] 646076866Skris 646176866Skris *) Fix PKCS#7 decode routines so they correctly update the length 646276866Skris after reading an EOC for the EXPLICIT tag. 646376866Skris [Steve Henson] 646476866Skris [This change does not apply to 0.9.7.] 646576866Skris 646676866Skris *) Fix bug in PKCS#12 key generation routines. This was triggered 646776866Skris if a 3DES key was generated with a 0 initial byte. Include 646876866Skris PKCS12_BROKEN_KEYGEN compilation option to retain the old 646976866Skris (but broken) behaviour. 647076866Skris [Steve Henson] 647176866Skris 647276866Skris *) Enhance bctest to search for a working bc along $PATH and print 647376866Skris it when found. 647476866Skris [Tim Rice <tim@multitalents.net> via Richard Levitte] 647576866Skris 647676866Skris *) Fix memory leaks in err.c: free err_data string if necessary; 647776866Skris don't write to the wrong index in ERR_set_error_data. 647876866Skris [Bodo Moeller] 647976866Skris 648076866Skris *) Implement ssl23_peek (analogous to ssl23_read), which previously 648176866Skris did not exist. 648276866Skris [Bodo Moeller] 648376866Skris 648476866Skris *) Replace rdtsc with _emit statements for VC++ version 5. 648576866Skris [Jeremy Cooper <jeremy@baymoo.org>] 648676866Skris 648776866Skris *) Make it possible to reuse SSLv2 sessions. 648876866Skris [Richard Levitte] 648976866Skris 649076866Skris *) In copy_email() check for >= 0 as a return value for 649176866Skris X509_NAME_get_index_by_NID() since 0 is a valid index. 649276866Skris [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 649376866Skris 649476866Skris *) Avoid coredump with unsupported or invalid public keys by checking if 649576866Skris X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 649676866Skris PKCS7_verify() fails with non detached data. 649776866Skris [Steve Henson] 649876866Skris 649976866Skris *) Don't use getenv in library functions when run as setuid/setgid. 650076866Skris New function OPENSSL_issetugid(). 650176866Skris [Ulf Moeller] 650276866Skris 650376866Skris *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 650476866Skris due to incorrect handling of multi-threading: 650576866Skris 650676866Skris 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 650776866Skris 650876866Skris 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 650976866Skris 651076866Skris 3. Count how many times MemCheck_off() has been called so that 651176866Skris nested use can be treated correctly. This also avoids 651276866Skris inband-signalling in the previous code (which relied on the 651376866Skris assumption that thread ID 0 is impossible). 651476866Skris [Bodo Moeller] 651576866Skris 651676866Skris *) Add "-rand" option also to s_client and s_server. 651776866Skris [Lutz Jaenicke] 651876866Skris 651976866Skris *) Fix CPU detection on Irix 6.x. 652076866Skris [Kurt Hockenbury <khockenb@stevens-tech.edu> and 652176866Skris "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 652276866Skris 652376866Skris *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 652476866Skris was empty. 652576866Skris [Steve Henson] 652676866Skris [This change does not apply to 0.9.7.] 652776866Skris 652876866Skris *) Use the cached encoding of an X509_NAME structure rather than 652976866Skris copying it. This is apparently the reason for the libsafe "errors" 653076866Skris but the code is actually correct. 653176866Skris [Steve Henson] 653276866Skris 653372613Skris *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 653472613Skris Bleichenbacher's DSA attack. 653576866Skris Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 653676866Skris to be set and top=0 forces the highest bit to be set; top=-1 is new 653776866Skris and leaves the highest bit random. 653876866Skris [Ulf Moeller, Bodo Moeller] 653972613Skris 654072613Skris *) In the NCONF_...-based implementations for CONF_... queries 654172613Skris (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 654272613Skris a temporary CONF structure with the data component set to NULL 654372613Skris (which gives segmentation faults in lh_retrieve). 654472613Skris Instead, use NULL for the CONF pointer in CONF_get_string and 654572613Skris CONF_get_number (which may use environment variables) and directly 654672613Skris return NULL from CONF_get_section. 654772613Skris [Bodo Moeller] 654872613Skris 654972613Skris *) Fix potential buffer overrun for EBCDIC. 655072613Skris [Ulf Moeller] 655172613Skris 655272613Skris *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 655372613Skris keyUsage if basicConstraints absent for a CA. 655472613Skris [Steve Henson] 655572613Skris 655672613Skris *) Make SMIME_write_PKCS7() write mail header values with a format that 655772613Skris is more generally accepted (no spaces before the semicolon), since 655872613Skris some programs can't parse those values properly otherwise. Also make 655972613Skris sure BIO's that break lines after each write do not create invalid 656072613Skris headers. 656172613Skris [Richard Levitte] 656272613Skris 656372613Skris *) Make the CRL encoding routines work with empty SEQUENCE OF. The 656472613Skris macros previously used would not encode an empty SEQUENCE OF 656572613Skris and break the signature. 656672613Skris [Steve Henson] 656776866Skris [This change does not apply to 0.9.7.] 656872613Skris 656972613Skris *) Zero the premaster secret after deriving the master secret in 657072613Skris DH ciphersuites. 657172613Skris [Steve Henson] 657272613Skris 657372613Skris *) Add some EVP_add_digest_alias registrations (as found in 657472613Skris OpenSSL_add_all_digests()) to SSL_library_init() 657572613Skris aka OpenSSL_add_ssl_algorithms(). This provides improved 657672613Skris compatibility with peers using X.509 certificates 657772613Skris with unconventional AlgorithmIdentifier OIDs. 657872613Skris [Bodo Moeller] 657972613Skris 658072613Skris *) Fix for Irix with NO_ASM. 658172613Skris ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 658272613Skris 658372613Skris *) ./config script fixes. 658472613Skris [Ulf Moeller, Richard Levitte] 658572613Skris 658672613Skris *) Fix 'openssl passwd -1'. 658772613Skris [Bodo Moeller] 658872613Skris 658972613Skris *) Change PKCS12_key_gen_asc() so it can cope with non null 659072613Skris terminated strings whose length is passed in the passlen 659172613Skris parameter, for example from PEM callbacks. This was done 659272613Skris by adding an extra length parameter to asc2uni(). 659372613Skris [Steve Henson, reported by <oddissey@samsung.co.kr>] 659472613Skris 659572613Skris *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 659672613Skris call failed, free the DSA structure. 659772613Skris [Bodo Moeller] 659872613Skris 659972613Skris *) Fix to uni2asc() to cope with zero length Unicode strings. 660072613Skris These are present in some PKCS#12 files. 660172613Skris [Steve Henson] 660272613Skris 660372613Skris *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 660472613Skris Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 660572613Skris when writing a 32767 byte record. 660672613Skris [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 660772613Skris 660872613Skris *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 660972613Skris obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 661072613Skris 661172613Skris (RSA objects have a reference count access to which is protected 661272613Skris by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 661372613Skris so they are meant to be shared between threads.) 661472613Skris [Bodo Moeller, Geoff Thorpe; original patch submitted by 661572613Skris "Reddie, Steven" <Steven.Reddie@ca.com>] 661672613Skris 661772613Skris *) Fix a deadlock in CRYPTO_mem_leaks(). 661872613Skris [Bodo Moeller] 661972613Skris 662076866Skris *) Use better test patterns in bntest. 662176866Skris [Ulf M�ller] 662276866Skris 662372613Skris *) rand_win.c fix for Borland C. 662472613Skris [Ulf M�ller] 662572613Skris 662672613Skris *) BN_rshift bugfix for n == 0. 662772613Skris [Bodo Moeller] 662872613Skris 662976866Skris *) Add a 'bctest' script that checks for some known 'bc' bugs 663076866Skris so that 'make test' does not abort just because 'bc' is broken. 663176866Skris [Bodo Moeller] 663276866Skris 663372613Skris *) Store verify_result within SSL_SESSION also for client side to 663472613Skris avoid potential security hole. (Re-used sessions on the client side 663572613Skris always resulted in verify_result==X509_V_OK, not using the original 663672613Skris result of the server certificate verification.) 663772613Skris [Lutz Jaenicke] 663872613Skris 663972613Skris *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 664072613Skris SSL3_RT_APPLICATION_DATA, return 0. 664172613Skris Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 664272613Skris [Bodo Moeller] 664372613Skris 664472613Skris *) Fix SSL_peek: 664572613Skris Both ssl2_peek and ssl3_peek, which were totally broken in earlier 664672613Skris releases, have been re-implemented by renaming the previous 664772613Skris implementations of ssl2_read and ssl3_read to ssl2_read_internal 664872613Skris and ssl3_read_internal, respectively, and adding 'peek' parameters 664972613Skris to them. The new ssl[23]_{read,peek} functions are calls to 665072613Skris ssl[23]_read_internal with the 'peek' flag set appropriately. 665172613Skris A 'peek' parameter has also been added to ssl3_read_bytes, which 665272613Skris does the actual work for ssl3_read_internal. 665372613Skris [Bodo Moeller] 665472613Skris 665576866Skris *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 665676866Skris the method-specific "init()" handler. Also clean up ex_data after 665776866Skris calling the method-specific "finish()" handler. Previously, this was 665876866Skris happening the other way round. 665976866Skris [Geoff Thorpe] 666076866Skris 666172613Skris *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 666272613Skris The previous value, 12, was not always sufficient for BN_mod_exp(). 666372613Skris [Bodo Moeller] 666472613Skris 666576866Skris *) Make sure that shared libraries get the internal name engine with 666676866Skris the full version number and not just 0. This should mark the 666776866Skris shared libraries as not backward compatible. Of course, this should 666876866Skris be changed again when we can guarantee backward binary compatibility. 666976866Skris [Richard Levitte] 667076866Skris 667172613Skris *) Fix typo in get_cert_by_subject() in by_dir.c 667272613Skris [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 667372613Skris 667476866Skris *) Rework the system to generate shared libraries: 667576866Skris 667676866Skris - Make note of the expected extension for the shared libraries and 667776866Skris if there is a need for symbolic links from for example libcrypto.so.0 667876866Skris to libcrypto.so.0.9.7. There is extended info in Configure for 667976866Skris that. 668076866Skris 668176866Skris - Make as few rebuilds of the shared libraries as possible. 668276866Skris 668376866Skris - Still avoid linking the OpenSSL programs with the shared libraries. 668476866Skris 668576866Skris - When installing, install the shared libraries separately from the 668676866Skris static ones. 668776866Skris [Richard Levitte] 668876866Skris 668972613Skris *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 669072613Skris 669172613Skris Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 669272613Skris and not in SSL_clear because the latter is also used by the 669372613Skris accept/connect functions; previously, the settings made by 669472613Skris SSL_set_read_ahead would be lost during the handshake. 669572613Skris [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 669672613Skris 669772613Skris *) Correct util/mkdef.pl to be selective about disabled algorithms. 669872613Skris Previously, it would create entries for disableed algorithms no 669972613Skris matter what. 670072613Skris [Richard Levitte] 670172613Skris 670272613Skris *) Added several new manual pages for SSL_* function. 670372613Skris [Lutz Jaenicke] 670472613Skris 670568651Skris Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 670668651Skris 670768651Skris *) In ssl23_get_client_hello, generate an error message when faced 670868651Skris with an initial SSL 3.0/TLS record that is too small to contain the 670968651Skris first two bytes of the ClientHello message, i.e. client_version. 671068651Skris (Note that this is a pathologic case that probably has never happened 671168651Skris in real life.) The previous approach was to use the version number 671268651Skris from the record header as a substitute; but our protocol choice 671368651Skris should not depend on that one because it is not authenticated 671468651Skris by the Finished messages. 671568651Skris [Bodo Moeller] 671668651Skris 671768651Skris *) More robust randomness gathering functions for Windows. 671868651Skris [Jeffrey Altman <jaltman@columbia.edu>] 671968651Skris 672068651Skris *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 672168651Skris not set then we don't setup the error code for issuer check errors 672268651Skris to avoid possibly overwriting other errors which the callback does 672368651Skris handle. If an application does set the flag then we assume it knows 672468651Skris what it is doing and can handle the new informational codes 672568651Skris appropriately. 672668651Skris [Steve Henson] 672768651Skris 672868651Skris *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 672968651Skris a general "ANY" type, as such it should be able to decode anything 673068651Skris including tagged types. However it didn't check the class so it would 673168651Skris wrongly interpret tagged types in the same way as their universal 673268651Skris counterpart and unknown types were just rejected. Changed so that the 673368651Skris tagged and unknown types are handled in the same way as a SEQUENCE: 673468651Skris that is the encoding is stored intact. There is also a new type 673568651Skris "V_ASN1_OTHER" which is used when the class is not universal, in this 673668651Skris case we have no idea what the actual type is so we just lump them all 673768651Skris together. 673868651Skris [Steve Henson] 673968651Skris 674068651Skris *) On VMS, stdout may very well lead to a file that is written to 674168651Skris in a record-oriented fashion. That means that every write() will 674268651Skris write a separate record, which will be read separately by the 674368651Skris programs trying to read from it. This can be very confusing. 674468651Skris 674568651Skris The solution is to put a BIO filter in the way that will buffer 674668651Skris text until a linefeed is reached, and then write everything a 674768651Skris line at a time, so every record written will be an actual line, 674868651Skris not chunks of lines and not (usually doesn't happen, but I've 674968651Skris seen it once) several lines in one record. BIO_f_linebuffer() is 675068651Skris the answer. 675168651Skris 675268651Skris Currently, it's a VMS-only method, because that's where it has 675368651Skris been tested well enough. 675468651Skris [Richard Levitte] 675568651Skris 675668651Skris *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 675768651Skris it can return incorrect results. 675868651Skris (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 675968651Skris but it was in 0.9.6-beta[12].) 676068651Skris [Bodo Moeller] 676168651Skris 676268651Skris *) Disable the check for content being present when verifying detached 676368651Skris signatures in pk7_smime.c. Some versions of Netscape (wrongly) 676468651Skris include zero length content when signing messages. 676568651Skris [Steve Henson] 676668651Skris 676768651Skris *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 676868651Skris BIO_ctrl (for BIO pairs). 676968651Skris [Bodo M�ller] 677068651Skris 677168651Skris *) Add DSO method for VMS. 677268651Skris [Richard Levitte] 677368651Skris 677468651Skris *) Bug fix: Montgomery multiplication could produce results with the 677568651Skris wrong sign. 677668651Skris [Ulf M�ller] 677768651Skris 677868651Skris *) Add RPM specification openssl.spec and modify it to build three 677968651Skris packages. The default package contains applications, application 678068651Skris documentation and run-time libraries. The devel package contains 678168651Skris include files, static libraries and function documentation. The 678268651Skris doc package contains the contents of the doc directory. The original 678368651Skris openssl.spec was provided by Damien Miller <djm@mindrot.org>. 678468651Skris [Richard Levitte] 678568651Skris 678668651Skris *) Add a large number of documentation files for many SSL routines. 678768651Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 678868651Skris 678968651Skris *) Add a configuration entry for Sony News 4. 679068651Skris [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 679168651Skris 679268651Skris *) Don't set the two most significant bits to one when generating a 679368651Skris random number < q in the DSA library. 679468651Skris [Ulf M�ller] 679568651Skris 679668651Skris *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 679768651Skris behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 679868651Skris the underlying transport is blocking) if a handshake took place. 679968651Skris (The default behaviour is needed by applications such as s_client 680068651Skris and s_server that use select() to determine when to use SSL_read; 680168651Skris but for applications that know in advance when to expect data, it 680268651Skris just makes things more complicated.) 680368651Skris [Bodo Moeller] 680468651Skris 680568651Skris *) Add RAND_egd_bytes(), which gives control over the number of bytes read 680668651Skris from EGD. 680768651Skris [Ben Laurie] 680868651Skris 680968651Skris *) Add a few more EBCDIC conditionals that make `req' and `x509' 681068651Skris work better on such systems. 681168651Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 681268651Skris 681368651Skris *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 681468651Skris Update PKCS12_parse() so it copies the friendlyName and the 681568651Skris keyid to the certificates aux info. 681668651Skris [Steve Henson] 681768651Skris 681868651Skris *) Fix bug in PKCS7_verify() which caused an infinite loop 681968651Skris if there was more than one signature. 682068651Skris [Sven Uszpelkat <su@celocom.de>] 682168651Skris 682268651Skris *) Major change in util/mkdef.pl to include extra information 682368651Skris about each symbol, as well as presentig variables as well 682468651Skris as functions. This change means that there's n more need 682568651Skris to rebuild the .num files when some algorithms are excluded. 682668651Skris [Richard Levitte] 682768651Skris 682868651Skris *) Allow the verify time to be set by an application, 682968651Skris rather than always using the current time. 683068651Skris [Steve Henson] 683168651Skris 683268651Skris *) Phase 2 verify code reorganisation. The certificate 683368651Skris verify code now looks up an issuer certificate by a 683468651Skris number of criteria: subject name, authority key id 683568651Skris and key usage. It also verifies self signed certificates 683668651Skris by the same criteria. The main comparison function is 683768651Skris X509_check_issued() which performs these checks. 683868651Skris 683968651Skris Lot of changes were necessary in order to support this 684068651Skris without completely rewriting the lookup code. 684168651Skris 684268651Skris Authority and subject key identifier are now cached. 684368651Skris 684468651Skris The LHASH 'certs' is X509_STORE has now been replaced 684568651Skris by a STACK_OF(X509_OBJECT). This is mainly because an 684668651Skris LHASH can't store or retrieve multiple objects with 684768651Skris the same hash value. 684868651Skris 684968651Skris As a result various functions (which were all internal 685068651Skris use only) have changed to handle the new X509_STORE 685168651Skris structure. This will break anything that messed round 685268651Skris with X509_STORE internally. 685368651Skris 685468651Skris The functions X509_STORE_add_cert() now checks for an 685568651Skris exact match, rather than just subject name. 685668651Skris 685768651Skris The X509_STORE API doesn't directly support the retrieval 685868651Skris of multiple certificates matching a given criteria, however 685968651Skris this can be worked round by performing a lookup first 686068651Skris (which will fill the cache with candidate certificates) 686168651Skris and then examining the cache for matches. This is probably 686268651Skris the best we can do without throwing out X509_LOOKUP 686368651Skris entirely (maybe later...). 686468651Skris 686568651Skris The X509_VERIFY_CTX structure has been enhanced considerably. 686668651Skris 686768651Skris All certificate lookup operations now go via a get_issuer() 686868651Skris callback. Although this currently uses an X509_STORE it 686968651Skris can be replaced by custom lookups. This is a simple way 687068651Skris to bypass the X509_STORE hackery necessary to make this 687168651Skris work and makes it possible to use more efficient techniques 687268651Skris in future. A very simple version which uses a simple 687368651Skris STACK for its trusted certificate store is also provided 687468651Skris using X509_STORE_CTX_trusted_stack(). 687568651Skris 687668651Skris The verify_cb() and verify() callbacks now have equivalents 687768651Skris in the X509_STORE_CTX structure. 687868651Skris 687968651Skris X509_STORE_CTX also has a 'flags' field which can be used 688068651Skris to customise the verify behaviour. 688168651Skris [Steve Henson] 688268651Skris 688368651Skris *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 688468651Skris excludes S/MIME capabilities. 688568651Skris [Steve Henson] 688668651Skris 688768651Skris *) When a certificate request is read in keep a copy of the 688868651Skris original encoding of the signed data and use it when outputing 688968651Skris again. Signatures then use the original encoding rather than 689068651Skris a decoded, encoded version which may cause problems if the 689168651Skris request is improperly encoded. 689268651Skris [Steve Henson] 689368651Skris 689468651Skris *) For consistency with other BIO_puts implementations, call 689568651Skris buffer_write(b, ...) directly in buffer_puts instead of calling 689668651Skris BIO_write(b, ...). 689768651Skris 689868651Skris In BIO_puts, increment b->num_write as in BIO_write. 689968651Skris [Peter.Sylvester@EdelWeb.fr] 690068651Skris 690168651Skris *) Fix BN_mul_word for the case where the word is 0. (We have to use 690268651Skris BN_zero, we may not return a BIGNUM with an array consisting of 690368651Skris words set to zero.) 690468651Skris [Bodo Moeller] 690568651Skris 690668651Skris *) Avoid calling abort() from within the library when problems are 690768651Skris detected, except if preprocessor symbols have been defined 690868651Skris (such as REF_CHECK, BN_DEBUG etc.). 690968651Skris [Bodo Moeller] 691068651Skris 691168651Skris *) New openssl application 'rsautl'. This utility can be 691268651Skris used for low level RSA operations. DER public key 691368651Skris BIO/fp routines also added. 691468651Skris [Steve Henson] 691568651Skris 691668651Skris *) New Configure entry and patches for compiling on QNX 4. 691768651Skris [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 691868651Skris 691968651Skris *) A demo state-machine implementation was sponsored by 692068651Skris Nuron (http://www.nuron.com/) and is now available in 692168651Skris demos/state_machine. 692268651Skris [Ben Laurie] 692368651Skris 692468651Skris *) New options added to the 'dgst' utility for signature 692568651Skris generation and verification. 692668651Skris [Steve Henson] 692768651Skris 692868651Skris *) Unrecognized PKCS#7 content types are now handled via a 692968651Skris catch all ASN1_TYPE structure. This allows unsupported 693068651Skris types to be stored as a "blob" and an application can 693168651Skris encode and decode it manually. 693268651Skris [Steve Henson] 693368651Skris 693468651Skris *) Fix various signed/unsigned issues to make a_strex.c 693568651Skris compile under VC++. 693668651Skris [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 693768651Skris 693868651Skris *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 693968651Skris length if passed a buffer. ASN1_INTEGER_to_BN failed 694068651Skris if passed a NULL BN and its argument was negative. 694168651Skris [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 694268651Skris 694368651Skris *) Modification to PKCS#7 encoding routines to output definite 694468651Skris length encoding. Since currently the whole structures are in 694568651Skris memory there's not real point in using indefinite length 694668651Skris constructed encoding. However if OpenSSL is compiled with 694768651Skris the flag PKCS7_INDEFINITE_ENCODING the old form is used. 694868651Skris [Steve Henson] 694968651Skris 695068651Skris *) Added BIO_vprintf() and BIO_vsnprintf(). 695168651Skris [Richard Levitte] 695268651Skris 695368651Skris *) Added more prefixes to parse for in the the strings written 695468651Skris through a logging bio, to cover all the levels that are available 695568651Skris through syslog. The prefixes are now: 695668651Skris 695768651Skris PANIC, EMERG, EMR => LOG_EMERG 695868651Skris ALERT, ALR => LOG_ALERT 695968651Skris CRIT, CRI => LOG_CRIT 696068651Skris ERROR, ERR => LOG_ERR 696168651Skris WARNING, WARN, WAR => LOG_WARNING 696268651Skris NOTICE, NOTE, NOT => LOG_NOTICE 696368651Skris INFO, INF => LOG_INFO 696468651Skris DEBUG, DBG => LOG_DEBUG 696568651Skris 696668651Skris and as before, if none of those prefixes are present at the 696768651Skris beginning of the string, LOG_ERR is chosen. 696868651Skris 696968651Skris On Win32, the LOG_* levels are mapped according to this: 697068651Skris 697168651Skris LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 697268651Skris LOG_WARNING => EVENTLOG_WARNING_TYPE 697368651Skris LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 697468651Skris 697568651Skris [Richard Levitte] 697668651Skris 697768651Skris *) Made it possible to reconfigure with just the configuration 697868651Skris argument "reconf" or "reconfigure". The command line arguments 697968651Skris are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 698068651Skris and are retrieved from there when reconfiguring. 698168651Skris [Richard Levitte] 698268651Skris 698368651Skris *) MD4 implemented. 698468651Skris [Assar Westerlund <assar@sics.se>, Richard Levitte] 698568651Skris 698668651Skris *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 698768651Skris [Richard Levitte] 698868651Skris 698968651Skris *) The obj_dat.pl script was messing up the sorting of object 699068651Skris names. The reason was that it compared the quoted version 699168651Skris of strings as a result "OCSP" > "OCSP Signing" because 699268651Skris " > SPACE. Changed script to store unquoted versions of 699368651Skris names and add quotes on output. It was also omitting some 699468651Skris names from the lookup table if they were given a default 699568651Skris value (that is if SN is missing it is given the same 699668651Skris value as LN and vice versa), these are now added on the 699768651Skris grounds that if an object has a name we should be able to 699868651Skris look it up. Finally added warning output when duplicate 699968651Skris short or long names are found. 700068651Skris [Steve Henson] 700168651Skris 700268651Skris *) Changes needed for Tandem NSK. 700368651Skris [Scott Uroff <scott@xypro.com>] 700468651Skris 700568651Skris *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 700668651Skris RSA_padding_check_SSLv23(), special padding was never detected 700768651Skris and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 700868651Skris version rollback attacks was not effective. 700968651Skris 701068651Skris In s23_clnt.c, don't use special rollback-attack detection padding 701168651Skris (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 701268651Skris client; similarly, in s23_srvr.c, don't do the rollback check if 701368651Skris SSL 2.0 is the only protocol enabled in the server. 701468651Skris [Bodo Moeller] 701568651Skris 701668651Skris *) Make it possible to get hexdumps of unprintable data with 'openssl 701768651Skris asn1parse'. By implication, the functions ASN1_parse_dump() and 701868651Skris BIO_dump_indent() are added. 701968651Skris [Richard Levitte] 702068651Skris 702168651Skris *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 702268651Skris these print out strings and name structures based on various 702368651Skris flags including RFC2253 support and proper handling of 702468651Skris multibyte characters. Added options to the 'x509' utility 702568651Skris to allow the various flags to be set. 702668651Skris [Steve Henson] 702768651Skris 702868651Skris *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 702968651Skris Also change the functions X509_cmp_current_time() and 703068651Skris X509_gmtime_adj() work with an ASN1_TIME structure, 703168651Skris this will enable certificates using GeneralizedTime in validity 703268651Skris dates to be checked. 703368651Skris [Steve Henson] 703468651Skris 703568651Skris *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 703668651Skris negative public key encodings) on by default, 703768651Skris NO_NEG_PUBKEY_BUG can be set to disable it. 703868651Skris [Steve Henson] 703968651Skris 704068651Skris *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 704168651Skris content octets. An i2c_ASN1_OBJECT is unnecessary because 704268651Skris the encoding can be trivially obtained from the structure. 704368651Skris [Steve Henson] 704468651Skris 704568651Skris *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 704668651Skris not read locks (CRYPTO_r_[un]lock). 704768651Skris [Bodo Moeller] 704868651Skris 704968651Skris *) A first attempt at creating official support for shared 705068651Skris libraries through configuration. I've kept it so the 705168651Skris default is static libraries only, and the OpenSSL programs 705268651Skris are always statically linked for now, but there are 705368651Skris preparations for dynamic linking in place. 705489837Skris This has been tested on Linux and Tru64. 705568651Skris [Richard Levitte] 705668651Skris 705768651Skris *) Randomness polling function for Win9x, as described in: 705868651Skris Peter Gutmann, Software Generation of Practically Strong 705968651Skris Random Numbers. 706068651Skris [Ulf M�ller] 706168651Skris 706268651Skris *) Fix so PRNG is seeded in req if using an already existing 706368651Skris DSA key. 706468651Skris [Steve Henson] 706568651Skris 706668651Skris *) New options to smime application. -inform and -outform 706768651Skris allow alternative formats for the S/MIME message including 706868651Skris PEM and DER. The -content option allows the content to be 706968651Skris specified separately. This should allow things like Netscape 707068651Skris form signing output easier to verify. 707168651Skris [Steve Henson] 707268651Skris 707368651Skris *) Fix the ASN1 encoding of tags using the 'long form'. 707468651Skris [Steve Henson] 707568651Skris 707668651Skris *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 707768651Skris STRING types. These convert content octets to and from the 707868651Skris underlying type. The actual tag and length octets are 707968651Skris already assumed to have been read in and checked. These 708068651Skris are needed because all other string types have virtually 708168651Skris identical handling apart from the tag. By having versions 708268651Skris of the ASN1 functions that just operate on content octets 708368651Skris IMPLICIT tagging can be handled properly. It also allows 708468651Skris the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 708568651Skris and ASN1_INTEGER are identical apart from the tag. 708668651Skris [Steve Henson] 708768651Skris 708868651Skris *) Change the handling of OID objects as follows: 708968651Skris 709068651Skris - New object identifiers are inserted in objects.txt, following 709168651Skris the syntax given in objects.README. 709268651Skris - objects.pl is used to process obj_mac.num and create a new 709368651Skris obj_mac.h. 709468651Skris - obj_dat.pl is used to create a new obj_dat.h, using the data in 709568651Skris obj_mac.h. 709668651Skris 709768651Skris This is currently kind of a hack, and the perl code in objects.pl 709868651Skris isn't very elegant, but it works as I intended. The simplest way 709968651Skris to check that it worked correctly is to look in obj_dat.h and 710068651Skris check the array nid_objs and make sure the objects haven't moved 710168651Skris around (this is important!). Additions are OK, as well as 710268651Skris consistent name changes. 710368651Skris [Richard Levitte] 710468651Skris 710568651Skris *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 710668651Skris [Bodo Moeller] 710768651Skris 710868651Skris *) Addition of the command line parameter '-rand file' to 'openssl req'. 710968651Skris The given file adds to whatever has already been seeded into the 711068651Skris random pool through the RANDFILE configuration file option or 711168651Skris environment variable, or the default random state file. 711268651Skris [Richard Levitte] 711368651Skris 711468651Skris *) mkstack.pl now sorts each macro group into lexical order. 711568651Skris Previously the output order depended on the order the files 711668651Skris appeared in the directory, resulting in needless rewriting 711768651Skris of safestack.h . 711868651Skris [Steve Henson] 711968651Skris 712068651Skris *) Patches to make OpenSSL compile under Win32 again. Mostly 712168651Skris work arounds for the VC++ problem that it treats func() as 712268651Skris func(void). Also stripped out the parts of mkdef.pl that 712368651Skris added extra typesafe functions: these no longer exist. 712468651Skris [Steve Henson] 712568651Skris 712668651Skris *) Reorganisation of the stack code. The macros are now all 712768651Skris collected in safestack.h . Each macro is defined in terms of 712868651Skris a "stack macro" of the form SKM_<name>(type, a, b). The 712968651Skris DEBUG_SAFESTACK is now handled in terms of function casts, 713068651Skris this has the advantage of retaining type safety without the 713168651Skris use of additional functions. If DEBUG_SAFESTACK is not defined 713268651Skris then the non typesafe macros are used instead. Also modified the 713368651Skris mkstack.pl script to handle the new form. Needs testing to see 713468651Skris if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 713568651Skris the default if no major problems. Similar behaviour for ASN1_SET_OF 713668651Skris and PKCS12_STACK_OF. 713768651Skris [Steve Henson] 713868651Skris 713968651Skris *) When some versions of IIS use the 'NET' form of private key the 714068651Skris key derivation algorithm is different. Normally MD5(password) is 714168651Skris used as a 128 bit RC4 key. In the modified case 714268651Skris MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 714368651Skris new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 714468651Skris as the old Netscape_RSA functions except they have an additional 714568651Skris 'sgckey' parameter which uses the modified algorithm. Also added 714668651Skris an -sgckey command line option to the rsa utility. Thanks to 714768651Skris Adrian Peck <bertie@ncipher.com> for posting details of the modified 714868651Skris algorithm to openssl-dev. 714968651Skris [Steve Henson] 715068651Skris 715168651Skris *) The evp_local.h macros were using 'c.##kname' which resulted in 715268651Skris invalid expansion on some systems (SCO 5.0.5 for example). 715368651Skris Corrected to 'c.kname'. 715468651Skris [Phillip Porch <root@theporch.com>] 715568651Skris 715668651Skris *) New X509_get1_email() and X509_REQ_get1_email() functions that return 715768651Skris a STACK of email addresses from a certificate or request, these look 715868651Skris in the subject name and the subject alternative name extensions and 715968651Skris omit any duplicate addresses. 716068651Skris [Steve Henson] 716168651Skris 716268651Skris *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 716368651Skris This makes DSA verification about 2 % faster. 716468651Skris [Bodo Moeller] 716568651Skris 716668651Skris *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 716768651Skris (meaning that now 2^5 values will be precomputed, which is only 4 KB 716868651Skris plus overhead for 1024 bit moduli). 716968651Skris This makes exponentiations about 0.5 % faster for 1024 bit 717068651Skris exponents (as measured by "openssl speed rsa2048"). 717168651Skris [Bodo Moeller] 717268651Skris 717368651Skris *) Rename memory handling macros to avoid conflicts with other 717468651Skris software: 717568651Skris Malloc => OPENSSL_malloc 717668651Skris Malloc_locked => OPENSSL_malloc_locked 717768651Skris Realloc => OPENSSL_realloc 717868651Skris Free => OPENSSL_free 717968651Skris [Richard Levitte] 718068651Skris 718168651Skris *) New function BN_mod_exp_mont_word for small bases (roughly 15% 718268651Skris faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 718368651Skris [Bodo Moeller] 718468651Skris 718568651Skris *) CygWin32 support. 718668651Skris [John Jarvie <jjarvie@newsguy.com>] 718768651Skris 718868651Skris *) The type-safe stack code has been rejigged. It is now only compiled 718968651Skris in when OpenSSL is configured with the DEBUG_SAFESTACK option and 719068651Skris by default all type-specific stack functions are "#define"d back to 719168651Skris standard stack functions. This results in more streamlined output 719268651Skris but retains the type-safety checking possibilities of the original 719368651Skris approach. 719468651Skris [Geoff Thorpe] 719568651Skris 719668651Skris *) The STACK code has been cleaned up, and certain type declarations 719768651Skris that didn't make a lot of sense have been brought in line. This has 719868651Skris also involved a cleanup of sorts in safestack.h to more correctly 719968651Skris map type-safe stack functions onto their plain stack counterparts. 720068651Skris This work has also resulted in a variety of "const"ifications of 720168651Skris lots of the code, especially "_cmp" operations which should normally 720268651Skris be prototyped with "const" parameters anyway. 720368651Skris [Geoff Thorpe] 720468651Skris 720568651Skris *) When generating bytes for the first time in md_rand.c, 'stir the pool' 720668651Skris by seeding with STATE_SIZE dummy bytes (with zero entropy count). 720768651Skris (The PRNG state consists of two parts, the large pool 'state' and 'md', 720868651Skris where all of 'md' is used each time the PRNG is used, but 'state' 720968651Skris is used only indexed by a cyclic counter. As entropy may not be 721068651Skris well distributed from the beginning, 'md' is important as a 721168651Skris chaining variable. However, the output function chains only half 721268651Skris of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 721368651Skris all of 'md', and seeding with STATE_SIZE dummy bytes will result 721468651Skris in all of 'state' being rewritten, with the new values depending 721568651Skris on virtually all of 'md'. This overcomes the 80 bit limitation.) 721668651Skris [Bodo Moeller] 721768651Skris 721868651Skris *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 721968651Skris the handshake is continued after ssl_verify_cert_chain(); 722068651Skris otherwise, if SSL_VERIFY_NONE is set, remaining error codes 722168651Skris can lead to 'unexplainable' connection aborts later. 722268651Skris [Bodo Moeller; problem tracked down by Lutz Jaenicke] 722368651Skris 722468651Skris *) Major EVP API cipher revision. 722568651Skris Add hooks for extra EVP features. This allows various cipher 722668651Skris parameters to be set in the EVP interface. Support added for variable 722768651Skris key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 722868651Skris setting of RC2 and RC5 parameters. 722968651Skris 723068651Skris Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 723168651Skris ciphers. 723268651Skris 723368651Skris Remove lots of duplicated code from the EVP library. For example *every* 723468651Skris cipher init() function handles the 'iv' in the same way according to the 723568651Skris cipher mode. They also all do nothing if the 'key' parameter is NULL and 723668651Skris for CFB and OFB modes they zero ctx->num. 723768651Skris 723868651Skris New functionality allows removal of S/MIME code RC2 hack. 723968651Skris 724068651Skris Most of the routines have the same form and so can be declared in terms 724168651Skris of macros. 724268651Skris 724368651Skris By shifting this to the top level EVP_CipherInit() it can be removed from 724468651Skris all individual ciphers. If the cipher wants to handle IVs or keys 724568651Skris differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 724668651Skris flags. 724768651Skris 724868651Skris Change lots of functions like EVP_EncryptUpdate() to now return a 724968651Skris value: although software versions of the algorithms cannot fail 725068651Skris any installed hardware versions can. 725168651Skris [Steve Henson] 725268651Skris 725368651Skris *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 725468651Skris this option is set, tolerate broken clients that send the negotiated 725568651Skris protocol version number instead of the requested protocol version 725668651Skris number. 725768651Skris [Bodo Moeller] 725868651Skris 725968651Skris *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 726068651Skris i.e. non-zero for export ciphersuites, zero otherwise. 726168651Skris Previous versions had this flag inverted, inconsistent with 726268651Skris rsa_tmp_cb (..._TMP_RSA_CB). 726368651Skris [Bodo Moeller; problem reported by Amit Chopra] 726468651Skris 726568651Skris *) Add missing DSA library text string. Work around for some IIS 726668651Skris key files with invalid SEQUENCE encoding. 726768651Skris [Steve Henson] 726868651Skris 726968651Skris *) Add a document (doc/standards.txt) that list all kinds of standards 727068651Skris and so on that are implemented in OpenSSL. 727168651Skris [Richard Levitte] 727268651Skris 727368651Skris *) Enhance c_rehash script. Old version would mishandle certificates 727468651Skris with the same subject name hash and wouldn't handle CRLs at all. 727568651Skris Added -fingerprint option to crl utility, to support new c_rehash 727668651Skris features. 727768651Skris [Steve Henson] 727868651Skris 727968651Skris *) Eliminate non-ANSI declarations in crypto.h and stack.h. 728068651Skris [Ulf M�ller] 728168651Skris 728268651Skris *) Fix for SSL server purpose checking. Server checking was 728368651Skris rejecting certificates which had extended key usage present 728468651Skris but no ssl client purpose. 728568651Skris [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 728668651Skris 728768651Skris *) Make PKCS#12 code work with no password. The PKCS#12 spec 728868651Skris is a little unclear about how a blank password is handled. 728968651Skris Since the password in encoded as a BMPString with terminating 729068651Skris double NULL a zero length password would end up as just the 729168651Skris double NULL. However no password at all is different and is 729268651Skris handled differently in the PKCS#12 key generation code. NS 729368651Skris treats a blank password as zero length. MSIE treats it as no 729468651Skris password on export: but it will try both on import. We now do 729568651Skris the same: PKCS12_parse() tries zero length and no password if 729668651Skris the password is set to "" or NULL (NULL is now a valid password: 729768651Skris it wasn't before) as does the pkcs12 application. 729868651Skris [Steve Henson] 729968651Skris 730068651Skris *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 730168651Skris perror when PEM_read_bio_X509_REQ fails, the error message must 730268651Skris be obtained from the error queue. 730368651Skris [Bodo Moeller] 730468651Skris 730568651Skris *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 730668651Skris it in ERR_remove_state if appropriate, and change ERR_get_state 730768651Skris accordingly to avoid race conditions (this is necessary because 730868651Skris thread_hash is no longer constant once set). 730968651Skris [Bodo Moeller] 731068651Skris 731168651Skris *) Bugfix for linux-elf makefile.one. 731268651Skris [Ulf M�ller] 731368651Skris 731468651Skris *) RSA_get_default_method() will now cause a default 731568651Skris RSA_METHOD to be chosen if one doesn't exist already. 731668651Skris Previously this was only set during a call to RSA_new() 731768651Skris or RSA_new_method(NULL) meaning it was possible for 731868651Skris RSA_get_default_method() to return NULL. 731968651Skris [Geoff Thorpe] 732068651Skris 732168651Skris *) Added native name translation to the existing DSO code 732268651Skris that will convert (if the flag to do so is set) filenames 732368651Skris that are sufficiently small and have no path information 732468651Skris into a canonical native form. Eg. "blah" converted to 732568651Skris "libblah.so" or "blah.dll" etc. 732668651Skris [Geoff Thorpe] 732768651Skris 732868651Skris *) New function ERR_error_string_n(e, buf, len) which is like 732968651Skris ERR_error_string(e, buf), but writes at most 'len' bytes 733068651Skris including the 0 terminator. For ERR_error_string_n, 'buf' 733168651Skris may not be NULL. 733268651Skris [Damien Miller <djm@mindrot.org>, Bodo Moeller] 733368651Skris 733468651Skris *) CONF library reworked to become more general. A new CONF 733568651Skris configuration file reader "class" is implemented as well as a 733668651Skris new functions (NCONF_*, for "New CONF") to handle it. The now 733768651Skris old CONF_* functions are still there, but are reimplemented to 733868651Skris work in terms of the new functions. Also, a set of functions 733968651Skris to handle the internal storage of the configuration data is 734068651Skris provided to make it easier to write new configuration file 734168651Skris reader "classes" (I can definitely see something reading a 734268651Skris configuration file in XML format, for example), called _CONF_*, 734368651Skris or "the configuration storage API"... 734468651Skris 734568651Skris The new configuration file reading functions are: 734668651Skris 734768651Skris NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 734868651Skris NCONF_get_section, NCONF_get_string, NCONF_get_numbre 734968651Skris 735068651Skris NCONF_default, NCONF_WIN32 735168651Skris 735268651Skris NCONF_dump_fp, NCONF_dump_bio 735368651Skris 735468651Skris NCONF_default and NCONF_WIN32 are method (or "class") choosers, 735568651Skris NCONF_new creates a new CONF object. This works in the same way 735668651Skris as other interfaces in OpenSSL, like the BIO interface. 735768651Skris NCONF_dump_* dump the internal storage of the configuration file, 735868651Skris which is useful for debugging. All other functions take the same 735968651Skris arguments as the old CONF_* functions wth the exception of the 736068651Skris first that must be a `CONF *' instead of a `LHASH *'. 736168651Skris 736268651Skris To make it easer to use the new classes with the old CONF_* functions, 736368651Skris the function CONF_set_default_method is provided. 736468651Skris [Richard Levitte] 736568651Skris 736668651Skris *) Add '-tls1' option to 'openssl ciphers', which was already 736768651Skris mentioned in the documentation but had not been implemented. 736868651Skris (This option is not yet really useful because even the additional 736968651Skris experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 737068651Skris [Bodo Moeller] 737168651Skris 737268651Skris *) Initial DSO code added into libcrypto for letting OpenSSL (and 737368651Skris OpenSSL-based applications) load shared libraries and bind to 737468651Skris them in a portable way. 737568651Skris [Geoff Thorpe, with contributions from Richard Levitte] 737668651Skris 737759191Skris Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 737859191Skris 737959191Skris *) Make sure _lrotl and _lrotr are only used with MSVC. 738059191Skris 738159191Skris *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 738259191Skris (the default implementation of RAND_status). 738359191Skris 738459191Skris *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 738559191Skris to '-clrext' (= clear extensions), as intended and documented. 738659191Skris [Bodo Moeller; inconsistency pointed out by Michael Attili 738759191Skris <attili@amaxo.com>] 738859191Skris 738959191Skris *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 739059191Skris was larger than the MD block size. 739159191Skris [Steve Henson, pointed out by Yost William <YostW@tce.com>] 739259191Skris 739359191Skris *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 739459191Skris fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 739559191Skris using the passed key: if the passed key was a private key the result 739659191Skris of X509_print(), for example, would be to print out all the private key 739759191Skris components. 739859191Skris [Steve Henson] 739959191Skris 740059191Skris *) des_quad_cksum() byte order bug fix. 740159191Skris [Ulf M�ller, using the problem description in krb4-0.9.7, where 740259191Skris the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 740359191Skris 740459191Skris *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 740559191Skris discouraged. 740659191Skris [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 740759191Skris 740859191Skris *) For easily testing in shell scripts whether some command 740959191Skris 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 741059191Skris returns with exit code 0 iff no command of the given name is available. 741159191Skris 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 741259191Skris the output goes to stdout and nothing is printed to stderr. 741359191Skris Additional arguments are always ignored. 741459191Skris 741559191Skris Since for each cipher there is a command of the same name, 741659191Skris the 'no-cipher' compilation switches can be tested this way. 741759191Skris 741859191Skris ('openssl no-XXX' is not able to detect pseudo-commands such 741959191Skris as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 742059191Skris [Bodo Moeller] 742159191Skris 742259191Skris *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 742359191Skris [Bodo Moeller] 742459191Skris 742559191Skris *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 742659191Skris is set; it will be thrown away anyway because each handshake creates 742759191Skris its own key. 742859191Skris ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 742959191Skris to parameters -- in previous versions (since OpenSSL 0.9.3) the 743059191Skris 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 743159191Skris you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 743259191Skris [Bodo Moeller] 743359191Skris 743459191Skris *) New s_client option -ign_eof: EOF at stdin is ignored, and 743559191Skris 'Q' and 'R' lose their special meanings (quit/renegotiate). 743659191Skris This is part of what -quiet does; unlike -quiet, -ign_eof 743759191Skris does not suppress any output. 743859191Skris [Richard Levitte] 743959191Skris 744059191Skris *) Add compatibility options to the purpose and trust code. The 744159191Skris purpose X509_PURPOSE_ANY is "any purpose" which automatically 744259191Skris accepts a certificate or CA, this was the previous behaviour, 744359191Skris with all the associated security issues. 744459191Skris 744559191Skris X509_TRUST_COMPAT is the old trust behaviour: only and 744659191Skris automatically trust self signed roots in certificate store. A 744759191Skris new trust setting X509_TRUST_DEFAULT is used to specify that 744859191Skris a purpose has no associated trust setting and it should instead 744959191Skris use the value in the default purpose. 745059191Skris [Steve Henson] 745159191Skris 745259191Skris *) Fix the PKCS#8 DSA private key code so it decodes keys again 745359191Skris and fix a memory leak. 745459191Skris [Steve Henson] 745559191Skris 745659191Skris *) In util/mkerr.pl (which implements 'make errors'), preserve 745759191Skris reason strings from the previous version of the .c file, as 745859191Skris the default to have only downcase letters (and digits) in 745959191Skris automatically generated reasons codes is not always appropriate. 746059191Skris [Bodo Moeller] 746159191Skris 746259191Skris *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 746359191Skris using strerror. Previously, ERR_reason_error_string() returned 746459191Skris library names as reason strings for SYSerr; but SYSerr is a special 746559191Skris case where small numbers are errno values, not library numbers. 746659191Skris [Bodo Moeller] 746759191Skris 746859191Skris *) Add '-dsaparam' option to 'openssl dhparam' application. This 746959191Skris converts DSA parameters into DH parameters. (When creating parameters, 747059191Skris DSA_generate_parameters is used.) 747159191Skris [Bodo Moeller] 747259191Skris 747359191Skris *) Include 'length' (recommended exponent length) in C code generated 747459191Skris by 'openssl dhparam -C'. 747559191Skris [Bodo Moeller] 747659191Skris 747759191Skris *) The second argument to set_label in perlasm was already being used 747859191Skris so couldn't be used as a "file scope" flag. Moved to third argument 747959191Skris which was free. 748059191Skris [Steve Henson] 748159191Skris 748259191Skris *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 748359191Skris instead of RAND_bytes for encryption IVs and salts. 748459191Skris [Bodo Moeller] 748559191Skris 748659191Skris *) Include RAND_status() into RAND_METHOD instead of implementing 748759191Skris it only for md_rand.c Otherwise replacing the PRNG by calling 748859191Skris RAND_set_rand_method would be impossible. 748959191Skris [Bodo Moeller] 749059191Skris 749159191Skris *) Don't let DSA_generate_key() enter an infinite loop if the random 749259191Skris number generation fails. 749359191Skris [Bodo Moeller] 749459191Skris 749559191Skris *) New 'rand' application for creating pseudo-random output. 749659191Skris [Bodo Moeller] 749759191Skris 749859191Skris *) Added configuration support for Linux/IA64 749959191Skris [Rolf Haberrecker <rolf@suse.de>] 750059191Skris 750159191Skris *) Assembler module support for Mingw32. 750259191Skris [Ulf M�ller] 750359191Skris 750459191Skris *) Shared library support for HPUX (in shlib/). 750559191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 750659191Skris 750759191Skris *) Shared library support for Solaris gcc. 750859191Skris [Lutz Behnke <behnke@trustcenter.de>] 750959191Skris 751059191Skris Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 751159191Skris 751259191Skris *) PKCS7_encrypt() was adding text MIME headers twice because they 751359191Skris were added manually and by SMIME_crlf_copy(). 751459191Skris [Steve Henson] 751559191Skris 751659191Skris *) In bntest.c don't call BN_rand with zero bits argument. 751759191Skris [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 751859191Skris 751959191Skris *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 752059191Skris case was implemented. This caused BN_div_recp() to fail occasionally. 752159191Skris [Ulf M�ller] 752259191Skris 752359191Skris *) Add an optional second argument to the set_label() in the perl 752459191Skris assembly language builder. If this argument exists and is set 752559191Skris to 1 it signals that the assembler should use a symbol whose 752659191Skris scope is the entire file, not just the current function. This 752759191Skris is needed with MASM which uses the format label:: for this scope. 752859191Skris [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 752959191Skris 753059191Skris *) Change the ASN1 types so they are typedefs by default. Before 753159191Skris almost all types were #define'd to ASN1_STRING which was causing 753259191Skris STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 753359191Skris for example. 753459191Skris [Steve Henson] 753559191Skris 753659191Skris *) Change names of new functions to the new get1/get0 naming 753759191Skris convention: After 'get1', the caller owns a reference count 753859191Skris and has to call ..._free; 'get0' returns a pointer to some 753959191Skris data structure without incrementing reference counters. 754059191Skris (Some of the existing 'get' functions increment a reference 754159191Skris counter, some don't.) 754259191Skris Similarly, 'set1' and 'add1' functions increase reference 754359191Skris counters or duplicate objects. 754459191Skris [Steve Henson] 754559191Skris 754659191Skris *) Allow for the possibility of temp RSA key generation failure: 754759191Skris the code used to assume it always worked and crashed on failure. 754859191Skris [Steve Henson] 754959191Skris 755059191Skris *) Fix potential buffer overrun problem in BIO_printf(). 755159191Skris [Ulf M�ller, using public domain code by Patrick Powell; problem 755259191Skris pointed out by David Sacerdote <das33@cornell.edu>] 755359191Skris 755459191Skris *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 755559191Skris RAND_egd() and RAND_status(). In the command line application, 755659191Skris the EGD socket can be specified like a seed file using RANDFILE 755759191Skris or -rand. 755859191Skris [Ulf M�ller] 755959191Skris 756059191Skris *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 756159191Skris Some CAs (e.g. Verisign) distribute certificates in this form. 756259191Skris [Steve Henson] 756359191Skris 756459191Skris *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 756559191Skris list to exclude them. This means that no special compilation option 756659191Skris is needed to use anonymous DH: it just needs to be included in the 756759191Skris cipher list. 756859191Skris [Steve Henson] 756959191Skris 757059191Skris *) Change the EVP_MD_CTX_type macro so its meaning consistent with 757159191Skris EVP_MD_type. The old functionality is available in a new macro called 757259191Skris EVP_MD_md(). Change code that uses it and update docs. 757359191Skris [Steve Henson] 757459191Skris 757559191Skris *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 757659191Skris where the 'void *' argument is replaced by a function pointer argument. 757759191Skris Previously 'void *' was abused to point to functions, which works on 757859191Skris many platforms, but is not correct. As these functions are usually 757959191Skris called by macros defined in OpenSSL header files, most source code 758059191Skris should work without changes. 758159191Skris [Richard Levitte] 758259191Skris 758359191Skris *) <openssl/opensslconf.h> (which is created by Configure) now contains 758459191Skris sections with information on -D... compiler switches used for 758559191Skris compiling the library so that applications can see them. To enable 758659191Skris one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 758759191Skris must be defined. E.g., 758859191Skris #define OPENSSL_ALGORITHM_DEFINES 758959191Skris #include <openssl/opensslconf.h> 759059191Skris defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 759159191Skris [Richard Levitte, Ulf and Bodo M�ller] 759259191Skris 759359191Skris *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 759459191Skris record layer. 759559191Skris [Bodo Moeller] 759659191Skris 759759191Skris *) Change the 'other' type in certificate aux info to a STACK_OF 759859191Skris X509_ALGOR. Although not an AlgorithmIdentifier as such it has 759959191Skris the required ASN1 format: arbitrary types determined by an OID. 760059191Skris [Steve Henson] 760159191Skris 760259191Skris *) Add some PEM_write_X509_REQ_NEW() functions and a command line 760359191Skris argument to 'req'. This is not because the function is newer or 760459191Skris better than others it just uses the work 'NEW' in the certificate 760559191Skris request header lines. Some software needs this. 760659191Skris [Steve Henson] 760759191Skris 760859191Skris *) Reorganise password command line arguments: now passwords can be 760959191Skris obtained from various sources. Delete the PEM_cb function and make 761059191Skris it the default behaviour: i.e. if the callback is NULL and the 761159191Skris usrdata argument is not NULL interpret it as a null terminated pass 761259191Skris phrase. If usrdata and the callback are NULL then the pass phrase 761359191Skris is prompted for as usual. 761459191Skris [Steve Henson] 761559191Skris 761659191Skris *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 761759191Skris the support is automatically enabled. The resulting binaries will 761859191Skris autodetect the card and use it if present. 761959191Skris [Ben Laurie and Compaq Inc.] 762059191Skris 762159191Skris *) Work around for Netscape hang bug. This sends certificate request 762259191Skris and server done in one record. Since this is perfectly legal in the 762359191Skris SSL/TLS protocol it isn't a "bug" option and is on by default. See 762459191Skris the bugs/SSLv3 entry for more info. 762559191Skris [Steve Henson] 762659191Skris 762759191Skris *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 762859191Skris [Andy Polyakov] 762959191Skris 763059191Skris *) Add -rand argument to smime and pkcs12 applications and read/write 763159191Skris of seed file. 763259191Skris [Steve Henson] 763359191Skris 763459191Skris *) New 'passwd' tool for crypt(3) and apr1 password hashes. 763559191Skris [Bodo Moeller] 763659191Skris 763759191Skris *) Add command line password options to the remaining applications. 763859191Skris [Steve Henson] 763959191Skris 764059191Skris *) Bug fix for BN_div_recp() for numerators with an even number of 764159191Skris bits. 764259191Skris [Ulf M�ller] 764359191Skris 764459191Skris *) More tests in bntest.c, and changed test_bn output. 764559191Skris [Ulf M�ller] 764659191Skris 764759191Skris *) ./config recognizes MacOS X now. 764859191Skris [Andy Polyakov] 764959191Skris 765059191Skris *) Bug fix for BN_div() when the first words of num and divsor are 765159191Skris equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 765259191Skris [Ulf M�ller] 765359191Skris 765459191Skris *) Add support for various broken PKCS#8 formats, and command line 765559191Skris options to produce them. 765659191Skris [Steve Henson] 765759191Skris 765859191Skris *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 765959191Skris get temporary BIGNUMs from a BN_CTX. 766059191Skris [Ulf M�ller] 766159191Skris 766259191Skris *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 766359191Skris for p == 0. 766459191Skris [Ulf M�ller] 766559191Skris 766659191Skris *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 766759191Skris include a #define from the old name to the new. The original intent 766859191Skris was that statically linked binaries could for example just call 766959191Skris SSLeay_add_all_ciphers() to just add ciphers to the table and not 767059191Skris link with digests. This never worked becayse SSLeay_add_all_digests() 767159191Skris and SSLeay_add_all_ciphers() were in the same source file so calling 767259191Skris one would link with the other. They are now in separate source files. 767359191Skris [Steve Henson] 767459191Skris 767559191Skris *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 767659191Skris [Steve Henson] 767759191Skris 767859191Skris *) Use a less unusual form of the Miller-Rabin primality test (it used 767959191Skris a binary algorithm for exponentiation integrated into the Miller-Rabin 768059191Skris loop, our standard modexp algorithms are faster). 768159191Skris [Bodo Moeller] 768259191Skris 768359191Skris *) Support for the EBCDIC character set completed. 768459191Skris [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 768559191Skris 768659191Skris *) Source code cleanups: use const where appropriate, eliminate casts, 768759191Skris use void * instead of char * in lhash. 768859191Skris [Ulf M�ller] 768959191Skris 769059191Skris *) Bugfix: ssl3_send_server_key_exchange was not restartable 769159191Skris (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 769259191Skris this the server could overwrite ephemeral keys that the client 769359191Skris has already seen). 769459191Skris [Bodo Moeller] 769559191Skris 769659191Skris *) Turn DSA_is_prime into a macro that calls BN_is_prime, 769759191Skris using 50 iterations of the Rabin-Miller test. 769859191Skris 769959191Skris DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 770059191Skris iterations of the Rabin-Miller test as required by the appendix 770159191Skris to FIPS PUB 186[-1]) instead of DSA_is_prime. 770259191Skris As BN_is_prime_fasttest includes trial division, DSA parameter 770359191Skris generation becomes much faster. 770459191Skris 770559191Skris This implies a change for the callback functions in DSA_is_prime 770659191Skris and DSA_generate_parameters: The callback function is called once 770759191Skris for each positive witness in the Rabin-Miller test, not just 770859191Skris occasionally in the inner loop; and the parameters to the 770959191Skris callback function now provide an iteration count for the outer 771059191Skris loop rather than for the current invocation of the inner loop. 771159191Skris DSA_generate_parameters additionally can call the callback 771259191Skris function with an 'iteration count' of -1, meaning that a 771359191Skris candidate has passed the trial division test (when q is generated 771459191Skris from an application-provided seed, trial division is skipped). 771559191Skris [Bodo Moeller] 771659191Skris 771759191Skris *) New function BN_is_prime_fasttest that optionally does trial 771859191Skris division before starting the Rabin-Miller test and has 771959191Skris an additional BN_CTX * argument (whereas BN_is_prime always 772059191Skris has to allocate at least one BN_CTX). 772159191Skris 'callback(1, -1, cb_arg)' is called when a number has passed the 772259191Skris trial division stage. 772359191Skris [Bodo Moeller] 772459191Skris 772559191Skris *) Fix for bug in CRL encoding. The validity dates weren't being handled 772659191Skris as ASN1_TIME. 772759191Skris [Steve Henson] 772859191Skris 772959191Skris *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 773059191Skris [Steve Henson] 773159191Skris 773259191Skris *) New function BN_pseudo_rand(). 773359191Skris [Ulf M�ller] 773459191Skris 773559191Skris *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 773659191Skris bignum version of BN_from_montgomery() with the working code from 773759191Skris SSLeay 0.9.0 (the word based version is faster anyway), and clean up 773859191Skris the comments. 773959191Skris [Ulf M�ller] 774059191Skris 774159191Skris *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 774259191Skris made it impossible to use the same SSL_SESSION data structure in 774359191Skris SSL2 clients in multiple threads. 774459191Skris [Bodo Moeller] 774559191Skris 774659191Skris *) The return value of RAND_load_file() no longer counts bytes obtained 774759191Skris by stat(). RAND_load_file(..., -1) is new and uses the complete file 774859191Skris to seed the PRNG (previously an explicit byte count was required). 774959191Skris [Ulf M�ller, Bodo M�ller] 775059191Skris 775159191Skris *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 775259191Skris used (char *) instead of (void *) and had casts all over the place. 775359191Skris [Steve Henson] 775459191Skris 775559191Skris *) Make BN_generate_prime() return NULL on error if ret!=NULL. 775659191Skris [Ulf M�ller] 775759191Skris 775859191Skris *) Retain source code compatibility for BN_prime_checks macro: 775959191Skris BN_is_prime(..., BN_prime_checks, ...) now uses 776059191Skris BN_prime_checks_for_size to determine the appropriate number of 776159191Skris Rabin-Miller iterations. 776259191Skris [Ulf M�ller] 776359191Skris 776459191Skris *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 776559191Skris DH_CHECK_P_NOT_SAFE_PRIME. 776659191Skris (Check if this is true? OpenPGP calls them "strong".) 776759191Skris [Ulf M�ller] 776859191Skris 776959191Skris *) Merge the functionality of "dh" and "gendh" programs into a new program 777059191Skris "dhparam". The old programs are retained for now but will handle DH keys 777159191Skris (instead of parameters) in future. 777259191Skris [Steve Henson] 777359191Skris 777459191Skris *) Make the ciphers, s_server and s_client programs check the return values 777559191Skris when a new cipher list is set. 777659191Skris [Steve Henson] 777759191Skris 777859191Skris *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 777959191Skris ciphers. Before when the 56bit ciphers were enabled the sorting was 778059191Skris wrong. 778159191Skris 778259191Skris The syntax for the cipher sorting has been extended to support sorting by 778359191Skris cipher-strength (using the strength_bits hard coded in the tables). 778459191Skris The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 778559191Skris 778659191Skris Fix a bug in the cipher-command parser: when supplying a cipher command 778759191Skris string with an "undefined" symbol (neither command nor alphanumeric 778859191Skris [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 778959191Skris an error is flagged. 779059191Skris 779159191Skris Due to the strength-sorting extension, the code of the 779259191Skris ssl_create_cipher_list() function was completely rearranged. I hope that 779359191Skris the readability was also increased :-) 779459191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 779559191Skris 779659191Skris *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 779759191Skris for the first serial number and places 2 in the serial number file. This 779859191Skris avoids problems when the root CA is created with serial number zero and 779959191Skris the first user certificate has the same issuer name and serial number 780059191Skris as the root CA. 780159191Skris [Steve Henson] 780259191Skris 780359191Skris *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 780459191Skris the new code. Add documentation for this stuff. 780559191Skris [Steve Henson] 780659191Skris 780759191Skris *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 780859191Skris X509_*() to X509at_*() on the grounds that they don't handle X509 780959191Skris structures and behave in an analagous way to the X509v3 functions: 781059191Skris they shouldn't be called directly but wrapper functions should be used 781159191Skris instead. 781259191Skris 781359191Skris So we also now have some wrapper functions that call the X509at functions 781459191Skris when passed certificate requests. (TO DO: similar things can be done with 781559191Skris PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 781659191Skris things. Some of these need some d2i or i2d and print functionality 781759191Skris because they handle more complex structures.) 781859191Skris [Steve Henson] 781959191Skris 782059191Skris *) Add missing #ifndefs that caused missing symbols when building libssl 782159191Skris as a shared library without RSA. Use #ifndef NO_SSL2 instead of 782259191Skris NO_RSA in ssl/s2*.c. 782359191Skris [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 782459191Skris 782559191Skris *) Precautions against using the PRNG uninitialized: RAND_bytes() now 782659191Skris has a return value which indicates the quality of the random data 782759191Skris (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 782859191Skris error queue. New function RAND_pseudo_bytes() generates output that is 782959191Skris guaranteed to be unique but not unpredictable. RAND_add is like 783059191Skris RAND_seed, but takes an extra argument for an entropy estimate 783159191Skris (RAND_seed always assumes full entropy). 783259191Skris [Ulf M�ller] 783359191Skris 783459191Skris *) Do more iterations of Rabin-Miller probable prime test (specifically, 783559191Skris 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 783659191Skris instead of only 2 for all lengths; see BN_prime_checks_for_size definition 783759191Skris in crypto/bn/bn_prime.c for the complete table). This guarantees a 783859191Skris false-positive rate of at most 2^-80 for random input. 783959191Skris [Bodo Moeller] 784059191Skris 784159191Skris *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 784259191Skris [Bodo Moeller] 784359191Skris 784459191Skris *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 784559191Skris in the 0.9.5 release), this returns the chain 784659191Skris from an X509_CTX structure with a dup of the stack and all 784759191Skris the X509 reference counts upped: so the stack will exist 784859191Skris after X509_CTX_cleanup() has been called. Modify pkcs12.c 784959191Skris to use this. 785059191Skris 785159191Skris Also make SSL_SESSION_print() print out the verify return 785259191Skris code. 785359191Skris [Steve Henson] 785459191Skris 785559191Skris *) Add manpage for the pkcs12 command. Also change the default 785659191Skris behaviour so MAC iteration counts are used unless the new 785759191Skris -nomaciter option is used. This improves file security and 785859191Skris only older versions of MSIE (4.0 for example) need it. 785959191Skris [Steve Henson] 786059191Skris 786159191Skris *) Honor the no-xxx Configure options when creating .DEF files. 786259191Skris [Ulf M�ller] 786359191Skris 786459191Skris *) Add PKCS#10 attributes to field table: challengePassword, 786559191Skris unstructuredName and unstructuredAddress. These are taken from 786659191Skris draft PKCS#9 v2.0 but are compatible with v1.2 provided no 786759191Skris international characters are used. 786859191Skris 786959191Skris More changes to X509_ATTRIBUTE code: allow the setting of types 787059191Skris based on strings. Remove the 'loc' parameter when adding 787159191Skris attributes because these will be a SET OF encoding which is sorted 787259191Skris in ASN1 order. 787359191Skris [Steve Henson] 787459191Skris 787559191Skris *) Initial changes to the 'req' utility to allow request generation 787659191Skris automation. This will allow an application to just generate a template 787759191Skris file containing all the field values and have req construct the 787859191Skris request. 787959191Skris 788059191Skris Initial support for X509_ATTRIBUTE handling. Stacks of these are 788159191Skris used all over the place including certificate requests and PKCS#7 788259191Skris structures. They are currently handled manually where necessary with 788359191Skris some primitive wrappers for PKCS#7. The new functions behave in a 788459191Skris manner analogous to the X509 extension functions: they allow 788559191Skris attributes to be looked up by NID and added. 788659191Skris 788759191Skris Later something similar to the X509V3 code would be desirable to 788859191Skris automatically handle the encoding, decoding and printing of the 788959191Skris more complex types. The string types like challengePassword can 789059191Skris be handled by the string table functions. 789159191Skris 789259191Skris Also modified the multi byte string table handling. Now there is 789359191Skris a 'global mask' which masks out certain types. The table itself 789459191Skris can use the flag STABLE_NO_MASK to ignore the mask setting: this 789559191Skris is useful when for example there is only one permissible type 789659191Skris (as in countryName) and using the mask might result in no valid 789759191Skris types at all. 789859191Skris [Steve Henson] 789959191Skris 790059191Skris *) Clean up 'Finished' handling, and add functions SSL_get_finished and 790159191Skris SSL_get_peer_finished to allow applications to obtain the latest 790259191Skris Finished messages sent to the peer or expected from the peer, 790359191Skris respectively. (SSL_get_peer_finished is usually the Finished message 790459191Skris actually received from the peer, otherwise the protocol will be aborted.) 790559191Skris 790659191Skris As the Finished message are message digests of the complete handshake 790759191Skris (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 790859191Skris be used for external authentication procedures when the authentication 790959191Skris provided by SSL/TLS is not desired or is not enough. 791059191Skris [Bodo Moeller] 791159191Skris 791259191Skris *) Enhanced support for Alpha Linux is added. Now ./config checks if 791359191Skris the host supports BWX extension and if Compaq C is present on the 791459191Skris $PATH. Just exploiting of the BWX extension results in 20-30% 791559191Skris performance kick for some algorithms, e.g. DES and RC4 to mention 791659191Skris a couple. Compaq C in turn generates ~20% faster code for MD5 and 791759191Skris SHA1. 791859191Skris [Andy Polyakov] 791959191Skris 792059191Skris *) Add support for MS "fast SGC". This is arguably a violation of the 792159191Skris SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 792259191Skris weak crypto and after checking the certificate is SGC a second one 792359191Skris with strong crypto. MS SGC stops the first handshake after receiving 792459191Skris the server certificate message and sends a second client hello. Since 792559191Skris a server will typically do all the time consuming operations before 792659191Skris expecting any further messages from the client (server key exchange 792759191Skris is the most expensive) there is little difference between the two. 792859191Skris 792959191Skris To get OpenSSL to support MS SGC we have to permit a second client 793059191Skris hello message after we have sent server done. In addition we have to 793159191Skris reset the MAC if we do get this second client hello. 793259191Skris [Steve Henson] 793359191Skris 793459191Skris *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 793559191Skris if a DER encoded private key is RSA or DSA traditional format. Changed 793659191Skris d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 793759191Skris format DER encoded private key. Newer code should use PKCS#8 format which 793859191Skris has the key type encoded in the ASN1 structure. Added DER private key 793959191Skris support to pkcs8 application. 794059191Skris [Steve Henson] 794159191Skris 794259191Skris *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 794359191Skris ciphersuites has been selected (as required by the SSL 3/TLS 1 794459191Skris specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 794559191Skris is set, we interpret this as a request to violate the specification 794659191Skris (the worst that can happen is a handshake failure, and 'correct' 794759191Skris behaviour would result in a handshake failure anyway). 794859191Skris [Bodo Moeller] 794959191Skris 795059191Skris *) In SSL_CTX_add_session, take into account that there might be multiple 795159191Skris SSL_SESSION structures with the same session ID (e.g. when two threads 795259191Skris concurrently obtain them from an external cache). 795359191Skris The internal cache can handle only one SSL_SESSION with a given ID, 795459191Skris so if there's a conflict, we now throw out the old one to achieve 795559191Skris consistency. 795659191Skris [Bodo Moeller] 795759191Skris 795859191Skris *) Add OIDs for idea and blowfish in CBC mode. This will allow both 795959191Skris to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 796059191Skris some routines that use cipher OIDs: some ciphers do not have OIDs 796159191Skris defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 796259191Skris example. 796359191Skris [Steve Henson] 796459191Skris 796559191Skris *) Simplify the trust setting structure and code. Now we just have 796659191Skris two sequences of OIDs for trusted and rejected settings. These will 796759191Skris typically have values the same as the extended key usage extension 796859191Skris and any application specific purposes. 796959191Skris 797059191Skris The trust checking code now has a default behaviour: it will just 797159191Skris check for an object with the same NID as the passed id. Functions can 797259191Skris be provided to override either the default behaviour or the behaviour 797359191Skris for a given id. SSL client, server and email already have functions 797459191Skris in place for compatibility: they check the NID and also return "trusted" 797559191Skris if the certificate is self signed. 797659191Skris [Steve Henson] 797759191Skris 797859191Skris *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 797959191Skris traditional format into an EVP_PKEY structure. 798059191Skris [Steve Henson] 798159191Skris 798259191Skris *) Add a password callback function PEM_cb() which either prompts for 798359191Skris a password if usr_data is NULL or otherwise assumes it is a null 798459191Skris terminated password. Allow passwords to be passed on command line 798559191Skris environment or config files in a few more utilities. 798659191Skris [Steve Henson] 798759191Skris 798859191Skris *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 798959191Skris keys. Add some short names for PKCS#8 PBE algorithms and allow them 799059191Skris to be specified on the command line for the pkcs8 and pkcs12 utilities. 799159191Skris Update documentation. 799259191Skris [Steve Henson] 799359191Skris 799459191Skris *) Support for ASN1 "NULL" type. This could be handled before by using 799559191Skris ASN1_TYPE but there wasn't any function that would try to read a NULL 799659191Skris and produce an error if it couldn't. For compatibility we also have 799759191Skris ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 799859191Skris don't allocate anything because they don't need to. 799959191Skris [Steve Henson] 800059191Skris 800159191Skris *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 800259191Skris for details. 800359191Skris [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 800459191Skris 800559191Skris *) Rebuild of the memory allocation routines used by OpenSSL code and 800659191Skris possibly others as well. The purpose is to make an interface that 800759191Skris provide hooks so anyone can build a separate set of allocation and 800859191Skris deallocation routines to be used by OpenSSL, for example memory 800959191Skris pool implementations, or something else, which was previously hard 801059191Skris since Malloc(), Realloc() and Free() were defined as macros having 801159191Skris the values malloc, realloc and free, respectively (except for Win32 801259191Skris compilations). The same is provided for memory debugging code. 801359191Skris OpenSSL already comes with functionality to find memory leaks, but 801459191Skris this gives people a chance to debug other memory problems. 801559191Skris 801659191Skris With these changes, a new set of functions and macros have appeared: 801759191Skris 801868651Skris CRYPTO_set_mem_debug_functions() [F] 801968651Skris CRYPTO_get_mem_debug_functions() [F] 802068651Skris CRYPTO_dbg_set_options() [F] 802168651Skris CRYPTO_dbg_get_options() [F] 802268651Skris CRYPTO_malloc_debug_init() [M] 802359191Skris 802459191Skris The memory debug functions are NULL by default, unless the library 802559191Skris is compiled with CRYPTO_MDEBUG or friends is defined. If someone 802659191Skris wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 802759191Skris gives the standard debugging functions that come with OpenSSL) or 802859191Skris CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 802959191Skris provided by the library user) must be used. When the standard 803059191Skris debugging functions are used, CRYPTO_dbg_set_options can be used to 803159191Skris request additional information: 803259191Skris CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 803359191Skris the CRYPTO_MDEBUG_xxx macro when compiling the library. 803459191Skris 803559191Skris Also, things like CRYPTO_set_mem_functions will always give the 803659191Skris expected result (the new set of functions is used for allocation 803759191Skris and deallocation) at all times, regardless of platform and compiler 803859191Skris options. 803959191Skris 804059191Skris To finish it up, some functions that were never use in any other 804159191Skris way than through macros have a new API and new semantic: 804259191Skris 804359191Skris CRYPTO_dbg_malloc() 804459191Skris CRYPTO_dbg_realloc() 804559191Skris CRYPTO_dbg_free() 804659191Skris 804759191Skris All macros of value have retained their old syntax. 804859191Skris [Richard Levitte and Bodo Moeller] 804959191Skris 805059191Skris *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 805159191Skris ordering of SMIMECapabilities wasn't in "strength order" and there 805259191Skris was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 805359191Skris algorithm. 805459191Skris [Steve Henson] 805559191Skris 805659191Skris *) Some ASN1 types with illegal zero length encoding (INTEGER, 805759191Skris ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 805859191Skris [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 805959191Skris 806059191Skris *) Merge in my S/MIME library for OpenSSL. This provides a simple 806159191Skris S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 806259191Skris functionality to handle multipart/signed properly) and a utility 806359191Skris called 'smime' to call all this stuff. This is based on code I 806459191Skris originally wrote for Celo who have kindly allowed it to be 806559191Skris included in OpenSSL. 806659191Skris [Steve Henson] 806759191Skris 806859191Skris *) Add variants des_set_key_checked and des_set_key_unchecked of 806959191Skris des_set_key (aka des_key_sched). Global variable des_check_key 807059191Skris decides which of these is called by des_set_key; this way 807159191Skris des_check_key behaves as it always did, but applications and 807259191Skris the library itself, which was buggy for des_check_key == 1, 807359191Skris have a cleaner way to pick the version they need. 807459191Skris [Bodo Moeller] 807559191Skris 807659191Skris *) New function PKCS12_newpass() which changes the password of a 807759191Skris PKCS12 structure. 807859191Skris [Steve Henson] 807959191Skris 808059191Skris *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 808159191Skris dynamic mix. In both cases the ids can be used as an index into the 808259191Skris table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 808359191Skris functions so they accept a list of the field values and the 808459191Skris application doesn't need to directly manipulate the X509_TRUST 808559191Skris structure. 808659191Skris [Steve Henson] 808759191Skris 808859191Skris *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 808959191Skris need initialising. 809059191Skris [Steve Henson] 809159191Skris 809259191Skris *) Modify the way the V3 extension code looks up extensions. This now 809359191Skris works in a similar way to the object code: we have some "standard" 809459191Skris extensions in a static table which is searched with OBJ_bsearch() 809559191Skris and the application can add dynamic ones if needed. The file 809659191Skris crypto/x509v3/ext_dat.h now has the info: this file needs to be 809759191Skris updated whenever a new extension is added to the core code and kept 809859191Skris in ext_nid order. There is a simple program 'tabtest.c' which checks 809959191Skris this. New extensions are not added too often so this file can readily 810059191Skris be maintained manually. 810159191Skris 810259191Skris There are two big advantages in doing things this way. The extensions 810359191Skris can be looked up immediately and no longer need to be "added" using 810459191Skris X509V3_add_standard_extensions(): this function now does nothing. 810559191Skris [Side note: I get *lots* of email saying the extension code doesn't 810659191Skris work because people forget to call this function] 810759191Skris Also no dynamic allocation is done unless new extensions are added: 810859191Skris so if we don't add custom extensions there is no need to call 810959191Skris X509V3_EXT_cleanup(). 811059191Skris [Steve Henson] 811159191Skris 811259191Skris *) Modify enc utility's salting as follows: make salting the default. Add a 811359191Skris magic header, so unsalted files fail gracefully instead of just decrypting 811459191Skris to garbage. This is because not salting is a big security hole, so people 811559191Skris should be discouraged from doing it. 811659191Skris [Ben Laurie] 811759191Skris 811859191Skris *) Fixes and enhancements to the 'x509' utility. It allowed a message 811959191Skris digest to be passed on the command line but it only used this 812059191Skris parameter when signing a certificate. Modified so all relevant 812159191Skris operations are affected by the digest parameter including the 812259191Skris -fingerprint and -x509toreq options. Also -x509toreq choked if a 812359191Skris DSA key was used because it didn't fix the digest. 812459191Skris [Steve Henson] 812559191Skris 812659191Skris *) Initial certificate chain verify code. Currently tests the untrusted 812759191Skris certificates for consistency with the verify purpose (which is set 812859191Skris when the X509_STORE_CTX structure is set up) and checks the pathlength. 812959191Skris 813059191Skris There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 813159191Skris this is because it will reject chains with invalid extensions whereas 813259191Skris every previous version of OpenSSL and SSLeay made no checks at all. 813359191Skris 813459191Skris Trust code: checks the root CA for the relevant trust settings. Trust 813559191Skris settings have an initial value consistent with the verify purpose: e.g. 813659191Skris if the verify purpose is for SSL client use it expects the CA to be 813759191Skris trusted for SSL client use. However the default value can be changed to 813859191Skris permit custom trust settings: one example of this would be to only trust 813959191Skris certificates from a specific "secure" set of CAs. 814059191Skris 814159191Skris Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 814259191Skris which should be used for version portability: especially since the 814359191Skris verify structure is likely to change more often now. 814459191Skris 814559191Skris SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 814659191Skris to set them. If not set then assume SSL clients will verify SSL servers 814759191Skris and vice versa. 814859191Skris 814959191Skris Two new options to the verify program: -untrusted allows a set of 815059191Skris untrusted certificates to be passed in and -purpose which sets the 815159191Skris intended purpose of the certificate. If a purpose is set then the 815259191Skris new chain verify code is used to check extension consistency. 815359191Skris [Steve Henson] 815459191Skris 815559191Skris *) Support for the authority information access extension. 815659191Skris [Steve Henson] 815759191Skris 815859191Skris *) Modify RSA and DSA PEM read routines to transparently handle 815959191Skris PKCS#8 format private keys. New *_PUBKEY_* functions that handle 816059191Skris public keys in a format compatible with certificate 816159191Skris SubjectPublicKeyInfo structures. Unfortunately there were already 816259191Skris functions called *_PublicKey_* which used various odd formats so 816359191Skris these are retained for compatibility: however the DSA variants were 816459191Skris never in a public release so they have been deleted. Changed dsa/rsa 816559191Skris utilities to handle the new format: note no releases ever handled public 816659191Skris keys so we should be OK. 816759191Skris 816859191Skris The primary motivation for this change is to avoid the same fiasco 816959191Skris that dogs private keys: there are several incompatible private key 817059191Skris formats some of which are standard and some OpenSSL specific and 817159191Skris require various evil hacks to allow partial transparent handling and 817259191Skris even then it doesn't work with DER formats. Given the option anything 817359191Skris other than PKCS#8 should be dumped: but the other formats have to 817459191Skris stay in the name of compatibility. 817559191Skris 817659191Skris With public keys and the benefit of hindsight one standard format 817759191Skris is used which works with EVP_PKEY, RSA or DSA structures: though 817859191Skris it clearly returns an error if you try to read the wrong kind of key. 817959191Skris 818059191Skris Added a -pubkey option to the 'x509' utility to output the public key. 818159191Skris Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 818259191Skris (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 818359191Skris EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 818459191Skris that do the same as the EVP_PKEY_assign_*() except they up the 818559191Skris reference count of the added key (they don't "swallow" the 818659191Skris supplied key). 818759191Skris [Steve Henson] 818859191Skris 818959191Skris *) Fixes to crypto/x509/by_file.c the code to read in certificates and 819059191Skris CRLs would fail if the file contained no certificates or no CRLs: 819159191Skris added a new function to read in both types and return the number 819259191Skris read: this means that if none are read it will be an error. The 819359191Skris DER versions of the certificate and CRL reader would always fail 819459191Skris because it isn't possible to mix certificates and CRLs in DER format 819559191Skris without choking one or the other routine. Changed this to just read 819659191Skris a certificate: this is the best we can do. Also modified the code 819759191Skris in apps/verify.c to take notice of return codes: it was previously 819859191Skris attempting to read in certificates from NULL pointers and ignoring 819959191Skris any errors: this is one reason why the cert and CRL reader seemed 820059191Skris to work. It doesn't check return codes from the default certificate 820159191Skris routines: these may well fail if the certificates aren't installed. 820259191Skris [Steve Henson] 820359191Skris 820459191Skris *) Code to support otherName option in GeneralName. 820559191Skris [Steve Henson] 820659191Skris 820759191Skris *) First update to verify code. Change the verify utility 820859191Skris so it warns if it is passed a self signed certificate: 820959191Skris for consistency with the normal behaviour. X509_verify 821059191Skris has been modified to it will now verify a self signed 821159191Skris certificate if *exactly* the same certificate appears 821259191Skris in the store: it was previously impossible to trust a 821359191Skris single self signed certificate. This means that: 821459191Skris openssl verify ss.pem 821559191Skris now gives a warning about a self signed certificate but 821659191Skris openssl verify -CAfile ss.pem ss.pem 821759191Skris is OK. 821859191Skris [Steve Henson] 821959191Skris 822059191Skris *) For servers, store verify_result in SSL_SESSION data structure 822159191Skris (and add it to external session representation). 822259191Skris This is needed when client certificate verifications fails, 822359191Skris but an application-provided verification callback (set by 822459191Skris SSL_CTX_set_cert_verify_callback) allows accepting the session 822559191Skris anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 822659191Skris but returns 1): When the session is reused, we have to set 822759191Skris ssl->verify_result to the appropriate error code to avoid 822859191Skris security holes. 822959191Skris [Bodo Moeller, problem pointed out by Lutz Jaenicke] 823059191Skris 823159191Skris *) Fix a bug in the new PKCS#7 code: it didn't consider the 823259191Skris case in PKCS7_dataInit() where the signed PKCS7 structure 823359191Skris didn't contain any existing data because it was being created. 823459191Skris [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 823559191Skris 823659191Skris *) Add a salt to the key derivation routines in enc.c. This 823759191Skris forms the first 8 bytes of the encrypted file. Also add a 823859191Skris -S option to allow a salt to be input on the command line. 823959191Skris [Steve Henson] 824059191Skris 824159191Skris *) New function X509_cmp(). Oddly enough there wasn't a function 824259191Skris to compare two certificates. We do this by working out the SHA1 824359191Skris hash and comparing that. X509_cmp() will be needed by the trust 824459191Skris code. 824559191Skris [Steve Henson] 824659191Skris 824759191Skris *) SSL_get1_session() is like SSL_get_session(), but increments 824859191Skris the reference count in the SSL_SESSION returned. 824959191Skris [Geoff Thorpe <geoff@eu.c2.net>] 825059191Skris 825159191Skris *) Fix for 'req': it was adding a null to request attributes. 825259191Skris Also change the X509_LOOKUP and X509_INFO code to handle 825359191Skris certificate auxiliary information. 825459191Skris [Steve Henson] 825559191Skris 825659191Skris *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 825759191Skris the 'enc' command. 825859191Skris [Steve Henson] 825959191Skris 826059191Skris *) Add the possibility to add extra information to the memory leak 826159191Skris detecting output, to form tracebacks, showing from where each 826259191Skris allocation was originated: CRYPTO_push_info("constant string") adds 826359191Skris the string plus current file name and line number to a per-thread 826459191Skris stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 826559191Skris is like calling CYRPTO_pop_info() until the stack is empty. 826659191Skris Also updated memory leak detection code to be multi-thread-safe. 826759191Skris [Richard Levitte] 826859191Skris 826959191Skris *) Add options -text and -noout to pkcs7 utility and delete the 827059191Skris encryption options which never did anything. Update docs. 827159191Skris [Steve Henson] 827259191Skris 827359191Skris *) Add options to some of the utilities to allow the pass phrase 827459191Skris to be included on either the command line (not recommended on 827559191Skris OSes like Unix) or read from the environment. Update the 827659191Skris manpages and fix a few bugs. 827759191Skris [Steve Henson] 827859191Skris 827959191Skris *) Add a few manpages for some of the openssl commands. 828059191Skris [Steve Henson] 828159191Skris 828259191Skris *) Fix the -revoke option in ca. It was freeing up memory twice, 828359191Skris leaking and not finding already revoked certificates. 828459191Skris [Steve Henson] 828559191Skris 828659191Skris *) Extensive changes to support certificate auxiliary information. 828759191Skris This involves the use of X509_CERT_AUX structure and X509_AUX 828859191Skris functions. An X509_AUX function such as PEM_read_X509_AUX() 828959191Skris can still read in a certificate file in the usual way but it 829059191Skris will also read in any additional "auxiliary information". By 829159191Skris doing things this way a fair degree of compatibility can be 829259191Skris retained: existing certificates can have this information added 829359191Skris using the new 'x509' options. 829459191Skris 829559191Skris Current auxiliary information includes an "alias" and some trust 829659191Skris settings. The trust settings will ultimately be used in enhanced 829759191Skris certificate chain verification routines: currently a certificate 829859191Skris can only be trusted if it is self signed and then it is trusted 829959191Skris for all purposes. 830059191Skris [Steve Henson] 830159191Skris 830259191Skris *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 830359191Skris The problem was that one of the replacement routines had not been working 830459191Skris since SSLeay releases. For now the offending routine has been replaced 830559191Skris with non-optimised assembler. Even so, this now gives around 95% 830659191Skris performance improvement for 1024 bit RSA signs. 830759191Skris [Mark Cox] 830859191Skris 830959191Skris *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 831059191Skris handling. Most clients have the effective key size in bits equal to 831159191Skris the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 831259191Skris A few however don't do this and instead use the size of the decrypted key 831359191Skris to determine the RC2 key length and the AlgorithmIdentifier to determine 831459191Skris the effective key length. In this case the effective key length can still 831559191Skris be 40 bits but the key length can be 168 bits for example. This is fixed 831659191Skris by manually forcing an RC2 key into the EVP_PKEY structure because the 831759191Skris EVP code can't currently handle unusual RC2 key sizes: it always assumes 831859191Skris the key length and effective key length are equal. 831959191Skris [Steve Henson] 832059191Skris 832159191Skris *) Add a bunch of functions that should simplify the creation of 832259191Skris X509_NAME structures. Now you should be able to do: 832359191Skris X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 832459191Skris and have it automatically work out the correct field type and fill in 832559191Skris the structures. The more adventurous can try: 832659191Skris X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 832759191Skris and it will (hopefully) work out the correct multibyte encoding. 832859191Skris [Steve Henson] 832959191Skris 833059191Skris *) Change the 'req' utility to use the new field handling and multibyte 833159191Skris copy routines. Before the DN field creation was handled in an ad hoc 833259191Skris way in req, ca, and x509 which was rather broken and didn't support 833359191Skris BMPStrings or UTF8Strings. Since some software doesn't implement 833459191Skris BMPStrings or UTF8Strings yet, they can be enabled using the config file 833559191Skris using the dirstring_type option. See the new comment in the default 833659191Skris openssl.cnf for more info. 833759191Skris [Steve Henson] 833859191Skris 833959191Skris *) Make crypto/rand/md_rand.c more robust: 834059191Skris - Assure unique random numbers after fork(). 834159191Skris - Make sure that concurrent threads access the global counter and 834259191Skris md serializably so that we never lose entropy in them 834359191Skris or use exactly the same state in multiple threads. 834459191Skris Access to the large state is not always serializable because 834559191Skris the additional locking could be a performance killer, and 834659191Skris md should be large enough anyway. 834759191Skris [Bodo Moeller] 834859191Skris 834959191Skris *) New file apps/app_rand.c with commonly needed functionality 835059191Skris for handling the random seed file. 835159191Skris 835259191Skris Use the random seed file in some applications that previously did not: 835359191Skris ca, 835459191Skris dsaparam -genkey (which also ignored its '-rand' option), 835559191Skris s_client, 835659191Skris s_server, 835759191Skris x509 (when signing). 835859191Skris Except on systems with /dev/urandom, it is crucial to have a random 835959191Skris seed file at least for key creation, DSA signing, and for DH exchanges; 836059191Skris for RSA signatures we could do without one. 836159191Skris 836259191Skris gendh and gendsa (unlike genrsa) used to read only the first byte 836359191Skris of each file listed in the '-rand' option. The function as previously 836459191Skris found in genrsa is now in app_rand.c and is used by all programs 836559191Skris that support '-rand'. 836659191Skris [Bodo Moeller] 836759191Skris 836859191Skris *) In RAND_write_file, use mode 0600 for creating files; 836959191Skris don't just chmod when it may be too late. 837059191Skris [Bodo Moeller] 837159191Skris 837259191Skris *) Report an error from X509_STORE_load_locations 837359191Skris when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 837459191Skris [Bill Perry] 837559191Skris 837659191Skris *) New function ASN1_mbstring_copy() this copies a string in either 837759191Skris ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 837859191Skris into an ASN1_STRING type. A mask of permissible types is passed 837959191Skris and it chooses the "minimal" type to use or an error if not type 838059191Skris is suitable. 838159191Skris [Steve Henson] 838259191Skris 838359191Skris *) Add function equivalents to the various macros in asn1.h. The old 838459191Skris macros are retained with an M_ prefix. Code inside the library can 838559191Skris use the M_ macros. External code (including the openssl utility) 838659191Skris should *NOT* in order to be "shared library friendly". 838759191Skris [Steve Henson] 838859191Skris 838959191Skris *) Add various functions that can check a certificate's extensions 839059191Skris to see if it usable for various purposes such as SSL client, 839159191Skris server or S/MIME and CAs of these types. This is currently 839259191Skris VERY EXPERIMENTAL but will ultimately be used for certificate chain 839359191Skris verification. Also added a -purpose flag to x509 utility to 839459191Skris print out all the purposes. 839559191Skris [Steve Henson] 839659191Skris 839759191Skris *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 839859191Skris functions. 839959191Skris [Steve Henson] 840059191Skris 840159191Skris *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 840259191Skris for, obtain and decode and extension and obtain its critical flag. 840359191Skris This allows all the necessary extension code to be handled in a 840459191Skris single function call. 840559191Skris [Steve Henson] 840659191Skris 840759191Skris *) RC4 tune-up featuring 30-40% performance improvement on most RISC 840859191Skris platforms. See crypto/rc4/rc4_enc.c for further details. 840959191Skris [Andy Polyakov] 841059191Skris 841159191Skris *) New -noout option to asn1parse. This causes no output to be produced 841259191Skris its main use is when combined with -strparse and -out to extract data 841359191Skris from a file (which may not be in ASN.1 format). 841459191Skris [Steve Henson] 841559191Skris 841659191Skris *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 841759191Skris when producing the local key id. 841859191Skris [Richard Levitte <levitte@stacken.kth.se>] 841959191Skris 842059191Skris *) New option -dhparam in s_server. This allows a DH parameter file to be 842159191Skris stated explicitly. If it is not stated then it tries the first server 842259191Skris certificate file. The previous behaviour hard coded the filename 842359191Skris "server.pem". 842459191Skris [Steve Henson] 842559191Skris 842659191Skris *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 842759191Skris a public key to be input or output. For example: 842859191Skris openssl rsa -in key.pem -pubout -out pubkey.pem 842959191Skris Also added necessary DSA public key functions to handle this. 843059191Skris [Steve Henson] 843159191Skris 843259191Skris *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 843359191Skris in the message. This was handled by allowing 843459191Skris X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 843559191Skris [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 843659191Skris 843759191Skris *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 843859191Skris to the end of the strings whereas this didn't. This would cause problems 843959191Skris if strings read with d2i_ASN1_bytes() were later modified. 844059191Skris [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 844159191Skris 844259191Skris *) Fix for base64 decode bug. When a base64 bio reads only one line of 844359191Skris data and it contains EOF it will end up returning an error. This is 844459191Skris caused by input 46 bytes long. The cause is due to the way base64 844559191Skris BIOs find the start of base64 encoded data. They do this by trying a 844659191Skris trial decode on each line until they find one that works. When they 844759191Skris do a flag is set and it starts again knowing it can pass all the 844859191Skris data directly through the decoder. Unfortunately it doesn't reset 844959191Skris the context it uses. This means that if EOF is reached an attempt 845059191Skris is made to pass two EOFs through the context and this causes the 845159191Skris resulting error. This can also cause other problems as well. As is 845259191Skris usual with these problems it takes *ages* to find and the fix is 845359191Skris trivial: move one line. 845459191Skris [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 845559191Skris 845659191Skris *) Ugly workaround to get s_client and s_server working under Windows. The 845759191Skris old code wouldn't work because it needed to select() on sockets and the 845859191Skris tty (for keypresses and to see if data could be written). Win32 only 845959191Skris supports select() on sockets so we select() with a 1s timeout on the 846059191Skris sockets and then see if any characters are waiting to be read, if none 846159191Skris are present then we retry, we also assume we can always write data to 846259191Skris the tty. This isn't nice because the code then blocks until we've 846359191Skris received a complete line of data and it is effectively polling the 846459191Skris keyboard at 1s intervals: however it's quite a bit better than not 846559191Skris working at all :-) A dedicated Windows application might handle this 846659191Skris with an event loop for example. 846759191Skris [Steve Henson] 846859191Skris 846959191Skris *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 847059191Skris and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 847159191Skris will be called when RSA_sign() and RSA_verify() are used. This is useful 847259191Skris if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 847359191Skris For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 847459191Skris should *not* be used: RSA_sign() and RSA_verify() must be used instead. 847559191Skris This necessitated the support of an extra signature type NID_md5_sha1 847659191Skris for SSL signatures and modifications to the SSL library to use it instead 847759191Skris of calling RSA_public_decrypt() and RSA_private_encrypt(). 847859191Skris [Steve Henson] 847959191Skris 848059191Skris *) Add new -verify -CAfile and -CApath options to the crl program, these 848159191Skris will lookup a CRL issuers certificate and verify the signature in a 848259191Skris similar way to the verify program. Tidy up the crl program so it 848359191Skris no longer accesses structures directly. Make the ASN1 CRL parsing a bit 848459191Skris less strict. It will now permit CRL extensions even if it is not 848559191Skris a V2 CRL: this will allow it to tolerate some broken CRLs. 848659191Skris [Steve Henson] 848759191Skris 848859191Skris *) Initialize all non-automatic variables each time one of the openssl 848959191Skris sub-programs is started (this is necessary as they may be started 849059191Skris multiple times from the "OpenSSL>" prompt). 849159191Skris [Lennart Bang, Bodo Moeller] 849259191Skris 849359191Skris *) Preliminary compilation option RSA_NULL which disables RSA crypto without 849459191Skris removing all other RSA functionality (this is what NO_RSA does). This 849559191Skris is so (for example) those in the US can disable those operations covered 849659191Skris by the RSA patent while allowing storage and parsing of RSA keys and RSA 849759191Skris key generation. 849859191Skris [Steve Henson] 849959191Skris 850059191Skris *) Non-copying interface to BIO pairs. 850159191Skris (still largely untested) 850259191Skris [Bodo Moeller] 850359191Skris 850459191Skris *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 850559191Skris ASCII string. This was handled independently in various places before. 850659191Skris [Steve Henson] 850759191Skris 850859191Skris *) New functions UTF8_getc() and UTF8_putc() that parse and generate 850959191Skris UTF8 strings a character at a time. 851059191Skris [Steve Henson] 851159191Skris 851259191Skris *) Use client_version from client hello to select the protocol 851359191Skris (s23_srvr.c) and for RSA client key exchange verification 851459191Skris (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 851559191Skris [Bodo Moeller] 851659191Skris 851759191Skris *) Add various utility functions to handle SPKACs, these were previously 851859191Skris handled by poking round in the structure internals. Added new function 851959191Skris NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 852059191Skris print, verify and generate SPKACs. Based on an original idea from 852159191Skris Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 852259191Skris [Steve Henson] 852359191Skris 852459191Skris *) RIPEMD160 is operational on all platforms and is back in 'make test'. 852559191Skris [Andy Polyakov] 852659191Skris 852759191Skris *) Allow the config file extension section to be overwritten on the 852859191Skris command line. Based on an original idea from Massimiliano Pala 852959191Skris <madwolf@comune.modena.it>. The new option is called -extensions 853059191Skris and can be applied to ca, req and x509. Also -reqexts to override 853159191Skris the request extensions in req and -crlexts to override the crl extensions 853259191Skris in ca. 853359191Skris [Steve Henson] 853459191Skris 853559191Skris *) Add new feature to the SPKAC handling in ca. Now you can include 853659191Skris the same field multiple times by preceding it by "XXXX." for example: 853759191Skris 1.OU="Unit name 1" 853859191Skris 2.OU="Unit name 2" 853959191Skris this is the same syntax as used in the req config file. 854059191Skris [Steve Henson] 854159191Skris 854259191Skris *) Allow certificate extensions to be added to certificate requests. These 854359191Skris are specified in a 'req_extensions' option of the req section of the 854459191Skris config file. They can be printed out with the -text option to req but 854559191Skris are otherwise ignored at present. 854659191Skris [Steve Henson] 854759191Skris 854859191Skris *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 854959191Skris data read consists of only the final block it would not decrypted because 855059191Skris EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 855159191Skris A misplaced 'break' also meant the decrypted final block might not be 855259191Skris copied until the next read. 855359191Skris [Steve Henson] 855459191Skris 855559191Skris *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 855659191Skris a few extra parameters to the DH structure: these will be useful if 855759191Skris for example we want the value of 'q' or implement X9.42 DH. 855859191Skris [Steve Henson] 855959191Skris 856059191Skris *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 856159191Skris provides hooks that allow the default DSA functions or functions on a 856259191Skris "per key" basis to be replaced. This allows hardware acceleration and 856359191Skris hardware key storage to be handled without major modification to the 856459191Skris library. Also added low level modexp hooks and CRYPTO_EX structure and 856559191Skris associated functions. 856659191Skris [Steve Henson] 856759191Skris 856859191Skris *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 856959191Skris as "read only": it can't be written to and the buffer it points to will 857059191Skris not be freed. Reading from a read only BIO is much more efficient than 857159191Skris a normal memory BIO. This was added because there are several times when 857259191Skris an area of memory needs to be read from a BIO. The previous method was 857359191Skris to create a memory BIO and write the data to it, this results in two 857459191Skris copies of the data and an O(n^2) reading algorithm. There is a new 857559191Skris function BIO_new_mem_buf() which creates a read only memory BIO from 857659191Skris an area of memory. Also modified the PKCS#7 routines to use read only 857759191Skris memory BIOs. 857859191Skris [Steve Henson] 857959191Skris 858059191Skris *) Bugfix: ssl23_get_client_hello did not work properly when called in 858159191Skris state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 858259191Skris a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 858359191Skris but a retry condition occured while trying to read the rest. 858459191Skris [Bodo Moeller] 858559191Skris 858659191Skris *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 858759191Skris NID_pkcs7_encrypted by default: this was wrong since this should almost 858859191Skris always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 858959191Skris the encrypted data type: this is a more sensible place to put it and it 859059191Skris allows the PKCS#12 code to be tidied up that duplicated this 859159191Skris functionality. 859259191Skris [Steve Henson] 859359191Skris 859459191Skris *) Changed obj_dat.pl script so it takes its input and output files on 859559191Skris the command line. This should avoid shell escape redirection problems 859659191Skris under Win32. 859759191Skris [Steve Henson] 859859191Skris 859959191Skris *) Initial support for certificate extension requests, these are included 860059191Skris in things like Xenroll certificate requests. Included functions to allow 860159191Skris extensions to be obtained and added. 860259191Skris [Steve Henson] 860359191Skris 860459191Skris *) -crlf option to s_client and s_server for sending newlines as 860559191Skris CRLF (as required by many protocols). 860659191Skris [Bodo Moeller] 860759191Skris 860855714Skris Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 860955714Skris 861055714Skris *) Install libRSAglue.a when OpenSSL is built with RSAref. 861155714Skris [Ralf S. Engelschall] 861255714Skris 861355714Skris *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 861455714Skris [Andrija Antonijevic <TheAntony2@bigfoot.com>] 861555714Skris 861655714Skris *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 861755714Skris program. 861855714Skris [Steve Henson] 861955714Skris 862055714Skris *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 862155714Skris DH parameters/keys (q is lost during that conversion, but the resulting 862255714Skris DH parameters contain its length). 862355714Skris 862455714Skris For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 862555714Skris much faster than DH_generate_parameters (which creates parameters 862655714Skris where p = 2*q + 1), and also the smaller q makes DH computations 862755714Skris much more efficient (160-bit exponentiation instead of 1024-bit 862855714Skris exponentiation); so this provides a convenient way to support DHE 862955714Skris ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 863055714Skris utter importance to use 863155714Skris SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 863255714Skris or 863355714Skris SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 863455714Skris when such DH parameters are used, because otherwise small subgroup 863555714Skris attacks may become possible! 863655714Skris [Bodo Moeller] 863755714Skris 863855714Skris *) Avoid memory leak in i2d_DHparams. 863955714Skris [Bodo Moeller] 864055714Skris 864155714Skris *) Allow the -k option to be used more than once in the enc program: 864255714Skris this allows the same encrypted message to be read by multiple recipients. 864355714Skris [Steve Henson] 864455714Skris 864555714Skris *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 864655714Skris an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 864755714Skris it will always use the numerical form of the OID, even if it has a short 864855714Skris or long name. 864955714Skris [Steve Henson] 865055714Skris 865155714Skris *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 865255714Skris method only got called if p,q,dmp1,dmq1,iqmp components were present, 865355714Skris otherwise bn_mod_exp was called. In the case of hardware keys for example 865455714Skris no private key components need be present and it might store extra data 865559191Skris in the RSA structure, which cannot be accessed from bn_mod_exp. 865659191Skris By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 865759191Skris private key operations. 865855714Skris [Steve Henson] 865955714Skris 866055714Skris *) Added support for SPARC Linux. 866155714Skris [Andy Polyakov] 866255714Skris 866355714Skris *) pem_password_cb function type incompatibly changed from 866455714Skris typedef int pem_password_cb(char *buf, int size, int rwflag); 866555714Skris to 866655714Skris ....(char *buf, int size, int rwflag, void *userdata); 866755714Skris so that applications can pass data to their callbacks: 866855714Skris The PEM[_ASN1]_{read,write}... functions and macros now take an 866955714Skris additional void * argument, which is just handed through whenever 867055714Skris the password callback is called. 867159191Skris [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 867255714Skris 867355714Skris New function SSL_CTX_set_default_passwd_cb_userdata. 867455714Skris 867555714Skris Compatibility note: As many C implementations push function arguments 867655714Skris onto the stack in reverse order, the new library version is likely to 867755714Skris interoperate with programs that have been compiled with the old 867855714Skris pem_password_cb definition (PEM_whatever takes some data that 867955714Skris happens to be on the stack as its last argument, and the callback 868055714Skris just ignores this garbage); but there is no guarantee whatsoever that 868155714Skris this will work. 868255714Skris 868355714Skris *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 868455714Skris (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 868555714Skris problems not only on Windows, but also on some Unix platforms. 868655714Skris To avoid problematic command lines, these definitions are now in an 868755714Skris auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 868855714Skris for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 868955714Skris [Bodo Moeller] 869055714Skris 869155714Skris *) MIPS III/IV assembler module is reimplemented. 869255714Skris [Andy Polyakov] 869355714Skris 869455714Skris *) More DES library cleanups: remove references to srand/rand and 869555714Skris delete an unused file. 869655714Skris [Ulf M�ller] 869755714Skris 869855714Skris *) Add support for the the free Netwide assembler (NASM) under Win32, 869955714Skris since not many people have MASM (ml) and it can be hard to obtain. 870055714Skris This is currently experimental but it seems to work OK and pass all 870155714Skris the tests. Check out INSTALL.W32 for info. 870255714Skris [Steve Henson] 870355714Skris 870455714Skris *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 870555714Skris without temporary keys kept an extra copy of the server key, 870655714Skris and connections with temporary keys did not free everything in case 870755714Skris of an error. 870855714Skris [Bodo Moeller] 870955714Skris 871055714Skris *) New function RSA_check_key and new openssl rsa option -check 871155714Skris for verifying the consistency of RSA keys. 871255714Skris [Ulf Moeller, Bodo Moeller] 871355714Skris 871455714Skris *) Various changes to make Win32 compile work: 871555714Skris 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 871655714Skris 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 871755714Skris comparison" warnings. 871855714Skris 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 871955714Skris [Steve Henson] 872055714Skris 872155714Skris *) Add a debugging option to PKCS#5 v2 key generation function: when 872255714Skris you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 872355714Skris derived keys are printed to stderr. 872455714Skris [Steve Henson] 872555714Skris 872655714Skris *) Copy the flags in ASN1_STRING_dup(). 872755714Skris [Roman E. Pavlov <pre@mo.msk.ru>] 872855714Skris 872955714Skris *) The x509 application mishandled signing requests containing DSA 873055714Skris keys when the signing key was also DSA and the parameters didn't match. 873155714Skris 873255714Skris It was supposed to omit the parameters when they matched the signing key: 873355714Skris the verifying software was then supposed to automatically use the CA's 873455714Skris parameters if they were absent from the end user certificate. 873555714Skris 873655714Skris Omitting parameters is no longer recommended. The test was also 873755714Skris the wrong way round! This was probably due to unusual behaviour in 873855714Skris EVP_cmp_parameters() which returns 1 if the parameters match. 873955714Skris This meant that parameters were omitted when they *didn't* match and 874055714Skris the certificate was useless. Certificates signed with 'ca' didn't have 874155714Skris this bug. 874255714Skris [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 874355714Skris 874455714Skris *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 874555714Skris The interface is as follows: 874655714Skris Applications can use 874755714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 874855714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 874955714Skris "off" is now the default. 875055714Skris The library internally uses 875155714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 875255714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 875355714Skris to disable memory-checking temporarily. 875455714Skris 875555714Skris Some inconsistent states that previously were possible (and were 875655714Skris even the default) are now avoided. 875755714Skris 875855714Skris -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 875955714Skris with each memory chunk allocated; this is occasionally more helpful 876055714Skris than just having a counter. 876155714Skris 876255714Skris -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 876355714Skris 876455714Skris -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 876555714Skris extensions. 876655714Skris [Bodo Moeller] 876755714Skris 876855714Skris *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 876955714Skris which largely parallels "options", but is for changing API behaviour, 877055714Skris whereas "options" are about protocol behaviour. 877155714Skris Initial "mode" flags are: 877255714Skris 877355714Skris SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 877455714Skris a single record has been written. 877555714Skris SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 877655714Skris retries use the same buffer location. 877755714Skris (But all of the contents must be 877855714Skris copied!) 877955714Skris [Bodo Moeller] 878055714Skris 878179998Skris *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 878255714Skris worked. 878355714Skris 878455714Skris *) Fix problems with no-hmac etc. 878555714Skris [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 878655714Skris 878755714Skris *) New functions RSA_get_default_method(), RSA_set_method() and 878855714Skris RSA_get_method(). These allows replacement of RSA_METHODs without having 878955714Skris to mess around with the internals of an RSA structure. 879055714Skris [Steve Henson] 879155714Skris 879255714Skris *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 879355714Skris Also really enable memory leak checks in openssl.c and in some 879455714Skris test programs. 879555714Skris [Chad C. Mulligan, Bodo Moeller] 879655714Skris 879755714Skris *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 879855714Skris up the length of negative integers. This has now been simplified to just 879955714Skris store the length when it is first determined and use it later, rather 880055714Skris than trying to keep track of where data is copied and updating it to 880155714Skris point to the end. 880255714Skris [Steve Henson, reported by Brien Wheeler 880355714Skris <bwheeler@authentica-security.com>] 880455714Skris 880555714Skris *) Add a new function PKCS7_signatureVerify. This allows the verification 880655714Skris of a PKCS#7 signature but with the signing certificate passed to the 880755714Skris function itself. This contrasts with PKCS7_dataVerify which assumes the 880855714Skris certificate is present in the PKCS#7 structure. This isn't always the 880955714Skris case: certificates can be omitted from a PKCS#7 structure and be 881055714Skris distributed by "out of band" means (such as a certificate database). 881155714Skris [Steve Henson] 881255714Skris 881355714Skris *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 881455714Skris function prototypes in pem.h, also change util/mkdef.pl to add the 881555714Skris necessary function names. 881655714Skris [Steve Henson] 881755714Skris 881855714Skris *) mk1mf.pl (used by Windows builds) did not properly read the 881955714Skris options set by Configure in the top level Makefile, and Configure 882055714Skris was not even able to write more than one option correctly. 882155714Skris Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 882255714Skris [Bodo Moeller] 882355714Skris 882455714Skris *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 882555714Skris file to be loaded from a BIO or FILE pointer. The BIO version will 882655714Skris for example allow memory BIOs to contain config info. 882755714Skris [Steve Henson] 882855714Skris 882955714Skris *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 883055714Skris Whoever hopes to achieve shared-library compatibility across versions 883155714Skris must use this, not the compile-time macro. 883255714Skris (Exercise 0.9.4: Which is the minimum library version required by 883355714Skris such programs?) 883455714Skris Note: All this applies only to multi-threaded programs, others don't 883555714Skris need locks. 883655714Skris [Bodo Moeller] 883755714Skris 883855714Skris *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 883955714Skris through a BIO pair triggered the default case, i.e. 884055714Skris SSLerr(...,SSL_R_UNKNOWN_STATE). 884155714Skris [Bodo Moeller] 884255714Skris 884355714Skris *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 884455714Skris can use the SSL library even if none of the specific BIOs is 884555714Skris appropriate. 884655714Skris [Bodo Moeller] 884755714Skris 884855714Skris *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 884955714Skris for the encoded length. 885055714Skris [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 885155714Skris 885255714Skris *) Add initial documentation of the X509V3 functions. 885355714Skris [Steve Henson] 885455714Skris 885555714Skris *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 885655714Skris PEM_write_bio_PKCS8PrivateKey() that are equivalent to 885755714Skris PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 885855714Skris secure PKCS#8 private key format with a high iteration count. 885955714Skris [Steve Henson] 886055714Skris 886155714Skris *) Fix determination of Perl interpreter: A perl or perl5 886255714Skris _directory_ in $PATH was also accepted as the interpreter. 886355714Skris [Ralf S. Engelschall] 886455714Skris 886555714Skris *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 886655714Skris wrong with it but it was very old and did things like calling 886755714Skris PEM_ASN1_read() directly and used MD5 for the hash not to mention some 886855714Skris unusual formatting. 886955714Skris [Steve Henson] 887055714Skris 887155714Skris *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 887255714Skris to use the new extension code. 887355714Skris [Steve Henson] 887455714Skris 887555714Skris *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 887655714Skris with macros. This should make it easier to change their form, add extra 887755714Skris arguments etc. Fix a few PEM prototypes which didn't have cipher as a 887855714Skris constant. 887955714Skris [Steve Henson] 888055714Skris 888155714Skris *) Add to configuration table a new entry that can specify an alternative 888255714Skris name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 888355714Skris according to Mark Crispin <MRC@Panda.COM>. 888455714Skris [Bodo Moeller] 888555714Skris 888655714Skris#if 0 888755714Skris *) DES CBC did not update the IV. Weird. 888855714Skris [Ben Laurie] 888955714Skris#else 889055714Skris des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 889155714Skris Changing the behaviour of the former might break existing programs -- 889255714Skris where IV updating is needed, des_ncbc_encrypt can be used. 889355714Skris#endif 889455714Skris 889555714Skris *) When bntest is run from "make test" it drives bc to check its 889655714Skris calculations, as well as internally checking them. If an internal check 889755714Skris fails, it needs to cause bc to give a non-zero result or make test carries 889855714Skris on without noticing the failure. Fixed. 889955714Skris [Ben Laurie] 890055714Skris 890155714Skris *) DES library cleanups. 890255714Skris [Ulf M�ller] 890355714Skris 890455714Skris *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 890555714Skris used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 890655714Skris ciphers. NOTE: although the key derivation function has been verified 890755714Skris against some published test vectors it has not been extensively tested 890855714Skris yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 890955714Skris of v2.0. 891055714Skris [Steve Henson] 891155714Skris 891255714Skris *) Instead of "mkdir -p", which is not fully portable, use new 891355714Skris Perl script "util/mkdir-p.pl". 891455714Skris [Bodo Moeller] 891555714Skris 891655714Skris *) Rewrite the way password based encryption (PBE) is handled. It used to 891755714Skris assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 891855714Skris structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 891955714Skris but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 892055714Skris the 'parameter' field of the AlgorithmIdentifier is passed to the 892155714Skris underlying key generation function so it must do its own ASN1 parsing. 892255714Skris This has also changed the EVP_PBE_CipherInit() function which now has a 892355714Skris 'parameter' argument instead of literal salt and iteration count values 892455714Skris and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 892555714Skris [Steve Henson] 892655714Skris 892755714Skris *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 892855714Skris and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 892955714Skris Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 893055714Skris KEY" because this clashed with PKCS#8 unencrypted string. Since this 893155714Skris value was just used as a "magic string" and not used directly its 893255714Skris value doesn't matter. 893355714Skris [Steve Henson] 893455714Skris 893555714Skris *) Introduce some semblance of const correctness to BN. Shame C doesn't 893655714Skris support mutable. 893755714Skris [Ben Laurie] 893855714Skris 893955714Skris *) "linux-sparc64" configuration (ultrapenguin). 894055714Skris [Ray Miller <ray.miller@oucs.ox.ac.uk>] 894155714Skris "linux-sparc" configuration. 894255714Skris [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 894355714Skris 894455714Skris *) config now generates no-xxx options for missing ciphers. 894555714Skris [Ulf M�ller] 894655714Skris 894755714Skris *) Support the EBCDIC character set (work in progress). 894855714Skris File ebcdic.c not yet included because it has a different license. 894955714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 895055714Skris 895155714Skris *) Support BS2000/OSD-POSIX. 895255714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 895355714Skris 895455714Skris *) Make callbacks for key generation use void * instead of char *. 895555714Skris [Ben Laurie] 895655714Skris 895755714Skris *) Make S/MIME samples compile (not yet tested). 895855714Skris [Ben Laurie] 895955714Skris 896055714Skris *) Additional typesafe stacks. 896155714Skris [Ben Laurie] 896255714Skris 896355714Skris *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 896455714Skris [Bodo Moeller] 896555714Skris 896655714Skris 896755714Skris Changes between 0.9.3 and 0.9.3a [29 May 1999] 896855714Skris 896955714Skris *) New configuration variant "sco5-gcc". 897055714Skris 897155714Skris *) Updated some demos. 897255714Skris [Sean O Riordain, Wade Scholine] 897355714Skris 897455714Skris *) Add missing BIO_free at exit of pkcs12 application. 897555714Skris [Wu Zhigang] 897655714Skris 897755714Skris *) Fix memory leak in conf.c. 897855714Skris [Steve Henson] 897955714Skris 898055714Skris *) Updates for Win32 to assembler version of MD5. 898155714Skris [Steve Henson] 898255714Skris 898355714Skris *) Set #! path to perl in apps/der_chop to where we found it 898455714Skris instead of using a fixed path. 898555714Skris [Bodo Moeller] 898655714Skris 898755714Skris *) SHA library changes for irix64-mips4-cc. 898855714Skris [Andy Polyakov] 898955714Skris 899055714Skris *) Improvements for VMS support. 899155714Skris [Richard Levitte] 899255714Skris 899355714Skris 899455714Skris Changes between 0.9.2b and 0.9.3 [24 May 1999] 899555714Skris 899655714Skris *) Bignum library bug fix. IRIX 6 passes "make test" now! 899755714Skris This also avoids the problems with SC4.2 and unpatched SC5. 899855714Skris [Andy Polyakov <appro@fy.chalmers.se>] 899955714Skris 900055714Skris *) New functions sk_num, sk_value and sk_set to replace the previous macros. 900155714Skris These are required because of the typesafe stack would otherwise break 900255714Skris existing code. If old code used a structure member which used to be STACK 900355714Skris and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 900455714Skris sk_num or sk_value it would produce an error because the num, data members 900555714Skris are not present in STACK_OF. Now it just produces a warning. sk_set 900655714Skris replaces the old method of assigning a value to sk_value 900755714Skris (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 900855714Skris that does this will no longer work (and should use sk_set instead) but 900955714Skris this could be regarded as a "questionable" behaviour anyway. 901055714Skris [Steve Henson] 901155714Skris 901255714Skris *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 901355714Skris correctly handle encrypted S/MIME data. 901455714Skris [Steve Henson] 901555714Skris 901655714Skris *) Change type of various DES function arguments from des_cblock 901755714Skris (which means, in function argument declarations, pointer to char) 901855714Skris to des_cblock * (meaning pointer to array with 8 char elements), 901955714Skris which allows the compiler to do more typechecking; it was like 902055714Skris that back in SSLeay, but with lots of ugly casts. 902155714Skris 902255714Skris Introduce new type const_des_cblock. 902355714Skris [Bodo Moeller] 902455714Skris 902555714Skris *) Reorganise the PKCS#7 library and get rid of some of the more obvious 902655714Skris problems: find RecipientInfo structure that matches recipient certificate 902755714Skris and initialise the ASN1 structures properly based on passed cipher. 902855714Skris [Steve Henson] 902955714Skris 903055714Skris *) Belatedly make the BN tests actually check the results. 903155714Skris [Ben Laurie] 903255714Skris 903355714Skris *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 903455714Skris to and from BNs: it was completely broken. New compilation option 903555714Skris NEG_PUBKEY_BUG to allow for some broken certificates that encode public 903655714Skris key elements as negative integers. 903755714Skris [Steve Henson] 903855714Skris 903955714Skris *) Reorganize and speed up MD5. 904055714Skris [Andy Polyakov <appro@fy.chalmers.se>] 904155714Skris 904255714Skris *) VMS support. 904355714Skris [Richard Levitte <richard@levitte.org>] 904455714Skris 904555714Skris *) New option -out to asn1parse to allow the parsed structure to be 904655714Skris output to a file. This is most useful when combined with the -strparse 904755714Skris option to examine the output of things like OCTET STRINGS. 904855714Skris [Steve Henson] 904955714Skris 905055714Skris *) Make SSL library a little more fool-proof by not requiring any longer 905155714Skris that SSL_set_{accept,connect}_state be called before 905255714Skris SSL_{accept,connect} may be used (SSL_set_..._state is omitted 905355714Skris in many applications because usually everything *appeared* to work as 905455714Skris intended anyway -- now it really works as intended). 905555714Skris [Bodo Moeller] 905655714Skris 905755714Skris *) Move openssl.cnf out of lib/. 905855714Skris [Ulf M�ller] 905955714Skris 906055714Skris *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 906155714Skris -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 906255714Skris -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 906355714Skris [Ralf S. Engelschall] 906455714Skris 906555714Skris *) Various fixes to the EVP and PKCS#7 code. It may now be able to 906655714Skris handle PKCS#7 enveloped data properly. 906755714Skris [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 906855714Skris 906955714Skris *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 907055714Skris copying pointers. The cert_st handling is changed by this in 907155714Skris various ways (and thus what used to be known as ctx->default_cert 907255714Skris is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 907355714Skris any longer when s->cert does not give us what we need). 907455714Skris ssl_cert_instantiate becomes obsolete by this change. 907555714Skris As soon as we've got the new code right (possibly it already is?), 907655714Skris we have solved a couple of bugs of the earlier code where s->cert 907755714Skris was used as if it could not have been shared with other SSL structures. 907855714Skris 907955714Skris Note that using the SSL API in certain dirty ways now will result 908055714Skris in different behaviour than observed with earlier library versions: 908155714Skris Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 908255714Skris does not influence s as it used to. 908355714Skris 908455714Skris In order to clean up things more thoroughly, inside SSL_SESSION 908555714Skris we don't use CERT any longer, but a new structure SESS_CERT 908655714Skris that holds per-session data (if available); currently, this is 908755714Skris the peer's certificate chain and, for clients, the server's certificate 908855714Skris and temporary key. CERT holds only those values that can have 908955714Skris meaningful defaults in an SSL_CTX. 909055714Skris [Bodo Moeller] 909155714Skris 909255714Skris *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 909355714Skris from the internal representation. Various PKCS#7 fixes: remove some 909455714Skris evil casts and set the enc_dig_alg field properly based on the signing 909555714Skris key type. 909655714Skris [Steve Henson] 909755714Skris 909855714Skris *) Allow PKCS#12 password to be set from the command line or the 909955714Skris environment. Let 'ca' get its config file name from the environment 910055714Skris variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 910155714Skris and 'x509'). 910255714Skris [Steve Henson] 910355714Skris 910455714Skris *) Allow certificate policies extension to use an IA5STRING for the 910555714Skris organization field. This is contrary to the PKIX definition but 910655714Skris VeriSign uses it and IE5 only recognises this form. Document 'x509' 910755714Skris extension option. 910855714Skris [Steve Henson] 910955714Skris 911055714Skris *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 911155714Skris without disallowing inline assembler and the like for non-pedantic builds. 911255714Skris [Ben Laurie] 911355714Skris 911455714Skris *) Support Borland C++ builder. 911555714Skris [Janez Jere <jj@void.si>, modified by Ulf M�ller] 911655714Skris 911755714Skris *) Support Mingw32. 911855714Skris [Ulf M�ller] 911955714Skris 912055714Skris *) SHA-1 cleanups and performance enhancements. 912155714Skris [Andy Polyakov <appro@fy.chalmers.se>] 912255714Skris 912355714Skris *) Sparc v8plus assembler for the bignum library. 912455714Skris [Andy Polyakov <appro@fy.chalmers.se>] 912555714Skris 912655714Skris *) Accept any -xxx and +xxx compiler options in Configure. 912755714Skris [Ulf M�ller] 912855714Skris 912955714Skris *) Update HPUX configuration. 913055714Skris [Anonymous] 913155714Skris 913255714Skris *) Add missing sk_<type>_unshift() function to safestack.h 913355714Skris [Ralf S. Engelschall] 913455714Skris 913555714Skris *) New function SSL_CTX_use_certificate_chain_file that sets the 913655714Skris "extra_cert"s in addition to the certificate. (This makes sense 913755714Skris only for "PEM" format files, as chains as a whole are not 913855714Skris DER-encoded.) 913955714Skris [Bodo Moeller] 914055714Skris 914155714Skris *) Support verify_depth from the SSL API. 914255714Skris x509_vfy.c had what can be considered an off-by-one-error: 914355714Skris Its depth (which was not part of the external interface) 914455714Skris was actually counting the number of certificates in a chain; 914555714Skris now it really counts the depth. 914655714Skris [Bodo Moeller] 914755714Skris 914855714Skris *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 914955714Skris instead of X509err, which often resulted in confusing error 915055714Skris messages since the error codes are not globally unique 915155714Skris (e.g. an alleged error in ssl3_accept when a certificate 915255714Skris didn't match the private key). 915355714Skris 915455714Skris *) New function SSL_CTX_set_session_id_context that allows to set a default 915555714Skris value (so that you don't need SSL_set_session_id_context for each 915655714Skris connection using the SSL_CTX). 915755714Skris [Bodo Moeller] 915855714Skris 915955714Skris *) OAEP decoding bug fix. 916055714Skris [Ulf M�ller] 916155714Skris 916255714Skris *) Support INSTALL_PREFIX for package builders, as proposed by 916355714Skris David Harris. 916455714Skris [Bodo Moeller] 916555714Skris 916655714Skris *) New Configure options "threads" and "no-threads". For systems 916755714Skris where the proper compiler options are known (currently Solaris 916855714Skris and Linux), "threads" is the default. 916955714Skris [Bodo Moeller] 917055714Skris 917155714Skris *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 917255714Skris [Bodo Moeller] 917355714Skris 917455714Skris *) Install various scripts to $(OPENSSLDIR)/misc, not to 917555714Skris $(INSTALLTOP)/bin -- they shouldn't clutter directories 917655714Skris such as /usr/local/bin. 917755714Skris [Bodo Moeller] 917855714Skris 917955714Skris *) "make linux-shared" to build shared libraries. 918055714Skris [Niels Poppe <niels@netbox.org>] 918155714Skris 918255714Skris *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 918355714Skris [Ulf M�ller] 918455714Skris 918555714Skris *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 918655714Skris extension adding in x509 utility. 918755714Skris [Steve Henson] 918855714Skris 918955714Skris *) Remove NOPROTO sections and error code comments. 919055714Skris [Ulf M�ller] 919155714Skris 919255714Skris *) Partial rewrite of the DEF file generator to now parse the ANSI 919355714Skris prototypes. 919455714Skris [Steve Henson] 919555714Skris 919655714Skris *) New Configure options --prefix=DIR and --openssldir=DIR. 919755714Skris [Ulf M�ller] 919855714Skris 919955714Skris *) Complete rewrite of the error code script(s). It is all now handled 920055714Skris by one script at the top level which handles error code gathering, 920155714Skris header rewriting and C source file generation. It should be much better 920255714Skris than the old method: it now uses a modified version of Ulf's parser to 920355714Skris read the ANSI prototypes in all header files (thus the old K&R definitions 920455714Skris aren't needed for error creation any more) and do a better job of 920555714Skris translating function codes into names. The old 'ASN1 error code imbedded 920655714Skris in a comment' is no longer necessary and it doesn't use .err files which 920755714Skris have now been deleted. Also the error code call doesn't have to appear all 920855714Skris on one line (which resulted in some large lines...). 920955714Skris [Steve Henson] 921055714Skris 921155714Skris *) Change #include filenames from <foo.h> to <openssl/foo.h>. 921255714Skris [Bodo Moeller] 921355714Skris 921455714Skris *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 921555714Skris 0 (which usually indicates a closed connection), but continue reading. 921655714Skris [Bodo Moeller] 921755714Skris 921855714Skris *) Fix some race conditions. 921955714Skris [Bodo Moeller] 922055714Skris 922155714Skris *) Add support for CRL distribution points extension. Add Certificate 922255714Skris Policies and CRL distribution points documentation. 922355714Skris [Steve Henson] 922455714Skris 922555714Skris *) Move the autogenerated header file parts to crypto/opensslconf.h. 922655714Skris [Ulf M�ller] 922755714Skris 922855714Skris *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 922955714Skris 8 of keying material. Merlin has also confirmed interop with this fix 923055714Skris between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 923155714Skris [Merlin Hughes <merlin@baltimore.ie>] 923255714Skris 923355714Skris *) Fix lots of warnings. 923455714Skris [Richard Levitte <levitte@stacken.kth.se>] 923555714Skris 923655714Skris *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 923755714Skris the directory spec didn't end with a LIST_SEPARATOR_CHAR. 923855714Skris [Richard Levitte <levitte@stacken.kth.se>] 923955714Skris 924055714Skris *) Fix problems with sizeof(long) == 8. 924155714Skris [Andy Polyakov <appro@fy.chalmers.se>] 924255714Skris 924355714Skris *) Change functions to ANSI C. 924455714Skris [Ulf M�ller] 924555714Skris 924655714Skris *) Fix typos in error codes. 924755714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 924855714Skris 924955714Skris *) Remove defunct assembler files from Configure. 925055714Skris [Ulf M�ller] 925155714Skris 925255714Skris *) SPARC v8 assembler BIGNUM implementation. 925355714Skris [Andy Polyakov <appro@fy.chalmers.se>] 925455714Skris 925555714Skris *) Support for Certificate Policies extension: both print and set. 925655714Skris Various additions to support the r2i method this uses. 925755714Skris [Steve Henson] 925855714Skris 925955714Skris *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 926055714Skris return a const string when you are expecting an allocated buffer. 926155714Skris [Ben Laurie] 926255714Skris 926355714Skris *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 926455714Skris types DirectoryString and DisplayText. 926555714Skris [Steve Henson] 926655714Skris 926755714Skris *) Add code to allow r2i extensions to access the configuration database, 926855714Skris add an LHASH database driver and add several ctx helper functions. 926955714Skris [Steve Henson] 927055714Skris 927155714Skris *) Fix an evil bug in bn_expand2() which caused various BN functions to 927255714Skris fail when they extended the size of a BIGNUM. 927355714Skris [Steve Henson] 927455714Skris 927555714Skris *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 927655714Skris support typesafe stack. 927755714Skris [Steve Henson] 927855714Skris 927955714Skris *) Fix typo in SSL_[gs]et_options(). 928055714Skris [Nils Frostberg <nils@medcom.se>] 928155714Skris 928255714Skris *) Delete various functions and files that belonged to the (now obsolete) 928355714Skris old X509V3 handling code. 928455714Skris [Steve Henson] 928555714Skris 928655714Skris *) New Configure option "rsaref". 928755714Skris [Ulf M�ller] 928855714Skris 928955714Skris *) Don't auto-generate pem.h. 929055714Skris [Bodo Moeller] 929155714Skris 929255714Skris *) Introduce type-safe ASN.1 SETs. 929355714Skris [Ben Laurie] 929455714Skris 929555714Skris *) Convert various additional casted stacks to type-safe STACK_OF() variants. 929655714Skris [Ben Laurie, Ralf S. Engelschall, Steve Henson] 929755714Skris 929855714Skris *) Introduce type-safe STACKs. This will almost certainly break lots of code 929955714Skris that links with OpenSSL (well at least cause lots of warnings), but fear 930055714Skris not: the conversion is trivial, and it eliminates loads of evil casts. A 930155714Skris few STACKed things have been converted already. Feel free to convert more. 930255714Skris In the fullness of time, I'll do away with the STACK type altogether. 930355714Skris [Ben Laurie] 930455714Skris 930555714Skris *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 930655714Skris specified in <certfile> by updating the entry in the index.txt file. 930755714Skris This way one no longer has to edit the index.txt file manually for 930855714Skris revoking a certificate. The -revoke option does the gory details now. 930955714Skris [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 931055714Skris 931155714Skris *) Fix `openssl crl -noout -text' combination where `-noout' killed the 931255714Skris `-text' option at all and this way the `-noout -text' combination was 931355714Skris inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 931455714Skris [Ralf S. Engelschall] 931555714Skris 931655714Skris *) Make sure a corresponding plain text error message exists for the 931755714Skris X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 931855714Skris verify callback function determined that a certificate was revoked. 931955714Skris [Ralf S. Engelschall] 932055714Skris 932155714Skris *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 932255714Skris ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 932355714Skris all available cipers including rc5, which was forgotten until now. 932455714Skris In order to let the testing shell script know which algorithms 932555714Skris are available, a new (up to now undocumented) command 932655714Skris "openssl list-cipher-commands" is used. 932755714Skris [Bodo Moeller] 932855714Skris 932955714Skris *) Bugfix: s_client occasionally would sleep in select() when 933055714Skris it should have checked SSL_pending() first. 933155714Skris [Bodo Moeller] 933255714Skris 933355714Skris *) New functions DSA_do_sign and DSA_do_verify to provide access to 933455714Skris the raw DSA values prior to ASN.1 encoding. 933555714Skris [Ulf M�ller] 933655714Skris 933755714Skris *) Tweaks to Configure 933855714Skris [Niels Poppe <niels@netbox.org>] 933955714Skris 934055714Skris *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 934155714Skris yet... 934255714Skris [Steve Henson] 934355714Skris 934455714Skris *) New variables $(RANLIB) and $(PERL) in the Makefiles. 934555714Skris [Ulf M�ller] 934655714Skris 934755714Skris *) New config option to avoid instructions that are illegal on the 80386. 934855714Skris The default code is faster, but requires at least a 486. 934955714Skris [Ulf M�ller] 935055714Skris 935155714Skris *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 935255714Skris SSL2_SERVER_VERSION (not used at all) macros, which are now the 935355714Skris same as SSL2_VERSION anyway. 935455714Skris [Bodo Moeller] 935555714Skris 935655714Skris *) New "-showcerts" option for s_client. 935755714Skris [Bodo Moeller] 935855714Skris 935955714Skris *) Still more PKCS#12 integration. Add pkcs12 application to openssl 936055714Skris application. Various cleanups and fixes. 936155714Skris [Steve Henson] 936255714Skris 936355714Skris *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 936455714Skris modify error routines to work internally. Add error codes and PBE init 936555714Skris to library startup routines. 936655714Skris [Steve Henson] 936755714Skris 936855714Skris *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 936955714Skris packing functions to asn1 and evp. Changed function names and error 937055714Skris codes along the way. 937155714Skris [Steve Henson] 937255714Skris 937355714Skris *) PKCS12 integration: and so it begins... First of several patches to 937455714Skris slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 937555714Skris objects to objects.h 937655714Skris [Steve Henson] 937755714Skris 937855714Skris *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 937955714Skris and display support for Thawte strong extranet extension. 938055714Skris [Steve Henson] 938155714Skris 938255714Skris *) Add LinuxPPC support. 938355714Skris [Jeff Dubrule <igor@pobox.org>] 938455714Skris 938555714Skris *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 938655714Skris bn_div_words in alpha.s. 938755714Skris [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 938855714Skris 938955714Skris *) Make sure the RSA OAEP test is skipped under -DRSAref because 939055714Skris OAEP isn't supported when OpenSSL is built with RSAref. 939155714Skris [Ulf Moeller <ulf@fitug.de>] 939255714Skris 939355714Skris *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 939455714Skris so they no longer are missing under -DNOPROTO. 939555714Skris [Soren S. Jorvang <soren@t.dk>] 939655714Skris 939755714Skris 939855714Skris Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 939955714Skris 940055714Skris *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 940155714Skris doesn't work when the session is reused. Coming soon! 940255714Skris [Ben Laurie] 940355714Skris 940455714Skris *) Fix a security hole, that allows sessions to be reused in the wrong 940555714Skris context thus bypassing client cert protection! All software that uses 940655714Skris client certs and session caches in multiple contexts NEEDS PATCHING to 940755714Skris allow session reuse! A fuller solution is in the works. 940855714Skris [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 940955714Skris 941055714Skris *) Some more source tree cleanups (removed obsolete files 941155714Skris crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 941255714Skris permission on "config" script to be executable) and a fix for the INSTALL 941355714Skris document. 941455714Skris [Ulf Moeller <ulf@fitug.de>] 941555714Skris 941655714Skris *) Remove some legacy and erroneous uses of malloc, free instead of 941755714Skris Malloc, Free. 941855714Skris [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 941955714Skris 942055714Skris *) Make rsa_oaep_test return non-zero on error. 942155714Skris [Ulf Moeller <ulf@fitug.de>] 942255714Skris 942355714Skris *) Add support for native Solaris shared libraries. Configure 942455714Skris solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 942555714Skris if someone would make that last step automatic. 942655714Skris [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 942755714Skris 942855714Skris *) ctx_size was not built with the right compiler during "make links". Fixed. 942955714Skris [Ben Laurie] 943055714Skris 943155714Skris *) Change the meaning of 'ALL' in the cipher list. It now means "everything 943255714Skris except NULL ciphers". This means the default cipher list will no longer 943355714Skris enable NULL ciphers. They need to be specifically enabled e.g. with 943455714Skris the string "DEFAULT:eNULL". 943555714Skris [Steve Henson] 943655714Skris 943755714Skris *) Fix to RSA private encryption routines: if p < q then it would 943855714Skris occasionally produce an invalid result. This will only happen with 943955714Skris externally generated keys because OpenSSL (and SSLeay) ensure p > q. 944055714Skris [Steve Henson] 944155714Skris 944255714Skris *) Be less restrictive and allow also `perl util/perlpath.pl 944355714Skris /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 944455714Skris because this way one can also use an interpreter named `perl5' (which is 944555714Skris usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 944655714Skris installed as `perl'). 944755714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 944855714Skris 944955714Skris *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 945055714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 945155714Skris 945255714Skris *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 945355714Skris advapi32.lib to Win32 build and change the pem test comparision 945455714Skris to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 945555714Skris suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 945655714Skris and crypto/des/ede_cbcm_enc.c. 945755714Skris [Steve Henson] 945855714Skris 945955714Skris *) DES quad checksum was broken on big-endian architectures. Fixed. 946055714Skris [Ben Laurie] 946155714Skris 946255714Skris *) Comment out two functions in bio.h that aren't implemented. Fix up the 946355714Skris Win32 test batch file so it (might) work again. The Win32 test batch file 946455714Skris is horrible: I feel ill.... 946555714Skris [Steve Henson] 946655714Skris 946755714Skris *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 946855714Skris in e_os.h. Audit of header files to check ANSI and non ANSI 946955714Skris sections: 10 functions were absent from non ANSI section and not exported 947055714Skris from Windows DLLs. Fixed up libeay.num for new functions. 947155714Skris [Steve Henson] 947255714Skris 947355714Skris *) Make `openssl version' output lines consistent. 947455714Skris [Ralf S. Engelschall] 947555714Skris 947655714Skris *) Fix Win32 symbol export lists for BIO functions: Added 947755714Skris BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 947855714Skris to ms/libeay{16,32}.def. 947955714Skris [Ralf S. Engelschall] 948055714Skris 948155714Skris *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 948255714Skris fine under Unix and passes some trivial tests I've now added. But the 948355714Skris whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 948455714Skris added to make sure no one expects that this stuff really works in the 948555714Skris OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 948655714Skris up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 948755714Skris openssl_bio.xs. 948855714Skris [Ralf S. Engelschall] 948955714Skris 949055714Skris *) Fix the generation of two part addresses in perl. 949155714Skris [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 949255714Skris 949355714Skris *) Add config entry for Linux on MIPS. 949455714Skris [John Tobey <jtobey@channel1.com>] 949555714Skris 949655714Skris *) Make links whenever Configure is run, unless we are on Windoze. 949755714Skris [Ben Laurie] 949855714Skris 949955714Skris *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 950055714Skris Currently only issuerAltName and AuthorityKeyIdentifier make any sense 950155714Skris in CRLs. 950255714Skris [Steve Henson] 950355714Skris 950455714Skris *) Add a useful kludge to allow package maintainers to specify compiler and 950555714Skris other platforms details on the command line without having to patch the 950655714Skris Configure script everytime: One now can use ``perl Configure 950755714Skris <id>:<details>'', i.e. platform ids are allowed to have details appended 950855714Skris to them (seperated by colons). This is treated as there would be a static 950955714Skris pre-configured entry in Configure's %table under key <id> with value 951055714Skris <details> and ``perl Configure <id>'' is called. So, when you want to 951155714Skris perform a quick test-compile under FreeBSD 3.1 with pgcc and without 951255714Skris assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 951355714Skris now, which overrides the FreeBSD-elf entry on-the-fly. 951455714Skris [Ralf S. Engelschall] 951555714Skris 951655714Skris *) Disable new TLS1 ciphersuites by default: they aren't official yet. 951755714Skris [Ben Laurie] 951855714Skris 951955714Skris *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 952055714Skris on the `perl Configure ...' command line. This way one can compile 952155714Skris OpenSSL libraries with Position Independent Code (PIC) which is needed 952255714Skris for linking it into DSOs. 952355714Skris [Ralf S. Engelschall] 952455714Skris 952555714Skris *) Remarkably, export ciphers were totally broken and no-one had noticed! 952655714Skris Fixed. 952755714Skris [Ben Laurie] 952855714Skris 952955714Skris *) Cleaned up the LICENSE document: The official contact for any license 953055714Skris questions now is the OpenSSL core team under openssl-core@openssl.org. 953155714Skris And add a paragraph about the dual-license situation to make sure people 953255714Skris recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 953355714Skris to the OpenSSL toolkit. 953455714Skris [Ralf S. Engelschall] 953555714Skris 953655714Skris *) General source tree makefile cleanups: Made `making xxx in yyy...' 953755714Skris display consistent in the source tree and replaced `/bin/rm' by `rm'. 953855714Skris Additonally cleaned up the `make links' target: Remove unnecessary 953955714Skris semicolons, subsequent redundant removes, inline point.sh into mklink.sh 954055714Skris to speed processing and no longer clutter the display with confusing 954155714Skris stuff. Instead only the actually done links are displayed. 954255714Skris [Ralf S. Engelschall] 954355714Skris 954455714Skris *) Permit null encryption ciphersuites, used for authentication only. It used 954555714Skris to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 954655714Skris It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 954755714Skris encryption. 954855714Skris [Ben Laurie] 954955714Skris 955055714Skris *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 955155714Skris signed attributes when verifying signatures (this would break them), 955255714Skris the detached data encoding was wrong and public keys obtained using 955355714Skris X509_get_pubkey() weren't freed. 955455714Skris [Steve Henson] 955555714Skris 955655714Skris *) Add text documentation for the BUFFER functions. Also added a work around 955755714Skris to a Win95 console bug. This was triggered by the password read stuff: the 955855714Skris last character typed gets carried over to the next fread(). If you were 955955714Skris generating a new cert request using 'req' for example then the last 956055714Skris character of the passphrase would be CR which would then enter the first 956155714Skris field as blank. 956255714Skris [Steve Henson] 956355714Skris 956455714Skris *) Added the new `Includes OpenSSL Cryptography Software' button as 956555714Skris doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 956655714Skris button and can be used by applications based on OpenSSL to show the 956755714Skris relationship to the OpenSSL project. 956855714Skris [Ralf S. Engelschall] 956955714Skris 957055714Skris *) Remove confusing variables in function signatures in files 957155714Skris ssl/ssl_lib.c and ssl/ssl.h. 957255714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 957355714Skris 957455714Skris *) Don't install bss_file.c under PREFIX/include/ 957555714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 957655714Skris 957755714Skris *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 957855714Skris functions that return function pointers and has support for NT specific 957955714Skris stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 958055714Skris #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 958155714Skris unsigned to signed types: this was killing the Win32 compile. 958255714Skris [Steve Henson] 958355714Skris 958455714Skris *) Add new certificate file to stack functions, 958555714Skris SSL_add_dir_cert_subjects_to_stack() and 958655714Skris SSL_add_file_cert_subjects_to_stack(). These largely supplant 958755714Skris SSL_load_client_CA_file(), and can be used to add multiple certs easily 958855714Skris to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 958955714Skris This means that Apache-SSL and similar packages don't have to mess around 959055714Skris to add as many CAs as they want to the preferred list. 959155714Skris [Ben Laurie] 959255714Skris 959355714Skris *) Experiment with doxygen documentation. Currently only partially applied to 959455714Skris ssl/ssl_lib.c. 959555714Skris See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 959655714Skris openssl.doxy as the configuration file. 959755714Skris [Ben Laurie] 959855714Skris 959955714Skris *) Get rid of remaining C++-style comments which strict C compilers hate. 960055714Skris [Ralf S. Engelschall, pointed out by Carlos Amengual] 960155714Skris 960255714Skris *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 960355714Skris compiled in by default: it has problems with large keys. 960455714Skris [Steve Henson] 960555714Skris 960655714Skris *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 960755714Skris DH private keys and/or callback functions which directly correspond to 960855714Skris their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 960955714Skris is needed for applications which have to configure certificates on a 961055714Skris per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 961155714Skris (e.g. s_server). 961255714Skris For the RSA certificate situation is makes no difference, but 961355714Skris for the DSA certificate situation this fixes the "no shared cipher" 961455714Skris problem where the OpenSSL cipher selection procedure failed because the 961555714Skris temporary keys were not overtaken from the context and the API provided 961655714Skris no way to reconfigure them. 961755714Skris The new functions now let applications reconfigure the stuff and they 961855714Skris are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 961955714Skris SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 962055714Skris non-public-API function ssl_cert_instantiate() is used as a helper 962155714Skris function and also to reduce code redundancy inside ssl_rsa.c. 962255714Skris [Ralf S. Engelschall] 962355714Skris 962455714Skris *) Move s_server -dcert and -dkey options out of the undocumented feature 962555714Skris area because they are useful for the DSA situation and should be 962655714Skris recognized by the users. 962755714Skris [Ralf S. Engelschall] 962855714Skris 962955714Skris *) Fix the cipher decision scheme for export ciphers: the export bits are 963055714Skris *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 963155714Skris SSL_EXP_MASK. So, the original variable has to be used instead of the 963255714Skris already masked variable. 963355714Skris [Richard Levitte <levitte@stacken.kth.se>] 963455714Skris 963555714Skris *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 963655714Skris [Richard Levitte <levitte@stacken.kth.se>] 963755714Skris 963855714Skris *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 963955714Skris from `int' to `unsigned int' because it's a length and initialized by 964055714Skris EVP_DigestFinal() which expects an `unsigned int *'. 964155714Skris [Richard Levitte <levitte@stacken.kth.se>] 964255714Skris 964355714Skris *) Don't hard-code path to Perl interpreter on shebang line of Configure 964455714Skris script. Instead use the usual Shell->Perl transition trick. 964555714Skris [Ralf S. Engelschall] 964655714Skris 964755714Skris *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 964855714Skris (in addition to RSA certificates) to match the behaviour of `openssl dsa 964955714Skris -noout -modulus' as it's already the case for `openssl rsa -noout 965055714Skris -modulus'. For RSA the -modulus is the real "modulus" while for DSA 965155714Skris currently the public key is printed (a decision which was already done by 965255714Skris `openssl dsa -modulus' in the past) which serves a similar purpose. 965355714Skris Additionally the NO_RSA no longer completely removes the whole -modulus 965455714Skris option; it now only avoids using the RSA stuff. Same applies to NO_DSA 965555714Skris now, too. 965655714Skris [Ralf S. Engelschall] 965755714Skris 965855714Skris *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 965955714Skris BIO. See the source (crypto/evp/bio_ok.c) for more info. 966055714Skris [Arne Ansper <arne@ats.cyber.ee>] 966155714Skris 966255714Skris *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 966355714Skris to be added. Now both 'req' and 'ca' can use new objects defined in the 966455714Skris config file. 966555714Skris [Steve Henson] 966655714Skris 966755714Skris *) Add cool BIO that does syslog (or event log on NT). 966855714Skris [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 966955714Skris 967055714Skris *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 967155714Skris TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 967255714Skris TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 967355714Skris Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 967455714Skris [Ben Laurie] 967555714Skris 967655714Skris *) Add preliminary config info for new extension code. 967755714Skris [Steve Henson] 967855714Skris 967955714Skris *) Make RSA_NO_PADDING really use no padding. 968055714Skris [Ulf Moeller <ulf@fitug.de>] 968155714Skris 968255714Skris *) Generate errors when private/public key check is done. 968355714Skris [Ben Laurie] 968455714Skris 968555714Skris *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 968655714Skris for some CRL extensions and new objects added. 968755714Skris [Steve Henson] 968855714Skris 968955714Skris *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 969055714Skris key usage extension and fuller support for authority key id. 969155714Skris [Steve Henson] 969255714Skris 969355714Skris *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 969455714Skris padding method for RSA, which is recommended for new applications in PKCS 969555714Skris #1 v2.0 (RFC 2437, October 1998). 969655714Skris OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 969755714Skris foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 969855714Skris against Bleichbacher's attack on RSA. 969955714Skris [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 970055714Skris Ben Laurie] 970155714Skris 970255714Skris *) Updates to the new SSL compression code 970355714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 970455714Skris 970555714Skris *) Fix so that the version number in the master secret, when passed 970655714Skris via RSA, checks that if TLS was proposed, but we roll back to SSLv3 970755714Skris (because the server will not accept higher), that the version number 970855714Skris is 0x03,0x01, not 0x03,0x00 970955714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 971055714Skris 971155714Skris *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 971255714Skris leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 971355714Skris in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 971455714Skris [Steve Henson] 971555714Skris 971655714Skris *) Support for RAW extensions where an arbitrary extension can be 971755714Skris created by including its DER encoding. See apps/openssl.cnf for 971855714Skris an example. 971955714Skris [Steve Henson] 972055714Skris 972155714Skris *) Make sure latest Perl versions don't interpret some generated C array 972255714Skris code as Perl array code in the crypto/err/err_genc.pl script. 972355714Skris [Lars Weber <3weber@informatik.uni-hamburg.de>] 972455714Skris 972555714Skris *) Modify ms/do_ms.bat to not generate assembly language makefiles since 972655714Skris not many people have the assembler. Various Win32 compilation fixes and 972755714Skris update to the INSTALL.W32 file with (hopefully) more accurate Win32 972855714Skris build instructions. 972955714Skris [Steve Henson] 973055714Skris 973155714Skris *) Modify configure script 'Configure' to automatically create crypto/date.h 973255714Skris file under Win32 and also build pem.h from pem.org. New script 973355714Skris util/mkfiles.pl to create the MINFO file on environments that can't do a 973455714Skris 'make files': perl util/mkfiles.pl >MINFO should work. 973555714Skris [Steve Henson] 973655714Skris 973755714Skris *) Major rework of DES function declarations, in the pursuit of correctness 973855714Skris and purity. As a result, many evil casts evaporated, and some weirdness, 973955714Skris too. You may find this causes warnings in your code. Zapping your evil 974055714Skris casts will probably fix them. Mostly. 974155714Skris [Ben Laurie] 974255714Skris 974355714Skris *) Fix for a typo in asn1.h. Bug fix to object creation script 974455714Skris obj_dat.pl. It considered a zero in an object definition to mean 974555714Skris "end of object": none of the objects in objects.h have any zeros 974655714Skris so it wasn't spotted. 974755714Skris [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 974855714Skris 974955714Skris *) Add support for Triple DES Cipher Block Chaining with Output Feedback 975055714Skris Masking (CBCM). In the absence of test vectors, the best I have been able 975155714Skris to do is check that the decrypt undoes the encrypt, so far. Send me test 975255714Skris vectors if you have them. 975355714Skris [Ben Laurie] 975455714Skris 975555714Skris *) Correct calculation of key length for export ciphers (too much space was 975655714Skris allocated for null ciphers). This has not been tested! 975755714Skris [Ben Laurie] 975855714Skris 975955714Skris *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 976055714Skris message is now correct (it understands "crypto" and "ssl" on its 976155714Skris command line). There is also now an "update" option. This will update 976255714Skris the util/ssleay.num and util/libeay.num files with any new functions. 976355714Skris If you do a: 976455714Skris perl util/mkdef.pl crypto ssl update 976555714Skris it will update them. 976655714Skris [Steve Henson] 976755714Skris 976855714Skris *) Overhauled the Perl interface (perl/*): 976955714Skris - ported BN stuff to OpenSSL's different BN library 977055714Skris - made the perl/ source tree CVS-aware 977155714Skris - renamed the package from SSLeay to OpenSSL (the files still contain 977255714Skris their history because I've copied them in the repository) 977355714Skris - removed obsolete files (the test scripts will be replaced 977455714Skris by better Test::Harness variants in the future) 977555714Skris [Ralf S. Engelschall] 977655714Skris 977755714Skris *) First cut for a very conservative source tree cleanup: 977855714Skris 1. merge various obsolete readme texts into doc/ssleay.txt 977955714Skris where we collect the old documents and readme texts. 978055714Skris 2. remove the first part of files where I'm already sure that we no 978155714Skris longer need them because of three reasons: either they are just temporary 978255714Skris files which were left by Eric or they are preserved original files where 978355714Skris I've verified that the diff is also available in the CVS via "cvs diff 978455714Skris -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 978555714Skris the crypto/md/ stuff). 978655714Skris [Ralf S. Engelschall] 978755714Skris 978855714Skris *) More extension code. Incomplete support for subject and issuer alt 978955714Skris name, issuer and authority key id. Change the i2v function parameters 979055714Skris and add an extra 'crl' parameter in the X509V3_CTX structure: guess 979155714Skris what that's for :-) Fix to ASN1 macro which messed up 979255714Skris IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 979355714Skris [Steve Henson] 979455714Skris 979555714Skris *) Preliminary support for ENUMERATED type. This is largely copied from the 979655714Skris INTEGER code. 979755714Skris [Steve Henson] 979855714Skris 979955714Skris *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 980055714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 980155714Skris 980255714Skris *) Make sure `make rehash' target really finds the `openssl' program. 980355714Skris [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 980455714Skris 980555714Skris *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 980655714Skris like to hear about it if this slows down other processors. 980755714Skris [Ben Laurie] 980855714Skris 980955714Skris *) Add CygWin32 platform information to Configure script. 981055714Skris [Alan Batie <batie@aahz.jf.intel.com>] 981155714Skris 981255714Skris *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 981355714Skris [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 981455714Skris 981555714Skris *) New program nseq to manipulate netscape certificate sequences 981655714Skris [Steve Henson] 981755714Skris 981855714Skris *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 981955714Skris few typos. 982055714Skris [Steve Henson] 982155714Skris 982255714Skris *) Fixes to BN code. Previously the default was to define BN_RECURSION 982355714Skris but the BN code had some problems that would cause failures when 982455714Skris doing certificate verification and some other functions. 982555714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 982655714Skris 982755714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 982855714Skris [Steve Henson] 982955714Skris 983055714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 983155714Skris [Steve Henson] 983255714Skris 983355714Skris *) Add several PKIX and private extended key usage OIDs. 983455714Skris [Steve Henson] 983555714Skris 983655714Skris *) Modify the 'ca' program to handle the new extension code. Modify 983755714Skris openssl.cnf for new extension format, add comments. 983855714Skris [Steve Henson] 983955714Skris 984055714Skris *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 984155714Skris and add a sample to openssl.cnf so req -x509 now adds appropriate 984255714Skris CA extensions. 984355714Skris [Steve Henson] 984455714Skris 984555714Skris *) Continued X509 V3 changes. Add to other makefiles, integrate with the 984655714Skris error code, add initial support to X509_print() and x509 application. 984755714Skris [Steve Henson] 984855714Skris 984955714Skris *) Takes a deep breath and start addding X509 V3 extension support code. Add 985055714Skris files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 985155714Skris stuff is currently isolated and isn't even compiled yet. 985255714Skris [Steve Henson] 985355714Skris 985455714Skris *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 985555714Skris ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 985655714Skris Removed the versions check from X509 routines when loading extensions: 985755714Skris this allows certain broken certificates that don't set the version 985855714Skris properly to be processed. 985955714Skris [Steve Henson] 986055714Skris 986155714Skris *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 986255714Skris Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 986355714Skris can still be regenerated with "make depend". 986455714Skris [Ben Laurie] 986555714Skris 986655714Skris *) Spelling mistake in C version of CAST-128. 986755714Skris [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 986855714Skris 986955714Skris *) Changes to the error generation code. The perl script err-code.pl 987055714Skris now reads in the old error codes and retains the old numbers, only 987155714Skris adding new ones if necessary. It also only changes the .err files if new 987255714Skris codes are added. The makefiles have been modified to only insert errors 987355714Skris when needed (to avoid needlessly modifying header files). This is done 987455714Skris by only inserting errors if the .err file is newer than the auto generated 987555714Skris C file. To rebuild all the error codes from scratch (the old behaviour) 987655714Skris either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 987755714Skris or delete all the .err files. 987855714Skris [Steve Henson] 987955714Skris 988055714Skris *) CAST-128 was incorrectly implemented for short keys. The C version has 988155714Skris been fixed, but is untested. The assembler versions are also fixed, but 988255714Skris new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 988355714Skris to regenerate it if needed. 988455714Skris [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 988555714Skris Hagino <itojun@kame.net>] 988655714Skris 988755714Skris *) File was opened incorrectly in randfile.c. 988855714Skris [Ulf M�ller <ulf@fitug.de>] 988955714Skris 989055714Skris *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 989155714Skris functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 989255714Skris GeneralizedTime. ASN1_TIME is the proper type used in certificates et 989355714Skris al: it's just almost always a UTCTime. Note this patch adds new error 989455714Skris codes so do a "make errors" if there are problems. 989555714Skris [Steve Henson] 989655714Skris 989755714Skris *) Correct Linux 1 recognition in config. 989855714Skris [Ulf M�ller <ulf@fitug.de>] 989955714Skris 990055714Skris *) Remove pointless MD5 hash when using DSA keys in ca. 990155714Skris [Anonymous <nobody@replay.com>] 990255714Skris 990355714Skris *) Generate an error if given an empty string as a cert directory. Also 990455714Skris generate an error if handed NULL (previously returned 0 to indicate an 990555714Skris error, but didn't set one). 990655714Skris [Ben Laurie, reported by Anonymous <nobody@replay.com>] 990755714Skris 990855714Skris *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 990955714Skris [Ben Laurie] 991055714Skris 991155714Skris *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 991255714Skris parameters. This was causing a warning which killed off the Win32 compile. 991355714Skris [Steve Henson] 991455714Skris 991555714Skris *) Remove C++ style comments from crypto/bn/bn_local.h. 991655714Skris [Neil Costigan <neil.costigan@celocom.com>] 991755714Skris 991855714Skris *) The function OBJ_txt2nid was broken. It was supposed to return a nid 991955714Skris based on a text string, looking up short and long names and finally 992055714Skris "dot" format. The "dot" format stuff didn't work. Added new function 992155714Skris OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 992255714Skris OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 992355714Skris OID is not part of the table. 992455714Skris [Steve Henson] 992555714Skris 992655714Skris *) Add prototypes to X509 lookup/verify methods, fixing a bug in 992755714Skris X509_LOOKUP_by_alias(). 992855714Skris [Ben Laurie] 992955714Skris 993055714Skris *) Sort openssl functions by name. 993155714Skris [Ben Laurie] 993255714Skris 993355714Skris *) Get the gendsa program working (hopefully) and add it to app list. Remove 993455714Skris encryption from sample DSA keys (in case anyone is interested the password 993555714Skris was "1234"). 993655714Skris [Steve Henson] 993755714Skris 993855714Skris *) Make _all_ *_free functions accept a NULL pointer. 993955714Skris [Frans Heymans <fheymans@isaserver.be>] 994055714Skris 994155714Skris *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 994255714Skris NULL pointers. 994355714Skris [Anonymous <nobody@replay.com>] 994455714Skris 994555714Skris *) s_server should send the CAfile as acceptable CAs, not its own cert. 994655714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 994755714Skris 994855714Skris *) Don't blow it for numeric -newkey arguments to apps/req. 994955714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 995055714Skris 995155714Skris *) Temp key "for export" tests were wrong in s3_srvr.c. 995255714Skris [Anonymous <nobody@replay.com>] 995355714Skris 995455714Skris *) Add prototype for temp key callback functions 995555714Skris SSL_CTX_set_tmp_{rsa,dh}_callback(). 995655714Skris [Ben Laurie] 995755714Skris 995855714Skris *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 995955714Skris DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 996055714Skris [Steve Henson] 996155714Skris 996255714Skris *) X509_name_add_entry() freed the wrong thing after an error. 996355714Skris [Arne Ansper <arne@ats.cyber.ee>] 996455714Skris 996555714Skris *) rsa_eay.c would attempt to free a NULL context. 996655714Skris [Arne Ansper <arne@ats.cyber.ee>] 996755714Skris 996855714Skris *) BIO_s_socket() had a broken should_retry() on Windoze. 996955714Skris [Arne Ansper <arne@ats.cyber.ee>] 997055714Skris 997155714Skris *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 997255714Skris [Arne Ansper <arne@ats.cyber.ee>] 997355714Skris 997455714Skris *) Make sure the already existing X509_STORE->depth variable is initialized 997555714Skris in X509_STORE_new(), but document the fact that this variable is still 997655714Skris unused in the certificate verification process. 997755714Skris [Ralf S. Engelschall] 997855714Skris 997955714Skris *) Fix the various library and apps files to free up pkeys obtained from 998055714Skris X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 998155714Skris [Steve Henson] 998255714Skris 998355714Skris *) Fix reference counting in X509_PUBKEY_get(). This makes 998455714Skris demos/maurice/example2.c work, amongst others, probably. 998555714Skris [Steve Henson and Ben Laurie] 998655714Skris 998755714Skris *) First cut of a cleanup for apps/. First the `ssleay' program is now named 998855714Skris `openssl' and second, the shortcut symlinks for the `openssl <command>' 998955714Skris are no longer created. This way we have a single and consistent command 999055714Skris line interface `openssl <command>', similar to `cvs <command>'. 999155714Skris [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 999255714Skris 999355714Skris *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 999455714Skris BIT STRING wrapper always have zero unused bits. 999555714Skris [Steve Henson] 999655714Skris 999755714Skris *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 999855714Skris [Steve Henson] 999955714Skris 1000055714Skris *) Make the top-level INSTALL documentation easier to understand. 1000155714Skris [Paul Sutton] 1000255714Skris 1000355714Skris *) Makefiles updated to exit if an error occurs in a sub-directory 1000455714Skris make (including if user presses ^C) [Paul Sutton] 1000555714Skris 1000655714Skris *) Make Montgomery context stuff explicit in RSA data structure. 1000755714Skris [Ben Laurie] 1000855714Skris 1000955714Skris *) Fix build order of pem and err to allow for generated pem.h. 1001055714Skris [Ben Laurie] 1001155714Skris 1001255714Skris *) Fix renumbering bug in X509_NAME_delete_entry(). 1001355714Skris [Ben Laurie] 1001455714Skris 1001555714Skris *) Enhanced the err-ins.pl script so it makes the error library number 1001655714Skris global and can add a library name. This is needed for external ASN1 and 1001755714Skris other error libraries. 1001855714Skris [Steve Henson] 1001955714Skris 1002055714Skris *) Fixed sk_insert which never worked properly. 1002155714Skris [Steve Henson] 1002255714Skris 1002355714Skris *) Fix ASN1 macros so they can handle indefinite length construted 1002455714Skris EXPLICIT tags. Some non standard certificates use these: they can now 1002555714Skris be read in. 1002655714Skris [Steve Henson] 1002755714Skris 1002855714Skris *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 1002955714Skris into a single doc/ssleay.txt bundle. This way the information is still 1003055714Skris preserved but no longer messes up this directory. Now it's new room for 1003155714Skris the new set of documenation files. 1003255714Skris [Ralf S. Engelschall] 1003355714Skris 1003455714Skris *) SETs were incorrectly DER encoded. This was a major pain, because they 1003555714Skris shared code with SEQUENCEs, which aren't coded the same. This means that 1003655714Skris almost everything to do with SETs or SEQUENCEs has either changed name or 1003755714Skris number of arguments. 1003855714Skris [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 1003955714Skris 1004055714Skris *) Fix test data to work with the above. 1004155714Skris [Ben Laurie] 1004255714Skris 1004355714Skris *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 1004455714Skris was already fixed by Eric for 0.9.1 it seems. 1004555714Skris [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 1004655714Skris 1004755714Skris *) Autodetect FreeBSD3. 1004855714Skris [Ben Laurie] 1004955714Skris 1005055714Skris *) Fix various bugs in Configure. This affects the following platforms: 1005155714Skris nextstep 1005255714Skris ncr-scde 1005355714Skris unixware-2.0 1005455714Skris unixware-2.0-pentium 1005555714Skris sco5-cc. 1005655714Skris [Ben Laurie] 1005755714Skris 1005855714Skris *) Eliminate generated files from CVS. Reorder tests to regenerate files 1005955714Skris before they are needed. 1006055714Skris [Ben Laurie] 1006155714Skris 1006255714Skris *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 1006355714Skris [Ben Laurie] 1006455714Skris 1006555714Skris 1006655714Skris Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 1006755714Skris 1006855714Skris *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 1006955714Skris changed SSLeay to OpenSSL in version strings. 1007055714Skris [Ralf S. Engelschall] 1007155714Skris 1007255714Skris *) Some fixups to the top-level documents. 1007355714Skris [Paul Sutton] 1007455714Skris 1007555714Skris *) Fixed the nasty bug where rsaref.h was not found under compile-time 1007655714Skris because the symlink to include/ was missing. 1007755714Skris [Ralf S. Engelschall] 1007855714Skris 1007955714Skris *) Incorporated the popular no-RSA/DSA-only patches 1008055714Skris which allow to compile a RSA-free SSLeay. 1008155714Skris [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 1008255714Skris 1008355714Skris *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 1008455714Skris when "ssleay" is still not found. 1008555714Skris [Ralf S. Engelschall] 1008655714Skris 1008755714Skris *) Added more platforms to Configure: Cray T3E, HPUX 11, 1008855714Skris [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 1008955714Skris 1009055714Skris *) Updated the README file. 1009155714Skris [Ralf S. Engelschall] 1009255714Skris 1009355714Skris *) Added various .cvsignore files in the CVS repository subdirs 1009455714Skris to make a "cvs update" really silent. 1009555714Skris [Ralf S. Engelschall] 1009655714Skris 1009755714Skris *) Recompiled the error-definition header files and added 1009855714Skris missing symbols to the Win32 linker tables. 1009955714Skris [Ralf S. Engelschall] 1010055714Skris 1010155714Skris *) Cleaned up the top-level documents; 1010255714Skris o new files: CHANGES and LICENSE 1010355714Skris o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 1010455714Skris o merged COPYRIGHT into LICENSE 1010555714Skris o removed obsolete TODO file 1010655714Skris o renamed MICROSOFT to INSTALL.W32 1010755714Skris [Ralf S. Engelschall] 1010855714Skris 1010955714Skris *) Removed dummy files from the 0.9.1b source tree: 1011055714Skris crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 1011155714Skris crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 1011255714Skris crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 1011355714Skris crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 1011455714Skris util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 1011555714Skris [Ralf S. Engelschall] 1011655714Skris 1011755714Skris *) Added various platform portability fixes. 1011855714Skris [Mark J. Cox] 1011955714Skris 1012055714Skris *) The Genesis of the OpenSSL rpject: 1012155714Skris We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 1012255714Skris Young and Tim J. Hudson created while they were working for C2Net until 1012355714Skris summer 1998. 1012455714Skris [The OpenSSL Project] 1012555714Skris 1012655714Skris 1012755714Skris Changes between 0.9.0b and 0.9.1b [not released] 1012855714Skris 1012955714Skris *) Updated a few CA certificates under certs/ 1013055714Skris [Eric A. Young] 1013155714Skris 1013255714Skris *) Changed some BIGNUM api stuff. 1013355714Skris [Eric A. Young] 1013455714Skris 1013555714Skris *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 1013655714Skris DGUX x86, Linux Alpha, etc. 1013755714Skris [Eric A. Young] 1013855714Skris 1013955714Skris *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 1014055714Skris RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 1014155714Skris available). 1014255714Skris [Eric A. Young] 1014355714Skris 1014455714Skris *) Add -strparse option to asn1pars program which parses nested 1014555714Skris binary structures 1014655714Skris [Dr Stephen Henson <shenson@bigfoot.com>] 1014755714Skris 1014855714Skris *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 1014955714Skris [Eric A. Young] 1015055714Skris 1015155714Skris *) DSA fix for "ca" program. 1015255714Skris [Eric A. Young] 1015355714Skris 1015455714Skris *) Added "-genkey" option to "dsaparam" program. 1015555714Skris [Eric A. Young] 1015655714Skris 1015755714Skris *) Added RIPE MD160 (rmd160) message digest. 1015855714Skris [Eric A. Young] 1015955714Skris 1016055714Skris *) Added -a (all) option to "ssleay version" command. 1016155714Skris [Eric A. Young] 1016255714Skris 1016355714Skris *) Added PLATFORM define which is the id given to Configure. 1016455714Skris [Eric A. Young] 1016555714Skris 1016655714Skris *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 1016755714Skris [Eric A. Young] 1016855714Skris 1016955714Skris *) Extended the ASN.1 parser routines. 1017055714Skris [Eric A. Young] 1017155714Skris 1017255714Skris *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 1017355714Skris [Eric A. Young] 1017455714Skris 1017555714Skris *) Added a BN_CTX to the BN library. 1017655714Skris [Eric A. Young] 1017755714Skris 1017855714Skris *) Fixed the weak key values in DES library 1017955714Skris [Eric A. Young] 1018055714Skris 1018155714Skris *) Changed API in EVP library for cipher aliases. 1018255714Skris [Eric A. Young] 1018355714Skris 1018455714Skris *) Added support for RC2/64bit cipher. 1018555714Skris [Eric A. Young] 1018655714Skris 1018755714Skris *) Converted the lhash library to the crypto/mem.c functions. 1018855714Skris [Eric A. Young] 1018955714Skris 1019055714Skris *) Added more recognized ASN.1 object ids. 1019155714Skris [Eric A. Young] 1019255714Skris 1019355714Skris *) Added more RSA padding checks for SSL/TLS. 1019455714Skris [Eric A. Young] 1019555714Skris 1019655714Skris *) Added BIO proxy/filter functionality. 1019755714Skris [Eric A. Young] 1019855714Skris 1019955714Skris *) Added extra_certs to SSL_CTX which can be used 1020055714Skris send extra CA certificates to the client in the CA cert chain sending 1020155714Skris process. It can be configured with SSL_CTX_add_extra_chain_cert(). 1020255714Skris [Eric A. Young] 1020355714Skris 1020455714Skris *) Now Fortezza is denied in the authentication phase because 1020555714Skris this is key exchange mechanism is not supported by SSLeay at all. 1020655714Skris [Eric A. Young] 1020755714Skris 1020855714Skris *) Additional PKCS1 checks. 1020955714Skris [Eric A. Young] 1021055714Skris 1021155714Skris *) Support the string "TLSv1" for all TLS v1 ciphers. 1021255714Skris [Eric A. Young] 1021355714Skris 1021455714Skris *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 1021555714Skris ex_data index of the SSL context in the X509_STORE_CTX ex_data. 1021655714Skris [Eric A. Young] 1021755714Skris 1021855714Skris *) Fixed a few memory leaks. 1021955714Skris [Eric A. Young] 1022055714Skris 1022155714Skris *) Fixed various code and comment typos. 1022255714Skris [Eric A. Young] 1022355714Skris 1022455714Skris *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 1022555714Skris bytes sent in the client random. 1022655714Skris [Edward Bishop <ebishop@spyglass.com>] 1022755714Skris 10228