CHANGES revision 246772
155714Skris 255714Skris OpenSSL CHANGES 355714Skris _______________ 455714Skris 5246772Sjkim Changes between 1.0.1d and 1.0.1e [11 Feb 2013] 6246772Sjkim 7246772Sjkim *) 8246772Sjkim 9246772Sjkim Changes between 1.0.1c and 1.0.1d [5 Feb 2013] 10246772Sjkim 11246772Sjkim *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time. 12246772Sjkim 13246772Sjkim This addresses the flaw in CBC record processing discovered by 14246772Sjkim Nadhem Alfardan and Kenny Paterson. Details of this attack can be found 15246772Sjkim at: http://www.isg.rhul.ac.uk/tls/ 16246772Sjkim 17246772Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 18246772Sjkim Security Group at Royal Holloway, University of London 19246772Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and 20246772Sjkim Emilia K�sper for the initial patch. 21246772Sjkim (CVE-2013-0169) 22246772Sjkim [Emilia K�sper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson] 23246772Sjkim 24246772Sjkim *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode 25246772Sjkim ciphersuites which can be exploited in a denial of service attack. 26246772Sjkim Thanks go to and to Adam Langley <agl@chromium.org> for discovering 27246772Sjkim and detecting this bug and to Wolfgang Ettlinger 28246772Sjkim <wolfgang.ettlinger@gmail.com> for independently discovering this issue. 29246772Sjkim (CVE-2012-2686) 30246772Sjkim [Adam Langley] 31246772Sjkim 32246772Sjkim *) Return an error when checking OCSP signatures when key is NULL. 33246772Sjkim This fixes a DoS attack. (CVE-2013-0166) 34246772Sjkim [Steve Henson] 35246772Sjkim 36246772Sjkim *) Make openssl verify return errors. 37246772Sjkim [Chris Palmer <palmer@google.com> and Ben Laurie] 38246772Sjkim 39246772Sjkim *) Call OCSP Stapling callback after ciphersuite has been chosen, so 40246772Sjkim the right response is stapled. Also change SSL_get_certificate() 41246772Sjkim so it returns the certificate actually sent. 42246772Sjkim See http://rt.openssl.org/Ticket/Display.html?id=2836. 43246772Sjkim [Rob Stradling <rob.stradling@comodo.com>] 44246772Sjkim 45246772Sjkim *) Fix possible deadlock when decoding public keys. 46246772Sjkim [Steve Henson] 47246772Sjkim 48246772Sjkim *) Don't use TLS 1.0 record version number in initial client hello 49246772Sjkim if renegotiating. 50246772Sjkim [Steve Henson] 51246772Sjkim 52238405Sjkim Changes between 1.0.1b and 1.0.1c [10 May 2012] 53237657Sjkim 54238405Sjkim *) Sanity check record length before skipping explicit IV in TLS 55238405Sjkim 1.2, 1.1 and DTLS to fix DoS attack. 56237657Sjkim 57237657Sjkim Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic 58237657Sjkim fuzzing as a service testing platform. 59237657Sjkim (CVE-2012-2333) 60237657Sjkim [Steve Henson] 61237657Sjkim 62237657Sjkim *) Initialise tkeylen properly when encrypting CMS messages. 63237657Sjkim Thanks to Solar Designer of Openwall for reporting this issue. 64237657Sjkim [Steve Henson] 65237657Sjkim 66238405Sjkim *) In FIPS mode don't try to use composite ciphers as they are not 67238405Sjkim approved. 68238405Sjkim [Steve Henson] 69237657Sjkim 70238405Sjkim Changes between 1.0.1a and 1.0.1b [26 Apr 2012] 71237657Sjkim 72238405Sjkim *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and 73238405Sjkim 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately 74238405Sjkim mean any application compiled against OpenSSL 1.0.0 headers setting 75238405Sjkim SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng 76238405Sjkim TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to 77238405Sjkim 0x10000000L Any application which was previously compiled against 78238405Sjkim OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1 79238405Sjkim will need to be recompiled as a result. Letting be results in 80238405Sjkim inability to disable specifically TLS 1.1 and in client context, 81238405Sjkim in unlike event, limit maximum offered version to TLS 1.0 [see below]. 82238405Sjkim [Steve Henson] 83237657Sjkim 84238405Sjkim *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not 85238405Sjkim disable just protocol X, but all protocols above X *if* there are 86238405Sjkim protocols *below* X still enabled. In more practical terms it means 87238405Sjkim that if application wants to disable TLS1.0 in favor of TLS1.1 and 88238405Sjkim above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass 89238405Sjkim SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to 90238405Sjkim client side. 91238405Sjkim [Andy Polyakov] 92238405Sjkim 93238405Sjkim Changes between 1.0.1 and 1.0.1a [19 Apr 2012] 94238405Sjkim 95237657Sjkim *) Check for potentially exploitable overflows in asn1_d2i_read_bio 96237657Sjkim BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer 97237657Sjkim in CRYPTO_realloc_clean. 98237657Sjkim 99237657Sjkim Thanks to Tavis Ormandy, Google Security Team, for discovering this 100237657Sjkim issue and to Adam Langley <agl@chromium.org> for fixing it. 101237657Sjkim (CVE-2012-2110) 102237657Sjkim [Adam Langley (Google), Tavis Ormandy, Google Security Team] 103237657Sjkim 104238405Sjkim *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections. 105238405Sjkim [Adam Langley] 106237657Sjkim 107238405Sjkim *) Workarounds for some broken servers that "hang" if a client hello 108238405Sjkim record length exceeds 255 bytes. 109238405Sjkim 110238405Sjkim 1. Do not use record version number > TLS 1.0 in initial client 111238405Sjkim hello: some (but not all) hanging servers will now work. 112238405Sjkim 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate 113238405Sjkim the number of ciphers sent in the client hello. This should be 114238405Sjkim set to an even number, such as 50, for example by passing: 115238405Sjkim -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure. 116238405Sjkim Most broken servers should now work. 117238405Sjkim 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable 118238405Sjkim TLS 1.2 client support entirely. 119238405Sjkim [Steve Henson] 120238405Sjkim 121238405Sjkim *) Fix SEGV in Vector Permutation AES module observed in OpenSSH. 122238405Sjkim [Andy Polyakov] 123238405Sjkim 124238405Sjkim Changes between 1.0.0h and 1.0.1 [14 Mar 2012] 125238405Sjkim 126238405Sjkim *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET 127238405Sjkim STRING form instead of a DigestInfo. 128238405Sjkim [Steve Henson] 129238405Sjkim 130238405Sjkim *) The format used for MDC2 RSA signatures is inconsistent between EVP 131238405Sjkim and the RSA_sign/RSA_verify functions. This was made more apparent when 132238405Sjkim OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular 133238405Sjkim those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect 134238405Sjkim the correct format in RSA_verify so both forms transparently work. 135238405Sjkim [Steve Henson] 136238405Sjkim 137238405Sjkim *) Some servers which support TLS 1.0 can choke if we initially indicate 138238405Sjkim support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA 139238405Sjkim encrypted premaster secret. As a workaround use the maximum pemitted 140238405Sjkim client version in client hello, this should keep such servers happy 141238405Sjkim and still work with previous versions of OpenSSL. 142238405Sjkim [Steve Henson] 143238405Sjkim 144238405Sjkim *) Add support for TLS/DTLS heartbeats. 145238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>] 146238405Sjkim 147238405Sjkim *) Add support for SCTP. 148238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>] 149238405Sjkim 150238405Sjkim *) Improved PRNG seeding for VOS. 151238405Sjkim [Paul Green <Paul.Green@stratus.com>] 152238405Sjkim 153238405Sjkim *) Extensive assembler packs updates, most notably: 154238405Sjkim 155238405Sjkim - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support; 156238405Sjkim - x86[_64]: SSSE3 support (SHA1, vector-permutation AES); 157238405Sjkim - x86_64: bit-sliced AES implementation; 158238405Sjkim - ARM: NEON support, contemporary platforms optimizations; 159238405Sjkim - s390x: z196 support; 160238405Sjkim - *: GHASH and GF(2^m) multiplication implementations; 161238405Sjkim 162238405Sjkim [Andy Polyakov] 163238405Sjkim 164238405Sjkim *) Make TLS-SRP code conformant with RFC 5054 API cleanup 165238405Sjkim (removal of unnecessary code) 166238405Sjkim [Peter Sylvester <peter.sylvester@edelweb.fr>] 167238405Sjkim 168238405Sjkim *) Add TLS key material exporter from RFC 5705. 169238405Sjkim [Eric Rescorla] 170238405Sjkim 171238405Sjkim *) Add DTLS-SRTP negotiation from RFC 5764. 172238405Sjkim [Eric Rescorla] 173238405Sjkim 174238405Sjkim *) Add Next Protocol Negotiation, 175238405Sjkim http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be 176238405Sjkim disabled with a no-npn flag to config or Configure. Code donated 177238405Sjkim by Google. 178238405Sjkim [Adam Langley <agl@google.com> and Ben Laurie] 179238405Sjkim 180238405Sjkim *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224, 181238405Sjkim NIST-P256, NIST-P521, with constant-time single point multiplication on 182238405Sjkim typical inputs. Compiler support for the nonstandard type __uint128_t is 183238405Sjkim required to use this (present in gcc 4.4 and later, for 64-bit builds). 184238405Sjkim Code made available under Apache License version 2.0. 185238405Sjkim 186238405Sjkim Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command 187238405Sjkim line to include this in your build of OpenSSL, and run "make depend" (or 188238405Sjkim "make update"). This enables the following EC_METHODs: 189238405Sjkim 190238405Sjkim EC_GFp_nistp224_method() 191238405Sjkim EC_GFp_nistp256_method() 192238405Sjkim EC_GFp_nistp521_method() 193238405Sjkim 194238405Sjkim EC_GROUP_new_by_curve_name() will automatically use these (while 195238405Sjkim EC_GROUP_new_curve_GFp() currently prefers the more flexible 196238405Sjkim implementations). 197238405Sjkim [Emilia K�sper, Adam Langley, Bodo Moeller (Google)] 198238405Sjkim 199238405Sjkim *) Use type ossl_ssize_t instad of ssize_t which isn't available on 200238405Sjkim all platforms. Move ssize_t definition from e_os.h to the public 201238405Sjkim header file e_os2.h as it now appears in public header file cms.h 202238405Sjkim [Steve Henson] 203238405Sjkim 204238405Sjkim *) New -sigopt option to the ca, req and x509 utilities. Additional 205238405Sjkim signature parameters can be passed using this option and in 206238405Sjkim particular PSS. 207238405Sjkim [Steve Henson] 208238405Sjkim 209238405Sjkim *) Add RSA PSS signing function. This will generate and set the 210238405Sjkim appropriate AlgorithmIdentifiers for PSS based on those in the 211238405Sjkim corresponding EVP_MD_CTX structure. No application support yet. 212238405Sjkim [Steve Henson] 213238405Sjkim 214238405Sjkim *) Support for companion algorithm specific ASN1 signing routines. 215238405Sjkim New function ASN1_item_sign_ctx() signs a pre-initialised 216238405Sjkim EVP_MD_CTX structure and sets AlgorithmIdentifiers based on 217238405Sjkim the appropriate parameters. 218238405Sjkim [Steve Henson] 219238405Sjkim 220238405Sjkim *) Add new algorithm specific ASN1 verification initialisation function 221238405Sjkim to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1 222238405Sjkim handling will be the same no matter what EVP_PKEY_METHOD is used. 223238405Sjkim Add a PSS handler to support verification of PSS signatures: checked 224238405Sjkim against a number of sample certificates. 225238405Sjkim [Steve Henson] 226238405Sjkim 227238405Sjkim *) Add signature printing for PSS. Add PSS OIDs. 228238405Sjkim [Steve Henson, Martin Kaiser <lists@kaiser.cx>] 229238405Sjkim 230238405Sjkim *) Add algorithm specific signature printing. An individual ASN1 method 231238405Sjkim can now print out signatures instead of the standard hex dump. 232238405Sjkim 233238405Sjkim More complex signatures (e.g. PSS) can print out more meaningful 234238405Sjkim information. Include DSA version that prints out the signature 235238405Sjkim parameters r, s. 236238405Sjkim [Steve Henson] 237238405Sjkim 238238405Sjkim *) Password based recipient info support for CMS library: implementing 239238405Sjkim RFC3211. 240238405Sjkim [Steve Henson] 241238405Sjkim 242238405Sjkim *) Split password based encryption into PBES2 and PBKDF2 functions. This 243238405Sjkim neatly separates the code into cipher and PBE sections and is required 244238405Sjkim for some algorithms that split PBES2 into separate pieces (such as 245238405Sjkim password based CMS). 246238405Sjkim [Steve Henson] 247238405Sjkim 248238405Sjkim *) Session-handling fixes: 249238405Sjkim - Fix handling of connections that are resuming with a session ID, 250238405Sjkim but also support Session Tickets. 251238405Sjkim - Fix a bug that suppressed issuing of a new ticket if the client 252238405Sjkim presented a ticket with an expired session. 253238405Sjkim - Try to set the ticket lifetime hint to something reasonable. 254238405Sjkim - Make tickets shorter by excluding irrelevant information. 255238405Sjkim - On the client side, don't ignore renewed tickets. 256238405Sjkim [Adam Langley, Bodo Moeller (Google)] 257238405Sjkim 258238405Sjkim *) Fix PSK session representation. 259238405Sjkim [Bodo Moeller] 260238405Sjkim 261238405Sjkim *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations. 262238405Sjkim 263238405Sjkim This work was sponsored by Intel. 264238405Sjkim [Andy Polyakov] 265238405Sjkim 266238405Sjkim *) Add GCM support to TLS library. Some custom code is needed to split 267238405Sjkim the IV between the fixed (from PRF) and explicit (from TLS record) 268238405Sjkim portions. This adds all GCM ciphersuites supported by RFC5288 and 269238405Sjkim RFC5289. Generalise some AES* cipherstrings to inlclude GCM and 270238405Sjkim add a special AESGCM string for GCM only. 271238405Sjkim [Steve Henson] 272238405Sjkim 273238405Sjkim *) Expand range of ctrls for AES GCM. Permit setting invocation 274238405Sjkim field on decrypt and retrieval of invocation field only on encrypt. 275238405Sjkim [Steve Henson] 276238405Sjkim 277238405Sjkim *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support. 278238405Sjkim As required by RFC5289 these ciphersuites cannot be used if for 279238405Sjkim versions of TLS earlier than 1.2. 280238405Sjkim [Steve Henson] 281238405Sjkim 282238405Sjkim *) For FIPS capable OpenSSL interpret a NULL default public key method 283238405Sjkim as unset and return the appopriate default but do *not* set the default. 284238405Sjkim This means we can return the appopriate method in applications that 285238405Sjkim swicth between FIPS and non-FIPS modes. 286238405Sjkim [Steve Henson] 287238405Sjkim 288238405Sjkim *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an 289238405Sjkim ENGINE is used then we cannot handle that in the FIPS module so we 290238405Sjkim keep original code iff non-FIPS operations are allowed. 291238405Sjkim [Steve Henson] 292238405Sjkim 293238405Sjkim *) Add -attime option to openssl utilities. 294238405Sjkim [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson] 295238405Sjkim 296238405Sjkim *) Redirect DSA and DH operations to FIPS module in FIPS mode. 297238405Sjkim [Steve Henson] 298238405Sjkim 299238405Sjkim *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use 300238405Sjkim FIPS EC methods unconditionally for now. 301238405Sjkim [Steve Henson] 302238405Sjkim 303238405Sjkim *) New build option no-ec2m to disable characteristic 2 code. 304238405Sjkim [Steve Henson] 305238405Sjkim 306238405Sjkim *) Backport libcrypto audit of return value checking from 1.1.0-dev; not 307238405Sjkim all cases can be covered as some introduce binary incompatibilities. 308238405Sjkim [Steve Henson] 309238405Sjkim 310238405Sjkim *) Redirect RSA operations to FIPS module including keygen, 311238405Sjkim encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods. 312238405Sjkim [Steve Henson] 313238405Sjkim 314238405Sjkim *) Add similar low level API blocking to ciphers. 315238405Sjkim [Steve Henson] 316238405Sjkim 317238405Sjkim *) Low level digest APIs are not approved in FIPS mode: any attempt 318238405Sjkim to use these will cause a fatal error. Applications that *really* want 319238405Sjkim to use them can use the private_* version instead. 320238405Sjkim [Steve Henson] 321238405Sjkim 322238405Sjkim *) Redirect cipher operations to FIPS module for FIPS builds. 323238405Sjkim [Steve Henson] 324238405Sjkim 325238405Sjkim *) Redirect digest operations to FIPS module for FIPS builds. 326238405Sjkim [Steve Henson] 327238405Sjkim 328238405Sjkim *) Update build system to add "fips" flag which will link in fipscanister.o 329238405Sjkim for static and shared library builds embedding a signature if needed. 330238405Sjkim [Steve Henson] 331238405Sjkim 332238405Sjkim *) Output TLS supported curves in preference order instead of numerical 333238405Sjkim order. This is currently hardcoded for the highest order curves first. 334238405Sjkim This should be configurable so applications can judge speed vs strength. 335238405Sjkim [Steve Henson] 336238405Sjkim 337238405Sjkim *) Add TLS v1.2 server support for client authentication. 338238405Sjkim [Steve Henson] 339238405Sjkim 340238405Sjkim *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers 341238405Sjkim and enable MD5. 342238405Sjkim [Steve Henson] 343238405Sjkim 344238405Sjkim *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying 345238405Sjkim FIPS modules versions. 346238405Sjkim [Steve Henson] 347238405Sjkim 348238405Sjkim *) Add TLS v1.2 client side support for client authentication. Keep cache 349238405Sjkim of handshake records longer as we don't know the hash algorithm to use 350238405Sjkim until after the certificate request message is received. 351238405Sjkim [Steve Henson] 352238405Sjkim 353238405Sjkim *) Initial TLS v1.2 client support. Add a default signature algorithms 354238405Sjkim extension including all the algorithms we support. Parse new signature 355238405Sjkim format in client key exchange. Relax some ECC signing restrictions for 356238405Sjkim TLS v1.2 as indicated in RFC5246. 357238405Sjkim [Steve Henson] 358238405Sjkim 359238405Sjkim *) Add server support for TLS v1.2 signature algorithms extension. Switch 360238405Sjkim to new signature format when needed using client digest preference. 361238405Sjkim All server ciphersuites should now work correctly in TLS v1.2. No client 362238405Sjkim support yet and no support for client certificates. 363238405Sjkim [Steve Henson] 364238405Sjkim 365238405Sjkim *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch 366238405Sjkim to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based 367238405Sjkim ciphersuites. At present only RSA key exchange ciphersuites work with 368238405Sjkim TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete 369238405Sjkim SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods 370238405Sjkim and version checking. 371238405Sjkim [Steve Henson] 372238405Sjkim 373238405Sjkim *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled 374238405Sjkim with this defined it will not be affected by any changes to ssl internal 375238405Sjkim structures. Add several utility functions to allow openssl application 376238405Sjkim to work with OPENSSL_NO_SSL_INTERN defined. 377238405Sjkim [Steve Henson] 378238405Sjkim 379238405Sjkim *) Add SRP support. 380238405Sjkim [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie] 381238405Sjkim 382238405Sjkim *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id. 383238405Sjkim [Steve Henson] 384238405Sjkim 385238405Sjkim *) Permit abbreviated handshakes when renegotiating using the function 386238405Sjkim SSL_renegotiate_abbreviated(). 387238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>] 388238405Sjkim 389238405Sjkim *) Add call to ENGINE_register_all_complete() to 390238405Sjkim ENGINE_load_builtin_engines(), so some implementations get used 391238405Sjkim automatically instead of needing explicit application support. 392238405Sjkim [Steve Henson] 393238405Sjkim 394238405Sjkim *) Add support for TLS key exporter as described in RFC5705. 395238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson] 396238405Sjkim 397238405Sjkim *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only 398238405Sjkim a few changes are required: 399238405Sjkim 400238405Sjkim Add SSL_OP_NO_TLSv1_1 flag. 401238405Sjkim Add TLSv1_1 methods. 402238405Sjkim Update version checking logic to handle version 1.1. 403238405Sjkim Add explicit IV handling (ported from DTLS code). 404238405Sjkim Add command line options to s_client/s_server. 405238405Sjkim [Steve Henson] 406238405Sjkim 407238405Sjkim Changes between 1.0.0g and 1.0.0h [12 Mar 2012] 408238405Sjkim 409237657Sjkim *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness 410237657Sjkim in CMS and PKCS7 code. When RSA decryption fails use a random key for 411237657Sjkim content decryption and always return the same error. Note: this attack 412237657Sjkim needs on average 2^20 messages so it only affects automated senders. The 413237657Sjkim old behaviour can be reenabled in the CMS code by setting the 414237657Sjkim CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where 415237657Sjkim an MMA defence is not necessary. 416237657Sjkim Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering 417237657Sjkim this issue. (CVE-2012-0884) 418237657Sjkim [Steve Henson] 419237657Sjkim 420237657Sjkim *) Fix CVE-2011-4619: make sure we really are receiving a 421237657Sjkim client hello before rejecting multiple SGC restarts. Thanks to 422237657Sjkim Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug. 423237657Sjkim [Steve Henson] 424237657Sjkim 425238405Sjkim Changes between 1.0.0f and 1.0.0g [18 Jan 2012] 426238405Sjkim 427238405Sjkim *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109. 428238405Sjkim Thanks to Antonio Martin, Enterprise Secure Access Research and 429238405Sjkim Development, Cisco Systems, Inc. for discovering this bug and 430238405Sjkim preparing a fix. (CVE-2012-0050) 431238405Sjkim [Antonio Martin] 432238405Sjkim 433238405Sjkim Changes between 1.0.0e and 1.0.0f [4 Jan 2012] 434238405Sjkim 435238405Sjkim *) Nadhem Alfardan and Kenny Paterson have discovered an extension 436238405Sjkim of the Vaudenay padding oracle attack on CBC mode encryption 437238405Sjkim which enables an efficient plaintext recovery attack against 438238405Sjkim the OpenSSL implementation of DTLS. Their attack exploits timing 439238405Sjkim differences arising during decryption processing. A research 440238405Sjkim paper describing this attack can be found at: 441238405Sjkim http://www.isg.rhul.ac.uk/~kp/dtls.pdf 442238405Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 443238405Sjkim Security Group at Royal Holloway, University of London 444238405Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann 445238405Sjkim <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de> 446238405Sjkim for preparing the fix. (CVE-2011-4108) 447238405Sjkim [Robin Seggelmann, Michael Tuexen] 448238405Sjkim 449238405Sjkim *) Clear bytes used for block padding of SSL 3.0 records. 450238405Sjkim (CVE-2011-4576) 451238405Sjkim [Adam Langley (Google)] 452238405Sjkim 453238405Sjkim *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George 454238405Sjkim Kadianakis <desnacked@gmail.com> for discovering this issue and 455238405Sjkim Adam Langley for preparing the fix. (CVE-2011-4619) 456238405Sjkim [Adam Langley (Google)] 457238405Sjkim 458238405Sjkim *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027) 459238405Sjkim [Andrey Kulikov <amdeich@gmail.com>] 460238405Sjkim 461238405Sjkim *) Prevent malformed RFC3779 data triggering an assertion failure. 462238405Sjkim Thanks to Andrew Chi, BBN Technologies, for discovering the flaw 463238405Sjkim and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577) 464238405Sjkim [Rob Austein <sra@hactrn.net>] 465238405Sjkim 466238405Sjkim *) Improved PRNG seeding for VOS. 467238405Sjkim [Paul Green <Paul.Green@stratus.com>] 468238405Sjkim 469238405Sjkim *) Fix ssl_ciph.c set-up race. 470238405Sjkim [Adam Langley (Google)] 471238405Sjkim 472238405Sjkim *) Fix spurious failures in ecdsatest.c. 473238405Sjkim [Emilia K�sper (Google)] 474238405Sjkim 475238405Sjkim *) Fix the BIO_f_buffer() implementation (which was mixing different 476238405Sjkim interpretations of the '..._len' fields). 477238405Sjkim [Adam Langley (Google)] 478238405Sjkim 479238405Sjkim *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than 480238405Sjkim BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent 481238405Sjkim threads won't reuse the same blinding coefficients. 482238405Sjkim 483238405Sjkim This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING 484238405Sjkim lock to call BN_BLINDING_invert_ex, and avoids one use of 485238405Sjkim BN_BLINDING_update for each BN_BLINDING structure (previously, 486238405Sjkim the last update always remained unused). 487238405Sjkim [Emilia K�sper (Google)] 488238405Sjkim 489238405Sjkim *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf. 490238405Sjkim [Bob Buckholz (Google)] 491238405Sjkim 492238405Sjkim Changes between 1.0.0d and 1.0.0e [6 Sep 2011] 493238405Sjkim 494238405Sjkim *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted 495238405Sjkim by initialising X509_STORE_CTX properly. (CVE-2011-3207) 496238405Sjkim [Kaspar Brand <ossl@velox.ch>] 497238405Sjkim 498238405Sjkim *) Fix SSL memory handling for (EC)DH ciphersuites, in particular 499238405Sjkim for multi-threaded use of ECDH. (CVE-2011-3210) 500238405Sjkim [Adam Langley (Google)] 501238405Sjkim 502238405Sjkim *) Fix x509_name_ex_d2i memory leak on bad inputs. 503238405Sjkim [Bodo Moeller] 504238405Sjkim 505238405Sjkim *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check 506238405Sjkim signature public key algorithm by using OID xref utilities instead. 507238405Sjkim Before this you could only use some ECC ciphersuites with SHA1 only. 508238405Sjkim [Steve Henson] 509238405Sjkim 510238405Sjkim *) Add protection against ECDSA timing attacks as mentioned in the paper 511238405Sjkim by Billy Bob Brumley and Nicola Tuveri, see: 512238405Sjkim 513238405Sjkim http://eprint.iacr.org/2011/232.pdf 514238405Sjkim 515238405Sjkim [Billy Bob Brumley and Nicola Tuveri] 516238405Sjkim 517238405Sjkim Changes between 1.0.0c and 1.0.0d [8 Feb 2011] 518238405Sjkim 519238405Sjkim *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014 520238405Sjkim [Neel Mehta, Adam Langley, Bodo Moeller (Google)] 521238405Sjkim 522238405Sjkim *) Fix bug in string printing code: if *any* escaping is enabled we must 523238405Sjkim escape the escape character (backslash) or the resulting string is 524238405Sjkim ambiguous. 525238405Sjkim [Steve Henson] 526238405Sjkim 527238405Sjkim Changes between 1.0.0b and 1.0.0c [2 Dec 2010] 528238405Sjkim 529238405Sjkim *) Disable code workaround for ancient and obsolete Netscape browsers 530238405Sjkim and servers: an attacker can use it in a ciphersuite downgrade attack. 531238405Sjkim Thanks to Martin Rex for discovering this bug. CVE-2010-4180 532238405Sjkim [Steve Henson] 533238405Sjkim 534238405Sjkim *) Fixed J-PAKE implementation error, originally discovered by 535238405Sjkim Sebastien Martini, further info and confirmation from Stefan 536238405Sjkim Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252 537238405Sjkim [Ben Laurie] 538238405Sjkim 539238405Sjkim Changes between 1.0.0a and 1.0.0b [16 Nov 2010] 540238405Sjkim 541238405Sjkim *) Fix extension code to avoid race conditions which can result in a buffer 542238405Sjkim overrun vulnerability: resumed sessions must not be modified as they can 543238405Sjkim be shared by multiple threads. CVE-2010-3864 544238405Sjkim [Steve Henson] 545238405Sjkim 546238405Sjkim *) Fix WIN32 build system to correctly link an ENGINE directory into 547238405Sjkim a DLL. 548238405Sjkim [Steve Henson] 549238405Sjkim 550238405Sjkim Changes between 1.0.0 and 1.0.0a [01 Jun 2010] 551238405Sjkim 552238405Sjkim *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover 553238405Sjkim (CVE-2010-1633) 554238405Sjkim [Steve Henson, Peter-Michael Hager <hager@dortmund.net>] 555238405Sjkim 556238405Sjkim Changes between 0.9.8n and 1.0.0 [29 Mar 2010] 557238405Sjkim 558238405Sjkim *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher 559238405Sjkim context. The operation can be customised via the ctrl mechanism in 560238405Sjkim case ENGINEs want to include additional functionality. 561238405Sjkim [Steve Henson] 562238405Sjkim 563238405Sjkim *) Tolerate yet another broken PKCS#8 key format: private key value negative. 564238405Sjkim [Steve Henson] 565238405Sjkim 566238405Sjkim *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to 567238405Sjkim output hashes compatible with older versions of OpenSSL. 568238405Sjkim [Willy Weisz <weisz@vcpc.univie.ac.at>] 569238405Sjkim 570238405Sjkim *) Fix compression algorithm handling: if resuming a session use the 571238405Sjkim compression algorithm of the resumed session instead of determining 572238405Sjkim it from client hello again. Don't allow server to change algorithm. 573238405Sjkim [Steve Henson] 574238405Sjkim 575238405Sjkim *) Add load_crls() function to apps tidying load_certs() too. Add option 576238405Sjkim to verify utility to allow additional CRLs to be included. 577238405Sjkim [Steve Henson] 578238405Sjkim 579238405Sjkim *) Update OCSP request code to permit adding custom headers to the request: 580238405Sjkim some responders need this. 581238405Sjkim [Steve Henson] 582238405Sjkim 583238405Sjkim *) The function EVP_PKEY_sign() returns <=0 on error: check return code 584238405Sjkim correctly. 585238405Sjkim [Julia Lawall <julia@diku.dk>] 586238405Sjkim 587238405Sjkim *) Update verify callback code in apps/s_cb.c and apps/verify.c, it 588238405Sjkim needlessly dereferenced structures, used obsolete functions and 589238405Sjkim didn't handle all updated verify codes correctly. 590238405Sjkim [Steve Henson] 591238405Sjkim 592238405Sjkim *) Disable MD2 in the default configuration. 593238405Sjkim [Steve Henson] 594238405Sjkim 595238405Sjkim *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to 596238405Sjkim indicate the initial BIO being pushed or popped. This makes it possible 597238405Sjkim to determine whether the BIO is the one explicitly called or as a result 598238405Sjkim of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so 599238405Sjkim it handles reference counts correctly and doesn't zero out the I/O bio 600238405Sjkim when it is not being explicitly popped. WARNING: applications which 601238405Sjkim included workarounds for the old buggy behaviour will need to be modified 602238405Sjkim or they could free up already freed BIOs. 603238405Sjkim [Steve Henson] 604238405Sjkim 605238405Sjkim *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni 606238405Sjkim renaming to all platforms (within the 0.9.8 branch, this was 607238405Sjkim done conditionally on Netware platforms to avoid a name clash). 608238405Sjkim [Guenter <lists@gknw.net>] 609238405Sjkim 610238405Sjkim *) Add ECDHE and PSK support to DTLS. 611238405Sjkim [Michael Tuexen <tuexen@fh-muenster.de>] 612238405Sjkim 613238405Sjkim *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't 614238405Sjkim be used on C++. 615238405Sjkim [Steve Henson] 616238405Sjkim 617238405Sjkim *) Add "missing" function EVP_MD_flags() (without this the only way to 618238405Sjkim retrieve a digest flags is by accessing the structure directly. Update 619238405Sjkim EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest 620238405Sjkim or cipher is registered as in the "from" argument. Print out all 621238405Sjkim registered digests in the dgst usage message instead of manually 622238405Sjkim attempting to work them out. 623238405Sjkim [Steve Henson] 624238405Sjkim 625238405Sjkim *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello: 626238405Sjkim this allows the use of compression and extensions. Change default cipher 627238405Sjkim string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2 628238405Sjkim by default unless an application cipher string requests it. 629238405Sjkim [Steve Henson] 630238405Sjkim 631238405Sjkim *) Alter match criteria in PKCS12_parse(). It used to try to use local 632238405Sjkim key ids to find matching certificates and keys but some PKCS#12 files 633238405Sjkim don't follow the (somewhat unwritten) rules and this strategy fails. 634238405Sjkim Now just gather all certificates together and the first private key 635238405Sjkim then look for the first certificate that matches the key. 636238405Sjkim [Steve Henson] 637238405Sjkim 638238405Sjkim *) Support use of registered digest and cipher names for dgst and cipher 639238405Sjkim commands instead of having to add each one as a special case. So now 640238405Sjkim you can do: 641238405Sjkim 642238405Sjkim openssl sha256 foo 643238405Sjkim 644238405Sjkim as well as: 645238405Sjkim 646238405Sjkim openssl dgst -sha256 foo 647238405Sjkim 648238405Sjkim and this works for ENGINE based algorithms too. 649238405Sjkim 650238405Sjkim [Steve Henson] 651238405Sjkim 652238405Sjkim *) Update Gost ENGINE to support parameter files. 653238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>] 654238405Sjkim 655238405Sjkim *) Support GeneralizedTime in ca utility. 656238405Sjkim [Oliver Martin <oliver@volatilevoid.net>, Steve Henson] 657238405Sjkim 658238405Sjkim *) Enhance the hash format used for certificate directory links. The new 659238405Sjkim form uses the canonical encoding (meaning equivalent names will work 660238405Sjkim even if they aren't identical) and uses SHA1 instead of MD5. This form 661238405Sjkim is incompatible with the older format and as a result c_rehash should 662238405Sjkim be used to rebuild symbolic links. 663238405Sjkim [Steve Henson] 664238405Sjkim 665238405Sjkim *) Make PKCS#8 the default write format for private keys, replacing the 666238405Sjkim traditional format. This form is standardised, more secure and doesn't 667238405Sjkim include an implicit MD5 dependency. 668238405Sjkim [Steve Henson] 669238405Sjkim 670238405Sjkim *) Add a $gcc_devteam_warn option to Configure. The idea is that any code 671238405Sjkim committed to OpenSSL should pass this lot as a minimum. 672238405Sjkim [Steve Henson] 673238405Sjkim 674238405Sjkim *) Add session ticket override functionality for use by EAP-FAST. 675238405Sjkim [Jouni Malinen <j@w1.fi>] 676238405Sjkim 677238405Sjkim *) Modify HMAC functions to return a value. Since these can be implemented 678238405Sjkim in an ENGINE errors can occur. 679238405Sjkim [Steve Henson] 680238405Sjkim 681238405Sjkim *) Type-checked OBJ_bsearch_ex. 682238405Sjkim [Ben Laurie] 683238405Sjkim 684238405Sjkim *) Type-checked OBJ_bsearch. Also some constification necessitated 685238405Sjkim by type-checking. Still to come: TXT_DB, bsearch(?), 686238405Sjkim OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING, 687238405Sjkim CONF_VALUE. 688238405Sjkim [Ben Laurie] 689238405Sjkim 690238405Sjkim *) New function OPENSSL_gmtime_adj() to add a specific number of days and 691238405Sjkim seconds to a tm structure directly, instead of going through OS 692238405Sjkim specific date routines. This avoids any issues with OS routines such 693238405Sjkim as the year 2038 bug. New *_adj() functions for ASN1 time structures 694238405Sjkim and X509_time_adj_ex() to cover the extended range. The existing 695238405Sjkim X509_time_adj() is still usable and will no longer have any date issues. 696238405Sjkim [Steve Henson] 697238405Sjkim 698238405Sjkim *) Delta CRL support. New use deltas option which will attempt to locate 699238405Sjkim and search any appropriate delta CRLs available. 700238405Sjkim 701238405Sjkim This work was sponsored by Google. 702238405Sjkim [Steve Henson] 703238405Sjkim 704238405Sjkim *) Support for CRLs partitioned by reason code. Reorganise CRL processing 705238405Sjkim code and add additional score elements. Validate alternate CRL paths 706238405Sjkim as part of the CRL checking and indicate a new error "CRL path validation 707238405Sjkim error" in this case. Applications wanting additional details can use 708238405Sjkim the verify callback and check the new "parent" field. If this is not 709238405Sjkim NULL CRL path validation is taking place. Existing applications wont 710238405Sjkim see this because it requires extended CRL support which is off by 711238405Sjkim default. 712238405Sjkim 713238405Sjkim This work was sponsored by Google. 714238405Sjkim [Steve Henson] 715238405Sjkim 716238405Sjkim *) Support for freshest CRL extension. 717238405Sjkim 718238405Sjkim This work was sponsored by Google. 719238405Sjkim [Steve Henson] 720238405Sjkim 721238405Sjkim *) Initial indirect CRL support. Currently only supported in the CRLs 722238405Sjkim passed directly and not via lookup. Process certificate issuer 723238405Sjkim CRL entry extension and lookup CRL entries by bother issuer name 724238405Sjkim and serial number. Check and process CRL issuer entry in IDP extension. 725238405Sjkim 726238405Sjkim This work was sponsored by Google. 727238405Sjkim [Steve Henson] 728238405Sjkim 729238405Sjkim *) Add support for distinct certificate and CRL paths. The CRL issuer 730238405Sjkim certificate is validated separately in this case. Only enabled if 731238405Sjkim an extended CRL support flag is set: this flag will enable additional 732238405Sjkim CRL functionality in future. 733238405Sjkim 734238405Sjkim This work was sponsored by Google. 735238405Sjkim [Steve Henson] 736238405Sjkim 737238405Sjkim *) Add support for policy mappings extension. 738238405Sjkim 739238405Sjkim This work was sponsored by Google. 740238405Sjkim [Steve Henson] 741238405Sjkim 742238405Sjkim *) Fixes to pathlength constraint, self issued certificate handling, 743238405Sjkim policy processing to align with RFC3280 and PKITS tests. 744238405Sjkim 745238405Sjkim This work was sponsored by Google. 746238405Sjkim [Steve Henson] 747238405Sjkim 748238405Sjkim *) Support for name constraints certificate extension. DN, email, DNS 749238405Sjkim and URI types are currently supported. 750238405Sjkim 751238405Sjkim This work was sponsored by Google. 752238405Sjkim [Steve Henson] 753238405Sjkim 754238405Sjkim *) To cater for systems that provide a pointer-based thread ID rather 755238405Sjkim than numeric, deprecate the current numeric thread ID mechanism and 756238405Sjkim replace it with a structure and associated callback type. This 757238405Sjkim mechanism allows a numeric "hash" to be extracted from a thread ID in 758238405Sjkim either case, and on platforms where pointers are larger than 'long', 759238405Sjkim mixing is done to help ensure the numeric 'hash' is usable even if it 760238405Sjkim can't be guaranteed unique. The default mechanism is to use "&errno" 761238405Sjkim as a pointer-based thread ID to distinguish between threads. 762238405Sjkim 763238405Sjkim Applications that want to provide their own thread IDs should now use 764238405Sjkim CRYPTO_THREADID_set_callback() to register a callback that will call 765238405Sjkim either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer(). 766238405Sjkim 767238405Sjkim Note that ERR_remove_state() is now deprecated, because it is tied 768238405Sjkim to the assumption that thread IDs are numeric. ERR_remove_state(0) 769238405Sjkim to free the current thread's error state should be replaced by 770238405Sjkim ERR_remove_thread_state(NULL). 771238405Sjkim 772238405Sjkim (This new approach replaces the functions CRYPTO_set_idptr_callback(), 773238405Sjkim CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in 774238405Sjkim OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an 775238405Sjkim application was previously providing a numeric thread callback that 776238405Sjkim was inappropriate for distinguishing threads, then uniqueness might 777238405Sjkim have been obtained with &errno that happened immediately in the 778238405Sjkim intermediate development versions of OpenSSL; this is no longer the 779238405Sjkim case, the numeric thread callback will now override the automatic use 780238405Sjkim of &errno.) 781238405Sjkim [Geoff Thorpe, with help from Bodo Moeller] 782238405Sjkim 783238405Sjkim *) Initial support for different CRL issuing certificates. This covers a 784238405Sjkim simple case where the self issued certificates in the chain exist and 785238405Sjkim the real CRL issuer is higher in the existing chain. 786238405Sjkim 787238405Sjkim This work was sponsored by Google. 788238405Sjkim [Steve Henson] 789238405Sjkim 790238405Sjkim *) Removed effectively defunct crypto/store from the build. 791238405Sjkim [Ben Laurie] 792238405Sjkim 793238405Sjkim *) Revamp of STACK to provide stronger type-checking. Still to come: 794238405Sjkim TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE, 795238405Sjkim ASN1_STRING, CONF_VALUE. 796238405Sjkim [Ben Laurie] 797238405Sjkim 798238405Sjkim *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer 799238405Sjkim RAM on SSL connections. This option can save about 34k per idle SSL. 800238405Sjkim [Nick Mathewson] 801238405Sjkim 802238405Sjkim *) Revamp of LHASH to provide stronger type-checking. Still to come: 803238405Sjkim STACK, TXT_DB, bsearch, qsort. 804238405Sjkim [Ben Laurie] 805238405Sjkim 806238405Sjkim *) Initial support for Cryptographic Message Syntax (aka CMS) based 807238405Sjkim on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility, 808238405Sjkim support for data, signedData, compressedData, digestedData and 809238405Sjkim encryptedData, envelopedData types included. Scripts to check against 810238405Sjkim RFC4134 examples draft and interop and consistency checks of many 811238405Sjkim content types and variants. 812238405Sjkim [Steve Henson] 813238405Sjkim 814238405Sjkim *) Add options to enc utility to support use of zlib compression BIO. 815238405Sjkim [Steve Henson] 816238405Sjkim 817238405Sjkim *) Extend mk1mf to support importing of options and assembly language 818238405Sjkim files from Configure script, currently only included in VC-WIN32. 819238405Sjkim The assembly language rules can now optionally generate the source 820238405Sjkim files from the associated perl scripts. 821238405Sjkim [Steve Henson] 822238405Sjkim 823238405Sjkim *) Implement remaining functionality needed to support GOST ciphersuites. 824238405Sjkim Interop testing has been performed using CryptoPro implementations. 825238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>] 826238405Sjkim 827238405Sjkim *) s390x assembler pack. 828238405Sjkim [Andy Polyakov] 829238405Sjkim 830238405Sjkim *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU 831238405Sjkim "family." 832238405Sjkim [Andy Polyakov] 833238405Sjkim 834238405Sjkim *) Implement Opaque PRF Input TLS extension as specified in 835238405Sjkim draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an 836238405Sjkim official specification yet and no extension type assignment by 837238405Sjkim IANA exists, this extension (for now) will have to be explicitly 838238405Sjkim enabled when building OpenSSL by providing the extension number 839238405Sjkim to use. For example, specify an option 840238405Sjkim 841238405Sjkim -DTLSEXT_TYPE_opaque_prf_input=0x9527 842238405Sjkim 843238405Sjkim to the "config" or "Configure" script to enable the extension, 844238405Sjkim assuming extension number 0x9527 (which is a completely arbitrary 845238405Sjkim and unofficial assignment based on the MD5 hash of the Internet 846238405Sjkim Draft). Note that by doing so, you potentially lose 847238405Sjkim interoperability with other TLS implementations since these might 848238405Sjkim be using the same extension number for other purposes. 849238405Sjkim 850238405Sjkim SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the 851238405Sjkim opaque PRF input value to use in the handshake. This will create 852238405Sjkim an interal copy of the length-'len' string at 'src', and will 853238405Sjkim return non-zero for success. 854238405Sjkim 855238405Sjkim To get more control and flexibility, provide a callback function 856238405Sjkim by using 857238405Sjkim 858238405Sjkim SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb) 859238405Sjkim SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg) 860238405Sjkim 861238405Sjkim where 862238405Sjkim 863238405Sjkim int (*cb)(SSL *, void *peerinput, size_t len, void *arg); 864238405Sjkim void *arg; 865238405Sjkim 866238405Sjkim Callback function 'cb' will be called in handshakes, and is 867238405Sjkim expected to use SSL_set_tlsext_opaque_prf_input() as appropriate. 868238405Sjkim Argument 'arg' is for application purposes (the value as given to 869238405Sjkim SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly 870238405Sjkim be provided to the callback function). The callback function 871238405Sjkim has to return non-zero to report success: usually 1 to use opaque 872238405Sjkim PRF input just if possible, or 2 to enforce use of the opaque PRF 873238405Sjkim input. In the latter case, the library will abort the handshake 874238405Sjkim if opaque PRF input is not successfully negotiated. 875238405Sjkim 876238405Sjkim Arguments 'peerinput' and 'len' given to the callback function 877238405Sjkim will always be NULL and 0 in the case of a client. A server will 878238405Sjkim see the client's opaque PRF input through these variables if 879238405Sjkim available (NULL and 0 otherwise). Note that if the server 880238405Sjkim provides an opaque PRF input, the length must be the same as the 881238405Sjkim length of the client's opaque PRF input. 882238405Sjkim 883238405Sjkim Note that the callback function will only be called when creating 884238405Sjkim a new session (session resumption can resume whatever was 885238405Sjkim previously negotiated), and will not be called in SSL 2.0 886238405Sjkim handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or 887238405Sjkim SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended 888238405Sjkim for applications that need to enforce opaque PRF input. 889238405Sjkim 890238405Sjkim [Bodo Moeller] 891238405Sjkim 892238405Sjkim *) Update ssl code to support digests other than SHA1+MD5 for handshake 893238405Sjkim MAC. 894238405Sjkim 895238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>] 896238405Sjkim 897238405Sjkim *) Add RFC4507 support to OpenSSL. This includes the corrections in 898238405Sjkim RFC4507bis. The encrypted ticket format is an encrypted encoded 899238405Sjkim SSL_SESSION structure, that way new session features are automatically 900238405Sjkim supported. 901238405Sjkim 902238405Sjkim If a client application caches session in an SSL_SESSION structure 903238405Sjkim support is transparent because tickets are now stored in the encoded 904238405Sjkim SSL_SESSION. 905238405Sjkim 906238405Sjkim The SSL_CTX structure automatically generates keys for ticket 907238405Sjkim protection in servers so again support should be possible 908238405Sjkim with no application modification. 909238405Sjkim 910238405Sjkim If a client or server wishes to disable RFC4507 support then the option 911238405Sjkim SSL_OP_NO_TICKET can be set. 912238405Sjkim 913238405Sjkim Add a TLS extension debugging callback to allow the contents of any client 914238405Sjkim or server extensions to be examined. 915238405Sjkim 916238405Sjkim This work was sponsored by Google. 917238405Sjkim [Steve Henson] 918238405Sjkim 919238405Sjkim *) Final changes to avoid use of pointer pointer casts in OpenSSL. 920238405Sjkim OpenSSL should now compile cleanly on gcc 4.2 921238405Sjkim [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson] 922238405Sjkim 923238405Sjkim *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC 924238405Sjkim support including streaming MAC support: this is required for GOST 925238405Sjkim ciphersuite support. 926238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson] 927238405Sjkim 928238405Sjkim *) Add option -stream to use PKCS#7 streaming in smime utility. New 929238405Sjkim function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream() 930238405Sjkim to output in BER and PEM format. 931238405Sjkim [Steve Henson] 932238405Sjkim 933238405Sjkim *) Experimental support for use of HMAC via EVP_PKEY interface. This 934238405Sjkim allows HMAC to be handled via the EVP_DigestSign*() interface. The 935238405Sjkim EVP_PKEY "key" in this case is the HMAC key, potentially allowing 936238405Sjkim ENGINE support for HMAC keys which are unextractable. New -mac and 937238405Sjkim -macopt options to dgst utility. 938238405Sjkim [Steve Henson] 939238405Sjkim 940238405Sjkim *) New option -sigopt to dgst utility. Update dgst to use 941238405Sjkim EVP_Digest{Sign,Verify}*. These two changes make it possible to use 942238405Sjkim alternative signing paramaters such as X9.31 or PSS in the dgst 943238405Sjkim utility. 944238405Sjkim [Steve Henson] 945238405Sjkim 946238405Sjkim *) Change ssl_cipher_apply_rule(), the internal function that does 947238405Sjkim the work each time a ciphersuite string requests enabling 948238405Sjkim ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or 949238405Sjkim removing ("!foo+bar") a class of ciphersuites: Now it maintains 950238405Sjkim the order of disabled ciphersuites such that those ciphersuites 951238405Sjkim that most recently went from enabled to disabled not only stay 952238405Sjkim in order with respect to each other, but also have higher priority 953238405Sjkim than other disabled ciphersuites the next time ciphersuites are 954238405Sjkim enabled again. 955238405Sjkim 956238405Sjkim This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable 957238405Sjkim the same ciphersuites as with "HIGH" alone, but in a specific 958238405Sjkim order where the PSK ciphersuites come first (since they are the 959238405Sjkim most recently disabled ciphersuites when "HIGH" is parsed). 960238405Sjkim 961238405Sjkim Also, change ssl_create_cipher_list() (using this new 962238405Sjkim funcionality) such that between otherwise identical 963238405Sjkim cihpersuites, ephemeral ECDH is preferred over ephemeral DH in 964238405Sjkim the default order. 965238405Sjkim [Bodo Moeller] 966238405Sjkim 967238405Sjkim *) Change ssl_create_cipher_list() so that it automatically 968238405Sjkim arranges the ciphersuites in reasonable order before starting 969238405Sjkim to process the rule string. Thus, the definition for "DEFAULT" 970238405Sjkim (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but 971238405Sjkim remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH". 972238405Sjkim This makes it much easier to arrive at a reasonable default order 973238405Sjkim in applications for which anonymous ciphers are OK (meaning 974238405Sjkim that you can't actually use DEFAULT). 975238405Sjkim [Bodo Moeller; suggested by Victor Duchovni] 976238405Sjkim 977238405Sjkim *) Split the SSL/TLS algorithm mask (as used for ciphersuite string 978238405Sjkim processing) into multiple integers instead of setting 979238405Sjkim "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK", 980238405Sjkim "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer. 981238405Sjkim (These masks as well as the individual bit definitions are hidden 982238405Sjkim away into the non-exported interface ssl/ssl_locl.h, so this 983238405Sjkim change to the definition of the SSL_CIPHER structure shouldn't 984238405Sjkim affect applications.) This give us more bits for each of these 985238405Sjkim categories, so there is no longer a need to coagulate AES128 and 986238405Sjkim AES256 into a single algorithm bit, and to coagulate Camellia128 987238405Sjkim and Camellia256 into a single algorithm bit, which has led to all 988238405Sjkim kinds of kludges. 989238405Sjkim 990238405Sjkim Thus, among other things, the kludge introduced in 0.9.7m and 991238405Sjkim 0.9.8e for masking out AES256 independently of AES128 or masking 992238405Sjkim out Camellia256 independently of AES256 is not needed here in 0.9.9. 993238405Sjkim 994238405Sjkim With the change, we also introduce new ciphersuite aliases that 995238405Sjkim so far were missing: "AES128", "AES256", "CAMELLIA128", and 996238405Sjkim "CAMELLIA256". 997238405Sjkim [Bodo Moeller] 998238405Sjkim 999238405Sjkim *) Add support for dsa-with-SHA224 and dsa-with-SHA256. 1000238405Sjkim Use the leftmost N bytes of the signature input if the input is 1001238405Sjkim larger than the prime q (with N being the size in bytes of q). 1002238405Sjkim [Nils Larsch] 1003238405Sjkim 1004238405Sjkim *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses 1005238405Sjkim it yet and it is largely untested. 1006238405Sjkim [Steve Henson] 1007238405Sjkim 1008238405Sjkim *) Add support for the ecdsa-with-SHA224/256/384/512 signature types. 1009238405Sjkim [Nils Larsch] 1010238405Sjkim 1011238405Sjkim *) Initial incomplete changes to avoid need for function casts in OpenSSL 1012238405Sjkim some compilers (gcc 4.2 and later) reject their use. Safestack is 1013238405Sjkim reimplemented. Update ASN1 to avoid use of legacy functions. 1014238405Sjkim [Steve Henson] 1015238405Sjkim 1016238405Sjkim *) Win32/64 targets are linked with Winsock2. 1017238405Sjkim [Andy Polyakov] 1018238405Sjkim 1019238405Sjkim *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected 1020238405Sjkim to external functions. This can be used to increase CRL handling 1021238405Sjkim efficiency especially when CRLs are very large by (for example) storing 1022238405Sjkim the CRL revoked certificates in a database. 1023238405Sjkim [Steve Henson] 1024238405Sjkim 1025238405Sjkim *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so 1026238405Sjkim new CRLs added to a directory can be used. New command line option 1027238405Sjkim -verify_return_error to s_client and s_server. This causes real errors 1028238405Sjkim to be returned by the verify callback instead of carrying on no matter 1029238405Sjkim what. This reflects the way a "real world" verify callback would behave. 1030238405Sjkim [Steve Henson] 1031238405Sjkim 1032238405Sjkim *) GOST engine, supporting several GOST algorithms and public key formats. 1033238405Sjkim Kindly donated by Cryptocom. 1034238405Sjkim [Cryptocom] 1035238405Sjkim 1036238405Sjkim *) Partial support for Issuing Distribution Point CRL extension. CRLs 1037238405Sjkim partitioned by DP are handled but no indirect CRL or reason partitioning 1038238405Sjkim (yet). Complete overhaul of CRL handling: now the most suitable CRL is 1039238405Sjkim selected via a scoring technique which handles IDP and AKID in CRLs. 1040238405Sjkim [Steve Henson] 1041238405Sjkim 1042238405Sjkim *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which 1043238405Sjkim will ultimately be used for all verify operations: this will remove the 1044238405Sjkim X509_STORE dependency on certificate verification and allow alternative 1045238405Sjkim lookup methods. X509_STORE based implementations of these two callbacks. 1046238405Sjkim [Steve Henson] 1047238405Sjkim 1048238405Sjkim *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names. 1049238405Sjkim Modify get_crl() to find a valid (unexpired) CRL if possible. 1050238405Sjkim [Steve Henson] 1051238405Sjkim 1052238405Sjkim *) New function X509_CRL_match() to check if two CRLs are identical. Normally 1053238405Sjkim this would be called X509_CRL_cmp() but that name is already used by 1054238405Sjkim a function that just compares CRL issuer names. Cache several CRL 1055238405Sjkim extensions in X509_CRL structure and cache CRLDP in X509. 1056238405Sjkim [Steve Henson] 1057238405Sjkim 1058238405Sjkim *) Store a "canonical" representation of X509_NAME structure (ASN1 Name) 1059238405Sjkim this maps equivalent X509_NAME structures into a consistent structure. 1060238405Sjkim Name comparison can then be performed rapidly using memcmp(). 1061238405Sjkim [Steve Henson] 1062238405Sjkim 1063238405Sjkim *) Non-blocking OCSP request processing. Add -timeout option to ocsp 1064238405Sjkim utility. 1065238405Sjkim [Steve Henson] 1066238405Sjkim 1067238405Sjkim *) Allow digests to supply their own micalg string for S/MIME type using 1068238405Sjkim the ctrl EVP_MD_CTRL_MICALG. 1069238405Sjkim [Steve Henson] 1070238405Sjkim 1071238405Sjkim *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the 1072238405Sjkim EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN 1073238405Sjkim ctrl. It can then customise the structure before and/or after signing 1074238405Sjkim if necessary. 1075238405Sjkim [Steve Henson] 1076238405Sjkim 1077238405Sjkim *) New function OBJ_add_sigid() to allow application defined signature OIDs 1078238405Sjkim to be added to OpenSSLs internal tables. New function OBJ_sigid_free() 1079238405Sjkim to free up any added signature OIDs. 1080238405Sjkim [Steve Henson] 1081238405Sjkim 1082238405Sjkim *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(), 1083238405Sjkim EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal 1084238405Sjkim digest and cipher tables. New options added to openssl utility: 1085238405Sjkim list-message-digest-algorithms and list-cipher-algorithms. 1086238405Sjkim [Steve Henson] 1087238405Sjkim 1088238405Sjkim *) Change the array representation of binary polynomials: the list 1089238405Sjkim of degrees of non-zero coefficients is now terminated with -1. 1090238405Sjkim Previously it was terminated with 0, which was also part of the 1091238405Sjkim value; thus, the array representation was not applicable to 1092238405Sjkim polynomials where t^0 has coefficient zero. This change makes 1093238405Sjkim the array representation useful in a more general context. 1094238405Sjkim [Douglas Stebila] 1095238405Sjkim 1096238405Sjkim *) Various modifications and fixes to SSL/TLS cipher string 1097238405Sjkim handling. For ECC, the code now distinguishes between fixed ECDH 1098238405Sjkim with RSA certificates on the one hand and with ECDSA certificates 1099238405Sjkim on the other hand, since these are separate ciphersuites. The 1100238405Sjkim unused code for Fortezza ciphersuites has been removed. 1101238405Sjkim 1102238405Sjkim For consistency with EDH, ephemeral ECDH is now called "EECDH" 1103238405Sjkim (not "ECDHE"). For consistency with the code for DH 1104238405Sjkim certificates, use of ECDH certificates is now considered ECDH 1105238405Sjkim authentication, not RSA or ECDSA authentication (the latter is 1106238405Sjkim merely the CA's signing algorithm and not actively used in the 1107238405Sjkim protocol). 1108238405Sjkim 1109238405Sjkim The temporary ciphersuite alias "ECCdraft" is no longer 1110238405Sjkim available, and ECC ciphersuites are no longer excluded from "ALL" 1111238405Sjkim and "DEFAULT". The following aliases now exist for RFC 4492 1112238405Sjkim ciphersuites, most of these by analogy with the DH case: 1113238405Sjkim 1114238405Sjkim kECDHr - ECDH cert, signed with RSA 1115238405Sjkim kECDHe - ECDH cert, signed with ECDSA 1116238405Sjkim kECDH - ECDH cert (signed with either RSA or ECDSA) 1117238405Sjkim kEECDH - ephemeral ECDH 1118238405Sjkim ECDH - ECDH cert or ephemeral ECDH 1119238405Sjkim 1120238405Sjkim aECDH - ECDH cert 1121238405Sjkim aECDSA - ECDSA cert 1122238405Sjkim ECDSA - ECDSA cert 1123238405Sjkim 1124238405Sjkim AECDH - anonymous ECDH 1125238405Sjkim EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH") 1126238405Sjkim 1127238405Sjkim [Bodo Moeller] 1128238405Sjkim 1129238405Sjkim *) Add additional S/MIME capabilities for AES and GOST ciphers if supported. 1130238405Sjkim Use correct micalg parameters depending on digest(s) in signed message. 1131238405Sjkim [Steve Henson] 1132238405Sjkim 1133238405Sjkim *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process 1134238405Sjkim an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code. 1135238405Sjkim [Steve Henson] 1136238405Sjkim 1137238405Sjkim *) Initial engine support for EVP_PKEY_METHOD. New functions to permit 1138238405Sjkim an engine to register a method. Add ENGINE lookups for methods and 1139238405Sjkim functional reference processing. 1140238405Sjkim [Steve Henson] 1141238405Sjkim 1142238405Sjkim *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of 1143238405Sjkim EVP_{Sign,Verify}* which allow an application to customise the signature 1144238405Sjkim process. 1145238405Sjkim [Steve Henson] 1146238405Sjkim 1147238405Sjkim *) New -resign option to smime utility. This adds one or more signers 1148238405Sjkim to an existing PKCS#7 signedData structure. Also -md option to use an 1149238405Sjkim alternative message digest algorithm for signing. 1150238405Sjkim [Steve Henson] 1151238405Sjkim 1152238405Sjkim *) Tidy up PKCS#7 routines and add new functions to make it easier to 1153238405Sjkim create PKCS7 structures containing multiple signers. Update smime 1154238405Sjkim application to support multiple signers. 1155238405Sjkim [Steve Henson] 1156238405Sjkim 1157238405Sjkim *) New -macalg option to pkcs12 utility to allow setting of an alternative 1158238405Sjkim digest MAC. 1159238405Sjkim [Steve Henson] 1160238405Sjkim 1161238405Sjkim *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC. 1162238405Sjkim Reorganize PBE internals to lookup from a static table using NIDs, 1163238405Sjkim add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl: 1164238405Sjkim EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative 1165238405Sjkim PRF which will be automatically used with PBES2. 1166238405Sjkim [Steve Henson] 1167238405Sjkim 1168238405Sjkim *) Replace the algorithm specific calls to generate keys in "req" with the 1169238405Sjkim new API. 1170238405Sjkim [Steve Henson] 1171238405Sjkim 1172238405Sjkim *) Update PKCS#7 enveloped data routines to use new API. This is now 1173238405Sjkim supported by any public key method supporting the encrypt operation. A 1174238405Sjkim ctrl is added to allow the public key algorithm to examine or modify 1175238405Sjkim the PKCS#7 RecipientInfo structure if it needs to: for RSA this is 1176238405Sjkim a no op. 1177238405Sjkim [Steve Henson] 1178238405Sjkim 1179238405Sjkim *) Add a ctrl to asn1 method to allow a public key algorithm to express 1180238405Sjkim a default digest type to use. In most cases this will be SHA1 but some 1181238405Sjkim algorithms (such as GOST) need to specify an alternative digest. The 1182238405Sjkim return value indicates how strong the prefernce is 1 means optional and 1183238405Sjkim 2 is mandatory (that is it is the only supported type). Modify 1184238405Sjkim ASN1_item_sign() to accept a NULL digest argument to indicate it should 1185238405Sjkim use the default md. Update openssl utilities to use the default digest 1186238405Sjkim type for signing if it is not explicitly indicated. 1187238405Sjkim [Steve Henson] 1188238405Sjkim 1189238405Sjkim *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New 1190238405Sjkim EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant 1191238405Sjkim signing method from the key type. This effectively removes the link 1192238405Sjkim between digests and public key types. 1193238405Sjkim [Steve Henson] 1194238405Sjkim 1195238405Sjkim *) Add an OID cross reference table and utility functions. Its purpose is to 1196238405Sjkim translate between signature OIDs such as SHA1WithrsaEncryption and SHA1, 1197238405Sjkim rsaEncryption. This will allow some of the algorithm specific hackery 1198238405Sjkim needed to use the correct OID to be removed. 1199238405Sjkim [Steve Henson] 1200238405Sjkim 1201238405Sjkim *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO 1202238405Sjkim structures for PKCS7_sign(). They are now set up by the relevant public 1203238405Sjkim key ASN1 method. 1204238405Sjkim [Steve Henson] 1205238405Sjkim 1206238405Sjkim *) Add provisional EC pkey method with support for ECDSA and ECDH. 1207238405Sjkim [Steve Henson] 1208238405Sjkim 1209238405Sjkim *) Add support for key derivation (agreement) in the API, DH method and 1210238405Sjkim pkeyutl. 1211238405Sjkim [Steve Henson] 1212238405Sjkim 1213238405Sjkim *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support 1214238405Sjkim public and private key formats. As a side effect these add additional 1215238405Sjkim command line functionality not previously available: DSA signatures can be 1216238405Sjkim generated and verified using pkeyutl and DH key support and generation in 1217238405Sjkim pkey, genpkey. 1218238405Sjkim [Steve Henson] 1219238405Sjkim 1220238405Sjkim *) BeOS support. 1221238405Sjkim [Oliver Tappe <zooey@hirschkaefer.de>] 1222238405Sjkim 1223238405Sjkim *) New make target "install_html_docs" installs HTML renditions of the 1224238405Sjkim manual pages. 1225238405Sjkim [Oliver Tappe <zooey@hirschkaefer.de>] 1226238405Sjkim 1227238405Sjkim *) New utility "genpkey" this is analagous to "genrsa" etc except it can 1228238405Sjkim generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to 1229238405Sjkim support key and parameter generation and add initial key generation 1230238405Sjkim functionality for RSA. 1231238405Sjkim [Steve Henson] 1232238405Sjkim 1233238405Sjkim *) Add functions for main EVP_PKEY_method operations. The undocumented 1234238405Sjkim functions EVP_PKEY_{encrypt,decrypt} have been renamed to 1235238405Sjkim EVP_PKEY_{encrypt,decrypt}_old. 1236238405Sjkim [Steve Henson] 1237238405Sjkim 1238238405Sjkim *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public 1239238405Sjkim key API, doesn't do much yet. 1240238405Sjkim [Steve Henson] 1241238405Sjkim 1242238405Sjkim *) New function EVP_PKEY_asn1_get0_info() to retrieve information about 1243238405Sjkim public key algorithms. New option to openssl utility: 1244238405Sjkim "list-public-key-algorithms" to print out info. 1245238405Sjkim [Steve Henson] 1246238405Sjkim 1247238405Sjkim *) Implement the Supported Elliptic Curves Extension for 1248238405Sjkim ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 1249238405Sjkim [Douglas Stebila] 1250238405Sjkim 1251238405Sjkim *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or 1252238405Sjkim EVP_CIPHER structures to avoid later problems in EVP_cleanup(). 1253238405Sjkim [Steve Henson] 1254238405Sjkim 1255238405Sjkim *) New utilities pkey and pkeyparam. These are similar to algorithm specific 1256238405Sjkim utilities such as rsa, dsa, dsaparam etc except they process any key 1257238405Sjkim type. 1258238405Sjkim [Steve Henson] 1259238405Sjkim 1260238405Sjkim *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New 1261238405Sjkim functions EVP_PKEY_print_public(), EVP_PKEY_print_private(), 1262238405Sjkim EVP_PKEY_print_param() to print public key data from an EVP_PKEY 1263238405Sjkim structure. 1264238405Sjkim [Steve Henson] 1265238405Sjkim 1266238405Sjkim *) Initial support for pluggable public key ASN1. 1267238405Sjkim De-spaghettify the public key ASN1 handling. Move public and private 1268238405Sjkim key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate 1269238405Sjkim algorithm specific handling to a single module within the relevant 1270238405Sjkim algorithm directory. Add functions to allow (near) opaque processing 1271238405Sjkim of public and private key structures. 1272238405Sjkim [Steve Henson] 1273238405Sjkim 1274238405Sjkim *) Implement the Supported Point Formats Extension for 1275238405Sjkim ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 1276238405Sjkim [Douglas Stebila] 1277238405Sjkim 1278238405Sjkim *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members 1279238405Sjkim for the psk identity [hint] and the psk callback functions to the 1280238405Sjkim SSL_SESSION, SSL and SSL_CTX structure. 1281238405Sjkim 1282238405Sjkim New ciphersuites: 1283238405Sjkim PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA, 1284238405Sjkim PSK-AES256-CBC-SHA 1285238405Sjkim 1286238405Sjkim New functions: 1287238405Sjkim SSL_CTX_use_psk_identity_hint 1288238405Sjkim SSL_get_psk_identity_hint 1289238405Sjkim SSL_get_psk_identity 1290238405Sjkim SSL_use_psk_identity_hint 1291238405Sjkim 1292238405Sjkim [Mika Kousa and Pasi Eronen of Nokia Corporation] 1293238405Sjkim 1294238405Sjkim *) Add RFC 3161 compliant time stamp request creation, response generation 1295238405Sjkim and response verification functionality. 1296238405Sjkim [Zolt�n Gl�zik <zglozik@opentsa.org>, The OpenTSA Project] 1297238405Sjkim 1298238405Sjkim *) Add initial support for TLS extensions, specifically for the server_name 1299238405Sjkim extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 1300238405Sjkim have new members for a host name. The SSL data structure has an 1301238405Sjkim additional member SSL_CTX *initial_ctx so that new sessions can be 1302238405Sjkim stored in that context to allow for session resumption, even after the 1303238405Sjkim SSL has been switched to a new SSL_CTX in reaction to a client's 1304238405Sjkim server_name extension. 1305238405Sjkim 1306238405Sjkim New functions (subject to change): 1307238405Sjkim 1308238405Sjkim SSL_get_servername() 1309238405Sjkim SSL_get_servername_type() 1310238405Sjkim SSL_set_SSL_CTX() 1311238405Sjkim 1312238405Sjkim New CTRL codes and macros (subject to change): 1313238405Sjkim 1314238405Sjkim SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 1315238405Sjkim - SSL_CTX_set_tlsext_servername_callback() 1316238405Sjkim SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 1317238405Sjkim - SSL_CTX_set_tlsext_servername_arg() 1318238405Sjkim SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 1319238405Sjkim 1320238405Sjkim openssl s_client has a new '-servername ...' option. 1321238405Sjkim 1322238405Sjkim openssl s_server has new options '-servername_host ...', '-cert2 ...', 1323238405Sjkim '-key2 ...', '-servername_fatal' (subject to change). This allows 1324238405Sjkim testing the HostName extension for a specific single host name ('-cert' 1325238405Sjkim and '-key' remain fallbacks for handshakes without HostName 1326238405Sjkim negotiation). If the unrecogninzed_name alert has to be sent, this by 1327238405Sjkim default is a warning; it becomes fatal with the '-servername_fatal' 1328238405Sjkim option. 1329238405Sjkim 1330238405Sjkim [Peter Sylvester, Remy Allais, Christophe Renou] 1331238405Sjkim 1332238405Sjkim *) Whirlpool hash implementation is added. 1333238405Sjkim [Andy Polyakov] 1334238405Sjkim 1335238405Sjkim *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to 1336238405Sjkim bn(64,32). Because of instruction set limitations it doesn't have 1337238405Sjkim any negative impact on performance. This was done mostly in order 1338238405Sjkim to make it possible to share assembler modules, such as bn_mul_mont 1339238405Sjkim implementations, between 32- and 64-bit builds without hassle. 1340238405Sjkim [Andy Polyakov] 1341238405Sjkim 1342238405Sjkim *) Move code previously exiled into file crypto/ec/ec2_smpt.c 1343238405Sjkim to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP 1344238405Sjkim macro. 1345238405Sjkim [Bodo Moeller] 1346238405Sjkim 1347238405Sjkim *) New candidate for BIGNUM assembler implementation, bn_mul_mont, 1348238405Sjkim dedicated Montgomery multiplication procedure, is introduced. 1349238405Sjkim BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher 1350238405Sjkim "64-bit" performance on certain 32-bit targets. 1351238405Sjkim [Andy Polyakov] 1352238405Sjkim 1353238405Sjkim *) New option SSL_OP_NO_COMP to disable use of compression selectively 1354238405Sjkim in SSL structures. New SSL ctrl to set maximum send fragment size. 1355238405Sjkim Save memory by seeting the I/O buffer sizes dynamically instead of 1356238405Sjkim using the maximum available value. 1357238405Sjkim [Steve Henson] 1358238405Sjkim 1359238405Sjkim *) New option -V for 'openssl ciphers'. This prints the ciphersuite code 1360238405Sjkim in addition to the text details. 1361238405Sjkim [Bodo Moeller] 1362238405Sjkim 1363238405Sjkim *) Very, very preliminary EXPERIMENTAL support for printing of general 1364238405Sjkim ASN1 structures. This currently produces rather ugly output and doesn't 1365238405Sjkim handle several customised structures at all. 1366238405Sjkim [Steve Henson] 1367238405Sjkim 1368238405Sjkim *) Integrated support for PVK file format and some related formats such 1369238405Sjkim as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support 1370238405Sjkim these in the 'rsa' and 'dsa' utilities. 1371238405Sjkim [Steve Henson] 1372238405Sjkim 1373238405Sjkim *) Support for PKCS#1 RSAPublicKey format on rsa utility command line. 1374238405Sjkim [Steve Henson] 1375238405Sjkim 1376238405Sjkim *) Remove the ancient ASN1_METHOD code. This was only ever used in one 1377238405Sjkim place for the (very old) "NETSCAPE" format certificates which are now 1378238405Sjkim handled using new ASN1 code equivalents. 1379238405Sjkim [Steve Henson] 1380238405Sjkim 1381238405Sjkim *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD 1382238405Sjkim pointer and make the SSL_METHOD parameter in SSL_CTX_new, 1383238405Sjkim SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'. 1384238405Sjkim [Nils Larsch] 1385238405Sjkim 1386238405Sjkim *) Modify CRL distribution points extension code to print out previously 1387238405Sjkim unsupported fields. Enhance extension setting code to allow setting of 1388238405Sjkim all fields. 1389238405Sjkim [Steve Henson] 1390238405Sjkim 1391238405Sjkim *) Add print and set support for Issuing Distribution Point CRL extension. 1392238405Sjkim [Steve Henson] 1393238405Sjkim 1394238405Sjkim *) Change 'Configure' script to enable Camellia by default. 1395238405Sjkim [NTT] 1396238405Sjkim 1397237657Sjkim Changes between 0.9.8s and 0.9.8t [18 Jan 2012] 1398237657Sjkim 1399237657Sjkim *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109. 1400237657Sjkim Thanks to Antonio Martin, Enterprise Secure Access Research and 1401237657Sjkim Development, Cisco Systems, Inc. for discovering this bug and 1402237657Sjkim preparing a fix. (CVE-2012-0050) 1403237657Sjkim [Antonio Martin] 1404238405Sjkim 1405237657Sjkim Changes between 0.9.8r and 0.9.8s [4 Jan 2012] 1406237657Sjkim 1407237657Sjkim *) Nadhem Alfardan and Kenny Paterson have discovered an extension 1408237657Sjkim of the Vaudenay padding oracle attack on CBC mode encryption 1409237657Sjkim which enables an efficient plaintext recovery attack against 1410237657Sjkim the OpenSSL implementation of DTLS. Their attack exploits timing 1411237657Sjkim differences arising during decryption processing. A research 1412237657Sjkim paper describing this attack can be found at: 1413237657Sjkim http://www.isg.rhul.ac.uk/~kp/dtls.pdf 1414237657Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 1415237657Sjkim Security Group at Royal Holloway, University of London 1416237657Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann 1417237657Sjkim <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de> 1418237657Sjkim for preparing the fix. (CVE-2011-4108) 1419237657Sjkim [Robin Seggelmann, Michael Tuexen] 1420237657Sjkim 1421237657Sjkim *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109) 1422237657Sjkim [Ben Laurie, Kasper <ekasper@google.com>] 1423237657Sjkim 1424237657Sjkim *) Clear bytes used for block padding of SSL 3.0 records. 1425237657Sjkim (CVE-2011-4576) 1426237657Sjkim [Adam Langley (Google)] 1427237657Sjkim 1428237657Sjkim *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George 1429237657Sjkim Kadianakis <desnacked@gmail.com> for discovering this issue and 1430237657Sjkim Adam Langley for preparing the fix. (CVE-2011-4619) 1431237657Sjkim [Adam Langley (Google)] 1432237657Sjkim 1433237657Sjkim *) Prevent malformed RFC3779 data triggering an assertion failure. 1434237657Sjkim Thanks to Andrew Chi, BBN Technologies, for discovering the flaw 1435237657Sjkim and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577) 1436237657Sjkim [Rob Austein <sra@hactrn.net>] 1437237657Sjkim 1438237657Sjkim *) Fix ssl_ciph.c set-up race. 1439237657Sjkim [Adam Langley (Google)] 1440237657Sjkim 1441237657Sjkim *) Fix spurious failures in ecdsatest.c. 1442237657Sjkim [Emilia K�sper (Google)] 1443237657Sjkim 1444237657Sjkim *) Fix the BIO_f_buffer() implementation (which was mixing different 1445237657Sjkim interpretations of the '..._len' fields). 1446237657Sjkim [Adam Langley (Google)] 1447237657Sjkim 1448237657Sjkim *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than 1449237657Sjkim BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent 1450237657Sjkim threads won't reuse the same blinding coefficients. 1451237657Sjkim 1452237657Sjkim This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING 1453237657Sjkim lock to call BN_BLINDING_invert_ex, and avoids one use of 1454237657Sjkim BN_BLINDING_update for each BN_BLINDING structure (previously, 1455237657Sjkim the last update always remained unused). 1456237657Sjkim [Emilia K�sper (Google)] 1457237657Sjkim 1458237657Sjkim *) Fix SSL memory handling for (EC)DH ciphersuites, in particular 1459237657Sjkim for multi-threaded use of ECDH. 1460237657Sjkim [Adam Langley (Google)] 1461237657Sjkim 1462237657Sjkim *) Fix x509_name_ex_d2i memory leak on bad inputs. 1463237657Sjkim [Bodo Moeller] 1464237657Sjkim 1465237657Sjkim *) Add protection against ECDSA timing attacks as mentioned in the paper 1466237657Sjkim by Billy Bob Brumley and Nicola Tuveri, see: 1467237657Sjkim 1468237657Sjkim http://eprint.iacr.org/2011/232.pdf 1469237657Sjkim 1470237657Sjkim [Billy Bob Brumley and Nicola Tuveri] 1471237657Sjkim 1472237657Sjkim Changes between 0.9.8q and 0.9.8r [8 Feb 2011] 1473237657Sjkim 1474237657Sjkim *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014 1475237657Sjkim [Neel Mehta, Adam Langley, Bodo Moeller (Google)] 1476237657Sjkim 1477237657Sjkim *) Fix bug in string printing code: if *any* escaping is enabled we must 1478237657Sjkim escape the escape character (backslash) or the resulting string is 1479237657Sjkim ambiguous. 1480237657Sjkim [Steve Henson] 1481237657Sjkim 1482216166Ssimon Changes between 0.9.8p and 0.9.8q [2 Dec 2010] 1483216166Ssimon 1484216166Ssimon *) Disable code workaround for ancient and obsolete Netscape browsers 1485216166Ssimon and servers: an attacker can use it in a ciphersuite downgrade attack. 1486216166Ssimon Thanks to Martin Rex for discovering this bug. CVE-2010-4180 1487216166Ssimon [Steve Henson] 1488216166Ssimon 1489216166Ssimon *) Fixed J-PAKE implementation error, originally discovered by 1490216166Ssimon Sebastien Martini, further info and confirmation from Stefan 1491216166Ssimon Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252 1492216166Ssimon [Ben Laurie] 1493216166Ssimon 1494215697Ssimon Changes between 0.9.8o and 0.9.8p [16 Nov 2010] 1495215697Ssimon 1496215697Ssimon *) Fix extension code to avoid race conditions which can result in a buffer 1497215697Ssimon overrun vulnerability: resumed sessions must not be modified as they can 1498215697Ssimon be shared by multiple threads. CVE-2010-3864 1499215697Ssimon [Steve Henson] 1500215697Ssimon 1501215697Ssimon *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939 1502215697Ssimon [Steve Henson] 1503215697Ssimon 1504215697Ssimon *) Don't reencode certificate when calculating signature: cache and use 1505215697Ssimon the original encoding instead. This makes signature verification of 1506215697Ssimon some broken encodings work correctly. 1507215697Ssimon [Steve Henson] 1508215697Ssimon 1509215697Ssimon *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT 1510215697Ssimon is also one of the inputs. 1511215697Ssimon [Emilia K�sper <emilia.kasper@esat.kuleuven.be> (Google)] 1512215697Ssimon 1513215697Ssimon *) Don't repeatedly append PBE algorithms to table if they already exist. 1514215697Ssimon Sort table on each new add. This effectively makes the table read only 1515215697Ssimon after all algorithms are added and subsequent calls to PKCS12_pbe_add 1516215697Ssimon etc are non-op. 1517215697Ssimon [Steve Henson] 1518215697Ssimon 1519215697Ssimon Changes between 0.9.8n and 0.9.8o [01 Jun 2010] 1520215697Ssimon 1521215697Ssimon [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after 1522215697Ssimon OpenSSL 1.0.0.] 1523215697Ssimon 1524215697Ssimon *) Correct a typo in the CMS ASN1 module which can result in invalid memory 1525215697Ssimon access or freeing data twice (CVE-2010-0742) 1526215697Ssimon [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>] 1527215697Ssimon 1528215697Ssimon *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more 1529215697Ssimon common in certificates and some applications which only call 1530215697Ssimon SSL_library_init and not OpenSSL_add_all_algorithms() will fail. 1531215697Ssimon [Steve Henson] 1532215697Ssimon 1533215697Ssimon *) VMS fixes: 1534215697Ssimon Reduce copying into .apps and .test in makevms.com 1535215697Ssimon Don't try to use blank CA certificate in CA.com 1536215697Ssimon Allow use of C files from original directories in maketests.com 1537215697Ssimon [Steven M. Schweda" <sms@antinode.info>] 1538215697Ssimon 1539206046Ssimon Changes between 0.9.8m and 0.9.8n [24 Mar 2010] 1540206046Ssimon 1541206046Ssimon *) When rejecting SSL/TLS records due to an incorrect version number, never 1542206046Ssimon update s->server with a new major version number. As of 1543206046Ssimon - OpenSSL 0.9.8m if 'short' is a 16-bit type, 1544206046Ssimon - OpenSSL 0.9.8f if 'short' is longer than 16 bits, 1545206046Ssimon the previous behavior could result in a read attempt at NULL when 1546206046Ssimon receiving specific incorrect SSL/TLS records once record payload 1547206046Ssimon protection is active. (CVE-2010-0740) 1548206046Ssimon [Bodo Moeller, Adam Langley <agl@chromium.org>] 1549206046Ssimon 1550206046Ssimon *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL 1551206046Ssimon could be crashed if the relevant tables were not present (e.g. chrooted). 1552206046Ssimon [Tomas Hoger <thoger@redhat.com>] 1553206046Ssimon 1554205128Ssimon Changes between 0.9.8l and 0.9.8m [25 Feb 2010] 1555205128Ssimon 1556205128Ssimon *) Always check bn_wexpend() return values for failure. (CVE-2009-3245) 1557205128Ssimon [Martin Olsson, Neel Mehta] 1558205128Ssimon 1559205128Ssimon *) Fix X509_STORE locking: Every 'objs' access requires a lock (to 1560205128Ssimon accommodate for stack sorting, always a write lock!). 1561205128Ssimon [Bodo Moeller] 1562205128Ssimon 1563205128Ssimon *) On some versions of WIN32 Heap32Next is very slow. This can cause 1564205128Ssimon excessive delays in the RAND_poll(): over a minute. As a workaround 1565205128Ssimon include a time check in the inner Heap32Next loop too. 1566205128Ssimon [Steve Henson] 1567205128Ssimon 1568205128Ssimon *) The code that handled flushing of data in SSL/TLS originally used the 1569205128Ssimon BIO_CTRL_INFO ctrl to see if any data was pending first. This caused 1570205128Ssimon the problem outlined in PR#1949. The fix suggested there however can 1571205128Ssimon trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions 1572205128Ssimon of Apache). So instead simplify the code to flush unconditionally. 1573205128Ssimon This should be fine since flushing with no data to flush is a no op. 1574205128Ssimon [Steve Henson] 1575205128Ssimon 1576205128Ssimon *) Handle TLS versions 2.0 and later properly and correctly use the 1577205128Ssimon highest version of TLS/SSL supported. Although TLS >= 2.0 is some way 1578205128Ssimon off ancient servers have a habit of sticking around for a while... 1579205128Ssimon [Steve Henson] 1580205128Ssimon 1581205128Ssimon *) Modify compression code so it frees up structures without using the 1582205128Ssimon ex_data callbacks. This works around a problem where some applications 1583205128Ssimon call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when 1584205128Ssimon restarting) then use compression (e.g. SSL with compression) later. 1585205128Ssimon This results in significant per-connection memory leaks and 1586205128Ssimon has caused some security issues including CVE-2008-1678 and 1587205128Ssimon CVE-2009-4355. 1588205128Ssimon [Steve Henson] 1589205128Ssimon 1590205128Ssimon *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't 1591205128Ssimon change when encrypting or decrypting. 1592205128Ssimon [Bodo Moeller] 1593205128Ssimon 1594205128Ssimon *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to 1595205128Ssimon connect and renegotiate with servers which do not support RI. 1596205128Ssimon Until RI is more widely deployed this option is enabled by default. 1597205128Ssimon [Steve Henson] 1598205128Ssimon 1599205128Ssimon *) Add "missing" ssl ctrls to clear options and mode. 1600205128Ssimon [Steve Henson] 1601205128Ssimon 1602205128Ssimon *) If client attempts to renegotiate and doesn't support RI respond with 1603205128Ssimon a no_renegotiation alert as required by RFC5746. Some renegotiating 1604205128Ssimon TLS clients will continue a connection gracefully when they receive 1605205128Ssimon the alert. Unfortunately OpenSSL mishandled this alert and would hang 1606205128Ssimon waiting for a server hello which it will never receive. Now we treat a 1607205128Ssimon received no_renegotiation alert as a fatal error. This is because 1608205128Ssimon applications requesting a renegotiation might well expect it to succeed 1609205128Ssimon and would have no code in place to handle the server denying it so the 1610205128Ssimon only safe thing to do is to terminate the connection. 1611205128Ssimon [Steve Henson] 1612205128Ssimon 1613205128Ssimon *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if 1614205128Ssimon peer supports secure renegotiation and 0 otherwise. Print out peer 1615205128Ssimon renegotiation support in s_client/s_server. 1616205128Ssimon [Steve Henson] 1617205128Ssimon 1618205128Ssimon *) Replace the highly broken and deprecated SPKAC certification method with 1619205128Ssimon the updated NID creation version. This should correctly handle UTF8. 1620205128Ssimon [Steve Henson] 1621205128Ssimon 1622205128Ssimon *) Implement RFC5746. Re-enable renegotiation but require the extension 1623205128Ssimon as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION 1624205128Ssimon turns out to be a bad idea. It has been replaced by 1625205128Ssimon SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with 1626205128Ssimon SSL_CTX_set_options(). This is really not recommended unless you 1627205128Ssimon know what you are doing. 1628205128Ssimon [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson] 1629205128Ssimon 1630205128Ssimon *) Fixes to stateless session resumption handling. Use initial_ctx when 1631205128Ssimon issuing and attempting to decrypt tickets in case it has changed during 1632205128Ssimon servername handling. Use a non-zero length session ID when attempting 1633205128Ssimon stateless session resumption: this makes it possible to determine if 1634205128Ssimon a resumption has occurred immediately after receiving server hello 1635205128Ssimon (several places in OpenSSL subtly assume this) instead of later in 1636205128Ssimon the handshake. 1637205128Ssimon [Steve Henson] 1638205128Ssimon 1639205128Ssimon *) The functions ENGINE_ctrl(), OPENSSL_isservice(), 1640205128Ssimon CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error 1641205128Ssimon fixes for a few places where the return code is not checked 1642205128Ssimon correctly. 1643205128Ssimon [Julia Lawall <julia@diku.dk>] 1644205128Ssimon 1645205128Ssimon *) Add --strict-warnings option to Configure script to include devteam 1646205128Ssimon warnings in other configurations. 1647205128Ssimon [Steve Henson] 1648205128Ssimon 1649205128Ssimon *) Add support for --libdir option and LIBDIR variable in makefiles. This 1650205128Ssimon makes it possible to install openssl libraries in locations which 1651205128Ssimon have names other than "lib", for example "/usr/lib64" which some 1652205128Ssimon systems need. 1653205128Ssimon [Steve Henson, based on patch from Jeremy Utley] 1654205128Ssimon 1655205128Ssimon *) Don't allow the use of leading 0x80 in OIDs. This is a violation of 1656205128Ssimon X690 8.9.12 and can produce some misleading textual output of OIDs. 1657205128Ssimon [Steve Henson, reported by Dan Kaminsky] 1658205128Ssimon 1659205128Ssimon *) Delete MD2 from algorithm tables. This follows the recommendation in 1660205128Ssimon several standards that it is not used in new applications due to 1661205128Ssimon several cryptographic weaknesses. For binary compatibility reasons 1662205128Ssimon the MD2 API is still compiled in by default. 1663205128Ssimon [Steve Henson] 1664205128Ssimon 1665205128Ssimon *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved 1666205128Ssimon and restored. 1667205128Ssimon [Steve Henson] 1668205128Ssimon 1669205128Ssimon *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and 1670205128Ssimon OPENSSL_asc2uni conditionally on Netware platforms to avoid a name 1671205128Ssimon clash. 1672205128Ssimon [Guenter <lists@gknw.net>] 1673205128Ssimon 1674205128Ssimon *) Fix the server certificate chain building code to use X509_verify_cert(), 1675205128Ssimon it used to have an ad-hoc builder which was unable to cope with anything 1676205128Ssimon other than a simple chain. 1677205128Ssimon [David Woodhouse <dwmw2@infradead.org>, Steve Henson] 1678205128Ssimon 1679205128Ssimon *) Don't check self signed certificate signatures in X509_verify_cert() 1680205128Ssimon by default (a flag can override this): it just wastes time without 1681205128Ssimon adding any security. As a useful side effect self signed root CAs 1682205128Ssimon with non-FIPS digests are now usable in FIPS mode. 1683205128Ssimon [Steve Henson] 1684205128Ssimon 1685205128Ssimon *) In dtls1_process_out_of_seq_message() the check if the current message 1686205128Ssimon is already buffered was missing. For every new message was memory 1687205128Ssimon allocated, allowing an attacker to perform an denial of service attack 1688205128Ssimon with sending out of seq handshake messages until there is no memory 1689205128Ssimon left. Additionally every future messege was buffered, even if the 1690205128Ssimon sequence number made no sense and would be part of another handshake. 1691205128Ssimon So only messages with sequence numbers less than 10 in advance will be 1692205128Ssimon buffered. (CVE-2009-1378) 1693205128Ssimon [Robin Seggelmann, discovered by Daniel Mentz] 1694205128Ssimon 1695205128Ssimon *) Records are buffered if they arrive with a future epoch to be 1696205128Ssimon processed after finishing the corresponding handshake. There is 1697205128Ssimon currently no limitation to this buffer allowing an attacker to perform 1698205128Ssimon a DOS attack with sending records with future epochs until there is no 1699205128Ssimon memory left. This patch adds the pqueue_size() function to detemine 1700205128Ssimon the size of a buffer and limits the record buffer to 100 entries. 1701205128Ssimon (CVE-2009-1377) 1702205128Ssimon [Robin Seggelmann, discovered by Daniel Mentz] 1703205128Ssimon 1704205128Ssimon *) Keep a copy of frag->msg_header.frag_len so it can be used after the 1705205128Ssimon parent structure is freed. (CVE-2009-1379) 1706205128Ssimon [Daniel Mentz] 1707205128Ssimon 1708205128Ssimon *) Handle non-blocking I/O properly in SSL_shutdown() call. 1709205128Ssimon [Darryl Miles <darryl-mailinglists@netbauds.net>] 1710205128Ssimon 1711205128Ssimon *) Add 2.5.4.* OIDs 1712205128Ssimon [Ilya O. <vrghost@gmail.com>] 1713205128Ssimon 1714205128Ssimon Changes between 0.9.8k and 0.9.8l [5 Nov 2009] 1715205128Ssimon 1716205128Ssimon *) Disable renegotiation completely - this fixes a severe security 1717205128Ssimon problem (CVE-2009-3555) at the cost of breaking all 1718205128Ssimon renegotiation. Renegotiation can be re-enabled by setting 1719205128Ssimon SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at 1720205128Ssimon run-time. This is really not recommended unless you know what 1721205128Ssimon you're doing. 1722205128Ssimon [Ben Laurie] 1723205128Ssimon 1724194206Ssimon Changes between 0.9.8j and 0.9.8k [25 Mar 2009] 1725167612Ssimon 1726194206Ssimon *) Don't set val to NULL when freeing up structures, it is freed up by 1727194206Ssimon underlying code. If sizeof(void *) > sizeof(long) this can result in 1728194206Ssimon zeroing past the valid field. (CVE-2009-0789) 1729194206Ssimon [Paolo Ganci <Paolo.Ganci@AdNovum.CH>] 1730194206Ssimon 1731194206Ssimon *) Fix bug where return value of CMS_SignerInfo_verify_content() was not 1732194206Ssimon checked correctly. This would allow some invalid signed attributes to 1733194206Ssimon appear to verify correctly. (CVE-2009-0591) 1734194206Ssimon [Ivan Nestlerode <inestlerode@us.ibm.com>] 1735194206Ssimon 1736194206Ssimon *) Reject UniversalString and BMPString types with invalid lengths. This 1737194206Ssimon prevents a crash in ASN1_STRING_print_ex() which assumes the strings have 1738194206Ssimon a legal length. (CVE-2009-0590) 1739194206Ssimon [Steve Henson] 1740194206Ssimon 1741194206Ssimon *) Set S/MIME signing as the default purpose rather than setting it 1742194206Ssimon unconditionally. This allows applications to override it at the store 1743194206Ssimon level. 1744194206Ssimon [Steve Henson] 1745194206Ssimon 1746194206Ssimon *) Permit restricted recursion of ASN1 strings. This is needed in practice 1747194206Ssimon to handle some structures. 1748194206Ssimon [Steve Henson] 1749194206Ssimon 1750194206Ssimon *) Improve efficiency of mem_gets: don't search whole buffer each time 1751194206Ssimon for a '\n' 1752194206Ssimon [Jeremy Shapiro <jnshapir@us.ibm.com>] 1753194206Ssimon 1754194206Ssimon *) New -hex option for openssl rand. 1755194206Ssimon [Matthieu Herrb] 1756194206Ssimon 1757194206Ssimon *) Print out UTF8String and NumericString when parsing ASN1. 1758194206Ssimon [Steve Henson] 1759194206Ssimon 1760194206Ssimon *) Support NumericString type for name components. 1761194206Ssimon [Steve Henson] 1762194206Ssimon 1763194206Ssimon *) Allow CC in the environment to override the automatically chosen 1764194206Ssimon compiler. Note that nothing is done to ensure flags work with the 1765194206Ssimon chosen compiler. 1766194206Ssimon [Ben Laurie] 1767194206Ssimon 1768194206Ssimon Changes between 0.9.8i and 0.9.8j [07 Jan 2009] 1769194206Ssimon 1770194206Ssimon *) Properly check EVP_VerifyFinal() and similar return values 1771194206Ssimon (CVE-2008-5077). 1772194206Ssimon [Ben Laurie, Bodo Moeller, Google Security Team] 1773194206Ssimon 1774194206Ssimon *) Enable TLS extensions by default. 1775194206Ssimon [Ben Laurie] 1776194206Ssimon 1777194206Ssimon *) Allow the CHIL engine to be loaded, whether the application is 1778194206Ssimon multithreaded or not. (This does not release the developer from the 1779194206Ssimon obligation to set up the dynamic locking callbacks.) 1780194206Ssimon [Sander Temme <sander@temme.net>] 1781194206Ssimon 1782194206Ssimon *) Use correct exit code if there is an error in dgst command. 1783194206Ssimon [Steve Henson; problem pointed out by Roland Dirlewanger] 1784194206Ssimon 1785194206Ssimon *) Tweak Configure so that you need to say "experimental-jpake" to enable 1786194206Ssimon JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications. 1787194206Ssimon [Bodo Moeller] 1788194206Ssimon 1789194206Ssimon *) Add experimental JPAKE support, including demo authentication in 1790194206Ssimon s_client and s_server. 1791194206Ssimon [Ben Laurie] 1792194206Ssimon 1793194206Ssimon *) Set the comparison function in v3_addr_canonize(). 1794194206Ssimon [Rob Austein <sra@hactrn.net>] 1795194206Ssimon 1796194206Ssimon *) Add support for XMPP STARTTLS in s_client. 1797194206Ssimon [Philip Paeps <philip@freebsd.org>] 1798194206Ssimon 1799194206Ssimon *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior 1800194206Ssimon to ensure that even with this option, only ciphersuites in the 1801194206Ssimon server's preference list will be accepted. (Note that the option 1802194206Ssimon applies only when resuming a session, so the earlier behavior was 1803194206Ssimon just about the algorithm choice for symmetric cryptography.) 1804194206Ssimon [Bodo Moeller] 1805194206Ssimon 1806194206Ssimon Changes between 0.9.8h and 0.9.8i [15 Sep 2008] 1807194206Ssimon 1808205128Ssimon *) Fix NULL pointer dereference if a DTLS server received 1809205128Ssimon ChangeCipherSpec as first record (CVE-2009-1386). 1810205128Ssimon [PR #1679] 1811205128Ssimon 1812194206Ssimon *) Fix a state transitition in s3_srvr.c and d1_srvr.c 1813194206Ssimon (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...). 1814194206Ssimon [Nagendra Modadugu] 1815194206Ssimon 1816194206Ssimon *) The fix in 0.9.8c that supposedly got rid of unsafe 1817194206Ssimon double-checked locking was incomplete for RSA blinding, 1818194206Ssimon addressing just one layer of what turns out to have been 1819194206Ssimon doubly unsafe triple-checked locking. 1820194206Ssimon 1821194206Ssimon So now fix this for real by retiring the MONT_HELPER macro 1822194206Ssimon in crypto/rsa/rsa_eay.c. 1823194206Ssimon 1824194206Ssimon [Bodo Moeller; problem pointed out by Marius Schilder] 1825194206Ssimon 1826194206Ssimon *) Various precautionary measures: 1827194206Ssimon 1828194206Ssimon - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h). 1829194206Ssimon 1830194206Ssimon - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c). 1831194206Ssimon (NB: This would require knowledge of the secret session ticket key 1832194206Ssimon to exploit, in which case you'd be SOL either way.) 1833194206Ssimon 1834194206Ssimon - Change bn_nist.c so that it will properly handle input BIGNUMs 1835194206Ssimon outside the expected range. 1836194206Ssimon 1837194206Ssimon - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG 1838194206Ssimon builds. 1839194206Ssimon 1840194206Ssimon [Neel Mehta, Bodo Moeller] 1841194206Ssimon 1842194206Ssimon *) Allow engines to be "soft loaded" - i.e. optionally don't die if 1843194206Ssimon the load fails. Useful for distros. 1844194206Ssimon [Ben Laurie and the FreeBSD team] 1845194206Ssimon 1846194206Ssimon *) Add support for Local Machine Keyset attribute in PKCS#12 files. 1847194206Ssimon [Steve Henson] 1848194206Ssimon 1849194206Ssimon *) Fix BN_GF2m_mod_arr() top-bit cleanup code. 1850194206Ssimon [Huang Ying] 1851194206Ssimon 1852194206Ssimon *) Expand ENGINE to support engine supplied SSL client certificate functions. 1853194206Ssimon 1854194206Ssimon This work was sponsored by Logica. 1855194206Ssimon [Steve Henson] 1856194206Ssimon 1857194206Ssimon *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows 1858194206Ssimon keystores. Support for SSL/TLS client authentication too. 1859194206Ssimon Not compiled unless enable-capieng specified to Configure. 1860194206Ssimon 1861194206Ssimon This work was sponsored by Logica. 1862194206Ssimon [Steve Henson] 1863194206Ssimon 1864194206Ssimon *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using 1865194206Ssimon ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain 1866194206Ssimon attribute creation routines such as certifcate requests and PKCS#12 1867194206Ssimon files. 1868194206Ssimon [Steve Henson] 1869194206Ssimon 1870194206Ssimon Changes between 0.9.8g and 0.9.8h [28 May 2008] 1871194206Ssimon 1872194206Ssimon *) Fix flaw if 'Server Key exchange message' is omitted from a TLS 1873194206Ssimon handshake which could lead to a cilent crash as found using the 1874194206Ssimon Codenomicon TLS test suite (CVE-2008-1672) 1875194206Ssimon [Steve Henson, Mark Cox] 1876194206Ssimon 1877194206Ssimon *) Fix double free in TLS server name extensions which could lead to 1878194206Ssimon a remote crash found by Codenomicon TLS test suite (CVE-2008-0891) 1879194206Ssimon [Joe Orton] 1880194206Ssimon 1881194206Ssimon *) Clear error queue in SSL_CTX_use_certificate_chain_file() 1882194206Ssimon 1883194206Ssimon Clear the error queue to ensure that error entries left from 1884194206Ssimon older function calls do not interfere with the correct operation. 1885194206Ssimon [Lutz Jaenicke, Erik de Castro Lopo] 1886194206Ssimon 1887194206Ssimon *) Remove root CA certificates of commercial CAs: 1888194206Ssimon 1889194206Ssimon The OpenSSL project does not recommend any specific CA and does not 1890194206Ssimon have any policy with respect to including or excluding any CA. 1891194206Ssimon Therefore it does not make any sense to ship an arbitrary selection 1892194206Ssimon of root CA certificates with the OpenSSL software. 1893194206Ssimon [Lutz Jaenicke] 1894194206Ssimon 1895194206Ssimon *) RSA OAEP patches to fix two separate invalid memory reads. 1896194206Ssimon The first one involves inputs when 'lzero' is greater than 1897194206Ssimon 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes 1898194206Ssimon before the beginning of from). The second one involves inputs where 1899194206Ssimon the 'db' section contains nothing but zeroes (there is a one-byte 1900194206Ssimon invalid read after the end of 'db'). 1901194206Ssimon [Ivan Nestlerode <inestlerode@us.ibm.com>] 1902194206Ssimon 1903194206Ssimon *) Partial backport from 0.9.9-dev: 1904194206Ssimon 1905194206Ssimon Introduce bn_mul_mont (dedicated Montgomery multiplication 1906194206Ssimon procedure) as a candidate for BIGNUM assembler implementation. 1907194206Ssimon While 0.9.9-dev uses assembler for various architectures, only 1908194206Ssimon x86_64 is available by default here in the 0.9.8 branch, and 1909194206Ssimon 32-bit x86 is available through a compile-time setting. 1910194206Ssimon 1911194206Ssimon To try the 32-bit x86 assembler implementation, use Configure 1912194206Ssimon option "enable-montasm" (which exists only for this backport). 1913194206Ssimon 1914194206Ssimon As "enable-montasm" for 32-bit x86 disclaims code stability 1915194206Ssimon anyway, in this constellation we activate additional code 1916194206Ssimon backported from 0.9.9-dev for further performance improvements, 1917194206Ssimon namely BN_from_montgomery_word. (To enable this otherwise, 1918194206Ssimon e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".) 1919194206Ssimon 1920194206Ssimon [Andy Polyakov (backport partially by Bodo Moeller)] 1921194206Ssimon 1922194206Ssimon *) Add TLS session ticket callback. This allows an application to set 1923194206Ssimon TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed 1924194206Ssimon values. This is useful for key rollover for example where several key 1925194206Ssimon sets may exist with different names. 1926194206Ssimon [Steve Henson] 1927194206Ssimon 1928194206Ssimon *) Reverse ENGINE-internal logic for caching default ENGINE handles. 1929194206Ssimon This was broken until now in 0.9.8 releases, such that the only way 1930194206Ssimon a registered ENGINE could be used (assuming it initialises 1931194206Ssimon successfully on the host) was to explicitly set it as the default 1932194206Ssimon for the relevant algorithms. This is in contradiction with 0.9.7 1933194206Ssimon behaviour and the documentation. With this fix, when an ENGINE is 1934194206Ssimon registered into a given algorithm's table of implementations, the 1935194206Ssimon 'uptodate' flag is reset so that auto-discovery will be used next 1936194206Ssimon time a new context for that algorithm attempts to select an 1937194206Ssimon implementation. 1938194206Ssimon [Ian Lister (tweaked by Geoff Thorpe)] 1939194206Ssimon 1940194206Ssimon *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9 1941194206Ssimon implemention in the following ways: 1942194206Ssimon 1943194206Ssimon Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be 1944194206Ssimon hard coded. 1945194206Ssimon 1946194206Ssimon Lack of BER streaming support means one pass streaming processing is 1947194206Ssimon only supported if data is detached: setting the streaming flag is 1948194206Ssimon ignored for embedded content. 1949194206Ssimon 1950194206Ssimon CMS support is disabled by default and must be explicitly enabled 1951194206Ssimon with the enable-cms configuration option. 1952194206Ssimon [Steve Henson] 1953194206Ssimon 1954194206Ssimon *) Update the GMP engine glue to do direct copies between BIGNUM and 1955194206Ssimon mpz_t when openssl and GMP use the same limb size. Otherwise the 1956194206Ssimon existing "conversion via a text string export" trick is still used. 1957194206Ssimon [Paul Sheer <paulsheer@gmail.com>] 1958194206Ssimon 1959194206Ssimon *) Zlib compression BIO. This is a filter BIO which compressed and 1960194206Ssimon uncompresses any data passed through it. 1961194206Ssimon [Steve Henson] 1962194206Ssimon 1963194206Ssimon *) Add AES_wrap_key() and AES_unwrap_key() functions to implement 1964194206Ssimon RFC3394 compatible AES key wrapping. 1965194206Ssimon [Steve Henson] 1966194206Ssimon 1967194206Ssimon *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0(): 1968194206Ssimon sets string data without copying. X509_ALGOR_set0() and 1969194206Ssimon X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier) 1970194206Ssimon data. Attribute function X509at_get0_data_by_OBJ(): retrieves data 1971194206Ssimon from an X509_ATTRIBUTE structure optionally checking it occurs only 1972194206Ssimon once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied 1973194206Ssimon data. 1974194206Ssimon [Steve Henson] 1975194206Ssimon 1976194206Ssimon *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set() 1977194206Ssimon to get the expected BN_FLG_CONSTTIME behavior. 1978194206Ssimon [Bodo Moeller (Google)] 1979194206Ssimon 1980194206Ssimon *) Netware support: 1981194206Ssimon 1982194206Ssimon - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets 1983194206Ssimon - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT) 1984194206Ssimon - added some more tests to do_tests.pl 1985194206Ssimon - fixed RunningProcess usage so that it works with newer LIBC NDKs too 1986194206Ssimon - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency 1987194206Ssimon - added new Configure targets netware-clib-bsdsock, netware-clib-gcc, 1988194206Ssimon netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc 1989194206Ssimon - various changes to netware.pl to enable gcc-cross builds on Win32 1990194206Ssimon platform 1991194206Ssimon - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD) 1992194206Ssimon - various changes to fix missing prototype warnings 1993194206Ssimon - fixed x86nasm.pl to create correct asm files for NASM COFF output 1994194206Ssimon - added AES, WHIRLPOOL and CPUID assembler code to build files 1995194206Ssimon - added missing AES assembler make rules to mk1mf.pl 1996194206Ssimon - fixed order of includes in apps/ocsp.c so that e_os.h settings apply 1997194206Ssimon [Guenter Knauf <eflash@gmx.net>] 1998194206Ssimon 1999194206Ssimon *) Implement certificate status request TLS extension defined in RFC3546. 2000194206Ssimon A client can set the appropriate parameters and receive the encoded 2001194206Ssimon OCSP response via a callback. A server can query the supplied parameters 2002194206Ssimon and set the encoded OCSP response in the callback. Add simplified examples 2003194206Ssimon to s_client and s_server. 2004194206Ssimon [Steve Henson] 2005194206Ssimon 2006194206Ssimon Changes between 0.9.8f and 0.9.8g [19 Oct 2007] 2007194206Ssimon 2008194206Ssimon *) Fix various bugs: 2009194206Ssimon + Binary incompatibility of ssl_ctx_st structure 2010194206Ssimon + DTLS interoperation with non-compliant servers 2011194206Ssimon + Don't call get_session_cb() without proposed session 2012194206Ssimon + Fix ia64 assembler code 2013194206Ssimon [Andy Polyakov, Steve Henson] 2014194206Ssimon 2015194206Ssimon Changes between 0.9.8e and 0.9.8f [11 Oct 2007] 2016194206Ssimon 2017194206Ssimon *) DTLS Handshake overhaul. There were longstanding issues with 2018194206Ssimon OpenSSL DTLS implementation, which were making it impossible for 2019194206Ssimon RFC 4347 compliant client to communicate with OpenSSL server. 2020194206Ssimon Unfortunately just fixing these incompatibilities would "cut off" 2021194206Ssimon pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e 2022194206Ssimon server keeps tolerating non RFC compliant syntax. The opposite is 2023194206Ssimon not true, 0.9.8f client can not communicate with earlier server. 2024194206Ssimon This update even addresses CVE-2007-4995. 2025194206Ssimon [Andy Polyakov] 2026194206Ssimon 2027194206Ssimon *) Changes to avoid need for function casts in OpenSSL: some compilers 2028194206Ssimon (gcc 4.2 and later) reject their use. 2029194206Ssimon [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>, 2030194206Ssimon Steve Henson] 2031194206Ssimon 2032194206Ssimon *) Add RFC4507 support to OpenSSL. This includes the corrections in 2033194206Ssimon RFC4507bis. The encrypted ticket format is an encrypted encoded 2034194206Ssimon SSL_SESSION structure, that way new session features are automatically 2035194206Ssimon supported. 2036194206Ssimon 2037194206Ssimon If a client application caches session in an SSL_SESSION structure 2038194206Ssimon support is transparent because tickets are now stored in the encoded 2039194206Ssimon SSL_SESSION. 2040194206Ssimon 2041194206Ssimon The SSL_CTX structure automatically generates keys for ticket 2042194206Ssimon protection in servers so again support should be possible 2043194206Ssimon with no application modification. 2044194206Ssimon 2045194206Ssimon If a client or server wishes to disable RFC4507 support then the option 2046194206Ssimon SSL_OP_NO_TICKET can be set. 2047194206Ssimon 2048194206Ssimon Add a TLS extension debugging callback to allow the contents of any client 2049194206Ssimon or server extensions to be examined. 2050194206Ssimon 2051194206Ssimon This work was sponsored by Google. 2052194206Ssimon [Steve Henson] 2053194206Ssimon 2054194206Ssimon *) Add initial support for TLS extensions, specifically for the server_name 2055194206Ssimon extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 2056194206Ssimon have new members for a host name. The SSL data structure has an 2057194206Ssimon additional member SSL_CTX *initial_ctx so that new sessions can be 2058194206Ssimon stored in that context to allow for session resumption, even after the 2059194206Ssimon SSL has been switched to a new SSL_CTX in reaction to a client's 2060194206Ssimon server_name extension. 2061194206Ssimon 2062194206Ssimon New functions (subject to change): 2063194206Ssimon 2064194206Ssimon SSL_get_servername() 2065194206Ssimon SSL_get_servername_type() 2066194206Ssimon SSL_set_SSL_CTX() 2067194206Ssimon 2068194206Ssimon New CTRL codes and macros (subject to change): 2069194206Ssimon 2070194206Ssimon SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 2071194206Ssimon - SSL_CTX_set_tlsext_servername_callback() 2072194206Ssimon SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 2073194206Ssimon - SSL_CTX_set_tlsext_servername_arg() 2074194206Ssimon SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 2075194206Ssimon 2076194206Ssimon openssl s_client has a new '-servername ...' option. 2077194206Ssimon 2078194206Ssimon openssl s_server has new options '-servername_host ...', '-cert2 ...', 2079194206Ssimon '-key2 ...', '-servername_fatal' (subject to change). This allows 2080194206Ssimon testing the HostName extension for a specific single host name ('-cert' 2081194206Ssimon and '-key' remain fallbacks for handshakes without HostName 2082194206Ssimon negotiation). If the unrecogninzed_name alert has to be sent, this by 2083194206Ssimon default is a warning; it becomes fatal with the '-servername_fatal' 2084194206Ssimon option. 2085194206Ssimon 2086194206Ssimon [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson] 2087194206Ssimon 2088194206Ssimon *) Add AES and SSE2 assembly language support to VC++ build. 2089194206Ssimon [Steve Henson] 2090194206Ssimon 2091194206Ssimon *) Mitigate attack on final subtraction in Montgomery reduction. 2092194206Ssimon [Andy Polyakov] 2093194206Ssimon 2094194206Ssimon *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0 2095194206Ssimon (which previously caused an internal error). 2096194206Ssimon [Bodo Moeller] 2097194206Ssimon 2098194206Ssimon *) Squeeze another 10% out of IGE mode when in != out. 2099194206Ssimon [Ben Laurie] 2100194206Ssimon 2101194206Ssimon *) AES IGE mode speedup. 2102194206Ssimon [Dean Gaudet (Google)] 2103194206Ssimon 2104194206Ssimon *) Add the Korean symmetric 128-bit cipher SEED (see 2105194206Ssimon http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and 2106194206Ssimon add SEED ciphersuites from RFC 4162: 2107194206Ssimon 2108194206Ssimon TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA" 2109194206Ssimon TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA" 2110194206Ssimon TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA" 2111194206Ssimon TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA" 2112194206Ssimon 2113194206Ssimon To minimize changes between patchlevels in the OpenSSL 0.9.8 2114194206Ssimon series, SEED remains excluded from compilation unless OpenSSL 2115194206Ssimon is configured with 'enable-seed'. 2116194206Ssimon [KISA, Bodo Moeller] 2117194206Ssimon 2118194206Ssimon *) Mitigate branch prediction attacks, which can be practical if a 2119194206Ssimon single processor is shared, allowing a spy process to extract 2120194206Ssimon information. For detailed background information, see 2121194206Ssimon http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron, 2122194206Ssimon J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL 2123194206Ssimon and Necessary Software Countermeasures"). The core of the change 2124194206Ssimon are new versions BN_div_no_branch() and 2125194206Ssimon BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(), 2126194206Ssimon respectively, which are slower, but avoid the security-relevant 2127194206Ssimon conditional branches. These are automatically called by BN_div() 2128194206Ssimon and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one 2129194206Ssimon of the input BIGNUMs. Also, BN_is_bit_set() has been changed to 2130194206Ssimon remove a conditional branch. 2131194206Ssimon 2132194206Ssimon BN_FLG_CONSTTIME is the new name for the previous 2133194206Ssimon BN_FLG_EXP_CONSTTIME flag, since it now affects more than just 2134194206Ssimon modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag 2135194206Ssimon in the exponent causes BN_mod_exp_mont() to use the alternative 2136194206Ssimon implementation in BN_mod_exp_mont_consttime().) The old name 2137194206Ssimon remains as a deprecated alias. 2138194206Ssimon 2139194206Ssimon Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general 2140194206Ssimon RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses 2141194206Ssimon constant-time implementations for more than just exponentiation. 2142194206Ssimon Here too the old name is kept as a deprecated alias. 2143194206Ssimon 2144194206Ssimon BN_BLINDING_new() will now use BN_dup() for the modulus so that 2145194206Ssimon the BN_BLINDING structure gets an independent copy of the 2146194206Ssimon modulus. This means that the previous "BIGNUM *m" argument to 2147194206Ssimon BN_BLINDING_new() and to BN_BLINDING_create_param() now 2148194206Ssimon essentially becomes "const BIGNUM *m", although we can't actually 2149194206Ssimon change this in the header file before 0.9.9. It allows 2150194206Ssimon RSA_setup_blinding() to use BN_with_flags() on the modulus to 2151194206Ssimon enable BN_FLG_CONSTTIME. 2152194206Ssimon 2153194206Ssimon [Matthew D Wood (Intel Corp)] 2154194206Ssimon 2155194206Ssimon *) In the SSL/TLS server implementation, be strict about session ID 2156194206Ssimon context matching (which matters if an application uses a single 2157194206Ssimon external cache for different purposes). Previously, 2158194206Ssimon out-of-context reuse was forbidden only if SSL_VERIFY_PEER was 2159194206Ssimon set. This did ensure strict client verification, but meant that, 2160194206Ssimon with applications using a single external cache for quite 2161194206Ssimon different requirements, clients could circumvent ciphersuite 2162194206Ssimon restrictions for a given session ID context by starting a session 2163194206Ssimon in a different context. 2164194206Ssimon [Bodo Moeller] 2165194206Ssimon 2166167612Ssimon *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 2167167612Ssimon a ciphersuite string such as "DEFAULT:RSA" cannot enable 2168167612Ssimon authentication-only ciphersuites. 2169167612Ssimon [Bodo Moeller] 2170167612Ssimon 2171194206Ssimon *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was 2172194206Ssimon not complete and could lead to a possible single byte overflow 2173194206Ssimon (CVE-2007-5135) [Ben Laurie] 2174194206Ssimon 2175194206Ssimon Changes between 0.9.8d and 0.9.8e [23 Feb 2007] 2176194206Ssimon 2177167612Ssimon *) Since AES128 and AES256 (and similarly Camellia128 and 2178167612Ssimon Camellia256) share a single mask bit in the logic of 2179167612Ssimon ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 2180167612Ssimon kludge to work properly if AES128 is available and AES256 isn't 2181167612Ssimon (or if Camellia128 is available and Camellia256 isn't). 2182167612Ssimon [Victor Duchovni] 2183167612Ssimon 2184167612Ssimon *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c 2185167612Ssimon (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters): 2186167612Ssimon When a point or a seed is encoded in a BIT STRING, we need to 2187167612Ssimon prevent the removal of trailing zero bits to get the proper DER 2188167612Ssimon encoding. (By default, crypto/asn1/a_bitstr.c assumes the case 2189167612Ssimon of a NamedBitList, for which trailing 0 bits need to be removed.) 2190167612Ssimon [Bodo Moeller] 2191167612Ssimon 2192167612Ssimon *) Have SSL/TLS server implementation tolerate "mismatched" record 2193167612Ssimon protocol version while receiving ClientHello even if the 2194167612Ssimon ClientHello is fragmented. (The server can't insist on the 2195167612Ssimon particular protocol version it has chosen before the ServerHello 2196167612Ssimon message has informed the client about his choice.) 2197167612Ssimon [Bodo Moeller] 2198167612Ssimon 2199167612Ssimon *) Add RFC 3779 support. 2200167612Ssimon [Rob Austein for ARIN, Ben Laurie] 2201167612Ssimon 2202167612Ssimon *) Load error codes if they are not already present instead of using a 2203167612Ssimon static variable. This allows them to be cleanly unloaded and reloaded. 2204167612Ssimon Improve header file function name parsing. 2205167612Ssimon [Steve Henson] 2206167612Ssimon 2207167612Ssimon *) extend SMTP and IMAP protocol emulation in s_client to use EHLO 2208167612Ssimon or CAPABILITY handshake as required by RFCs. 2209167612Ssimon [Goetz Babin-Ebell] 2210167612Ssimon 2211162911Ssimon Changes between 0.9.8c and 0.9.8d [28 Sep 2006] 2212162911Ssimon 2213162911Ssimon *) Introduce limits to prevent malicious keys being able to 2214162911Ssimon cause a denial of service. (CVE-2006-2940) 2215162911Ssimon [Steve Henson, Bodo Moeller] 2216162911Ssimon 2217162911Ssimon *) Fix ASN.1 parsing of certain invalid structures that can result 2218162911Ssimon in a denial of service. (CVE-2006-2937) [Steve Henson] 2219162911Ssimon 2220162911Ssimon *) Fix buffer overflow in SSL_get_shared_ciphers() function. 2221162911Ssimon (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 2222162911Ssimon 2223162911Ssimon *) Fix SSL client code which could crash if connecting to a 2224162911Ssimon malicious SSLv2 server. (CVE-2006-4343) 2225162911Ssimon [Tavis Ormandy and Will Drewry, Google Security Team] 2226162911Ssimon 2227162911Ssimon *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites 2228162911Ssimon match only those. Before that, "AES256-SHA" would be interpreted 2229162911Ssimon as a pattern and match "AES128-SHA" too (since AES128-SHA got 2230162911Ssimon the same strength classification in 0.9.7h) as we currently only 2231162911Ssimon have a single AES bit in the ciphersuite description bitmap. 2232162911Ssimon That change, however, also applied to ciphersuite strings such as 2233162911Ssimon "RC4-MD5" that intentionally matched multiple ciphersuites -- 2234162911Ssimon namely, SSL 2.0 ciphersuites in addition to the more common ones 2235162911Ssimon from SSL 3.0/TLS 1.0. 2236162911Ssimon 2237162911Ssimon So we change the selection algorithm again: Naming an explicit 2238162911Ssimon ciphersuite selects this one ciphersuite, and any other similar 2239162911Ssimon ciphersuite (same bitmap) from *other* protocol versions. 2240162911Ssimon Thus, "RC4-MD5" again will properly select both the SSL 2.0 2241162911Ssimon ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite. 2242162911Ssimon 2243162911Ssimon Since SSL 2.0 does not have any ciphersuites for which the 2244162911Ssimon 128/256 bit distinction would be relevant, this works for now. 2245162911Ssimon The proper fix will be to use different bits for AES128 and 2246162911Ssimon AES256, which would have avoided the problems from the beginning; 2247162911Ssimon however, bits are scarce, so we can only do this in a new release 2248162911Ssimon (not just a patchlevel) when we can change the SSL_CIPHER 2249162911Ssimon definition to split the single 'unsigned long mask' bitmap into 2250162911Ssimon multiple values to extend the available space. 2251162911Ssimon 2252162911Ssimon [Bodo Moeller] 2253162911Ssimon 2254162911Ssimon Changes between 0.9.8b and 0.9.8c [05 Sep 2006] 2255162911Ssimon 2256162911Ssimon *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 2257162911Ssimon (CVE-2006-4339) [Ben Laurie and Google Security Team] 2258162911Ssimon 2259162911Ssimon *) Add AES IGE and biIGE modes. 2260162911Ssimon [Ben Laurie] 2261162911Ssimon 2262162911Ssimon *) Change the Unix randomness entropy gathering to use poll() when 2263162911Ssimon possible instead of select(), since the latter has some 2264162911Ssimon undesirable limitations. 2265162911Ssimon [Darryl Miles via Richard Levitte and Bodo Moeller] 2266162911Ssimon 2267162911Ssimon *) Disable "ECCdraft" ciphersuites more thoroughly. Now special 2268162911Ssimon treatment in ssl/ssl_ciph.s makes sure that these ciphersuites 2269162911Ssimon cannot be implicitly activated as part of, e.g., the "AES" alias. 2270162911Ssimon However, please upgrade to OpenSSL 0.9.9[-dev] for 2271162911Ssimon non-experimental use of the ECC ciphersuites to get TLS extension 2272162911Ssimon support, which is required for curve and point format negotiation 2273162911Ssimon to avoid potential handshake problems. 2274162911Ssimon [Bodo Moeller] 2275162911Ssimon 2276162911Ssimon *) Disable rogue ciphersuites: 2277162911Ssimon 2278162911Ssimon - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 2279162911Ssimon - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 2280162911Ssimon - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 2281162911Ssimon 2282162911Ssimon The latter two were purportedly from 2283162911Ssimon draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 2284162911Ssimon appear there. 2285162911Ssimon 2286167612Ssimon Also deactivate the remaining ciphersuites from 2287162911Ssimon draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 2288162911Ssimon unofficial, and the ID has long expired. 2289162911Ssimon [Bodo Moeller] 2290162911Ssimon 2291162911Ssimon *) Fix RSA blinding Heisenbug (problems sometimes occured on 2292162911Ssimon dual-core machines) and other potential thread-safety issues. 2293162911Ssimon [Bodo Moeller] 2294162911Ssimon 2295162911Ssimon *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key 2296162911Ssimon versions), which is now available for royalty-free use 2297162911Ssimon (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html). 2298162911Ssimon Also, add Camellia TLS ciphersuites from RFC 4132. 2299162911Ssimon 2300162911Ssimon To minimize changes between patchlevels in the OpenSSL 0.9.8 2301162911Ssimon series, Camellia remains excluded from compilation unless OpenSSL 2302162911Ssimon is configured with 'enable-camellia'. 2303162911Ssimon [NTT] 2304162911Ssimon 2305162911Ssimon *) Disable the padding bug check when compression is in use. The padding 2306162911Ssimon bug check assumes the first packet is of even length, this is not 2307162911Ssimon necessarily true if compresssion is enabled and can result in false 2308162911Ssimon positives causing handshake failure. The actual bug test is ancient 2309162911Ssimon code so it is hoped that implementations will either have fixed it by 2310162911Ssimon now or any which still have the bug do not support compression. 2311162911Ssimon [Steve Henson] 2312162911Ssimon 2313160814Ssimon Changes between 0.9.8a and 0.9.8b [04 May 2006] 2314160814Ssimon 2315160814Ssimon *) When applying a cipher rule check to see if string match is an explicit 2316160814Ssimon cipher suite and only match that one cipher suite if it is. 2317160814Ssimon [Steve Henson] 2318160814Ssimon 2319160814Ssimon *) Link in manifests for VC++ if needed. 2320160814Ssimon [Austin Ziegler <halostatue@gmail.com>] 2321160814Ssimon 2322160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 2323160814Ssimon draft-ietf-tls-ecc-12.txt with proposed changes (but without 2324160814Ssimon TLS extensions, which are supported starting with the 0.9.9 2325160814Ssimon branch, not in the OpenSSL 0.9.8 branch). 2326160814Ssimon [Douglas Stebila] 2327160814Ssimon 2328160814Ssimon *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support 2329160814Ssimon opaque EVP_CIPHER_CTX handling. 2330160814Ssimon [Steve Henson] 2331160814Ssimon 2332160814Ssimon *) Fixes and enhancements to zlib compression code. We now only use 2333160814Ssimon "zlib1.dll" and use the default __cdecl calling convention on Win32 2334160814Ssimon to conform with the standards mentioned here: 2335160814Ssimon http://www.zlib.net/DLL_FAQ.txt 2336160814Ssimon Static zlib linking now works on Windows and the new --with-zlib-include 2337160814Ssimon --with-zlib-lib options to Configure can be used to supply the location 2338160814Ssimon of the headers and library. Gracefully handle case where zlib library 2339160814Ssimon can't be loaded. 2340160814Ssimon [Steve Henson] 2341160814Ssimon 2342160814Ssimon *) Several fixes and enhancements to the OID generation code. The old code 2343160814Ssimon sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't 2344160814Ssimon handle numbers larger than ULONG_MAX, truncated printing and had a 2345160814Ssimon non standard OBJ_obj2txt() behaviour. 2346160814Ssimon [Steve Henson] 2347160814Ssimon 2348160814Ssimon *) Add support for building of engines under engine/ as shared libraries 2349160814Ssimon under VC++ build system. 2350160814Ssimon [Steve Henson] 2351160814Ssimon 2352160814Ssimon *) Corrected the numerous bugs in the Win32 path splitter in DSO. 2353160814Ssimon Hopefully, we will not see any false combination of paths any more. 2354160814Ssimon [Richard Levitte] 2355160814Ssimon 2356160814Ssimon Changes between 0.9.8 and 0.9.8a [11 Oct 2005] 2357160814Ssimon 2358160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 2359160814Ssimon (part of SSL_OP_ALL). This option used to disable the 2360160814Ssimon countermeasure against man-in-the-middle protocol-version 2361160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 2362160814Ssimon idea. (CVE-2005-2969) 2363160814Ssimon 2364160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 2365160814Ssimon for Information Security, National Institute of Advanced Industrial 2366160814Ssimon Science and Technology [AIST], Japan)] 2367160814Ssimon 2368160814Ssimon *) Add two function to clear and return the verify parameter flags. 2369160814Ssimon [Steve Henson] 2370160814Ssimon 2371160814Ssimon *) Keep cipherlists sorted in the source instead of sorting them at 2372160814Ssimon runtime, thus removing the need for a lock. 2373160814Ssimon [Nils Larsch] 2374160814Ssimon 2375160814Ssimon *) Avoid some small subgroup attacks in Diffie-Hellman. 2376160814Ssimon [Nick Mathewson and Ben Laurie] 2377160814Ssimon 2378160814Ssimon *) Add functions for well-known primes. 2379160814Ssimon [Nick Mathewson] 2380160814Ssimon 2381160814Ssimon *) Extended Windows CE support. 2382160814Ssimon [Satoshi Nakamura and Andy Polyakov] 2383160814Ssimon 2384160814Ssimon *) Initialize SSL_METHOD structures at compile time instead of during 2385160814Ssimon runtime, thus removing the need for a lock. 2386160814Ssimon [Steve Henson] 2387160814Ssimon 2388160814Ssimon *) Make PKCS7_decrypt() work even if no certificate is supplied by 2389160814Ssimon attempting to decrypt each encrypted key in turn. Add support to 2390160814Ssimon smime utility. 2391160814Ssimon [Steve Henson] 2392160814Ssimon 2393160814Ssimon Changes between 0.9.7h and 0.9.8 [05 Jul 2005] 2394160814Ssimon 2395162911Ssimon [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after 2396162911Ssimon OpenSSL 0.9.8.] 2397162911Ssimon 2398160814Ssimon *) Add libcrypto.pc and libssl.pc for those who feel they need them. 2399160814Ssimon [Richard Levitte] 2400160814Ssimon 2401160814Ssimon *) Change CA.sh and CA.pl so they don't bundle the CSR and the private 2402160814Ssimon key into the same file any more. 2403160814Ssimon [Richard Levitte] 2404160814Ssimon 2405160814Ssimon *) Add initial support for Win64, both IA64 and AMD64/x64 flavors. 2406160814Ssimon [Andy Polyakov] 2407160814Ssimon 2408160814Ssimon *) Add -utf8 command line and config file option to 'ca'. 2409160814Ssimon [Stefan <stf@udoma.org] 2410160814Ssimon 2411160814Ssimon *) Removed the macro des_crypt(), as it seems to conflict with some 2412160814Ssimon libraries. Use DES_crypt(). 2413160814Ssimon [Richard Levitte] 2414160814Ssimon 2415160814Ssimon *) Correct naming of the 'chil' and '4758cca' ENGINEs. This 2416160814Ssimon involves renaming the source and generated shared-libs for 2417160814Ssimon both. The engines will accept the corrected or legacy ids 2418160814Ssimon ('ncipher' and '4758_cca' respectively) when binding. NB, 2419160814Ssimon this only applies when building 'shared'. 2420160814Ssimon [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe] 2421160814Ssimon 2422160814Ssimon *) Add attribute functions to EVP_PKEY structure. Modify 2423160814Ssimon PKCS12_create() to recognize a CSP name attribute and 2424160814Ssimon use it. Make -CSP option work again in pkcs12 utility. 2425160814Ssimon [Steve Henson] 2426160814Ssimon 2427160814Ssimon *) Add new functionality to the bn blinding code: 2428160814Ssimon - automatic re-creation of the BN_BLINDING parameters after 2429160814Ssimon a fixed number of uses (currently 32) 2430160814Ssimon - add new function for parameter creation 2431160814Ssimon - introduce flags to control the update behaviour of the 2432160814Ssimon BN_BLINDING parameters 2433160814Ssimon - hide BN_BLINDING structure 2434160814Ssimon Add a second BN_BLINDING slot to the RSA structure to improve 2435160814Ssimon performance when a single RSA object is shared among several 2436160814Ssimon threads. 2437160814Ssimon [Nils Larsch] 2438160814Ssimon 2439160814Ssimon *) Add support for DTLS. 2440160814Ssimon [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie] 2441160814Ssimon 2442160814Ssimon *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1) 2443160814Ssimon to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file() 2444160814Ssimon [Walter Goulet] 2445160814Ssimon 2446160814Ssimon *) Remove buggy and incompletet DH cert support from 2447160814Ssimon ssl/ssl_rsa.c and ssl/s3_both.c 2448160814Ssimon [Nils Larsch] 2449160814Ssimon 2450160814Ssimon *) Use SHA-1 instead of MD5 as the default digest algorithm for 2451160814Ssimon the apps/openssl applications. 2452160814Ssimon [Nils Larsch] 2453160814Ssimon 2454160814Ssimon *) Compile clean with "-Wall -Wmissing-prototypes 2455160814Ssimon -Wstrict-prototypes -Wmissing-declarations -Werror". Currently 2456160814Ssimon DEBUG_SAFESTACK must also be set. 2457160814Ssimon [Ben Laurie] 2458160814Ssimon 2459160814Ssimon *) Change ./Configure so that certain algorithms can be disabled by default. 2460160814Ssimon The new counterpiece to "no-xxx" is "enable-xxx". 2461160814Ssimon 2462160814Ssimon The patented RC5 and MDC2 algorithms will now be disabled unless 2463160814Ssimon "enable-rc5" and "enable-mdc2", respectively, are specified. 2464160814Ssimon 2465160814Ssimon (IDEA remains enabled despite being patented. This is because IDEA 2466160814Ssimon is frequently required for interoperability, and there is no license 2467160814Ssimon fee for non-commercial use. As before, "no-idea" can be used to 2468160814Ssimon avoid this algorithm.) 2469160814Ssimon 2470160814Ssimon [Bodo Moeller] 2471160814Ssimon 2472160814Ssimon *) Add processing of proxy certificates (see RFC 3820). This work was 2473160814Ssimon sponsored by KTH (The Royal Institute of Technology in Stockholm) and 2474160814Ssimon EGEE (Enabling Grids for E-science in Europe). 2475160814Ssimon [Richard Levitte] 2476160814Ssimon 2477160814Ssimon *) RC4 performance overhaul on modern architectures/implementations, such 2478160814Ssimon as Intel P4, IA-64 and AMD64. 2479160814Ssimon [Andy Polyakov] 2480160814Ssimon 2481160814Ssimon *) New utility extract-section.pl. This can be used specify an alternative 2482160814Ssimon section number in a pod file instead of having to treat each file as 2483160814Ssimon a separate case in Makefile. This can be done by adding two lines to the 2484160814Ssimon pod file: 2485160814Ssimon 2486160814Ssimon =for comment openssl_section:XXX 2487160814Ssimon 2488160814Ssimon The blank line is mandatory. 2489160814Ssimon 2490160814Ssimon [Steve Henson] 2491160814Ssimon 2492160814Ssimon *) New arguments -certform, -keyform and -pass for s_client and s_server 2493160814Ssimon to allow alternative format key and certificate files and passphrase 2494160814Ssimon sources. 2495160814Ssimon [Steve Henson] 2496160814Ssimon 2497160814Ssimon *) New structure X509_VERIFY_PARAM which combines current verify parameters, 2498160814Ssimon update associated structures and add various utility functions. 2499160814Ssimon 2500160814Ssimon Add new policy related verify parameters, include policy checking in 2501160814Ssimon standard verify code. Enhance 'smime' application with extra parameters 2502160814Ssimon to support policy checking and print out. 2503160814Ssimon [Steve Henson] 2504160814Ssimon 2505160814Ssimon *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3 2506160814Ssimon Nehemiah processors. These extensions support AES encryption in hardware 2507160814Ssimon as well as RNG (though RNG support is currently disabled). 2508160814Ssimon [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov] 2509160814Ssimon 2510160814Ssimon *) Deprecate BN_[get|set]_params() functions (they were ignored internally). 2511160814Ssimon [Geoff Thorpe] 2512160814Ssimon 2513160814Ssimon *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented. 2514160814Ssimon [Andy Polyakov and a number of other people] 2515160814Ssimon 2516160814Ssimon *) Improved PowerPC platform support. Most notably BIGNUM assembler 2517160814Ssimon implementation contributed by IBM. 2518160814Ssimon [Suresh Chari, Peter Waltenberg, Andy Polyakov] 2519160814Ssimon 2520160814Ssimon *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public 2521160814Ssimon exponent rather than 'unsigned long'. There is a corresponding change to 2522160814Ssimon the new 'rsa_keygen' element of the RSA_METHOD structure. 2523160814Ssimon [Jelte Jansen, Geoff Thorpe] 2524160814Ssimon 2525160814Ssimon *) Functionality for creating the initial serial number file is now 2526160814Ssimon moved from CA.pl to the 'ca' utility with a new option -create_serial. 2527160814Ssimon 2528160814Ssimon (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial 2529160814Ssimon number file to 1, which is bound to cause problems. To avoid 2530160814Ssimon the problems while respecting compatibility between different 0.9.7 2531160814Ssimon patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in 2532160814Ssimon CA.pl for serial number initialization. With the new release 0.9.8, 2533160814Ssimon we can fix the problem directly in the 'ca' utility.) 2534160814Ssimon [Steve Henson] 2535160814Ssimon 2536160814Ssimon *) Reduced header interdepencies by declaring more opaque objects in 2537160814Ssimon ossl_typ.h. As a consequence, including some headers (eg. engine.h) will 2538160814Ssimon give fewer recursive includes, which could break lazy source code - so 2539160814Ssimon this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always, 2540160814Ssimon developers should define this symbol when building and using openssl to 2541160814Ssimon ensure they track the recommended behaviour, interfaces, [etc], but 2542160814Ssimon backwards-compatible behaviour prevails when this isn't defined. 2543160814Ssimon [Geoff Thorpe] 2544160814Ssimon 2545160814Ssimon *) New function X509_POLICY_NODE_print() which prints out policy nodes. 2546160814Ssimon [Steve Henson] 2547160814Ssimon 2548160814Ssimon *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality. 2549160814Ssimon This will generate a random key of the appropriate length based on the 2550160814Ssimon cipher context. The EVP_CIPHER can provide its own random key generation 2551160814Ssimon routine to support keys of a specific form. This is used in the des and 2552160814Ssimon 3des routines to generate a key of the correct parity. Update S/MIME 2553160814Ssimon code to use new functions and hence generate correct parity DES keys. 2554160814Ssimon Add EVP_CHECK_DES_KEY #define to return an error if the key is not 2555160814Ssimon valid (weak or incorrect parity). 2556160814Ssimon [Steve Henson] 2557160814Ssimon 2558160814Ssimon *) Add a local set of CRLs that can be used by X509_verify_cert() as well 2559160814Ssimon as looking them up. This is useful when the verified structure may contain 2560160814Ssimon CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs 2561160814Ssimon present unless the new PKCS7_NO_CRL flag is asserted. 2562160814Ssimon [Steve Henson] 2563160814Ssimon 2564160814Ssimon *) Extend ASN1 oid configuration module. It now additionally accepts the 2565160814Ssimon syntax: 2566160814Ssimon 2567160814Ssimon shortName = some long name, 1.2.3.4 2568160814Ssimon [Steve Henson] 2569160814Ssimon 2570160814Ssimon *) Reimplemented the BN_CTX implementation. There is now no more static 2571160814Ssimon limitation on the number of variables it can handle nor the depth of the 2572160814Ssimon "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack 2573160814Ssimon information can now expand as required, and rather than having a single 2574160814Ssimon static array of bignums, BN_CTX now uses a linked-list of such arrays 2575160814Ssimon allowing it to expand on demand whilst maintaining the usefulness of 2576160814Ssimon BN_CTX's "bundling". 2577160814Ssimon [Geoff Thorpe] 2578160814Ssimon 2579160814Ssimon *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD 2580160814Ssimon to allow all RSA operations to function using a single BN_CTX. 2581160814Ssimon [Geoff Thorpe] 2582160814Ssimon 2583160814Ssimon *) Preliminary support for certificate policy evaluation and checking. This 2584160814Ssimon is initially intended to pass the tests outlined in "Conformance Testing 2585160814Ssimon of Relying Party Client Certificate Path Processing Logic" v1.07. 2586160814Ssimon [Steve Henson] 2587160814Ssimon 2588160814Ssimon *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and 2589160814Ssimon remained unused and not that useful. A variety of other little bignum 2590160814Ssimon tweaks and fixes have also been made continuing on from the audit (see 2591160814Ssimon below). 2592160814Ssimon [Geoff Thorpe] 2593160814Ssimon 2594160814Ssimon *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with 2595160814Ssimon associated ASN1, EVP and SSL functions and old ASN1 macros. 2596160814Ssimon [Richard Levitte] 2597160814Ssimon 2598160814Ssimon *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results, 2599160814Ssimon and this should never fail. So the return value from the use of 2600160814Ssimon BN_set_word() (which can fail due to needless expansion) is now deprecated; 2601160814Ssimon if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro. 2602160814Ssimon [Geoff Thorpe] 2603160814Ssimon 2604160814Ssimon *) BN_CTX_get() should return zero-valued bignums, providing the same 2605160814Ssimon initialised value as BN_new(). 2606160814Ssimon [Geoff Thorpe, suggested by Ulf M�ller] 2607160814Ssimon 2608160814Ssimon *) Support for inhibitAnyPolicy certificate extension. 2609160814Ssimon [Steve Henson] 2610160814Ssimon 2611160814Ssimon *) An audit of the BIGNUM code is underway, for which debugging code is 2612160814Ssimon enabled when BN_DEBUG is defined. This makes stricter enforcements on what 2613160814Ssimon is considered valid when processing BIGNUMs, and causes execution to 2614160814Ssimon assert() when a problem is discovered. If BN_DEBUG_RAND is defined, 2615160814Ssimon further steps are taken to deliberately pollute unused data in BIGNUM 2616160814Ssimon structures to try and expose faulty code further on. For now, openssl will 2617160814Ssimon (in its default mode of operation) continue to tolerate the inconsistent 2618160814Ssimon forms that it has tolerated in the past, but authors and packagers should 2619160814Ssimon consider trying openssl and their own applications when compiled with 2620160814Ssimon these debugging symbols defined. It will help highlight potential bugs in 2621160814Ssimon their own code, and will improve the test coverage for OpenSSL itself. At 2622160814Ssimon some point, these tighter rules will become openssl's default to improve 2623160814Ssimon maintainability, though the assert()s and other overheads will remain only 2624160814Ssimon in debugging configurations. See bn.h for more details. 2625160814Ssimon [Geoff Thorpe, Nils Larsch, Ulf M�ller] 2626160814Ssimon 2627160814Ssimon *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure 2628160814Ssimon that can only be obtained through BN_CTX_new() (which implicitly 2629160814Ssimon initialises it). The presence of this function only made it possible 2630160814Ssimon to overwrite an existing structure (and cause memory leaks). 2631160814Ssimon [Geoff Thorpe] 2632160814Ssimon 2633160814Ssimon *) Because of the callback-based approach for implementing LHASH as a 2634160814Ssimon template type, lh_insert() adds opaque objects to hash-tables and 2635160814Ssimon lh_doall() or lh_doall_arg() are typically used with a destructor callback 2636160814Ssimon to clean up those corresponding objects before destroying the hash table 2637160814Ssimon (and losing the object pointers). So some over-zealous constifications in 2638160814Ssimon LHASH have been relaxed so that lh_insert() does not take (nor store) the 2639160814Ssimon objects as "const" and the lh_doall[_arg] callback wrappers are not 2640160814Ssimon prototyped to have "const" restrictions on the object pointers they are 2641160814Ssimon given (and so aren't required to cast them away any more). 2642160814Ssimon [Geoff Thorpe] 2643160814Ssimon 2644160814Ssimon *) The tmdiff.h API was so ugly and minimal that our own timing utility 2645160814Ssimon (speed) prefers to use its own implementation. The two implementations 2646160814Ssimon haven't been consolidated as yet (volunteers?) but the tmdiff API has had 2647160814Ssimon its object type properly exposed (MS_TM) instead of casting to/from "char 2648160814Ssimon *". This may still change yet if someone realises MS_TM and "ms_time_***" 2649160814Ssimon aren't necessarily the greatest nomenclatures - but this is what was used 2650160814Ssimon internally to the implementation so I've used that for now. 2651160814Ssimon [Geoff Thorpe] 2652160814Ssimon 2653160814Ssimon *) Ensure that deprecated functions do not get compiled when 2654160814Ssimon OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of 2655160814Ssimon the self-tests were still using deprecated key-generation functions so 2656160814Ssimon these have been updated also. 2657160814Ssimon [Geoff Thorpe] 2658160814Ssimon 2659160814Ssimon *) Reorganise PKCS#7 code to separate the digest location functionality 2660160814Ssimon into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest(). 2661160814Ssimon New function PKCS7_set_digest() to set the digest type for PKCS#7 2662160814Ssimon digestedData type. Add additional code to correctly generate the 2663160814Ssimon digestedData type and add support for this type in PKCS7 initialization 2664160814Ssimon functions. 2665160814Ssimon [Steve Henson] 2666160814Ssimon 2667160814Ssimon *) New function PKCS7_set0_type_other() this initializes a PKCS7 2668160814Ssimon structure of type "other". 2669160814Ssimon [Steve Henson] 2670160814Ssimon 2671160814Ssimon *) Fix prime generation loop in crypto/bn/bn_prime.pl by making 2672160814Ssimon sure the loop does correctly stop and breaking ("division by zero") 2673160814Ssimon modulus operations are not performed. The (pre-generated) prime 2674160814Ssimon table crypto/bn/bn_prime.h was already correct, but it could not be 2675160814Ssimon re-generated on some platforms because of the "division by zero" 2676160814Ssimon situation in the script. 2677160814Ssimon [Ralf S. Engelschall] 2678160814Ssimon 2679160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 2680160814Ssimon draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with 2681160814Ssimon SHA-1 now is only used for "small" curves (where the 2682160814Ssimon representation of a field element takes up to 24 bytes); for 2683160814Ssimon larger curves, the field element resulting from ECDH is directly 2684160814Ssimon used as premaster secret. 2685160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 2686160814Ssimon 2687160814Ssimon *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2 2688160814Ssimon curve secp160r1 to the tests. 2689160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 2690160814Ssimon 2691160814Ssimon *) Add the possibility to load symbols globally with DSO. 2692160814Ssimon [G�tz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte] 2693160814Ssimon 2694160814Ssimon *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better 2695160814Ssimon control of the error stack. 2696160814Ssimon [Richard Levitte] 2697160814Ssimon 2698160814Ssimon *) Add support for STORE in ENGINE. 2699160814Ssimon [Richard Levitte] 2700160814Ssimon 2701160814Ssimon *) Add the STORE type. The intention is to provide a common interface 2702160814Ssimon to certificate and key stores, be they simple file-based stores, or 2703160814Ssimon HSM-type store, or LDAP stores, or... 2704160814Ssimon NOTE: The code is currently UNTESTED and isn't really used anywhere. 2705160814Ssimon [Richard Levitte] 2706160814Ssimon 2707160814Ssimon *) Add a generic structure called OPENSSL_ITEM. This can be used to 2708160814Ssimon pass a list of arguments to any function as well as provide a way 2709160814Ssimon for a function to pass data back to the caller. 2710160814Ssimon [Richard Levitte] 2711160814Ssimon 2712160814Ssimon *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup() 2713160814Ssimon works like BUF_strdup() but can be used to duplicate a portion of 2714160814Ssimon a string. The copy gets NUL-terminated. BUF_memdup() duplicates 2715160814Ssimon a memory area. 2716160814Ssimon [Richard Levitte] 2717160814Ssimon 2718160814Ssimon *) Add the function sk_find_ex() which works like sk_find(), but will 2719160814Ssimon return an index to an element even if an exact match couldn't be 2720160814Ssimon found. The index is guaranteed to point at the element where the 2721160814Ssimon searched-for key would be inserted to preserve sorting order. 2722160814Ssimon [Richard Levitte] 2723160814Ssimon 2724160814Ssimon *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but 2725160814Ssimon takes an extra flags argument for optional functionality. Currently, 2726160814Ssimon the following flags are defined: 2727160814Ssimon 2728160814Ssimon OBJ_BSEARCH_VALUE_ON_NOMATCH 2729160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 2730160814Ssimon element where the comparing function returns a negative or zero 2731160814Ssimon number. 2732160814Ssimon 2733160814Ssimon OBJ_BSEARCH_FIRST_VALUE_ON_MATCH 2734160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 2735160814Ssimon element where the comparing function returns zero. This is useful 2736160814Ssimon if there are more than one element where the comparing function 2737160814Ssimon returns zero. 2738160814Ssimon [Richard Levitte] 2739160814Ssimon 2740160814Ssimon *) Make it possible to create self-signed certificates with 'openssl ca' 2741160814Ssimon in such a way that the self-signed certificate becomes part of the 2742160814Ssimon CA database and uses the same mechanisms for serial number generation 2743160814Ssimon as all other certificate signing. The new flag '-selfsign' enables 2744160814Ssimon this functionality. Adapt CA.sh and CA.pl.in. 2745160814Ssimon [Richard Levitte] 2746160814Ssimon 2747160814Ssimon *) Add functionality to check the public key of a certificate request 2748160814Ssimon against a given private. This is useful to check that a certificate 2749160814Ssimon request can be signed by that key (self-signing). 2750160814Ssimon [Richard Levitte] 2751160814Ssimon 2752160814Ssimon *) Make it possible to have multiple active certificates with the same 2753160814Ssimon subject in the CA index file. This is done only if the keyword 2754160814Ssimon 'unique_subject' is set to 'no' in the main CA section (default 2755160814Ssimon if 'CA_default') of the configuration file. The value is saved 2756160814Ssimon with the database itself in a separate index attribute file, 2757160814Ssimon named like the index file with '.attr' appended to the name. 2758160814Ssimon [Richard Levitte] 2759160814Ssimon 2760160814Ssimon *) Generate muti valued AVAs using '+' notation in config files for 2761160814Ssimon req and dirName. 2762160814Ssimon [Steve Henson] 2763160814Ssimon 2764160814Ssimon *) Support for nameConstraints certificate extension. 2765160814Ssimon [Steve Henson] 2766160814Ssimon 2767160814Ssimon *) Support for policyConstraints certificate extension. 2768160814Ssimon [Steve Henson] 2769160814Ssimon 2770160814Ssimon *) Support for policyMappings certificate extension. 2771160814Ssimon [Steve Henson] 2772160814Ssimon 2773160814Ssimon *) Make sure the default DSA_METHOD implementation only uses its 2774160814Ssimon dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL, 2775160814Ssimon and change its own handlers to be NULL so as to remove unnecessary 2776160814Ssimon indirection. This lets alternative implementations fallback to the 2777160814Ssimon default implementation more easily. 2778160814Ssimon [Geoff Thorpe] 2779160814Ssimon 2780160814Ssimon *) Support for directoryName in GeneralName related extensions 2781160814Ssimon in config files. 2782160814Ssimon [Steve Henson] 2783160814Ssimon 2784160814Ssimon *) Make it possible to link applications using Makefile.shared. 2785160814Ssimon Make that possible even when linking against static libraries! 2786160814Ssimon [Richard Levitte] 2787160814Ssimon 2788160814Ssimon *) Support for single pass processing for S/MIME signing. This now 2789160814Ssimon means that S/MIME signing can be done from a pipe, in addition 2790160814Ssimon cleartext signing (multipart/signed type) is effectively streaming 2791160814Ssimon and the signed data does not need to be all held in memory. 2792160814Ssimon 2793160814Ssimon This is done with a new flag PKCS7_STREAM. When this flag is set 2794160814Ssimon PKCS7_sign() only initializes the PKCS7 structure and the actual signing 2795160814Ssimon is done after the data is output (and digests calculated) in 2796160814Ssimon SMIME_write_PKCS7(). 2797160814Ssimon [Steve Henson] 2798160814Ssimon 2799160814Ssimon *) Add full support for -rpath/-R, both in shared libraries and 2800160814Ssimon applications, at least on the platforms where it's known how 2801160814Ssimon to do it. 2802160814Ssimon [Richard Levitte] 2803160814Ssimon 2804160814Ssimon *) In crypto/ec/ec_mult.c, implement fast point multiplication with 2805160814Ssimon precomputation, based on wNAF splitting: EC_GROUP_precompute_mult() 2806160814Ssimon will now compute a table of multiples of the generator that 2807160814Ssimon makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul() 2808160814Ssimon faster (notably in the case of a single point multiplication, 2809160814Ssimon scalar * generator). 2810160814Ssimon [Nils Larsch, Bodo Moeller] 2811160814Ssimon 2812160814Ssimon *) IPv6 support for certificate extensions. The various extensions 2813160814Ssimon which use the IP:a.b.c.d can now take IPv6 addresses using the 2814160814Ssimon formats of RFC1884 2.2 . IPv6 addresses are now also displayed 2815160814Ssimon correctly. 2816160814Ssimon [Steve Henson] 2817160814Ssimon 2818160814Ssimon *) Added an ENGINE that implements RSA by performing private key 2819160814Ssimon exponentiations with the GMP library. The conversions to and from 2820160814Ssimon GMP's mpz_t format aren't optimised nor are any montgomery forms 2821160814Ssimon cached, and on x86 it appears OpenSSL's own performance has caught up. 2822160814Ssimon However there are likely to be other architectures where GMP could 2823160814Ssimon provide a boost. This ENGINE is not built in by default, but it can be 2824160814Ssimon specified at Configure time and should be accompanied by the necessary 2825160814Ssimon linker additions, eg; 2826160814Ssimon ./config -DOPENSSL_USE_GMP -lgmp 2827160814Ssimon [Geoff Thorpe] 2828160814Ssimon 2829160814Ssimon *) "openssl engine" will not display ENGINE/DSO load failure errors when 2830160814Ssimon testing availability of engines with "-t" - the old behaviour is 2831160814Ssimon produced by increasing the feature's verbosity with "-tt". 2832160814Ssimon [Geoff Thorpe] 2833160814Ssimon 2834160814Ssimon *) ECDSA routines: under certain error conditions uninitialized BN objects 2835160814Ssimon could be freed. Solution: make sure initialization is performed early 2836160814Ssimon enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de> 2837160814Ssimon via PR#459) 2838160814Ssimon [Lutz Jaenicke] 2839160814Ssimon 2840160814Ssimon *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD 2841160814Ssimon and DH_METHOD (eg. by ENGINE implementations) to override the normal 2842160814Ssimon software implementations. For DSA and DH, parameter generation can 2843160814Ssimon also be overriden by providing the appropriate method callbacks. 2844160814Ssimon [Geoff Thorpe] 2845160814Ssimon 2846160814Ssimon *) Change the "progress" mechanism used in key-generation and 2847160814Ssimon primality testing to functions that take a new BN_GENCB pointer in 2848160814Ssimon place of callback/argument pairs. The new API functions have "_ex" 2849160814Ssimon postfixes and the older functions are reimplemented as wrappers for 2850160814Ssimon the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide 2851160814Ssimon declarations of the old functions to help (graceful) attempts to 2852160814Ssimon migrate to the new functions. Also, the new key-generation API 2853160814Ssimon functions operate on a caller-supplied key-structure and return 2854160814Ssimon success/failure rather than returning a key or NULL - this is to 2855160814Ssimon help make "keygen" another member function of RSA_METHOD etc. 2856160814Ssimon 2857160814Ssimon Example for using the new callback interface: 2858160814Ssimon 2859160814Ssimon int (*my_callback)(int a, int b, BN_GENCB *cb) = ...; 2860160814Ssimon void *my_arg = ...; 2861160814Ssimon BN_GENCB my_cb; 2862160814Ssimon 2863160814Ssimon BN_GENCB_set(&my_cb, my_callback, my_arg); 2864160814Ssimon 2865160814Ssimon return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb); 2866160814Ssimon /* For the meaning of a, b in calls to my_callback(), see the 2867160814Ssimon * documentation of the function that calls the callback. 2868160814Ssimon * cb will point to my_cb; my_arg can be retrieved as cb->arg. 2869160814Ssimon * my_callback should return 1 if it wants BN_is_prime_ex() 2870160814Ssimon * to continue, or 0 to stop. 2871160814Ssimon */ 2872160814Ssimon 2873160814Ssimon [Geoff Thorpe] 2874160814Ssimon 2875160814Ssimon *) Change the ZLIB compression method to be stateful, and make it 2876160814Ssimon available to TLS with the number defined in 2877160814Ssimon draft-ietf-tls-compression-04.txt. 2878160814Ssimon [Richard Levitte] 2879160814Ssimon 2880160814Ssimon *) Add the ASN.1 structures and functions for CertificatePair, which 2881160814Ssimon is defined as follows (according to X.509_4thEditionDraftV6.pdf): 2882160814Ssimon 2883160814Ssimon CertificatePair ::= SEQUENCE { 2884160814Ssimon forward [0] Certificate OPTIONAL, 2885160814Ssimon reverse [1] Certificate OPTIONAL, 2886160814Ssimon -- at least one of the pair shall be present -- } 2887160814Ssimon 2888160814Ssimon Also implement the PEM functions to read and write certificate 2889160814Ssimon pairs, and defined the PEM tag as "CERTIFICATE PAIR". 2890160814Ssimon 2891160814Ssimon This needed to be defined, mostly for the sake of the LDAP 2892160814Ssimon attribute crossCertificatePair, but may prove useful elsewhere as 2893160814Ssimon well. 2894160814Ssimon [Richard Levitte] 2895160814Ssimon 2896160814Ssimon *) Make it possible to inhibit symlinking of shared libraries in 2897160814Ssimon Makefile.shared, for Cygwin's sake. 2898160814Ssimon [Richard Levitte] 2899160814Ssimon 2900160814Ssimon *) Extend the BIGNUM API by creating a function 2901160814Ssimon void BN_set_negative(BIGNUM *a, int neg); 2902160814Ssimon and a macro that behave like 2903160814Ssimon int BN_is_negative(const BIGNUM *a); 2904160814Ssimon 2905160814Ssimon to avoid the need to access 'a->neg' directly in applications. 2906160814Ssimon [Nils Larsch] 2907160814Ssimon 2908160814Ssimon *) Implement fast modular reduction for pseudo-Mersenne primes 2909160814Ssimon used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c). 2910160814Ssimon EC_GROUP_new_curve_GFp() will now automatically use this 2911160814Ssimon if applicable. 2912160814Ssimon [Nils Larsch <nla@trustcenter.de>] 2913160814Ssimon 2914160814Ssimon *) Add new lock type (CRYPTO_LOCK_BN). 2915160814Ssimon [Bodo Moeller] 2916160814Ssimon 2917160814Ssimon *) Change the ENGINE framework to automatically load engines 2918160814Ssimon dynamically from specific directories unless they could be 2919160814Ssimon found to already be built in or loaded. Move all the 2920160814Ssimon current engines except for the cryptodev one to a new 2921160814Ssimon directory engines/. 2922160814Ssimon The engines in engines/ are built as shared libraries if 2923160814Ssimon the "shared" options was given to ./Configure or ./config. 2924160814Ssimon Otherwise, they are inserted in libcrypto.a. 2925160814Ssimon /usr/local/ssl/engines is the default directory for dynamic 2926160814Ssimon engines, but that can be overriden at configure time through 2927160814Ssimon the usual use of --prefix and/or --openssldir, and at run 2928160814Ssimon time with the environment variable OPENSSL_ENGINES. 2929160814Ssimon [Geoff Thorpe and Richard Levitte] 2930160814Ssimon 2931160814Ssimon *) Add Makefile.shared, a helper makefile to build shared 2932160814Ssimon libraries. Addapt Makefile.org. 2933160814Ssimon [Richard Levitte] 2934160814Ssimon 2935160814Ssimon *) Add version info to Win32 DLLs. 2936160814Ssimon [Peter 'Luna' Runestig" <peter@runestig.com>] 2937160814Ssimon 2938160814Ssimon *) Add new 'medium level' PKCS#12 API. Certificates and keys 2939160814Ssimon can be added using this API to created arbitrary PKCS#12 2940160814Ssimon files while avoiding the low level API. 2941160814Ssimon 2942160814Ssimon New options to PKCS12_create(), key or cert can be NULL and 2943160814Ssimon will then be omitted from the output file. The encryption 2944160814Ssimon algorithm NIDs can be set to -1 for no encryption, the mac 2945160814Ssimon iteration count can be set to 0 to omit the mac. 2946160814Ssimon 2947160814Ssimon Enhance pkcs12 utility by making the -nokeys and -nocerts 2948160814Ssimon options work when creating a PKCS#12 file. New option -nomac 2949160814Ssimon to omit the mac, NONE can be set for an encryption algorithm. 2950160814Ssimon New code is modified to use the enhanced PKCS12_create() 2951160814Ssimon instead of the low level API. 2952160814Ssimon [Steve Henson] 2953160814Ssimon 2954160814Ssimon *) Extend ASN1 encoder to support indefinite length constructed 2955160814Ssimon encoding. This can output sequences tags and octet strings in 2956160814Ssimon this form. Modify pk7_asn1.c to support indefinite length 2957160814Ssimon encoding. This is experimental and needs additional code to 2958160814Ssimon be useful, such as an ASN1 bio and some enhanced streaming 2959160814Ssimon PKCS#7 code. 2960160814Ssimon 2961160814Ssimon Extend template encode functionality so that tagging is passed 2962160814Ssimon down to the template encoder. 2963160814Ssimon [Steve Henson] 2964160814Ssimon 2965160814Ssimon *) Let 'openssl req' fail if an argument to '-newkey' is not 2966160814Ssimon recognized instead of using RSA as a default. 2967160814Ssimon [Bodo Moeller] 2968160814Ssimon 2969160814Ssimon *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt. 2970160814Ssimon As these are not official, they are not included in "ALL"; 2971160814Ssimon the "ECCdraft" ciphersuite group alias can be used to select them. 2972160814Ssimon [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)] 2973160814Ssimon 2974160814Ssimon *) Add ECDH engine support. 2975160814Ssimon [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)] 2976160814Ssimon 2977160814Ssimon *) Add ECDH in new directory crypto/ecdh/. 2978160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 2979160814Ssimon 2980160814Ssimon *) Let BN_rand_range() abort with an error after 100 iterations 2981160814Ssimon without success (which indicates a broken PRNG). 2982160814Ssimon [Bodo Moeller] 2983160814Ssimon 2984160814Ssimon *) Change BN_mod_sqrt() so that it verifies that the input value 2985160814Ssimon is really the square of the return value. (Previously, 2986160814Ssimon BN_mod_sqrt would show GIGO behaviour.) 2987160814Ssimon [Bodo Moeller] 2988160814Ssimon 2989160814Ssimon *) Add named elliptic curves over binary fields from X9.62, SECG, 2990160814Ssimon and WAP/WTLS; add OIDs that were still missing. 2991160814Ssimon 2992160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 2993160814Ssimon (Sun Microsystems Laboratories)] 2994160814Ssimon 2995160814Ssimon *) Extend the EC library for elliptic curves over binary fields 2996160814Ssimon (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/). 2997160814Ssimon New EC_METHOD: 2998160814Ssimon 2999160814Ssimon EC_GF2m_simple_method 3000160814Ssimon 3001160814Ssimon New API functions: 3002160814Ssimon 3003160814Ssimon EC_GROUP_new_curve_GF2m 3004160814Ssimon EC_GROUP_set_curve_GF2m 3005160814Ssimon EC_GROUP_get_curve_GF2m 3006160814Ssimon EC_POINT_set_affine_coordinates_GF2m 3007160814Ssimon EC_POINT_get_affine_coordinates_GF2m 3008160814Ssimon EC_POINT_set_compressed_coordinates_GF2m 3009160814Ssimon 3010160814Ssimon Point compression for binary fields is disabled by default for 3011160814Ssimon patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to 3012160814Ssimon enable it). 3013160814Ssimon 3014160814Ssimon As binary polynomials are represented as BIGNUMs, various members 3015160814Ssimon of the EC_GROUP and EC_POINT data structures can be shared 3016160814Ssimon between the implementations for prime fields and binary fields; 3017160814Ssimon the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m) 3018160814Ssimon are essentially identical to their ..._GFp counterparts. 3019160814Ssimon (For simplicity, the '..._GFp' prefix has been dropped from 3020160814Ssimon various internal method names.) 3021160814Ssimon 3022160814Ssimon An internal 'field_div' method (similar to 'field_mul' and 3023160814Ssimon 'field_sqr') has been added; this is used only for binary fields. 3024160814Ssimon 3025160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3026160814Ssimon (Sun Microsystems Laboratories)] 3027160814Ssimon 3028160814Ssimon *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult() 3029160814Ssimon through methods ('mul', 'precompute_mult'). 3030160814Ssimon 3031160814Ssimon The generic implementations (now internally called 'ec_wNAF_mul' 3032160814Ssimon and 'ec_wNAF_precomputed_mult') remain the default if these 3033160814Ssimon methods are undefined. 3034160814Ssimon 3035160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3036160814Ssimon (Sun Microsystems Laboratories)] 3037160814Ssimon 3038160814Ssimon *) New function EC_GROUP_get_degree, which is defined through 3039160814Ssimon EC_METHOD. For curves over prime fields, this returns the bit 3040160814Ssimon length of the modulus. 3041160814Ssimon 3042160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3043160814Ssimon (Sun Microsystems Laboratories)] 3044160814Ssimon 3045160814Ssimon *) New functions EC_GROUP_dup, EC_POINT_dup. 3046160814Ssimon (These simply call ..._new and ..._copy). 3047160814Ssimon 3048160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3049160814Ssimon (Sun Microsystems Laboratories)] 3050160814Ssimon 3051160814Ssimon *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c. 3052160814Ssimon Polynomials are represented as BIGNUMs (where the sign bit is not 3053160814Ssimon used) in the following functions [macros]: 3054160814Ssimon 3055160814Ssimon BN_GF2m_add 3056160814Ssimon BN_GF2m_sub [= BN_GF2m_add] 3057160814Ssimon BN_GF2m_mod [wrapper for BN_GF2m_mod_arr] 3058160814Ssimon BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr] 3059160814Ssimon BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr] 3060160814Ssimon BN_GF2m_mod_inv 3061160814Ssimon BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr] 3062160814Ssimon BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr] 3063160814Ssimon BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr] 3064160814Ssimon BN_GF2m_cmp [= BN_ucmp] 3065160814Ssimon 3066160814Ssimon (Note that only the 'mod' functions are actually for fields GF(2^m). 3067160814Ssimon BN_GF2m_add() is misnomer, but this is for the sake of consistency.) 3068160814Ssimon 3069160814Ssimon For some functions, an the irreducible polynomial defining a 3070160814Ssimon field can be given as an 'unsigned int[]' with strictly 3071160814Ssimon decreasing elements giving the indices of those bits that are set; 3072160814Ssimon i.e., p[] represents the polynomial 3073160814Ssimon f(t) = t^p[0] + t^p[1] + ... + t^p[k] 3074160814Ssimon where 3075160814Ssimon p[0] > p[1] > ... > p[k] = 0. 3076160814Ssimon This applies to the following functions: 3077160814Ssimon 3078160814Ssimon BN_GF2m_mod_arr 3079160814Ssimon BN_GF2m_mod_mul_arr 3080160814Ssimon BN_GF2m_mod_sqr_arr 3081160814Ssimon BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv] 3082160814Ssimon BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div] 3083160814Ssimon BN_GF2m_mod_exp_arr 3084160814Ssimon BN_GF2m_mod_sqrt_arr 3085160814Ssimon BN_GF2m_mod_solve_quad_arr 3086160814Ssimon BN_GF2m_poly2arr 3087160814Ssimon BN_GF2m_arr2poly 3088160814Ssimon 3089160814Ssimon Conversion can be performed by the following functions: 3090160814Ssimon 3091160814Ssimon BN_GF2m_poly2arr 3092160814Ssimon BN_GF2m_arr2poly 3093160814Ssimon 3094160814Ssimon bntest.c has additional tests for binary polynomial arithmetic. 3095160814Ssimon 3096160814Ssimon Two implementations for BN_GF2m_mod_div() are available. 3097160814Ssimon The default algorithm simply uses BN_GF2m_mod_inv() and 3098160814Ssimon BN_GF2m_mod_mul(). The alternative algorithm is compiled in only 3099160814Ssimon if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the 3100160814Ssimon copyright notice in crypto/bn/bn_gf2m.c before enabling it). 3101160814Ssimon 3102160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3103160814Ssimon (Sun Microsystems Laboratories)] 3104160814Ssimon 3105160814Ssimon *) Add new error code 'ERR_R_DISABLED' that can be used when some 3106160814Ssimon functionality is disabled at compile-time. 3107160814Ssimon [Douglas Stebila <douglas.stebila@sun.com>] 3108160814Ssimon 3109160814Ssimon *) Change default behaviour of 'openssl asn1parse' so that more 3110160814Ssimon information is visible when viewing, e.g., a certificate: 3111160814Ssimon 3112160814Ssimon Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump' 3113160814Ssimon mode the content of non-printable OCTET STRINGs is output in a 3114160814Ssimon style similar to INTEGERs, but with '[HEX DUMP]' prepended to 3115160814Ssimon avoid the appearance of a printable string. 3116160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3117160814Ssimon 3118160814Ssimon *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access 3119160814Ssimon functions 3120160814Ssimon EC_GROUP_set_asn1_flag() 3121160814Ssimon EC_GROUP_get_asn1_flag() 3122160814Ssimon EC_GROUP_set_point_conversion_form() 3123160814Ssimon EC_GROUP_get_point_conversion_form() 3124160814Ssimon These control ASN1 encoding details: 3125160814Ssimon - Curves (i.e., groups) are encoded explicitly unless asn1_flag 3126160814Ssimon has been set to OPENSSL_EC_NAMED_CURVE. 3127160814Ssimon - Points are encoded in uncompressed form by default; options for 3128160814Ssimon asn1_for are as for point2oct, namely 3129160814Ssimon POINT_CONVERSION_COMPRESSED 3130160814Ssimon POINT_CONVERSION_UNCOMPRESSED 3131160814Ssimon POINT_CONVERSION_HYBRID 3132160814Ssimon 3133160814Ssimon Also add 'seed' and 'seed_len' members to EC_GROUP with access 3134160814Ssimon functions 3135160814Ssimon EC_GROUP_set_seed() 3136160814Ssimon EC_GROUP_get0_seed() 3137160814Ssimon EC_GROUP_get_seed_len() 3138160814Ssimon This is used only for ASN1 purposes (so far). 3139160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3140160814Ssimon 3141160814Ssimon *) Add 'field_type' member to EC_METHOD, which holds the NID 3142160814Ssimon of the appropriate field type OID. The new function 3143160814Ssimon EC_METHOD_get_field_type() returns this value. 3144160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3145160814Ssimon 3146160814Ssimon *) Add functions 3147160814Ssimon EC_POINT_point2bn() 3148160814Ssimon EC_POINT_bn2point() 3149160814Ssimon EC_POINT_point2hex() 3150160814Ssimon EC_POINT_hex2point() 3151160814Ssimon providing useful interfaces to EC_POINT_point2oct() and 3152160814Ssimon EC_POINT_oct2point(). 3153160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3154160814Ssimon 3155160814Ssimon *) Change internals of the EC library so that the functions 3156160814Ssimon EC_GROUP_set_generator() 3157160814Ssimon EC_GROUP_get_generator() 3158160814Ssimon EC_GROUP_get_order() 3159160814Ssimon EC_GROUP_get_cofactor() 3160160814Ssimon are implemented directly in crypto/ec/ec_lib.c and not dispatched 3161160814Ssimon to methods, which would lead to unnecessary code duplication when 3162160814Ssimon adding different types of curves. 3163160814Ssimon [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller] 3164160814Ssimon 3165160814Ssimon *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM 3166160814Ssimon arithmetic, and such that modified wNAFs are generated 3167160814Ssimon (which avoid length expansion in many cases). 3168160814Ssimon [Bodo Moeller] 3169160814Ssimon 3170160814Ssimon *) Add a function EC_GROUP_check_discriminant() (defined via 3171160814Ssimon EC_METHOD) that verifies that the curve discriminant is non-zero. 3172160814Ssimon 3173160814Ssimon Add a function EC_GROUP_check() that makes some sanity tests 3174160814Ssimon on a EC_GROUP, its generator and order. This includes 3175160814Ssimon EC_GROUP_check_discriminant(). 3176160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3177160814Ssimon 3178160814Ssimon *) Add ECDSA in new directory crypto/ecdsa/. 3179160814Ssimon 3180160814Ssimon Add applications 'openssl ecparam' and 'openssl ecdsa' 3181160814Ssimon (these are based on 'openssl dsaparam' and 'openssl dsa'). 3182160814Ssimon 3183160814Ssimon ECDSA support is also included in various other files across the 3184160814Ssimon library. Most notably, 3185160814Ssimon - 'openssl req' now has a '-newkey ecdsa:file' option; 3186160814Ssimon - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA; 3187160814Ssimon - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and 3188160814Ssimon d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make 3189160814Ssimon them suitable for ECDSA where domain parameters must be 3190160814Ssimon extracted before the specific public key; 3191160814Ssimon - ECDSA engine support has been added. 3192160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3193160814Ssimon 3194160814Ssimon *) Include some named elliptic curves, and add OIDs from X9.62, 3195160814Ssimon SECG, and WAP/WTLS. Each curve can be obtained from the new 3196160814Ssimon function 3197160814Ssimon EC_GROUP_new_by_curve_name(), 3198160814Ssimon and the list of available named curves can be obtained with 3199160814Ssimon EC_get_builtin_curves(). 3200160814Ssimon Also add a 'curve_name' member to EC_GROUP objects, which can be 3201160814Ssimon accessed via 3202160814Ssimon EC_GROUP_set_curve_name() 3203160814Ssimon EC_GROUP_get_curve_name() 3204160814Ssimon [Nils Larsch <larsch@trustcenter.de, Bodo Moeller] 3205160814Ssimon 3206160814Ssimon *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 3207160814Ssimon was actually never needed) and in BN_mul(). The removal in BN_mul() 3208160814Ssimon required a small change in bn_mul_part_recursive() and the addition 3209160814Ssimon of the functions bn_cmp_part_words(), bn_sub_part_words() and 3210160814Ssimon bn_add_part_words(), which do the same thing as bn_cmp_words(), 3211160814Ssimon bn_sub_words() and bn_add_words() except they take arrays with 3212160814Ssimon differing sizes. 3213160814Ssimon [Richard Levitte] 3214160814Ssimon 3215194206Ssimon Changes between 0.9.7l and 0.9.7m [23 Feb 2007] 3216194206Ssimon 3217167612Ssimon *) Cleanse PEM buffers before freeing them since they may contain 3218167612Ssimon sensitive data. 3219167612Ssimon [Benjamin Bennett <ben@psc.edu>] 3220167612Ssimon 3221167612Ssimon *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 3222167612Ssimon a ciphersuite string such as "DEFAULT:RSA" cannot enable 3223167612Ssimon authentication-only ciphersuites. 3224167612Ssimon [Bodo Moeller] 3225167612Ssimon 3226167612Ssimon *) Since AES128 and AES256 share a single mask bit in the logic of 3227167612Ssimon ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 3228167612Ssimon kludge to work properly if AES128 is available and AES256 isn't. 3229167612Ssimon [Victor Duchovni] 3230167612Ssimon 3231194206Ssimon *) Expand security boundary to match 1.1.1 module. 3232194206Ssimon [Steve Henson] 3233194206Ssimon 3234194206Ssimon *) Remove redundant features: hash file source, editing of test vectors 3235194206Ssimon modify fipsld to use external fips_premain.c signature. 3236194206Ssimon [Steve Henson] 3237194206Ssimon 3238194206Ssimon *) New perl script mkfipsscr.pl to create shell scripts or batch files to 3239194206Ssimon run algorithm test programs. 3240194206Ssimon [Steve Henson] 3241194206Ssimon 3242194206Ssimon *) Make algorithm test programs more tolerant of whitespace. 3243194206Ssimon [Steve Henson] 3244194206Ssimon 3245167612Ssimon *) Have SSL/TLS server implementation tolerate "mismatched" record 3246167612Ssimon protocol version while receiving ClientHello even if the 3247167612Ssimon ClientHello is fragmented. (The server can't insist on the 3248167612Ssimon particular protocol version it has chosen before the ServerHello 3249167612Ssimon message has informed the client about his choice.) 3250167612Ssimon [Bodo Moeller] 3251167612Ssimon 3252167612Ssimon *) Load error codes if they are not already present instead of using a 3253167612Ssimon static variable. This allows them to be cleanly unloaded and reloaded. 3254167612Ssimon [Steve Henson] 3255167612Ssimon 3256167612Ssimon Changes between 0.9.7k and 0.9.7l [28 Sep 2006] 3257167612Ssimon 3258167612Ssimon *) Introduce limits to prevent malicious keys being able to 3259167612Ssimon cause a denial of service. (CVE-2006-2940) 3260167612Ssimon [Steve Henson, Bodo Moeller] 3261167612Ssimon 3262167612Ssimon *) Fix ASN.1 parsing of certain invalid structures that can result 3263167612Ssimon in a denial of service. (CVE-2006-2937) [Steve Henson] 3264167612Ssimon 3265167612Ssimon *) Fix buffer overflow in SSL_get_shared_ciphers() function. 3266167612Ssimon (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 3267167612Ssimon 3268167612Ssimon *) Fix SSL client code which could crash if connecting to a 3269167612Ssimon malicious SSLv2 server. (CVE-2006-4343) 3270167612Ssimon [Tavis Ormandy and Will Drewry, Google Security Team] 3271167612Ssimon 3272162911Ssimon *) Change ciphersuite string processing so that an explicit 3273162911Ssimon ciphersuite selects this one ciphersuite (so that "AES256-SHA" 3274162911Ssimon will no longer include "AES128-SHA"), and any other similar 3275162911Ssimon ciphersuite (same bitmap) from *other* protocol versions (so that 3276162911Ssimon "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the 3277162911Ssimon SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining 3278162911Ssimon changes from 0.9.8b and 0.9.8d. 3279162911Ssimon [Bodo Moeller] 3280162911Ssimon 3281162911Ssimon Changes between 0.9.7j and 0.9.7k [05 Sep 2006] 3282162911Ssimon 3283162911Ssimon *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 3284162911Ssimon (CVE-2006-4339) [Ben Laurie and Google Security Team] 3285162911Ssimon 3286162911Ssimon *) Change the Unix randomness entropy gathering to use poll() when 3287162911Ssimon possible instead of select(), since the latter has some 3288162911Ssimon undesirable limitations. 3289162911Ssimon [Darryl Miles via Richard Levitte and Bodo Moeller] 3290162911Ssimon 3291162911Ssimon *) Disable rogue ciphersuites: 3292162911Ssimon 3293162911Ssimon - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 3294162911Ssimon - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 3295162911Ssimon - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 3296162911Ssimon 3297162911Ssimon The latter two were purportedly from 3298162911Ssimon draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 3299162911Ssimon appear there. 3300162911Ssimon 3301162911Ssimon Also deactive the remaining ciphersuites from 3302162911Ssimon draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 3303162911Ssimon unofficial, and the ID has long expired. 3304162911Ssimon [Bodo Moeller] 3305162911Ssimon 3306162911Ssimon *) Fix RSA blinding Heisenbug (problems sometimes occured on 3307162911Ssimon dual-core machines) and other potential thread-safety issues. 3308162911Ssimon [Bodo Moeller] 3309162911Ssimon 3310162911Ssimon Changes between 0.9.7i and 0.9.7j [04 May 2006] 3311162911Ssimon 3312162911Ssimon *) Adapt fipsld and the build system to link against the validated FIPS 3313162911Ssimon module in FIPS mode. 3314162911Ssimon [Steve Henson] 3315162911Ssimon 3316162911Ssimon *) Fixes for VC++ 2005 build under Windows. 3317162911Ssimon [Steve Henson] 3318162911Ssimon 3319162911Ssimon *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make 3320162911Ssimon from a Windows bash shell such as MSYS. It is autodetected from the 3321162911Ssimon "config" script when run from a VC++ environment. Modify standard VC++ 3322162911Ssimon build to use fipscanister.o from the GNU make build. 3323162911Ssimon [Steve Henson] 3324162911Ssimon 3325160814Ssimon Changes between 0.9.7h and 0.9.7i [14 Oct 2005] 3326160814Ssimon 3327160814Ssimon *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS. 3328160814Ssimon The value now differs depending on if you build for FIPS or not. 3329160814Ssimon BEWARE! A program linked with a shared FIPSed libcrypto can't be 3330160814Ssimon safely run with a non-FIPSed libcrypto, as it may crash because of 3331160814Ssimon the difference induced by this change. 3332160814Ssimon [Andy Polyakov] 3333160814Ssimon 3334160814Ssimon Changes between 0.9.7g and 0.9.7h [11 Oct 2005] 3335160814Ssimon 3336160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 3337160814Ssimon (part of SSL_OP_ALL). This option used to disable the 3338160814Ssimon countermeasure against man-in-the-middle protocol-version 3339160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 3340160814Ssimon idea. (CVE-2005-2969) 3341160814Ssimon 3342160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 3343160814Ssimon for Information Security, National Institute of Advanced Industrial 3344160814Ssimon Science and Technology [AIST], Japan)] 3345160814Ssimon 3346160814Ssimon *) Minimal support for X9.31 signatures and PSS padding modes. This is 3347160814Ssimon mainly for FIPS compliance and not fully integrated at this stage. 3348160814Ssimon [Steve Henson] 3349160814Ssimon 3350160814Ssimon *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform 3351160814Ssimon the exponentiation using a fixed-length exponent. (Otherwise, 3352160814Ssimon the information leaked through timing could expose the secret key 3353160814Ssimon after many signatures; cf. Bleichenbacher's attack on DSA with 3354160814Ssimon biased k.) 3355160814Ssimon [Bodo Moeller] 3356160814Ssimon 3357160814Ssimon *) Make a new fixed-window mod_exp implementation the default for 3358160814Ssimon RSA, DSA, and DH private-key operations so that the sequence of 3359160814Ssimon squares and multiplies and the memory access pattern are 3360160814Ssimon independent of the particular secret key. This will mitigate 3361160814Ssimon cache-timing and potential related attacks. 3362160814Ssimon 3363160814Ssimon BN_mod_exp_mont_consttime() is the new exponentiation implementation, 3364160814Ssimon and this is automatically used by BN_mod_exp_mont() if the new flag 3365160814Ssimon BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH 3366160814Ssimon will use this BN flag for private exponents unless the flag 3367160814Ssimon RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or 3368160814Ssimon DH_FLAG_NO_EXP_CONSTTIME, respectively, is set. 3369160814Ssimon 3370160814Ssimon [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller] 3371160814Ssimon 3372160814Ssimon *) Change the client implementation for SSLv23_method() and 3373160814Ssimon SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0 3374160814Ssimon Client Hello message format if the SSL_OP_NO_SSLv2 option is set. 3375160814Ssimon (Previously, the SSL 2.0 backwards compatible Client Hello 3376160814Ssimon message format would be used even with SSL_OP_NO_SSLv2.) 3377160814Ssimon [Bodo Moeller] 3378160814Ssimon 3379160814Ssimon *) Add support for smime-type MIME parameter in S/MIME messages which some 3380160814Ssimon clients need. 3381160814Ssimon [Steve Henson] 3382160814Ssimon 3383160814Ssimon *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in 3384160814Ssimon a threadsafe manner. Modify rsa code to use new function and add calls 3385160814Ssimon to dsa and dh code (which had race conditions before). 3386160814Ssimon [Steve Henson] 3387160814Ssimon 3388160814Ssimon *) Include the fixed error library code in the C error file definitions 3389160814Ssimon instead of fixing them up at runtime. This keeps the error code 3390160814Ssimon structures constant. 3391160814Ssimon [Steve Henson] 3392160814Ssimon 3393160814Ssimon Changes between 0.9.7f and 0.9.7g [11 Apr 2005] 3394160814Ssimon 3395160814Ssimon [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after 3396160814Ssimon OpenSSL 0.9.8.] 3397160814Ssimon 3398160814Ssimon *) Fixes for newer kerberos headers. NB: the casts are needed because 3399160814Ssimon the 'length' field is signed on one version and unsigned on another 3400160814Ssimon with no (?) obvious way to tell the difference, without these VC++ 3401160814Ssimon complains. Also the "definition" of FAR (blank) is no longer included 3402160814Ssimon nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up 3403160814Ssimon some needed definitions. 3404160814Ssimon [Steve Henson] 3405160814Ssimon 3406160814Ssimon *) Undo Cygwin change. 3407160814Ssimon [Ulf M�ller] 3408160814Ssimon 3409160814Ssimon *) Added support for proxy certificates according to RFC 3820. 3410160814Ssimon Because they may be a security thread to unaware applications, 3411160814Ssimon they must be explicitely allowed in run-time. See 3412160814Ssimon docs/HOWTO/proxy_certificates.txt for further information. 3413160814Ssimon [Richard Levitte] 3414160814Ssimon 3415160814Ssimon Changes between 0.9.7e and 0.9.7f [22 Mar 2005] 3416160814Ssimon 3417160814Ssimon *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating 3418160814Ssimon server and client random values. Previously 3419160814Ssimon (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in 3420160814Ssimon less random data when sizeof(time_t) > 4 (some 64 bit platforms). 3421160814Ssimon 3422160814Ssimon This change has negligible security impact because: 3423160814Ssimon 3424160814Ssimon 1. Server and client random values still have 24 bytes of pseudo random 3425160814Ssimon data. 3426160814Ssimon 3427160814Ssimon 2. Server and client random values are sent in the clear in the initial 3428160814Ssimon handshake. 3429160814Ssimon 3430160814Ssimon 3. The master secret is derived using the premaster secret (48 bytes in 3431160814Ssimon size for static RSA ciphersuites) as well as client server and random 3432160814Ssimon values. 3433160814Ssimon 3434160814Ssimon The OpenSSL team would like to thank the UK NISCC for bringing this issue 3435160814Ssimon to our attention. 3436160814Ssimon 3437160814Ssimon [Stephen Henson, reported by UK NISCC] 3438160814Ssimon 3439160814Ssimon *) Use Windows randomness collection on Cygwin. 3440160814Ssimon [Ulf M�ller] 3441160814Ssimon 3442160814Ssimon *) Fix hang in EGD/PRNGD query when communication socket is closed 3443160814Ssimon prematurely by EGD/PRNGD. 3444160814Ssimon [Darren Tucker <dtucker@zip.com.au> via Lutz J�nicke, resolves #1014] 3445160814Ssimon 3446160814Ssimon *) Prompt for pass phrases when appropriate for PKCS12 input format. 3447160814Ssimon [Steve Henson] 3448160814Ssimon 3449160814Ssimon *) Back-port of selected performance improvements from development 3450160814Ssimon branch, as well as improved support for PowerPC platforms. 3451160814Ssimon [Andy Polyakov] 3452160814Ssimon 3453160814Ssimon *) Add lots of checks for memory allocation failure, error codes to indicate 3454160814Ssimon failure and freeing up memory if a failure occurs. 3455160814Ssimon [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson] 3456160814Ssimon 3457160814Ssimon *) Add new -passin argument to dgst. 3458160814Ssimon [Steve Henson] 3459160814Ssimon 3460160814Ssimon *) Perform some character comparisons of different types in X509_NAME_cmp: 3461160814Ssimon this is needed for some certificates that reencode DNs into UTF8Strings 3462160814Ssimon (in violation of RFC3280) and can't or wont issue name rollover 3463160814Ssimon certificates. 3464160814Ssimon [Steve Henson] 3465160814Ssimon 3466160814Ssimon *) Make an explicit check during certificate validation to see that 3467160814Ssimon the CA setting in each certificate on the chain is correct. As a 3468160814Ssimon side effect always do the following basic checks on extensions, 3469160814Ssimon not just when there's an associated purpose to the check: 3470160814Ssimon 3471160814Ssimon - if there is an unhandled critical extension (unless the user 3472160814Ssimon has chosen to ignore this fault) 3473160814Ssimon - if the path length has been exceeded (if one is set at all) 3474160814Ssimon - that certain extensions fit the associated purpose (if one has 3475160814Ssimon been given) 3476160814Ssimon [Richard Levitte] 3477160814Ssimon 3478142425Snectar Changes between 0.9.7d and 0.9.7e [25 Oct 2004] 3479142425Snectar 3480142425Snectar *) Avoid a race condition when CRLs are checked in a multi threaded 3481142425Snectar environment. This would happen due to the reordering of the revoked 3482142425Snectar entries during signature checking and serial number lookup. Now the 3483142425Snectar encoding is cached and the serial number sort performed under a lock. 3484142425Snectar Add new STACK function sk_is_sorted(). 3485142425Snectar [Steve Henson] 3486142425Snectar 3487142425Snectar *) Add Delta CRL to the extension code. 3488142425Snectar [Steve Henson] 3489142425Snectar 3490142425Snectar *) Various fixes to s3_pkt.c so alerts are sent properly. 3491142425Snectar [David Holmes <d.holmes@f5.com>] 3492142425Snectar 3493142425Snectar *) Reduce the chances of duplicate issuer name and serial numbers (in 3494142425Snectar violation of RFC3280) using the OpenSSL certificate creation utilities. 3495142425Snectar This is done by creating a random 64 bit value for the initial serial 3496142425Snectar number when a serial number file is created or when a self signed 3497142425Snectar certificate is created using 'openssl req -x509'. The initial serial 3498142425Snectar number file is created using 'openssl x509 -next_serial' in CA.pl 3499142425Snectar rather than being initialized to 1. 3500142425Snectar [Steve Henson] 3501142425Snectar 3502127128Snectar Changes between 0.9.7c and 0.9.7d [17 Mar 2004] 3503127128Snectar 3504127128Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 3505160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 3506127128Snectar [Joe Orton, Steve Henson] 3507127128Snectar 3508127128Snectar *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites 3509160814Ssimon (CVE-2004-0112) 3510127128Snectar [Joe Orton, Steve Henson] 3511127128Snectar 3512127128Snectar *) Make it possible to have multiple active certificates with the same 3513127128Snectar subject in the CA index file. This is done only if the keyword 3514127128Snectar 'unique_subject' is set to 'no' in the main CA section (default 3515127128Snectar if 'CA_default') of the configuration file. The value is saved 3516127128Snectar with the database itself in a separate index attribute file, 3517127128Snectar named like the index file with '.attr' appended to the name. 3518127128Snectar [Richard Levitte] 3519127128Snectar 3520127128Snectar *) X509 verify fixes. Disable broken certificate workarounds when 3521127128Snectar X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if 3522127128Snectar keyUsage extension present. Don't accept CRLs with unhandled critical 3523127128Snectar extensions: since verify currently doesn't process CRL extensions this 3524127128Snectar rejects a CRL with *any* critical extensions. Add new verify error codes 3525127128Snectar for these cases. 3526127128Snectar [Steve Henson] 3527127128Snectar 3528127128Snectar *) When creating an OCSP nonce use an OCTET STRING inside the extnValue. 3529127128Snectar A clarification of RFC2560 will require the use of OCTET STRINGs and 3530127128Snectar some implementations cannot handle the current raw format. Since OpenSSL 3531127128Snectar copies and compares OCSP nonces as opaque blobs without any attempt at 3532127128Snectar parsing them this should not create any compatibility issues. 3533127128Snectar [Steve Henson] 3534127128Snectar 3535127128Snectar *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when 3536127128Snectar calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without 3537127128Snectar this HMAC (and other) operations are several times slower than OpenSSL 3538127128Snectar < 0.9.7. 3539127128Snectar [Steve Henson] 3540127128Snectar 3541127128Snectar *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex(). 3542127128Snectar [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 3543127128Snectar 3544127128Snectar *) Use the correct content when signing type "other". 3545127128Snectar [Steve Henson] 3546127128Snectar 3547120631Snectar Changes between 0.9.7b and 0.9.7c [30 Sep 2003] 3548120631Snectar 3549120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 3550120631Snectar 3551120631Snectar Stop out of bounds reads in the ASN1 code when presented with 3552160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 3553120631Snectar 3554160814Ssimon Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545). 3555120631Snectar 3556120631Snectar If verify callback ignores invalid public key errors don't try to check 3557120631Snectar certificate signature with the NULL public key. 3558120631Snectar 3559120631Snectar [Steve Henson] 3560120631Snectar 3561120631Snectar *) New -ignore_err option in ocsp application to stop the server 3562120631Snectar exiting on the first error in a request. 3563120631Snectar [Steve Henson] 3564120631Snectar 3565120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 3566120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 3567120631Snectar specifications. 3568120631Snectar [Steve Henson] 3569120631Snectar 3570120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 3571120631Snectar extra data after the compression methods not only for TLS 1.0 3572120631Snectar but also for SSL 3.0 (as required by the specification). 3573120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 3574120631Snectar 3575120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 3576120631Snectar when it's 512 *bits* long, not 512 bytes. 3577120631Snectar [Richard Levitte] 3578120631Snectar 3579120631Snectar *) Change AES_cbc_encrypt() so it outputs exact multiple of 3580120631Snectar blocks during encryption. 3581120631Snectar [Richard Levitte] 3582120631Snectar 3583120631Snectar *) Various fixes to base64 BIO and non blocking I/O. On write 3584120631Snectar flushes were not handled properly if the BIO retried. On read 3585120631Snectar data was not being buffered properly and had various logic bugs. 3586120631Snectar This also affects blocking I/O when the data being decoded is a 3587120631Snectar certain size. 3588120631Snectar [Steve Henson] 3589120631Snectar 3590120631Snectar *) Various S/MIME bugfixes and compatibility changes: 3591120631Snectar output correct application/pkcs7 MIME type if 3592120631Snectar PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures. 3593120631Snectar Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening 3594120631Snectar of files as .eml work). Correctly handle very long lines in MIME 3595120631Snectar parser. 3596120631Snectar [Steve Henson] 3597120631Snectar 3598120631Snectar Changes between 0.9.7a and 0.9.7b [10 Apr 2003] 3599120631Snectar 3600120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 3601120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 3602120631Snectar a protocol version number mismatch like a decryption error 3603120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 3604120631Snectar [Bodo Moeller] 3605120631Snectar 3606120631Snectar *) Turn on RSA blinding by default in the default implementation 3607120631Snectar to avoid a timing attack. Applications that don't want it can call 3608120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 3609120631Snectar They would be ill-advised to do so in most cases. 3610120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 3611120631Snectar 3612120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 3613120631Snectar seeded (in this case, the secret RSA exponent is abused as 3614120631Snectar an unpredictable seed -- if it is not unpredictable, there 3615120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 3616120631Snectar by remembering the creator's thread ID in rsa->blinding and 3617120631Snectar having all other threads use local one-time blinding factors 3618120631Snectar (this requires more computation than sharing rsa->blinding, but 3619120631Snectar avoids excessive locking; and if an RSA object is not shared 3620120631Snectar between threads, blinding will still be very fast). 3621120631Snectar [Bodo Moeller] 3622120631Snectar 3623120631Snectar *) Fixed a typo bug that would cause ENGINE_set_default() to set an 3624120631Snectar ENGINE as defaults for all supported algorithms irrespective of 3625120631Snectar the 'flags' parameter. 'flags' is now honoured, so applications 3626120631Snectar should make sure they are passing it correctly. 3627120631Snectar [Geoff Thorpe] 3628120631Snectar 3629120631Snectar *) Target "mingw" now allows native Windows code to be generated in 3630120631Snectar the Cygwin environment as well as with the MinGW compiler. 3631120631Snectar [Ulf Moeller] 3632120631Snectar 3633111147Snectar Changes between 0.9.7 and 0.9.7a [19 Feb 2003] 3634111147Snectar 3635111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 3636111147Snectar via timing by performing a MAC computation even if incorrrect 3637111147Snectar block cipher padding has been found. This is a countermeasure 3638111147Snectar against active attacks where the attacker has to distinguish 3639160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 3640111147Snectar 3641111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 3642111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 3643111147Snectar Martin Vuagnoux (EPFL, Ilion)] 3644111147Snectar 3645111147Snectar *) Make the no-err option work as intended. The intention with no-err 3646111147Snectar is not to have the whole error stack handling routines removed from 3647111147Snectar libcrypto, it's only intended to remove all the function name and 3648111147Snectar reason texts, thereby removing some of the footprint that may not 3649111147Snectar be interesting if those errors aren't displayed anyway. 3650111147Snectar 3651111147Snectar NOTE: it's still possible for any application or module to have it's 3652111147Snectar own set of error texts inserted. The routines are there, just not 3653111147Snectar used by default when no-err is given. 3654111147Snectar [Richard Levitte] 3655111147Snectar 3656111147Snectar *) Add support for FreeBSD on IA64. 3657111147Snectar [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454] 3658111147Snectar 3659111147Snectar *) Adjust DES_cbc_cksum() so it returns the same value as the MIT 3660111147Snectar Kerberos function mit_des_cbc_cksum(). Before this change, 3661111147Snectar the value returned by DES_cbc_cksum() was like the one from 3662111147Snectar mit_des_cbc_cksum(), except the bytes were swapped. 3663111147Snectar [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte] 3664111147Snectar 3665111147Snectar *) Allow an application to disable the automatic SSL chain building. 3666111147Snectar Before this a rather primitive chain build was always performed in 3667111147Snectar ssl3_output_cert_chain(): an application had no way to send the 3668111147Snectar correct chain if the automatic operation produced an incorrect result. 3669111147Snectar 3670111147Snectar Now the chain builder is disabled if either: 3671111147Snectar 3672111147Snectar 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert(). 3673111147Snectar 3674111147Snectar 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set. 3675111147Snectar 3676111147Snectar The reasoning behind this is that an application would not want the 3677111147Snectar auto chain building to take place if extra chain certificates are 3678111147Snectar present and it might also want a means of sending no additional 3679111147Snectar certificates (for example the chain has two certificates and the 3680111147Snectar root is omitted). 3681111147Snectar [Steve Henson] 3682111147Snectar 3683111147Snectar *) Add the possibility to build without the ENGINE framework. 3684111147Snectar [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 3685111147Snectar 3686111147Snectar *) Under Win32 gmtime() can return NULL: check return value in 3687111147Snectar OPENSSL_gmtime(). Add error code for case where gmtime() fails. 3688111147Snectar [Steve Henson] 3689111147Snectar 3690111147Snectar *) DSA routines: under certain error conditions uninitialized BN objects 3691111147Snectar could be freed. Solution: make sure initialization is performed early 3692111147Snectar enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>, 3693111147Snectar Nils Larsch <nla@trustcenter.de> via PR#459) 3694111147Snectar [Lutz Jaenicke] 3695111147Snectar 3696111147Snectar *) Another fix for SSLv2 session ID handling: the session ID was incorrectly 3697111147Snectar checked on reconnect on the client side, therefore session resumption 3698111147Snectar could still fail with a "ssl session id is different" error. This 3699111147Snectar behaviour is masked when SSL_OP_ALL is used due to 3700111147Snectar SSL_OP_MICROSOFT_SESS_ID_BUG being set. 3701111147Snectar Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 3702111147Snectar followup to PR #377. 3703111147Snectar [Lutz Jaenicke] 3704111147Snectar 3705111147Snectar *) IA-32 assembler support enhancements: unified ELF targets, support 3706111147Snectar for SCO/Caldera platforms, fix for Cygwin shared build. 3707111147Snectar [Andy Polyakov] 3708111147Snectar 3709111147Snectar *) Add support for FreeBSD on sparc64. As a consequence, support for 3710111147Snectar FreeBSD on non-x86 processors is separate from x86 processors on 3711111147Snectar the config script, much like the NetBSD support. 3712111147Snectar [Richard Levitte & Kris Kennaway <kris@obsecurity.org>] 3713111147Snectar 3714109998Smarkm Changes between 0.9.6h and 0.9.7 [31 Dec 2002] 3715109998Smarkm 3716120631Snectar [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after 3717120631Snectar OpenSSL 0.9.7.] 3718120631Snectar 3719109998Smarkm *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED 3720109998Smarkm code (06) was taken as the first octet of the session ID and the last 3721109998Smarkm octet was ignored consequently. As a result SSLv2 client side session 3722109998Smarkm caching could not have worked due to the session ID mismatch between 3723109998Smarkm client and server. 3724109998Smarkm Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 3725109998Smarkm PR #377. 3726109998Smarkm [Lutz Jaenicke] 3727109998Smarkm 3728109998Smarkm *) Change the declaration of needed Kerberos libraries to use EX_LIBS 3729109998Smarkm instead of the special (and badly supported) LIBKRB5. LIBKRB5 is 3730109998Smarkm removed entirely. 3731109998Smarkm [Richard Levitte] 3732109998Smarkm 3733109998Smarkm *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it 3734109998Smarkm seems that in spite of existing for more than a year, many application 3735109998Smarkm author have done nothing to provide the necessary callbacks, which 3736109998Smarkm means that this particular engine will not work properly anywhere. 3737109998Smarkm This is a very unfortunate situation which forces us, in the name 3738109998Smarkm of usability, to give the hw_ncipher.c a static lock, which is part 3739109998Smarkm of libcrypto. 3740109998Smarkm NOTE: This is for the 0.9.7 series ONLY. This hack will never 3741109998Smarkm appear in 0.9.8 or later. We EXPECT application authors to have 3742109998Smarkm dealt properly with this when 0.9.8 is released (unless we actually 3743109998Smarkm make such changes in the libcrypto locking code that changes will 3744109998Smarkm have to be made anyway). 3745109998Smarkm [Richard Levitte] 3746109998Smarkm 3747109998Smarkm *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content 3748109998Smarkm octets have been read, EOF or an error occurs. Without this change 3749109998Smarkm some truncated ASN1 structures will not produce an error. 3750109998Smarkm [Steve Henson] 3751109998Smarkm 3752109998Smarkm *) Disable Heimdal support, since it hasn't been fully implemented. 3753109998Smarkm Still give the possibility to force the use of Heimdal, but with 3754109998Smarkm warnings and a request that patches get sent to openssl-dev. 3755109998Smarkm [Richard Levitte] 3756109998Smarkm 3757109998Smarkm *) Add the VC-CE target, introduce the WINCE sysname, and add 3758109998Smarkm INSTALL.WCE and appropriate conditionals to make it build. 3759109998Smarkm [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 3760109998Smarkm 3761109998Smarkm *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and 3762109998Smarkm cygssl-x.y.z.dll, where x, y and z are the major, minor and 3763109998Smarkm edit numbers of the version. 3764109998Smarkm [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 3765109998Smarkm 3766109998Smarkm *) Introduce safe string copy and catenation functions 3767109998Smarkm (BUF_strlcpy() and BUF_strlcat()). 3768109998Smarkm [Ben Laurie (CHATS) and Richard Levitte] 3769109998Smarkm 3770109998Smarkm *) Avoid using fixed-size buffers for one-line DNs. 3771109998Smarkm [Ben Laurie (CHATS)] 3772109998Smarkm 3773109998Smarkm *) Add BUF_MEM_grow_clean() to avoid information leakage when 3774109998Smarkm resizing buffers containing secrets, and use where appropriate. 3775109998Smarkm [Ben Laurie (CHATS)] 3776109998Smarkm 3777109998Smarkm *) Avoid using fixed size buffers for configuration file location. 3778109998Smarkm [Ben Laurie (CHATS)] 3779109998Smarkm 3780109998Smarkm *) Avoid filename truncation for various CA files. 3781109998Smarkm [Ben Laurie (CHATS)] 3782109998Smarkm 3783109998Smarkm *) Use sizeof in preference to magic numbers. 3784109998Smarkm [Ben Laurie (CHATS)] 3785109998Smarkm 3786109998Smarkm *) Avoid filename truncation in cert requests. 3787109998Smarkm [Ben Laurie (CHATS)] 3788109998Smarkm 3789109998Smarkm *) Add assertions to check for (supposedly impossible) buffer 3790109998Smarkm overflows. 3791109998Smarkm [Ben Laurie (CHATS)] 3792109998Smarkm 3793109998Smarkm *) Don't cache truncated DNS entries in the local cache (this could 3794109998Smarkm potentially lead to a spoofing attack). 3795109998Smarkm [Ben Laurie (CHATS)] 3796109998Smarkm 3797109998Smarkm *) Fix various buffers to be large enough for hex/decimal 3798109998Smarkm representations in a platform independent manner. 3799109998Smarkm [Ben Laurie (CHATS)] 3800109998Smarkm 3801109998Smarkm *) Add CRYPTO_realloc_clean() to avoid information leakage when 3802109998Smarkm resizing buffers containing secrets, and use where appropriate. 3803109998Smarkm [Ben Laurie (CHATS)] 3804109998Smarkm 3805109998Smarkm *) Add BIO_indent() to avoid much slightly worrying code to do 3806109998Smarkm indents. 3807109998Smarkm [Ben Laurie (CHATS)] 3808109998Smarkm 3809109998Smarkm *) Convert sprintf()/BIO_puts() to BIO_printf(). 3810109998Smarkm [Ben Laurie (CHATS)] 3811109998Smarkm 3812109998Smarkm *) buffer_gets() could terminate with the buffer only half 3813109998Smarkm full. Fixed. 3814109998Smarkm [Ben Laurie (CHATS)] 3815109998Smarkm 3816109998Smarkm *) Add assertions to prevent user-supplied crypto functions from 3817109998Smarkm overflowing internal buffers by having large block sizes, etc. 3818109998Smarkm [Ben Laurie (CHATS)] 3819109998Smarkm 3820109998Smarkm *) New OPENSSL_assert() macro (similar to assert(), but enabled 3821109998Smarkm unconditionally). 3822109998Smarkm [Ben Laurie (CHATS)] 3823109998Smarkm 3824109998Smarkm *) Eliminate unused copy of key in RC4. 3825109998Smarkm [Ben Laurie (CHATS)] 3826109998Smarkm 3827109998Smarkm *) Eliminate unused and incorrectly sized buffers for IV in pem.h. 3828109998Smarkm [Ben Laurie (CHATS)] 3829109998Smarkm 3830109998Smarkm *) Fix off-by-one error in EGD path. 3831109998Smarkm [Ben Laurie (CHATS)] 3832109998Smarkm 3833109998Smarkm *) If RANDFILE path is too long, ignore instead of truncating. 3834109998Smarkm [Ben Laurie (CHATS)] 3835109998Smarkm 3836109998Smarkm *) Eliminate unused and incorrectly sized X.509 structure 3837109998Smarkm CBCParameter. 3838109998Smarkm [Ben Laurie (CHATS)] 3839109998Smarkm 3840109998Smarkm *) Eliminate unused and dangerous function knumber(). 3841109998Smarkm [Ben Laurie (CHATS)] 3842109998Smarkm 3843109998Smarkm *) Eliminate unused and dangerous structure, KSSL_ERR. 3844109998Smarkm [Ben Laurie (CHATS)] 3845109998Smarkm 3846109998Smarkm *) Protect against overlong session ID context length in an encoded 3847109998Smarkm session object. Since these are local, this does not appear to be 3848109998Smarkm exploitable. 3849109998Smarkm [Ben Laurie (CHATS)] 3850109998Smarkm 3851109998Smarkm *) Change from security patch (see 0.9.6e below) that did not affect 3852109998Smarkm the 0.9.6 release series: 3853109998Smarkm 3854109998Smarkm Remote buffer overflow in SSL3 protocol - an attacker could 3855109998Smarkm supply an oversized master key in Kerberos-enabled versions. 3856160814Ssimon (CVE-2002-0657) 3857109998Smarkm [Ben Laurie (CHATS)] 3858109998Smarkm 3859109998Smarkm *) Change the SSL kerb5 codes to match RFC 2712. 3860109998Smarkm [Richard Levitte] 3861109998Smarkm 3862109998Smarkm *) Make -nameopt work fully for req and add -reqopt switch. 3863109998Smarkm [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson] 3864109998Smarkm 3865109998Smarkm *) The "block size" for block ciphers in CFB and OFB mode should be 1. 3866109998Smarkm [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>] 3867109998Smarkm 3868109998Smarkm *) Make sure tests can be performed even if the corresponding algorithms 3869109998Smarkm have been removed entirely. This was also the last step to make 3870109998Smarkm OpenSSL compilable with DJGPP under all reasonable conditions. 3871109998Smarkm [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>] 3872109998Smarkm 3873109998Smarkm *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT 3874109998Smarkm to allow version independent disabling of normally unselected ciphers, 3875109998Smarkm which may be activated as a side-effect of selecting a single cipher. 3876109998Smarkm 3877109998Smarkm (E.g., cipher list string "RSA" enables ciphersuites that are left 3878109998Smarkm out of "ALL" because they do not provide symmetric encryption. 3879109998Smarkm "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) 3880109998Smarkm [Lutz Jaenicke, Bodo Moeller] 3881109998Smarkm 3882109998Smarkm *) Add appropriate support for separate platform-dependent build 3883109998Smarkm directories. The recommended way to make a platform-dependent 3884109998Smarkm build directory is the following (tested on Linux), maybe with 3885109998Smarkm some local tweaks: 3886109998Smarkm 3887109998Smarkm # Place yourself outside of the OpenSSL source tree. In 3888109998Smarkm # this example, the environment variable OPENSSL_SOURCE 3889109998Smarkm # is assumed to contain the absolute OpenSSL source directory. 3890109998Smarkm mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" 3891109998Smarkm cd objtree/"`uname -s`-`uname -r`-`uname -m`" 3892111147Snectar (cd $OPENSSL_SOURCE; find . -type f) | while read F; do 3893109998Smarkm mkdir -p `dirname $F` 3894109998Smarkm ln -s $OPENSSL_SOURCE/$F $F 3895109998Smarkm done 3896109998Smarkm 3897109998Smarkm To be absolutely sure not to disturb the source tree, a "make clean" 3898109998Smarkm is a good thing. If it isn't successfull, don't worry about it, 3899109998Smarkm it probably means the source directory is very clean. 3900109998Smarkm [Richard Levitte] 3901109998Smarkm 3902109998Smarkm *) Make sure any ENGINE control commands make local copies of string 3903109998Smarkm pointers passed to them whenever necessary. Otherwise it is possible 3904109998Smarkm the caller may have overwritten (or deallocated) the original string 3905109998Smarkm data when a later ENGINE operation tries to use the stored values. 3906109998Smarkm [G�tz Babin-Ebell <babinebell@trustcenter.de>] 3907109998Smarkm 3908109998Smarkm *) Improve diagnostics in file reading and command-line digests. 3909109998Smarkm [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>] 3910109998Smarkm 3911109998Smarkm *) Add AES modes CFB and OFB to the object database. Correct an 3912109998Smarkm error in AES-CFB decryption. 3913109998Smarkm [Richard Levitte] 3914109998Smarkm 3915109998Smarkm *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this 3916109998Smarkm allows existing EVP_CIPHER_CTX structures to be reused after 3917109998Smarkm calling EVP_*Final(). This behaviour is used by encryption 3918109998Smarkm BIOs and some applications. This has the side effect that 3919109998Smarkm applications must explicitly clean up cipher contexts with 3920109998Smarkm EVP_CIPHER_CTX_cleanup() or they will leak memory. 3921109998Smarkm [Steve Henson] 3922109998Smarkm 3923109998Smarkm *) Check the values of dna and dnb in bn_mul_recursive before calling 3924109998Smarkm bn_mul_comba (a non zero value means the a or b arrays do not contain 3925109998Smarkm n2 elements) and fallback to bn_mul_normal if either is not zero. 3926109998Smarkm [Steve Henson] 3927109998Smarkm 3928109998Smarkm *) Fix escaping of non-ASCII characters when using the -subj option 3929109998Smarkm of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) 3930109998Smarkm [Lutz Jaenicke] 3931109998Smarkm 3932109998Smarkm *) Make object definitions compliant to LDAP (RFC2256): SN is the short 3933109998Smarkm form for "surname", serialNumber has no short form. 3934109998Smarkm Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; 3935109998Smarkm therefore remove "mail" short name for "internet 7". 3936109998Smarkm The OID for unique identifiers in X509 certificates is 3937109998Smarkm x500UniqueIdentifier, not uniqueIdentifier. 3938109998Smarkm Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) 3939109998Smarkm [Lutz Jaenicke] 3940109998Smarkm 3941109998Smarkm *) Add an "init" command to the ENGINE config module and auto initialize 3942109998Smarkm ENGINEs. Without any "init" command the ENGINE will be initialized 3943109998Smarkm after all ctrl commands have been executed on it. If init=1 the 3944109998Smarkm ENGINE is initailized at that point (ctrls before that point are run 3945109998Smarkm on the uninitialized ENGINE and after on the initialized one). If 3946109998Smarkm init=0 then the ENGINE will not be iniatialized at all. 3947109998Smarkm [Steve Henson] 3948109998Smarkm 3949109998Smarkm *) Fix the 'app_verify_callback' interface so that the user-defined 3950109998Smarkm argument is actually passed to the callback: In the 3951109998Smarkm SSL_CTX_set_cert_verify_callback() prototype, the callback 3952109998Smarkm declaration has been changed from 3953109998Smarkm int (*cb)() 3954109998Smarkm into 3955109998Smarkm int (*cb)(X509_STORE_CTX *,void *); 3956109998Smarkm in ssl_verify_cert_chain (ssl/ssl_cert.c), the call 3957109998Smarkm i=s->ctx->app_verify_callback(&ctx) 3958109998Smarkm has been changed into 3959109998Smarkm i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). 3960109998Smarkm 3961109998Smarkm To update applications using SSL_CTX_set_cert_verify_callback(), 3962109998Smarkm a dummy argument can be added to their callback functions. 3963109998Smarkm [D. K. Smetters <smetters@parc.xerox.com>] 3964109998Smarkm 3965109998Smarkm *) Added the '4758cca' ENGINE to support IBM 4758 cards. 3966109998Smarkm [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe] 3967109998Smarkm 3968109998Smarkm *) Add and OPENSSL_LOAD_CONF define which will cause 3969109998Smarkm OpenSSL_add_all_algorithms() to load the openssl.cnf config file. 3970109998Smarkm This allows older applications to transparently support certain 3971109998Smarkm OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. 3972109998Smarkm Two new functions OPENSSL_add_all_algorithms_noconf() which will never 3973109998Smarkm load the config file and OPENSSL_add_all_algorithms_conf() which will 3974109998Smarkm always load it have also been added. 3975109998Smarkm [Steve Henson] 3976109998Smarkm 3977109998Smarkm *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. 3978109998Smarkm Adjust NIDs and EVP layer. 3979109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 3980109998Smarkm 3981109998Smarkm *) Config modules support in openssl utility. 3982109998Smarkm 3983109998Smarkm Most commands now load modules from the config file, 3984109998Smarkm though in a few (such as version) this isn't done 3985109998Smarkm because it couldn't be used for anything. 3986109998Smarkm 3987109998Smarkm In the case of ca and req the config file used is 3988109998Smarkm the same as the utility itself: that is the -config 3989109998Smarkm command line option can be used to specify an 3990109998Smarkm alternative file. 3991109998Smarkm [Steve Henson] 3992109998Smarkm 3993109998Smarkm *) Move default behaviour from OPENSSL_config(). If appname is NULL 3994109998Smarkm use "openssl_conf" if filename is NULL use default openssl config file. 3995109998Smarkm [Steve Henson] 3996109998Smarkm 3997109998Smarkm *) Add an argument to OPENSSL_config() to allow the use of an alternative 3998109998Smarkm config section name. Add a new flag to tolerate a missing config file 3999109998Smarkm and move code to CONF_modules_load_file(). 4000109998Smarkm [Steve Henson] 4001109998Smarkm 4002109998Smarkm *) Support for crypto accelerator cards from Accelerated Encryption 4003109998Smarkm Processing, www.aep.ie. (Use engine 'aep') 4004109998Smarkm The support was copied from 0.9.6c [engine] and adapted/corrected 4005109998Smarkm to work with the new engine framework. 4006109998Smarkm [AEP Inc. and Richard Levitte] 4007109998Smarkm 4008109998Smarkm *) Support for SureWare crypto accelerator cards from Baltimore 4009109998Smarkm Technologies. (Use engine 'sureware') 4010109998Smarkm The support was copied from 0.9.6c [engine] and adapted 4011109998Smarkm to work with the new engine framework. 4012109998Smarkm [Richard Levitte] 4013109998Smarkm 4014109998Smarkm *) Have the CHIL engine fork-safe (as defined by nCipher) and actually 4015109998Smarkm make the newer ENGINE framework commands for the CHIL engine work. 4016109998Smarkm [Toomas Kiisk <vix@cyber.ee> and Richard Levitte] 4017109998Smarkm 4018109998Smarkm *) Make it possible to produce shared libraries on ReliantUNIX. 4019109998Smarkm [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte] 4020109998Smarkm 4021109998Smarkm *) Add the configuration target debug-linux-ppro. 4022109998Smarkm Make 'openssl rsa' use the general key loading routines 4023109998Smarkm implemented in apps.c, and make those routines able to 4024109998Smarkm handle the key format FORMAT_NETSCAPE and the variant 4025109998Smarkm FORMAT_IISSGC. 4026109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 4027109998Smarkm 4028109998Smarkm *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 4029109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 4030109998Smarkm 4031109998Smarkm *) Add -keyform to rsautl, and document -engine. 4032109998Smarkm [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>] 4033109998Smarkm 4034109998Smarkm *) Change BIO_new_file (crypto/bio/bss_file.c) to use new 4035109998Smarkm BIO_R_NO_SUCH_FILE error code rather than the generic 4036109998Smarkm ERR_R_SYS_LIB error code if fopen() fails with ENOENT. 4037109998Smarkm [Ben Laurie] 4038109998Smarkm 4039109998Smarkm *) Add new functions 4040109998Smarkm ERR_peek_last_error 4041109998Smarkm ERR_peek_last_error_line 4042109998Smarkm ERR_peek_last_error_line_data. 4043109998Smarkm These are similar to 4044109998Smarkm ERR_peek_error 4045109998Smarkm ERR_peek_error_line 4046109998Smarkm ERR_peek_error_line_data, 4047109998Smarkm but report on the latest error recorded rather than the first one 4048109998Smarkm still in the error queue. 4049109998Smarkm [Ben Laurie, Bodo Moeller] 4050109998Smarkm 4051109998Smarkm *) default_algorithms option in ENGINE config module. This allows things 4052109998Smarkm like: 4053109998Smarkm default_algorithms = ALL 4054109998Smarkm default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS 4055109998Smarkm [Steve Henson] 4056109998Smarkm 4057109998Smarkm *) Prelminary ENGINE config module. 4058109998Smarkm [Steve Henson] 4059109998Smarkm 4060109998Smarkm *) New experimental application configuration code. 4061109998Smarkm [Steve Henson] 4062109998Smarkm 4063109998Smarkm *) Change the AES code to follow the same name structure as all other 4064109998Smarkm symmetric ciphers, and behave the same way. Move everything to 4065109998Smarkm the directory crypto/aes, thereby obsoleting crypto/rijndael. 4066109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 4067109998Smarkm 4068109998Smarkm *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. 4069109998Smarkm [Ben Laurie and Theo de Raadt] 4070109998Smarkm 4071109998Smarkm *) Add option to output public keys in req command. 4072109998Smarkm [Massimiliano Pala madwolf@openca.org] 4073109998Smarkm 4074109998Smarkm *) Use wNAFs in EC_POINTs_mul() for improved efficiency 4075109998Smarkm (up to about 10% better than before for P-192 and P-224). 4076109998Smarkm [Bodo Moeller] 4077109998Smarkm 4078109998Smarkm *) New functions/macros 4079109998Smarkm 4080109998Smarkm SSL_CTX_set_msg_callback(ctx, cb) 4081109998Smarkm SSL_CTX_set_msg_callback_arg(ctx, arg) 4082109998Smarkm SSL_set_msg_callback(ssl, cb) 4083109998Smarkm SSL_set_msg_callback_arg(ssl, arg) 4084109998Smarkm 4085109998Smarkm to request calling a callback function 4086109998Smarkm 4087109998Smarkm void cb(int write_p, int version, int content_type, 4088109998Smarkm const void *buf, size_t len, SSL *ssl, void *arg) 4089109998Smarkm 4090109998Smarkm whenever a protocol message has been completely received 4091109998Smarkm (write_p == 0) or sent (write_p == 1). Here 'version' is the 4092109998Smarkm protocol version according to which the SSL library interprets 4093109998Smarkm the current protocol message (SSL2_VERSION, SSL3_VERSION, or 4094109998Smarkm TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or 4095109998Smarkm the content type as defined in the SSL 3.0/TLS 1.0 protocol 4096109998Smarkm specification (change_cipher_spec(20), alert(21), handshake(22)). 4097109998Smarkm 'buf' and 'len' point to the actual message, 'ssl' to the 4098109998Smarkm SSL object, and 'arg' is the application-defined value set by 4099109998Smarkm SSL[_CTX]_set_msg_callback_arg(). 4100109998Smarkm 4101109998Smarkm 'openssl s_client' and 'openssl s_server' have new '-msg' options 4102109998Smarkm to enable a callback that displays all protocol messages. 4103109998Smarkm [Bodo Moeller] 4104109998Smarkm 4105109998Smarkm *) Change the shared library support so shared libraries are built as 4106109998Smarkm soon as the corresponding static library is finished, and thereby get 4107109998Smarkm openssl and the test programs linked against the shared library. 4108109998Smarkm This still only happens when the keyword "shard" has been given to 4109109998Smarkm the configuration scripts. 4110109998Smarkm 4111109998Smarkm NOTE: shared library support is still an experimental thing, and 4112109998Smarkm backward binary compatibility is still not guaranteed. 4113109998Smarkm ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte] 4114109998Smarkm 4115109998Smarkm *) Add support for Subject Information Access extension. 4116109998Smarkm [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 4117109998Smarkm 4118109998Smarkm *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero 4119109998Smarkm additional bytes when new memory had to be allocated, not just 4120109998Smarkm when reusing an existing buffer. 4121109998Smarkm [Bodo Moeller] 4122109998Smarkm 4123109998Smarkm *) New command line and configuration option 'utf8' for the req command. 4124109998Smarkm This allows field values to be specified as UTF8 strings. 4125109998Smarkm [Steve Henson] 4126109998Smarkm 4127109998Smarkm *) Add -multi and -mr options to "openssl speed" - giving multiple parallel 4128109998Smarkm runs for the former and machine-readable output for the latter. 4129109998Smarkm [Ben Laurie] 4130109998Smarkm 4131109998Smarkm *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion 4132109998Smarkm of the e-mail address in the DN (i.e., it will go into a certificate 4133109998Smarkm extension only). The new configuration file option 'email_in_dn = no' 4134109998Smarkm has the same effect. 4135109998Smarkm [Massimiliano Pala madwolf@openca.org] 4136109998Smarkm 4137109998Smarkm *) Change all functions with names starting with des_ to be starting 4138109998Smarkm with DES_ instead. Add wrappers that are compatible with libdes, 4139109998Smarkm but are named _ossl_old_des_*. Finally, add macros that map the 4140109998Smarkm des_* symbols to the corresponding _ossl_old_des_* if libdes 4141109998Smarkm compatibility is desired. If OpenSSL 0.9.6c compatibility is 4142109998Smarkm desired, the des_* symbols will be mapped to DES_*, with one 4143109998Smarkm exception. 4144109998Smarkm 4145109998Smarkm Since we provide two compatibility mappings, the user needs to 4146109998Smarkm define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes 4147109998Smarkm compatibility is desired. The default (i.e., when that macro 4148109998Smarkm isn't defined) is OpenSSL 0.9.6c compatibility. 4149109998Smarkm 4150109998Smarkm There are also macros that enable and disable the support of old 4151109998Smarkm des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT 4152109998Smarkm and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those 4153109998Smarkm are defined, the default will apply: to support the old des routines. 4154109998Smarkm 4155109998Smarkm In either case, one must include openssl/des.h to get the correct 4156109998Smarkm definitions. Do not try to just include openssl/des_old.h, that 4157109998Smarkm won't work. 4158109998Smarkm 4159109998Smarkm NOTE: This is a major break of an old API into a new one. Software 4160109998Smarkm authors are encouraged to switch to the DES_ style functions. Some 4161109998Smarkm time in the future, des_old.h and the libdes compatibility functions 4162109998Smarkm will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the 4163109998Smarkm default), and then completely removed. 4164109998Smarkm [Richard Levitte] 4165109998Smarkm 4166109998Smarkm *) Test for certificates which contain unsupported critical extensions. 4167109998Smarkm If such a certificate is found during a verify operation it is 4168109998Smarkm rejected by default: this behaviour can be overridden by either 4169109998Smarkm handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or 4170109998Smarkm by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function 4171109998Smarkm X509_supported_extension() has also been added which returns 1 if a 4172109998Smarkm particular extension is supported. 4173109998Smarkm [Steve Henson] 4174109998Smarkm 4175109998Smarkm *) Modify the behaviour of EVP cipher functions in similar way to digests 4176109998Smarkm to retain compatibility with existing code. 4177109998Smarkm [Steve Henson] 4178109998Smarkm 4179109998Smarkm *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain 4180109998Smarkm compatibility with existing code. In particular the 'ctx' parameter does 4181109998Smarkm not have to be to be initialized before the call to EVP_DigestInit() and 4182109998Smarkm it is tidied up after a call to EVP_DigestFinal(). New function 4183109998Smarkm EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function 4184109998Smarkm EVP_MD_CTX_copy() changed to not require the destination to be 4185109998Smarkm initialized valid and new function EVP_MD_CTX_copy_ex() added which 4186109998Smarkm requires the destination to be valid. 4187109998Smarkm 4188109998Smarkm Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), 4189109998Smarkm EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). 4190109998Smarkm [Steve Henson] 4191109998Smarkm 4192109998Smarkm *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it 4193109998Smarkm so that complete 'Handshake' protocol structures are kept in memory 4194109998Smarkm instead of overwriting 'msg_type' and 'length' with 'body' data. 4195109998Smarkm [Bodo Moeller] 4196109998Smarkm 4197109998Smarkm *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. 4198109998Smarkm [Massimo Santin via Richard Levitte] 4199109998Smarkm 4200109998Smarkm *) Major restructuring to the underlying ENGINE code. This includes 4201109998Smarkm reduction of linker bloat, separation of pure "ENGINE" manipulation 4202109998Smarkm (initialisation, etc) from functionality dealing with implementations 4203109998Smarkm of specific crypto iterfaces. This change also introduces integrated 4204109998Smarkm support for symmetric ciphers and digest implementations - so ENGINEs 4205109998Smarkm can now accelerate these by providing EVP_CIPHER and EVP_MD 4206109998Smarkm implementations of their own. This is detailed in crypto/engine/README 4207109998Smarkm as it couldn't be adequately described here. However, there are a few 4208109998Smarkm API changes worth noting - some RSA, DSA, DH, and RAND functions that 4209109998Smarkm were changed in the original introduction of ENGINE code have now 4210109998Smarkm reverted back - the hooking from this code to ENGINE is now a good 4211109998Smarkm deal more passive and at run-time, operations deal directly with 4212109998Smarkm RSA_METHODs, DSA_METHODs (etc) as they did before, rather than 4213109998Smarkm dereferencing through an ENGINE pointer any more. Also, the ENGINE 4214109998Smarkm functions dealing with BN_MOD_EXP[_CRT] handlers have been removed - 4215109998Smarkm they were not being used by the framework as there is no concept of a 4216109998Smarkm BIGNUM_METHOD and they could not be generalised to the new 4217109998Smarkm 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, 4218109998Smarkm ENGINE_cpy() has been removed as it cannot be consistently defined in 4219109998Smarkm the new code. 4220109998Smarkm [Geoff Thorpe] 4221109998Smarkm 4222109998Smarkm *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. 4223109998Smarkm [Steve Henson] 4224109998Smarkm 4225109998Smarkm *) Change mkdef.pl to sort symbols that get the same entry number, 4226109998Smarkm and make sure the automatically generated functions ERR_load_* 4227109998Smarkm become part of libeay.num as well. 4228109998Smarkm [Richard Levitte] 4229109998Smarkm 4230109998Smarkm *) New function SSL_renegotiate_pending(). This returns true once 4231109998Smarkm renegotiation has been requested (either SSL_renegotiate() call 4232109998Smarkm or HelloRequest/ClientHello receveived from the peer) and becomes 4233109998Smarkm false once a handshake has been completed. 4234109998Smarkm (For servers, SSL_renegotiate() followed by SSL_do_handshake() 4235109998Smarkm sends a HelloRequest, but does not ensure that a handshake takes 4236109998Smarkm place. SSL_renegotiate_pending() is useful for checking if the 4237109998Smarkm client has followed the request.) 4238109998Smarkm [Bodo Moeller] 4239109998Smarkm 4240109998Smarkm *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. 4241109998Smarkm By default, clients may request session resumption even during 4242109998Smarkm renegotiation (if session ID contexts permit); with this option, 4243109998Smarkm session resumption is possible only in the first handshake. 4244109998Smarkm 4245109998Smarkm SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes 4246109998Smarkm more bits available for options that should not be part of 4247109998Smarkm SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). 4248109998Smarkm [Bodo Moeller] 4249109998Smarkm 4250109998Smarkm *) Add some demos for certificate and certificate request creation. 4251109998Smarkm [Steve Henson] 4252109998Smarkm 4253109998Smarkm *) Make maximum certificate chain size accepted from the peer application 4254109998Smarkm settable (SSL*_get/set_max_cert_list()), as proposed by 4255109998Smarkm "Douglas E. Engert" <deengert@anl.gov>. 4256109998Smarkm [Lutz Jaenicke] 4257109998Smarkm 4258109998Smarkm *) Add support for shared libraries for Unixware-7 4259109998Smarkm (Boyd Lynn Gerber <gerberb@zenez.com>). 4260109998Smarkm [Lutz Jaenicke] 4261109998Smarkm 4262109998Smarkm *) Add a "destroy" handler to ENGINEs that allows structural cleanup to 4263109998Smarkm be done prior to destruction. Use this to unload error strings from 4264109998Smarkm ENGINEs that load their own error strings. NB: This adds two new API 4265109998Smarkm functions to "get" and "set" this destroy handler in an ENGINE. 4266109998Smarkm [Geoff Thorpe] 4267109998Smarkm 4268109998Smarkm *) Alter all existing ENGINE implementations (except "openssl" and 4269109998Smarkm "openbsd") to dynamically instantiate their own error strings. This 4270109998Smarkm makes them more flexible to be built both as statically-linked ENGINEs 4271109998Smarkm and self-contained shared-libraries loadable via the "dynamic" ENGINE. 4272109998Smarkm Also, add stub code to each that makes building them as self-contained 4273109998Smarkm shared-libraries easier (see README.ENGINE). 4274109998Smarkm [Geoff Thorpe] 4275109998Smarkm 4276109998Smarkm *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE 4277109998Smarkm implementations into applications that are completely implemented in 4278109998Smarkm self-contained shared-libraries. The "dynamic" ENGINE exposes control 4279109998Smarkm commands that can be used to configure what shared-library to load and 4280109998Smarkm to control aspects of the way it is handled. Also, made an update to 4281109998Smarkm the README.ENGINE file that brings its information up-to-date and 4282109998Smarkm provides some information and instructions on the "dynamic" ENGINE 4283109998Smarkm (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). 4284109998Smarkm [Geoff Thorpe] 4285109998Smarkm 4286109998Smarkm *) Make it possible to unload ranges of ERR strings with a new 4287109998Smarkm "ERR_unload_strings" function. 4288109998Smarkm [Geoff Thorpe] 4289109998Smarkm 4290109998Smarkm *) Add a copy() function to EVP_MD. 4291109998Smarkm [Ben Laurie] 4292109998Smarkm 4293109998Smarkm *) Make EVP_MD routines take a context pointer instead of just the 4294109998Smarkm md_data void pointer. 4295109998Smarkm [Ben Laurie] 4296109998Smarkm 4297109998Smarkm *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates 4298109998Smarkm that the digest can only process a single chunk of data 4299109998Smarkm (typically because it is provided by a piece of 4300109998Smarkm hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application 4301109998Smarkm is only going to provide a single chunk of data, and hence the 4302109998Smarkm framework needn't accumulate the data for oneshot drivers. 4303109998Smarkm [Ben Laurie] 4304109998Smarkm 4305109998Smarkm *) As with "ERR", make it possible to replace the underlying "ex_data" 4306109998Smarkm functions. This change also alters the storage and management of global 4307109998Smarkm ex_data state - it's now all inside ex_data.c and all "class" code (eg. 4308109998Smarkm RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class 4309109998Smarkm index counters. The API functions that use this state have been changed 4310109998Smarkm to take a "class_index" rather than pointers to the class's local STACK 4311109998Smarkm and counter, and there is now an API function to dynamically create new 4312109998Smarkm classes. This centralisation allows us to (a) plug a lot of the 4313109998Smarkm thread-safety problems that existed, and (b) makes it possible to clean 4314109998Smarkm up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) 4315109998Smarkm such data would previously have always leaked in application code and 4316109998Smarkm workarounds were in place to make the memory debugging turn a blind eye 4317109998Smarkm to it. Application code that doesn't use this new function will still 4318109998Smarkm leak as before, but their memory debugging output will announce it now 4319109998Smarkm rather than letting it slide. 4320109998Smarkm 4321109998Smarkm Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change 4322109998Smarkm induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now 4323109998Smarkm has a return value to indicate success or failure. 4324109998Smarkm [Geoff Thorpe] 4325109998Smarkm 4326109998Smarkm *) Make it possible to replace the underlying "ERR" functions such that the 4327109998Smarkm global state (2 LHASH tables and 2 locks) is only used by the "default" 4328109998Smarkm implementation. This change also adds two functions to "get" and "set" 4329109998Smarkm the implementation prior to it being automatically set the first time 4330109998Smarkm any other ERR function takes place. Ie. an application can call "get", 4331109998Smarkm pass the return value to a module it has just loaded, and that module 4332109998Smarkm can call its own "set" function using that value. This means the 4333109998Smarkm module's "ERR" operations will use (and modify) the error state in the 4334109998Smarkm application and not in its own statically linked copy of OpenSSL code. 4335109998Smarkm [Geoff Thorpe] 4336109998Smarkm 4337109998Smarkm *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment 4338109998Smarkm reference counts. This performs normal REF_PRINT/REF_CHECK macros on 4339109998Smarkm the operation, and provides a more encapsulated way for external code 4340109998Smarkm (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code 4341109998Smarkm to use these functions rather than manually incrementing the counts. 4342109998Smarkm 4343109998Smarkm Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". 4344109998Smarkm [Geoff Thorpe] 4345109998Smarkm 4346109998Smarkm *) Add EVP test program. 4347109998Smarkm [Ben Laurie] 4348109998Smarkm 4349109998Smarkm *) Add symmetric cipher support to ENGINE. Expect the API to change! 4350109998Smarkm [Ben Laurie] 4351109998Smarkm 4352109998Smarkm *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() 4353109998Smarkm X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), 4354109998Smarkm X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). 4355109998Smarkm These allow a CRL to be built without having to access X509_CRL fields 4356109998Smarkm directly. Modify 'ca' application to use new functions. 4357109998Smarkm [Steve Henson] 4358109998Smarkm 4359109998Smarkm *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended 4360109998Smarkm bug workarounds. Rollback attack detection is a security feature. 4361109998Smarkm The problem will only arise on OpenSSL servers when TLSv1 is not 4362109998Smarkm available (sslv3_server_method() or SSL_OP_NO_TLSv1). 4363109998Smarkm Software authors not wanting to support TLSv1 will have special reasons 4364109998Smarkm for their choice and can explicitly enable this option. 4365109998Smarkm [Bodo Moeller, Lutz Jaenicke] 4366109998Smarkm 4367109998Smarkm *) Rationalise EVP so it can be extended: don't include a union of 4368109998Smarkm cipher/digest structures, add init/cleanup functions for EVP_MD_CTX 4369109998Smarkm (similar to those existing for EVP_CIPHER_CTX). 4370109998Smarkm Usage example: 4371109998Smarkm 4372109998Smarkm EVP_MD_CTX md; 4373109998Smarkm 4374109998Smarkm EVP_MD_CTX_init(&md); /* new function call */ 4375109998Smarkm EVP_DigestInit(&md, EVP_sha1()); 4376109998Smarkm EVP_DigestUpdate(&md, in, len); 4377109998Smarkm EVP_DigestFinal(&md, out, NULL); 4378109998Smarkm EVP_MD_CTX_cleanup(&md); /* new function call */ 4379109998Smarkm 4380109998Smarkm [Ben Laurie] 4381109998Smarkm 4382109998Smarkm *) Make DES key schedule conform to the usual scheme, as well as 4383109998Smarkm correcting its structure. This means that calls to DES functions 4384109998Smarkm now have to pass a pointer to a des_key_schedule instead of a 4385109998Smarkm plain des_key_schedule (which was actually always a pointer 4386109998Smarkm anyway): E.g., 4387109998Smarkm 4388109998Smarkm des_key_schedule ks; 4389109998Smarkm 4390109998Smarkm des_set_key_checked(..., &ks); 4391109998Smarkm des_ncbc_encrypt(..., &ks, ...); 4392109998Smarkm 4393109998Smarkm (Note that a later change renames 'des_...' into 'DES_...'.) 4394109998Smarkm [Ben Laurie] 4395109998Smarkm 4396109998Smarkm *) Initial reduction of linker bloat: the use of some functions, such as 4397109998Smarkm PEM causes large amounts of unused functions to be linked in due to 4398109998Smarkm poor organisation. For example pem_all.c contains every PEM function 4399109998Smarkm which has a knock on effect of linking in large amounts of (unused) 4400109998Smarkm ASN1 code. Grouping together similar functions and splitting unrelated 4401109998Smarkm functions prevents this. 4402109998Smarkm [Steve Henson] 4403109998Smarkm 4404109998Smarkm *) Cleanup of EVP macros. 4405109998Smarkm [Ben Laurie] 4406109998Smarkm 4407109998Smarkm *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the 4408109998Smarkm correct _ecb suffix. 4409109998Smarkm [Ben Laurie] 4410109998Smarkm 4411109998Smarkm *) Add initial OCSP responder support to ocsp application. The 4412109998Smarkm revocation information is handled using the text based index 4413109998Smarkm use by the ca application. The responder can either handle 4414109998Smarkm requests generated internally, supplied in files (for example 4415109998Smarkm via a CGI script) or using an internal minimal server. 4416109998Smarkm [Steve Henson] 4417109998Smarkm 4418109998Smarkm *) Add configuration choices to get zlib compression for TLS. 4419109998Smarkm [Richard Levitte] 4420109998Smarkm 4421109998Smarkm *) Changes to Kerberos SSL for RFC 2712 compliance: 4422109998Smarkm 1. Implemented real KerberosWrapper, instead of just using 4423109998Smarkm KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] 4424109998Smarkm 2. Implemented optional authenticator field of KerberosWrapper. 4425109998Smarkm 4426109998Smarkm Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, 4427109998Smarkm and authenticator structs; see crypto/krb5/. 4428109998Smarkm 4429109998Smarkm Generalized Kerberos calls to support multiple Kerberos libraries. 4430109998Smarkm [Vern Staats <staatsvr@asc.hpc.mil>, 4431109998Smarkm Jeffrey Altman <jaltman@columbia.edu> 4432109998Smarkm via Richard Levitte] 4433109998Smarkm 4434109998Smarkm *) Cause 'openssl speed' to use fully hard-coded DSA keys as it 4435109998Smarkm already does with RSA. testdsa.h now has 'priv_key/pub_key' 4436109998Smarkm values for each of the key sizes rather than having just 4437109998Smarkm parameters (and 'speed' generating keys each time). 4438109998Smarkm [Geoff Thorpe] 4439109998Smarkm 4440109998Smarkm *) Speed up EVP routines. 4441109998Smarkm Before: 4442109998Smarkmencrypt 4443109998Smarkmtype 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 4444109998Smarkmdes-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k 4445109998Smarkmdes-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k 4446109998Smarkmdes-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k 4447109998Smarkmdecrypt 4448109998Smarkmdes-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k 4449109998Smarkmdes-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k 4450109998Smarkmdes-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k 4451109998Smarkm After: 4452109998Smarkmencrypt 4453109998Smarkmdes-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k 4454109998Smarkmdecrypt 4455109998Smarkmdes-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k 4456109998Smarkm [Ben Laurie] 4457109998Smarkm 4458109998Smarkm *) Added the OS2-EMX target. 4459109998Smarkm ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte] 4460109998Smarkm 4461109998Smarkm *) Rewrite apps to use NCONF routines instead of the old CONF. New functions 4462109998Smarkm to support NCONF routines in extension code. New function CONF_set_nconf() 4463109998Smarkm to allow functions which take an NCONF to also handle the old LHASH 4464109998Smarkm structure: this means that the old CONF compatible routines can be 4465109998Smarkm retained (in particular wrt extensions) without having to duplicate the 4466109998Smarkm code. New function X509V3_add_ext_nconf_sk to add extensions to a stack. 4467109998Smarkm [Steve Henson] 4468109998Smarkm 4469109998Smarkm *) Enhance the general user interface with mechanisms for inner control 4470109998Smarkm and with possibilities to have yes/no kind of prompts. 4471109998Smarkm [Richard Levitte] 4472109998Smarkm 4473109998Smarkm *) Change all calls to low level digest routines in the library and 4474109998Smarkm applications to use EVP. Add missing calls to HMAC_cleanup() and 4475109998Smarkm don't assume HMAC_CTX can be copied using memcpy(). 4476109998Smarkm [Verdon Walker <VWalker@novell.com>, Steve Henson] 4477109998Smarkm 4478109998Smarkm *) Add the possibility to control engines through control names but with 4479109998Smarkm arbitrary arguments instead of just a string. 4480109998Smarkm Change the key loaders to take a UI_METHOD instead of a callback 4481109998Smarkm function pointer. NOTE: this breaks binary compatibility with earlier 4482109998Smarkm versions of OpenSSL [engine]. 4483109998Smarkm Adapt the nCipher code for these new conditions and add a card insertion 4484109998Smarkm callback. 4485109998Smarkm [Richard Levitte] 4486109998Smarkm 4487109998Smarkm *) Enhance the general user interface with mechanisms to better support 4488109998Smarkm dialog box interfaces, application-defined prompts, the possibility 4489109998Smarkm to use defaults (for example default passwords from somewhere else) 4490109998Smarkm and interrupts/cancellations. 4491109998Smarkm [Richard Levitte] 4492109998Smarkm 4493109998Smarkm *) Tidy up PKCS#12 attribute handling. Add support for the CSP name 4494109998Smarkm attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. 4495109998Smarkm [Steve Henson] 4496109998Smarkm 4497109998Smarkm *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also 4498109998Smarkm tidy up some unnecessarily weird code in 'sk_new()'). 4499109998Smarkm [Geoff, reported by Diego Tartara <dtartara@novamens.com>] 4500109998Smarkm 4501109998Smarkm *) Change the key loading routines for ENGINEs to use the same kind 4502109998Smarkm callback (pem_password_cb) as all other routines that need this 4503109998Smarkm kind of callback. 4504109998Smarkm [Richard Levitte] 4505109998Smarkm 4506109998Smarkm *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with 4507109998Smarkm 256 bit (=32 byte) keys. Of course seeding with more entropy bytes 4508109998Smarkm than this minimum value is recommended. 4509109998Smarkm [Lutz Jaenicke] 4510109998Smarkm 4511109998Smarkm *) New random seeder for OpenVMS, using the system process statistics 4512109998Smarkm that are easily reachable. 4513109998Smarkm [Richard Levitte] 4514109998Smarkm 4515109998Smarkm *) Windows apparently can't transparently handle global 4516109998Smarkm variables defined in DLLs. Initialisations such as: 4517109998Smarkm 4518109998Smarkm const ASN1_ITEM *it = &ASN1_INTEGER_it; 4519109998Smarkm 4520109998Smarkm wont compile. This is used by the any applications that need to 4521109998Smarkm declare their own ASN1 modules. This was fixed by adding the option 4522109998Smarkm EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly 4523109998Smarkm needed for static libraries under Win32. 4524109998Smarkm [Steve Henson] 4525109998Smarkm 4526109998Smarkm *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle 4527109998Smarkm setting of purpose and trust fields. New X509_STORE trust and 4528109998Smarkm purpose functions and tidy up setting in other SSL functions. 4529109998Smarkm [Steve Henson] 4530109998Smarkm 4531109998Smarkm *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE 4532109998Smarkm structure. These are inherited by X509_STORE_CTX when it is 4533109998Smarkm initialised. This allows various defaults to be set in the 4534109998Smarkm X509_STORE structure (such as flags for CRL checking and custom 4535109998Smarkm purpose or trust settings) for functions which only use X509_STORE_CTX 4536109998Smarkm internally such as S/MIME. 4537109998Smarkm 4538109998Smarkm Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and 4539109998Smarkm trust settings if they are not set in X509_STORE. This allows X509_STORE 4540109998Smarkm purposes and trust (in S/MIME for example) to override any set by default. 4541109998Smarkm 4542109998Smarkm Add command line options for CRL checking to smime, s_client and s_server 4543109998Smarkm applications. 4544109998Smarkm [Steve Henson] 4545109998Smarkm 4546109998Smarkm *) Initial CRL based revocation checking. If the CRL checking flag(s) 4547109998Smarkm are set then the CRL is looked up in the X509_STORE structure and 4548109998Smarkm its validity and signature checked, then if the certificate is found 4549109998Smarkm in the CRL the verify fails with a revoked error. 4550109998Smarkm 4551109998Smarkm Various new CRL related callbacks added to X509_STORE_CTX structure. 4552109998Smarkm 4553109998Smarkm Command line options added to 'verify' application to support this. 4554109998Smarkm 4555109998Smarkm This needs some additional work, such as being able to handle multiple 4556109998Smarkm CRLs with different times, extension based lookup (rather than just 4557109998Smarkm by subject name) and ultimately more complete V2 CRL extension 4558109998Smarkm handling. 4559109998Smarkm [Steve Henson] 4560109998Smarkm 4561109998Smarkm *) Add a general user interface API (crypto/ui/). This is designed 4562109998Smarkm to replace things like des_read_password and friends (backward 4563109998Smarkm compatibility functions using this new API are provided). 4564109998Smarkm The purpose is to remove prompting functions from the DES code 4565109998Smarkm section as well as provide for prompting through dialog boxes in 4566109998Smarkm a window system and the like. 4567109998Smarkm [Richard Levitte] 4568109998Smarkm 4569109998Smarkm *) Add "ex_data" support to ENGINE so implementations can add state at a 4570109998Smarkm per-structure level rather than having to store it globally. 4571109998Smarkm [Geoff] 4572109998Smarkm 4573109998Smarkm *) Make it possible for ENGINE structures to be copied when retrieved by 4574109998Smarkm ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. 4575109998Smarkm This causes the "original" ENGINE structure to act like a template, 4576109998Smarkm analogous to the RSA vs. RSA_METHOD type of separation. Because of this 4577109998Smarkm operational state can be localised to each ENGINE structure, despite the 4578109998Smarkm fact they all share the same "methods". New ENGINE structures returned in 4579109998Smarkm this case have no functional references and the return value is the single 4580109998Smarkm structural reference. This matches the single structural reference returned 4581109998Smarkm by ENGINE_by_id() normally, when it is incremented on the pre-existing 4582109998Smarkm ENGINE structure. 4583109998Smarkm [Geoff] 4584109998Smarkm 4585109998Smarkm *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this 4586109998Smarkm needs to match any other type at all we need to manually clear the 4587109998Smarkm tag cache. 4588109998Smarkm [Steve Henson] 4589109998Smarkm 4590109998Smarkm *) Changes to the "openssl engine" utility to include; 4591109998Smarkm - verbosity levels ('-v', '-vv', and '-vvv') that provide information 4592109998Smarkm about an ENGINE's available control commands. 4593109998Smarkm - executing control commands from command line arguments using the 4594109998Smarkm '-pre' and '-post' switches. '-post' is only used if '-t' is 4595109998Smarkm specified and the ENGINE is successfully initialised. The syntax for 4596109998Smarkm the individual commands are colon-separated, for example; 4597109998Smarkm openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so 4598109998Smarkm [Geoff] 4599109998Smarkm 4600109998Smarkm *) New dynamic control command support for ENGINEs. ENGINEs can now 4601109998Smarkm declare their own commands (numbers), names (strings), descriptions, 4602109998Smarkm and input types for run-time discovery by calling applications. A 4603109998Smarkm subset of these commands are implicitly classed as "executable" 4604109998Smarkm depending on their input type, and only these can be invoked through 4605109998Smarkm the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this 4606109998Smarkm can be based on user input, config files, etc). The distinction is 4607109998Smarkm that "executable" commands cannot return anything other than a boolean 4608109998Smarkm result and can only support numeric or string input, whereas some 4609109998Smarkm discoverable commands may only be for direct use through 4610109998Smarkm ENGINE_ctrl(), eg. supporting the exchange of binary data, function 4611109998Smarkm pointers, or other custom uses. The "executable" commands are to 4612109998Smarkm support parameterisations of ENGINE behaviour that can be 4613109998Smarkm unambiguously defined by ENGINEs and used consistently across any 4614109998Smarkm OpenSSL-based application. Commands have been added to all the 4615109998Smarkm existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow 4616109998Smarkm control over shared-library paths without source code alterations. 4617109998Smarkm [Geoff] 4618109998Smarkm 4619109998Smarkm *) Changed all ENGINE implementations to dynamically allocate their 4620109998Smarkm ENGINEs rather than declaring them statically. Apart from this being 4621109998Smarkm necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, 4622109998Smarkm this also allows the implementations to compile without using the 4623109998Smarkm internal engine_int.h header. 4624109998Smarkm [Geoff] 4625109998Smarkm 4626109998Smarkm *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a 4627109998Smarkm 'const' value. Any code that should be able to modify a RAND_METHOD 4628109998Smarkm should already have non-const pointers to it (ie. they should only 4629109998Smarkm modify their own ones). 4630109998Smarkm [Geoff] 4631109998Smarkm 4632109998Smarkm *) Made a variety of little tweaks to the ENGINE code. 4633109998Smarkm - "atalla" and "ubsec" string definitions were moved from header files 4634109998Smarkm to C code. "nuron" string definitions were placed in variables 4635109998Smarkm rather than hard-coded - allowing parameterisation of these values 4636109998Smarkm later on via ctrl() commands. 4637109998Smarkm - Removed unused "#if 0"'d code. 4638109998Smarkm - Fixed engine list iteration code so it uses ENGINE_free() to release 4639109998Smarkm structural references. 4640109998Smarkm - Constified the RAND_METHOD element of ENGINE structures. 4641109998Smarkm - Constified various get/set functions as appropriate and added 4642109998Smarkm missing functions (including a catch-all ENGINE_cpy that duplicates 4643109998Smarkm all ENGINE values onto a new ENGINE except reference counts/state). 4644109998Smarkm - Removed NULL parameter checks in get/set functions. Setting a method 4645109998Smarkm or function to NULL is a way of cancelling out a previously set 4646109998Smarkm value. Passing a NULL ENGINE parameter is just plain stupid anyway 4647109998Smarkm and doesn't justify the extra error symbols and code. 4648109998Smarkm - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for 4649109998Smarkm flags from engine_int.h to engine.h. 4650109998Smarkm - Changed prototypes for ENGINE handler functions (init(), finish(), 4651109998Smarkm ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. 4652109998Smarkm [Geoff] 4653109998Smarkm 4654109998Smarkm *) Implement binary inversion algorithm for BN_mod_inverse in addition 4655109998Smarkm to the algorithm using long division. The binary algorithm can be 4656109998Smarkm used only if the modulus is odd. On 32-bit systems, it is faster 4657109998Smarkm only for relatively small moduli (roughly 20-30% for 128-bit moduli, 4658109998Smarkm roughly 5-15% for 256-bit moduli), so we use it only for moduli 4659109998Smarkm up to 450 bits. In 64-bit environments, the binary algorithm 4660109998Smarkm appears to be advantageous for much longer moduli; here we use it 4661109998Smarkm for moduli up to 2048 bits. 4662109998Smarkm [Bodo Moeller] 4663109998Smarkm 4664109998Smarkm *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code 4665109998Smarkm could not support the combine flag in choice fields. 4666109998Smarkm [Steve Henson] 4667109998Smarkm 4668109998Smarkm *) Add a 'copy_extensions' option to the 'ca' utility. This copies 4669109998Smarkm extensions from a certificate request to the certificate. 4670109998Smarkm [Steve Henson] 4671109998Smarkm 4672109998Smarkm *) Allow multiple 'certopt' and 'nameopt' options to be separated 4673109998Smarkm by commas. Add 'namopt' and 'certopt' options to the 'ca' config 4674109998Smarkm file: this allows the display of the certificate about to be 4675109998Smarkm signed to be customised, to allow certain fields to be included 4676109998Smarkm or excluded and extension details. The old system didn't display 4677109998Smarkm multicharacter strings properly, omitted fields not in the policy 4678109998Smarkm and couldn't display additional details such as extensions. 4679109998Smarkm [Steve Henson] 4680109998Smarkm 4681109998Smarkm *) Function EC_POINTs_mul for multiple scalar multiplication 4682109998Smarkm of an arbitrary number of elliptic curve points 4683109998Smarkm \sum scalars[i]*points[i], 4684109998Smarkm optionally including the generator defined for the EC_GROUP: 4685109998Smarkm scalar*generator + \sum scalars[i]*points[i]. 4686109998Smarkm 4687109998Smarkm EC_POINT_mul is a simple wrapper function for the typical case 4688109998Smarkm that the point list has just one item (besides the optional 4689109998Smarkm generator). 4690109998Smarkm [Bodo Moeller] 4691109998Smarkm 4692109998Smarkm *) First EC_METHODs for curves over GF(p): 4693109998Smarkm 4694109998Smarkm EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr 4695109998Smarkm operations and provides various method functions that can also 4696109998Smarkm operate with faster implementations of modular arithmetic. 4697109998Smarkm 4698109998Smarkm EC_GFp_mont_method() reuses most functions that are part of 4699109998Smarkm EC_GFp_simple_method, but uses Montgomery arithmetic. 4700109998Smarkm 4701109998Smarkm [Bodo Moeller; point addition and point doubling 4702109998Smarkm implementation directly derived from source code provided by 4703109998Smarkm Lenka Fibikova <fibikova@exp-math.uni-essen.de>] 4704109998Smarkm 4705109998Smarkm *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, 4706109998Smarkm crypto/ec/ec_lib.c): 4707109998Smarkm 4708109998Smarkm Curves are EC_GROUP objects (with an optional group generator) 4709109998Smarkm based on EC_METHODs that are built into the library. 4710109998Smarkm 4711109998Smarkm Points are EC_POINT objects based on EC_GROUP objects. 4712109998Smarkm 4713109998Smarkm Most of the framework would be able to handle curves over arbitrary 4714109998Smarkm finite fields, but as there are no obvious types for fields other 4715109998Smarkm than GF(p), some functions are limited to that for now. 4716109998Smarkm [Bodo Moeller] 4717109998Smarkm 4718109998Smarkm *) Add the -HTTP option to s_server. It is similar to -WWW, but requires 4719109998Smarkm that the file contains a complete HTTP response. 4720109998Smarkm [Richard Levitte] 4721109998Smarkm 4722109998Smarkm *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl 4723109998Smarkm change the def and num file printf format specifier from "%-40sXXX" 4724109998Smarkm to "%-39s XXX". The latter will always guarantee a space after the 4725109998Smarkm field while the former will cause them to run together if the field 4726109998Smarkm is 40 of more characters long. 4727109998Smarkm [Steve Henson] 4728109998Smarkm 4729109998Smarkm *) Constify the cipher and digest 'method' functions and structures 4730109998Smarkm and modify related functions to take constant EVP_MD and EVP_CIPHER 4731109998Smarkm pointers. 4732109998Smarkm [Steve Henson] 4733109998Smarkm 4734109998Smarkm *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them 4735109998Smarkm in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. 4736109998Smarkm [Bodo Moeller] 4737109998Smarkm 4738109998Smarkm *) Modify EVP_Digest*() routines so they now return values. Although the 4739109998Smarkm internal software routines can never fail additional hardware versions 4740109998Smarkm might. 4741109998Smarkm [Steve Henson] 4742109998Smarkm 4743109998Smarkm *) Clean up crypto/err/err.h and change some error codes to avoid conflicts: 4744109998Smarkm 4745109998Smarkm Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 4746109998Smarkm (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. 4747109998Smarkm 4748109998Smarkm ASN1 error codes 4749109998Smarkm ERR_R_NESTED_ASN1_ERROR 4750109998Smarkm ... 4751109998Smarkm ERR_R_MISSING_ASN1_EOS 4752109998Smarkm were 4 .. 9, conflicting with 4753109998Smarkm ERR_LIB_RSA (= ERR_R_RSA_LIB) 4754109998Smarkm ... 4755109998Smarkm ERR_LIB_PEM (= ERR_R_PEM_LIB). 4756109998Smarkm They are now 58 .. 63 (i.e., just below ERR_R_FATAL). 4757109998Smarkm 4758109998Smarkm Add new error code 'ERR_R_INTERNAL_ERROR'. 4759109998Smarkm [Bodo Moeller] 4760109998Smarkm 4761109998Smarkm *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock 4762109998Smarkm suffices. 4763109998Smarkm [Bodo Moeller] 4764109998Smarkm 4765109998Smarkm *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This 4766109998Smarkm sets the subject name for a new request or supersedes the 4767109998Smarkm subject name in a given request. Formats that can be parsed are 4768109998Smarkm 'CN=Some Name, OU=myOU, C=IT' 4769109998Smarkm and 4770109998Smarkm 'CN=Some Name/OU=myOU/C=IT'. 4771109998Smarkm 4772109998Smarkm Add options '-batch' and '-verbose' to 'openssl req'. 4773109998Smarkm [Massimiliano Pala <madwolf@hackmasters.net>] 4774109998Smarkm 4775109998Smarkm *) Introduce the possibility to access global variables through 4776109998Smarkm functions on platform were that's the best way to handle exporting 4777109998Smarkm global variables in shared libraries. To enable this functionality, 4778109998Smarkm one must configure with "EXPORT_VAR_AS_FN" or defined the C macro 4779109998Smarkm "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter 4780109998Smarkm is normally done by Configure or something similar). 4781109998Smarkm 4782109998Smarkm To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL 4783109998Smarkm in the source file (foo.c) like this: 4784109998Smarkm 4785109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; 4786109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(double,bar); 4787109998Smarkm 4788109998Smarkm To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL 4789109998Smarkm and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: 4790109998Smarkm 4791109998Smarkm OPENSSL_DECLARE_GLOBAL(int,foo); 4792109998Smarkm #define foo OPENSSL_GLOBAL_REF(foo) 4793109998Smarkm OPENSSL_DECLARE_GLOBAL(double,bar); 4794109998Smarkm #define bar OPENSSL_GLOBAL_REF(bar) 4795109998Smarkm 4796109998Smarkm The #defines are very important, and therefore so is including the 4797109998Smarkm header file everywhere where the defined globals are used. 4798109998Smarkm 4799109998Smarkm The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition 4800109998Smarkm of ASN.1 items, but that structure is a bit different. 4801109998Smarkm 4802109998Smarkm The largest change is in util/mkdef.pl which has been enhanced with 4803109998Smarkm better and easier to understand logic to choose which symbols should 4804109998Smarkm go into the Windows .def files as well as a number of fixes and code 4805109998Smarkm cleanup (among others, algorithm keywords are now sorted 4806109998Smarkm lexicographically to avoid constant rewrites). 4807109998Smarkm [Richard Levitte] 4808109998Smarkm 4809109998Smarkm *) In BN_div() keep a copy of the sign of 'num' before writing the 4810109998Smarkm result to 'rm' because if rm==num the value will be overwritten 4811109998Smarkm and produce the wrong result if 'num' is negative: this caused 4812109998Smarkm problems with BN_mod() and BN_nnmod(). 4813109998Smarkm [Steve Henson] 4814109998Smarkm 4815109998Smarkm *) Function OCSP_request_verify(). This checks the signature on an 4816109998Smarkm OCSP request and verifies the signer certificate. The signer 4817109998Smarkm certificate is just checked for a generic purpose and OCSP request 4818109998Smarkm trust settings. 4819109998Smarkm [Steve Henson] 4820109998Smarkm 4821109998Smarkm *) Add OCSP_check_validity() function to check the validity of OCSP 4822109998Smarkm responses. OCSP responses are prepared in real time and may only 4823109998Smarkm be a few seconds old. Simply checking that the current time lies 4824109998Smarkm between thisUpdate and nextUpdate max reject otherwise valid responses 4825109998Smarkm caused by either OCSP responder or client clock inaccuracy. Instead 4826109998Smarkm we allow thisUpdate and nextUpdate to fall within a certain period of 4827109998Smarkm the current time. The age of the response can also optionally be 4828109998Smarkm checked. Two new options -validity_period and -status_age added to 4829109998Smarkm ocsp utility. 4830109998Smarkm [Steve Henson] 4831109998Smarkm 4832109998Smarkm *) If signature or public key algorithm is unrecognized print out its 4833109998Smarkm OID rather that just UNKNOWN. 4834109998Smarkm [Steve Henson] 4835109998Smarkm 4836109998Smarkm *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and 4837109998Smarkm OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate 4838109998Smarkm ID to be generated from the issuer certificate alone which can then be 4839109998Smarkm passed to OCSP_id_issuer_cmp(). 4840109998Smarkm [Steve Henson] 4841109998Smarkm 4842109998Smarkm *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new 4843109998Smarkm ASN1 modules to export functions returning ASN1_ITEM pointers 4844109998Smarkm instead of the ASN1_ITEM structures themselves. This adds several 4845109998Smarkm new macros which allow the underlying ASN1 function/structure to 4846109998Smarkm be accessed transparently. As a result code should not use ASN1_ITEM 4847109998Smarkm references directly (such as &X509_it) but instead use the relevant 4848109998Smarkm macros (such as ASN1_ITEM_rptr(X509)). This option is to allow 4849109998Smarkm use of the new ASN1 code on platforms where exporting structures 4850109998Smarkm is problematical (for example in shared libraries) but exporting 4851109998Smarkm functions returning pointers to structures is not. 4852109998Smarkm [Steve Henson] 4853109998Smarkm 4854109998Smarkm *) Add support for overriding the generation of SSL/TLS session IDs. 4855109998Smarkm These callbacks can be registered either in an SSL_CTX or per SSL. 4856109998Smarkm The purpose of this is to allow applications to control, if they wish, 4857109998Smarkm the arbitrary values chosen for use as session IDs, particularly as it 4858109998Smarkm can be useful for session caching in multiple-server environments. A 4859109998Smarkm command-line switch for testing this (and any client code that wishes 4860109998Smarkm to use such a feature) has been added to "s_server". 4861109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 4862109998Smarkm 4863109998Smarkm *) Modify mkdef.pl to recognise and parse preprocessor conditionals 4864109998Smarkm of the form '#if defined(...) || defined(...) || ...' and 4865109998Smarkm '#if !defined(...) && !defined(...) && ...'. This also avoids 4866109998Smarkm the growing number of special cases it was previously handling. 4867109998Smarkm [Richard Levitte] 4868109998Smarkm 4869109998Smarkm *) Make all configuration macros available for application by making 4870109998Smarkm sure they are available in opensslconf.h, by giving them names starting 4871109998Smarkm with "OPENSSL_" to avoid conflicts with other packages and by making 4872109998Smarkm sure e_os2.h will cover all platform-specific cases together with 4873109998Smarkm opensslconf.h. 4874109998Smarkm Additionally, it is now possible to define configuration/platform- 4875109998Smarkm specific names (called "system identities"). In the C code, these 4876109998Smarkm are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another 4877109998Smarkm macro with the name beginning with "OPENSSL_SYS_", which is determined 4878109998Smarkm from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on 4879109998Smarkm what is available. 4880109998Smarkm [Richard Levitte] 4881109998Smarkm 4882109998Smarkm *) New option -set_serial to 'req' and 'x509' this allows the serial 4883109998Smarkm number to use to be specified on the command line. Previously self 4884109998Smarkm signed certificates were hard coded with serial number 0 and the 4885109998Smarkm CA options of 'x509' had to use a serial number in a file which was 4886109998Smarkm auto incremented. 4887109998Smarkm [Steve Henson] 4888109998Smarkm 4889109998Smarkm *) New options to 'ca' utility to support V2 CRL entry extensions. 4890109998Smarkm Currently CRL reason, invalidity date and hold instruction are 4891109998Smarkm supported. Add new CRL extensions to V3 code and some new objects. 4892109998Smarkm [Steve Henson] 4893109998Smarkm 4894109998Smarkm *) New function EVP_CIPHER_CTX_set_padding() this is used to 4895109998Smarkm disable standard block padding (aka PKCS#5 padding) in the EVP 4896109998Smarkm API, which was previously mandatory. This means that the data is 4897109998Smarkm not padded in any way and so the total length much be a multiple 4898109998Smarkm of the block size, otherwise an error occurs. 4899109998Smarkm [Steve Henson] 4900109998Smarkm 4901109998Smarkm *) Initial (incomplete) OCSP SSL support. 4902109998Smarkm [Steve Henson] 4903109998Smarkm 4904109998Smarkm *) New function OCSP_parse_url(). This splits up a URL into its host, 4905109998Smarkm port and path components: primarily to parse OCSP URLs. New -url 4906109998Smarkm option to ocsp utility. 4907109998Smarkm [Steve Henson] 4908109998Smarkm 4909109998Smarkm *) New nonce behavior. The return value of OCSP_check_nonce() now 4910109998Smarkm reflects the various checks performed. Applications can decide 4911109998Smarkm whether to tolerate certain situations such as an absent nonce 4912109998Smarkm in a response when one was present in a request: the ocsp application 4913109998Smarkm just prints out a warning. New function OCSP_add1_basic_nonce() 4914109998Smarkm this is to allow responders to include a nonce in a response even if 4915109998Smarkm the request is nonce-less. 4916109998Smarkm [Steve Henson] 4917109998Smarkm 4918109998Smarkm *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are 4919109998Smarkm skipped when using openssl x509 multiple times on a single input file, 4920109998Smarkm e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs". 4921109998Smarkm [Bodo Moeller] 4922109998Smarkm 4923109998Smarkm *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() 4924109998Smarkm set string type: to handle setting ASN1_TIME structures. Fix ca 4925109998Smarkm utility to correctly initialize revocation date of CRLs. 4926109998Smarkm [Steve Henson] 4927109998Smarkm 4928109998Smarkm *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override 4929109998Smarkm the clients preferred ciphersuites and rather use its own preferences. 4930109998Smarkm Should help to work around M$ SGC (Server Gated Cryptography) bug in 4931109998Smarkm Internet Explorer by ensuring unchanged hash method during stepup. 4932109998Smarkm (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) 4933109998Smarkm [Lutz Jaenicke] 4934109998Smarkm 4935109998Smarkm *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael 4936109998Smarkm to aes and add a new 'exist' option to print out symbols that don't 4937109998Smarkm appear to exist. 4938109998Smarkm [Steve Henson] 4939109998Smarkm 4940109998Smarkm *) Additional options to ocsp utility to allow flags to be set and 4941109998Smarkm additional certificates supplied. 4942109998Smarkm [Steve Henson] 4943109998Smarkm 4944109998Smarkm *) Add the option -VAfile to 'openssl ocsp', so the user can give the 4945109998Smarkm OCSP client a number of certificate to only verify the response 4946109998Smarkm signature against. 4947109998Smarkm [Richard Levitte] 4948109998Smarkm 4949109998Smarkm *) Update Rijndael code to version 3.0 and change EVP AES ciphers to 4950109998Smarkm handle the new API. Currently only ECB, CBC modes supported. Add new 4951109998Smarkm AES OIDs. 4952109998Smarkm 4953109998Smarkm Add TLS AES ciphersuites as described in RFC3268, "Advanced 4954109998Smarkm Encryption Standard (AES) Ciphersuites for Transport Layer 4955109998Smarkm Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were 4956109998Smarkm not enabled by default and were not part of the "ALL" ciphersuite 4957109998Smarkm alias because they were not yet official; they could be 4958109998Smarkm explicitly requested by specifying the "AESdraft" ciphersuite 4959109998Smarkm group alias. In the final release of OpenSSL 0.9.7, the group 4960109998Smarkm alias is called "AES" and is part of "ALL".) 4961109998Smarkm [Ben Laurie, Steve Henson, Bodo Moeller] 4962109998Smarkm 4963109998Smarkm *) New function OCSP_copy_nonce() to copy nonce value (if present) from 4964109998Smarkm request to response. 4965109998Smarkm [Steve Henson] 4966109998Smarkm 4967109998Smarkm *) Functions for OCSP responders. OCSP_request_onereq_count(), 4968109998Smarkm OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() 4969109998Smarkm extract information from a certificate request. OCSP_response_create() 4970109998Smarkm creates a response and optionally adds a basic response structure. 4971109998Smarkm OCSP_basic_add1_status() adds a complete single response to a basic 4972109998Smarkm response and returns the OCSP_SINGLERESP structure just added (to allow 4973109998Smarkm extensions to be included for example). OCSP_basic_add1_cert() adds a 4974109998Smarkm certificate to a basic response and OCSP_basic_sign() signs a basic 4975109998Smarkm response with various flags. New helper functions ASN1_TIME_check() 4976109998Smarkm (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() 4977109998Smarkm (converts ASN1_TIME to GeneralizedTime). 4978109998Smarkm [Steve Henson] 4979109998Smarkm 4980109998Smarkm *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() 4981109998Smarkm in a single operation. X509_get0_pubkey_bitstr() extracts the public_key 4982109998Smarkm structure from a certificate. X509_pubkey_digest() digests the public_key 4983109998Smarkm contents: this is used in various key identifiers. 4984109998Smarkm [Steve Henson] 4985109998Smarkm 4986109998Smarkm *) Make sk_sort() tolerate a NULL argument. 4987109998Smarkm [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>] 4988109998Smarkm 4989109998Smarkm *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates 4990109998Smarkm passed by the function are trusted implicitly. If any of them signed the 4991109998Smarkm response then it is assumed to be valid and is not verified. 4992109998Smarkm [Steve Henson] 4993109998Smarkm 4994109998Smarkm *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT 4995109998Smarkm to data. This was previously part of the PKCS7 ASN1 code. This 4996109998Smarkm was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. 4997109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 4998109998Smarkm <support@securenetterm.com>] 4999109998Smarkm 5000109998Smarkm *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 5001109998Smarkm routines: without these tracing memory leaks is very painful. 5002109998Smarkm Fix leaks in PKCS12 and PKCS7 routines. 5003109998Smarkm [Steve Henson] 5004109998Smarkm 5005109998Smarkm *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). 5006109998Smarkm Previously it initialised the 'type' argument to V_ASN1_UTCTIME which 5007109998Smarkm effectively meant GeneralizedTime would never be used. Now it 5008109998Smarkm is initialised to -1 but X509_time_adj() now has to check the value 5009109998Smarkm and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or 5010109998Smarkm V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. 5011109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 5012109998Smarkm <support@securenetterm.com>] 5013109998Smarkm 5014109998Smarkm *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously 5015109998Smarkm result in a zero length in the ASN1_INTEGER structure which was 5016109998Smarkm not consistent with the structure when d2i_ASN1_INTEGER() was used 5017109998Smarkm and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() 5018109998Smarkm to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() 5019109998Smarkm where it did not print out a minus for negative ASN1_INTEGER. 5020109998Smarkm [Steve Henson] 5021109998Smarkm 5022109998Smarkm *) Add summary printout to ocsp utility. The various functions which 5023109998Smarkm convert status values to strings have been renamed to: 5024109998Smarkm OCSP_response_status_str(), OCSP_cert_status_str() and 5025109998Smarkm OCSP_crl_reason_str() and are no longer static. New options 5026109998Smarkm to verify nonce values and to disable verification. OCSP response 5027109998Smarkm printout format cleaned up. 5028109998Smarkm [Steve Henson] 5029109998Smarkm 5030109998Smarkm *) Add additional OCSP certificate checks. These are those specified 5031109998Smarkm in RFC2560. This consists of two separate checks: the CA of the 5032109998Smarkm certificate being checked must either be the OCSP signer certificate 5033109998Smarkm or the issuer of the OCSP signer certificate. In the latter case the 5034109998Smarkm OCSP signer certificate must contain the OCSP signing extended key 5035109998Smarkm usage. This check is performed by attempting to match the OCSP 5036109998Smarkm signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash 5037109998Smarkm in the OCSP_CERTID structures of the response. 5038109998Smarkm [Steve Henson] 5039109998Smarkm 5040109998Smarkm *) Initial OCSP certificate verification added to OCSP_basic_verify() 5041109998Smarkm and related routines. This uses the standard OpenSSL certificate 5042109998Smarkm verify routines to perform initial checks (just CA validity) and 5043109998Smarkm to obtain the certificate chain. Then additional checks will be 5044109998Smarkm performed on the chain. Currently the root CA is checked to see 5045109998Smarkm if it is explicitly trusted for OCSP signing. This is used to set 5046109998Smarkm a root CA as a global signing root: that is any certificate that 5047109998Smarkm chains to that CA is an acceptable OCSP signing certificate. 5048109998Smarkm [Steve Henson] 5049109998Smarkm 5050109998Smarkm *) New '-extfile ...' option to 'openssl ca' for reading X.509v3 5051109998Smarkm extensions from a separate configuration file. 5052109998Smarkm As when reading extensions from the main configuration file, 5053109998Smarkm the '-extensions ...' option may be used for specifying the 5054109998Smarkm section to use. 5055109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 5056109998Smarkm 5057109998Smarkm *) New OCSP utility. Allows OCSP requests to be generated or 5058109998Smarkm read. The request can be sent to a responder and the output 5059109998Smarkm parsed, outputed or printed in text form. Not complete yet: 5060109998Smarkm still needs to check the OCSP response validity. 5061109998Smarkm [Steve Henson] 5062109998Smarkm 5063109998Smarkm *) New subcommands for 'openssl ca': 5064109998Smarkm 'openssl ca -status <serial>' prints the status of the cert with 5065109998Smarkm the given serial number (according to the index file). 5066109998Smarkm 'openssl ca -updatedb' updates the expiry status of certificates 5067109998Smarkm in the index file. 5068109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 5069109998Smarkm 5070109998Smarkm *) New '-newreq-nodes' command option to CA.pl. This is like 5071109998Smarkm '-newreq', but calls 'openssl req' with the '-nodes' option 5072109998Smarkm so that the resulting key is not encrypted. 5073109998Smarkm [Damien Miller <djm@mindrot.org>] 5074109998Smarkm 5075109998Smarkm *) New configuration for the GNU Hurd. 5076109998Smarkm [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte] 5077109998Smarkm 5078109998Smarkm *) Initial code to implement OCSP basic response verify. This 5079109998Smarkm is currently incomplete. Currently just finds the signer's 5080109998Smarkm certificate and verifies the signature on the response. 5081109998Smarkm [Steve Henson] 5082109998Smarkm 5083109998Smarkm *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in 5084109998Smarkm value of OPENSSLDIR. This is available via the new '-d' option 5085109998Smarkm to 'openssl version', and is also included in 'openssl version -a'. 5086109998Smarkm [Bodo Moeller] 5087109998Smarkm 5088109998Smarkm *) Allowing defining memory allocation callbacks that will be given 5089109998Smarkm file name and line number information in additional arguments 5090109998Smarkm (a const char* and an int). The basic functionality remains, as 5091109998Smarkm well as the original possibility to just replace malloc(), 5092109998Smarkm realloc() and free() by functions that do not know about these 5093109998Smarkm additional arguments. To register and find out the current 5094109998Smarkm settings for extended allocation functions, the following 5095109998Smarkm functions are provided: 5096109998Smarkm 5097109998Smarkm CRYPTO_set_mem_ex_functions 5098109998Smarkm CRYPTO_set_locked_mem_ex_functions 5099109998Smarkm CRYPTO_get_mem_ex_functions 5100109998Smarkm CRYPTO_get_locked_mem_ex_functions 5101109998Smarkm 5102109998Smarkm These work the same way as CRYPTO_set_mem_functions and friends. 5103109998Smarkm CRYPTO_get_[locked_]mem_functions now writes 0 where such an 5104109998Smarkm extended allocation function is enabled. 5105109998Smarkm Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where 5106109998Smarkm a conventional allocation function is enabled. 5107109998Smarkm [Richard Levitte, Bodo Moeller] 5108109998Smarkm 5109109998Smarkm *) Finish off removing the remaining LHASH function pointer casts. 5110109998Smarkm There should no longer be any prototype-casting required when using 5111109998Smarkm the LHASH abstraction, and any casts that remain are "bugs". See 5112109998Smarkm the callback types and macros at the head of lhash.h for details 5113109998Smarkm (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). 5114109998Smarkm [Geoff Thorpe] 5115109998Smarkm 5116109998Smarkm *) Add automatic query of EGD sockets in RAND_poll() for the unix variant. 5117109998Smarkm If /dev/[u]random devices are not available or do not return enough 5118109998Smarkm entropy, EGD style sockets (served by EGD or PRNGD) will automatically 5119109998Smarkm be queried. 5120109998Smarkm The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and 5121109998Smarkm /etc/entropy will be queried once each in this sequence, quering stops 5122109998Smarkm when enough entropy was collected without querying more sockets. 5123109998Smarkm [Lutz Jaenicke] 5124109998Smarkm 5125109998Smarkm *) Change the Unix RAND_poll() variant to be able to poll several 5126109998Smarkm random devices, as specified by DEVRANDOM, until a sufficient amount 5127109998Smarkm of data has been collected. We spend at most 10 ms on each file 5128109998Smarkm (select timeout) and read in non-blocking mode. DEVRANDOM now 5129109998Smarkm defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" 5130109998Smarkm (previously it was just the string "/dev/urandom"), so on typical 5131109998Smarkm platforms the 10 ms delay will never occur. 5132109998Smarkm Also separate out the Unix variant to its own file, rand_unix.c. 5133109998Smarkm For VMS, there's a currently-empty rand_vms.c. 5134109998Smarkm [Richard Levitte] 5135109998Smarkm 5136109998Smarkm *) Move OCSP client related routines to ocsp_cl.c. These 5137109998Smarkm provide utility functions which an application needing 5138109998Smarkm to issue a request to an OCSP responder and analyse the 5139109998Smarkm response will typically need: as opposed to those which an 5140109998Smarkm OCSP responder itself would need which will be added later. 5141109998Smarkm 5142109998Smarkm OCSP_request_sign() signs an OCSP request with an API similar 5143109998Smarkm to PKCS7_sign(). OCSP_response_status() returns status of OCSP 5144109998Smarkm response. OCSP_response_get1_basic() extracts basic response 5145109998Smarkm from response. OCSP_resp_find_status(): finds and extracts status 5146109998Smarkm information from an OCSP_CERTID structure (which will be created 5147109998Smarkm when the request structure is built). These are built from lower 5148109998Smarkm level functions which work on OCSP_SINGLERESP structures but 5149109998Smarkm wont normally be used unless the application wishes to examine 5150109998Smarkm extensions in the OCSP response for example. 5151109998Smarkm 5152109998Smarkm Replace nonce routines with a pair of functions. 5153109998Smarkm OCSP_request_add1_nonce() adds a nonce value and optionally 5154109998Smarkm generates a random value. OCSP_check_nonce() checks the 5155109998Smarkm validity of the nonce in an OCSP response. 5156109998Smarkm [Steve Henson] 5157109998Smarkm 5158109998Smarkm *) Change function OCSP_request_add() to OCSP_request_add0_id(). 5159109998Smarkm This doesn't copy the supplied OCSP_CERTID and avoids the 5160109998Smarkm need to free up the newly created id. Change return type 5161109998Smarkm to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. 5162109998Smarkm This can then be used to add extensions to the request. 5163109998Smarkm Deleted OCSP_request_new(), since most of its functionality 5164109998Smarkm is now in OCSP_REQUEST_new() (and the case insensitive name 5165109998Smarkm clash) apart from the ability to set the request name which 5166109998Smarkm will be added elsewhere. 5167109998Smarkm [Steve Henson] 5168109998Smarkm 5169109998Smarkm *) Update OCSP API. Remove obsolete extensions argument from 5170109998Smarkm various functions. Extensions are now handled using the new 5171109998Smarkm OCSP extension code. New simple OCSP HTTP function which 5172109998Smarkm can be used to send requests and parse the response. 5173109998Smarkm [Steve Henson] 5174109998Smarkm 5175109998Smarkm *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new 5176109998Smarkm ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN 5177109998Smarkm uses the special reorder version of SET OF to sort the attributes 5178109998Smarkm and reorder them to match the encoded order. This resolves a long 5179109998Smarkm standing problem: a verify on a PKCS7 structure just after signing 5180109998Smarkm it used to fail because the attribute order did not match the 5181109998Smarkm encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: 5182109998Smarkm it uses the received order. This is necessary to tolerate some broken 5183109998Smarkm software that does not order SET OF. This is handled by encoding 5184109998Smarkm as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) 5185109998Smarkm to produce the required SET OF. 5186109998Smarkm [Steve Henson] 5187109998Smarkm 5188109998Smarkm *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and 5189109998Smarkm OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header 5190109998Smarkm files to get correct declarations of the ASN.1 item variables. 5191109998Smarkm [Richard Levitte] 5192109998Smarkm 5193109998Smarkm *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many 5194109998Smarkm PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: 5195109998Smarkm asn1_check_tlen() would sometimes attempt to use 'ctx' when it was 5196109998Smarkm NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). 5197109998Smarkm New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant 5198109998Smarkm ASN1_ITEM and no wrapper functions. 5199109998Smarkm [Steve Henson] 5200109998Smarkm 5201109998Smarkm *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These 5202109998Smarkm replace the old function pointer based I/O routines. Change most of 5203109998Smarkm the *_d2i_bio() and *_d2i_fp() functions to use these. 5204109998Smarkm [Steve Henson] 5205109998Smarkm 5206109998Smarkm *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor 5207109998Smarkm lines, recognice more "algorithms" that can be deselected, and make 5208109998Smarkm it complain about algorithm deselection that isn't recognised. 5209109998Smarkm [Richard Levitte] 5210109998Smarkm 5211109998Smarkm *) New ASN1 functions to handle dup, sign, verify, digest, pack and 5212109998Smarkm unpack operations in terms of ASN1_ITEM. Modify existing wrappers 5213109998Smarkm to use new functions. Add NO_ASN1_OLD which can be set to remove 5214109998Smarkm some old style ASN1 functions: this can be used to determine if old 5215109998Smarkm code will still work when these eventually go away. 5216109998Smarkm [Steve Henson] 5217109998Smarkm 5218109998Smarkm *) New extension functions for OCSP structures, these follow the 5219109998Smarkm same conventions as certificates and CRLs. 5220109998Smarkm [Steve Henson] 5221109998Smarkm 5222109998Smarkm *) New function X509V3_add1_i2d(). This automatically encodes and 5223109998Smarkm adds an extension. Its behaviour can be customised with various 5224109998Smarkm flags to append, replace or delete. Various wrappers added for 5225109998Smarkm certifcates and CRLs. 5226109998Smarkm [Steve Henson] 5227109998Smarkm 5228109998Smarkm *) Fix to avoid calling the underlying ASN1 print routine when 5229109998Smarkm an extension cannot be parsed. Correct a typo in the 5230109998Smarkm OCSP_SERVICELOC extension. Tidy up print OCSP format. 5231109998Smarkm [Steve Henson] 5232109998Smarkm 5233109998Smarkm *) Make mkdef.pl parse some of the ASN1 macros and add apropriate 5234109998Smarkm entries for variables. 5235109998Smarkm [Steve Henson] 5236109998Smarkm 5237109998Smarkm *) Add functionality to apps/openssl.c for detecting locking 5238109998Smarkm problems: As the program is single-threaded, all we have 5239109998Smarkm to do is register a locking callback using an array for 5240109998Smarkm storing which locks are currently held by the program. 5241109998Smarkm [Bodo Moeller] 5242109998Smarkm 5243109998Smarkm *) Use a lock around the call to CRYPTO_get_ex_new_index() in 5244109998Smarkm SSL_get_ex_data_X509_STORE_idx(), which is used in 5245109998Smarkm ssl_verify_cert_chain() and thus can be called at any time 5246109998Smarkm during TLS/SSL handshakes so that thread-safety is essential. 5247109998Smarkm Unfortunately, the ex_data design is not at all suited 5248109998Smarkm for multi-threaded use, so it probably should be abolished. 5249109998Smarkm [Bodo Moeller] 5250109998Smarkm 5251109998Smarkm *) Added Broadcom "ubsec" ENGINE to OpenSSL. 5252109998Smarkm [Broadcom, tweaked and integrated by Geoff Thorpe] 5253109998Smarkm 5254109998Smarkm *) Move common extension printing code to new function 5255109998Smarkm X509V3_print_extensions(). Reorganise OCSP print routines and 5256109998Smarkm implement some needed OCSP ASN1 functions. Add OCSP extensions. 5257109998Smarkm [Steve Henson] 5258109998Smarkm 5259109998Smarkm *) New function X509_signature_print() to remove duplication in some 5260109998Smarkm print routines. 5261109998Smarkm [Steve Henson] 5262109998Smarkm 5263109998Smarkm *) Add a special meaning when SET OF and SEQUENCE OF flags are both 5264109998Smarkm set (this was treated exactly the same as SET OF previously). This 5265109998Smarkm is used to reorder the STACK representing the structure to match the 5266109998Smarkm encoding. This will be used to get round a problem where a PKCS7 5267109998Smarkm structure which was signed could not be verified because the STACK 5268109998Smarkm order did not reflect the encoded order. 5269109998Smarkm [Steve Henson] 5270109998Smarkm 5271109998Smarkm *) Reimplement the OCSP ASN1 module using the new code. 5272109998Smarkm [Steve Henson] 5273109998Smarkm 5274109998Smarkm *) Update the X509V3 code to permit the use of an ASN1_ITEM structure 5275109998Smarkm for its ASN1 operations. The old style function pointers still exist 5276109998Smarkm for now but they will eventually go away. 5277109998Smarkm [Steve Henson] 5278109998Smarkm 5279109998Smarkm *) Merge in replacement ASN1 code from the ASN1 branch. This almost 5280109998Smarkm completely replaces the old ASN1 functionality with a table driven 5281109998Smarkm encoder and decoder which interprets an ASN1_ITEM structure describing 5282109998Smarkm the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is 5283109998Smarkm largely maintained. Almost all of the old asn1_mac.h macro based ASN1 5284109998Smarkm has also been converted to the new form. 5285109998Smarkm [Steve Henson] 5286109998Smarkm 5287109998Smarkm *) Change BN_mod_exp_recp so that negative moduli are tolerated 5288109998Smarkm (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set 5289109998Smarkm so that BN_mod_exp_mont and BN_mod_exp_mont_word work 5290109998Smarkm for negative moduli. 5291109998Smarkm [Bodo Moeller] 5292109998Smarkm 5293109998Smarkm *) Fix BN_uadd and BN_usub: Always return non-negative results instead 5294109998Smarkm of not touching the result's sign bit. 5295109998Smarkm [Bodo Moeller] 5296109998Smarkm 5297109998Smarkm *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be 5298109998Smarkm set. 5299109998Smarkm [Bodo Moeller] 5300109998Smarkm 5301109998Smarkm *) Changed the LHASH code to use prototypes for callbacks, and created 5302109998Smarkm macros to declare and implement thin (optionally static) functions 5303109998Smarkm that provide type-safety and avoid function pointer casting for the 5304109998Smarkm type-specific callbacks. 5305109998Smarkm [Geoff Thorpe] 5306109998Smarkm 5307109998Smarkm *) Added Kerberos Cipher Suites to be used with TLS, as written in 5308109998Smarkm RFC 2712. 5309109998Smarkm [Veers Staats <staatsvr@asc.hpc.mil>, 5310109998Smarkm Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte] 5311109998Smarkm 5312109998Smarkm *) Reformat the FAQ so the different questions and answers can be divided 5313109998Smarkm in sections depending on the subject. 5314109998Smarkm [Richard Levitte] 5315109998Smarkm 5316109998Smarkm *) Have the zlib compression code load ZLIB.DLL dynamically under 5317109998Smarkm Windows. 5318109998Smarkm [Richard Levitte] 5319109998Smarkm 5320109998Smarkm *) New function BN_mod_sqrt for computing square roots modulo a prime 5321109998Smarkm (using the probabilistic Tonelli-Shanks algorithm unless 5322109998Smarkm p == 3 (mod 4) or p == 5 (mod 8), which are cases that can 5323109998Smarkm be handled deterministically). 5324109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 5325109998Smarkm 5326109998Smarkm *) Make BN_mod_inverse faster by explicitly handling small quotients 5327109998Smarkm in the Euclid loop. (Speed gain about 20% for small moduli [256 or 5328109998Smarkm 512 bits], about 30% for larger ones [1024 or 2048 bits].) 5329109998Smarkm [Bodo Moeller] 5330109998Smarkm 5331109998Smarkm *) New function BN_kronecker. 5332109998Smarkm [Bodo Moeller] 5333109998Smarkm 5334109998Smarkm *) Fix BN_gcd so that it works on negative inputs; the result is 5335109998Smarkm positive unless both parameters are zero. 5336109998Smarkm Previously something reasonably close to an infinite loop was 5337109998Smarkm possible because numbers could be growing instead of shrinking 5338109998Smarkm in the implementation of Euclid's algorithm. 5339109998Smarkm [Bodo Moeller] 5340109998Smarkm 5341109998Smarkm *) Fix BN_is_word() and BN_is_one() macros to take into account the 5342109998Smarkm sign of the number in question. 5343109998Smarkm 5344109998Smarkm Fix BN_is_word(a,w) to work correctly for w == 0. 5345109998Smarkm 5346109998Smarkm The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) 5347109998Smarkm because its test if the absolute value of 'a' equals 'w'. 5348109998Smarkm Note that BN_abs_is_word does *not* handle w == 0 reliably; 5349109998Smarkm it exists mostly for use in the implementations of BN_is_zero(), 5350109998Smarkm BN_is_one(), and BN_is_word(). 5351109998Smarkm [Bodo Moeller] 5352109998Smarkm 5353109998Smarkm *) New function BN_swap. 5354109998Smarkm [Bodo Moeller] 5355109998Smarkm 5356109998Smarkm *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that 5357109998Smarkm the exponentiation functions are more likely to produce reasonable 5358109998Smarkm results on negative inputs. 5359109998Smarkm [Bodo Moeller] 5360109998Smarkm 5361109998Smarkm *) Change BN_mod_mul so that the result is always non-negative. 5362109998Smarkm Previously, it could be negative if one of the factors was negative; 5363109998Smarkm I don't think anyone really wanted that behaviour. 5364109998Smarkm [Bodo Moeller] 5365109998Smarkm 5366109998Smarkm *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c 5367109998Smarkm (except for exponentiation, which stays in crypto/bn/bn_exp.c, 5368109998Smarkm and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c) 5369109998Smarkm and add new functions: 5370109998Smarkm 5371109998Smarkm BN_nnmod 5372109998Smarkm BN_mod_sqr 5373109998Smarkm BN_mod_add 5374109998Smarkm BN_mod_add_quick 5375109998Smarkm BN_mod_sub 5376109998Smarkm BN_mod_sub_quick 5377109998Smarkm BN_mod_lshift1 5378109998Smarkm BN_mod_lshift1_quick 5379109998Smarkm BN_mod_lshift 5380109998Smarkm BN_mod_lshift_quick 5381109998Smarkm 5382109998Smarkm These functions always generate non-negative results. 5383109998Smarkm 5384109998Smarkm BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r 5385109998Smarkm such that |m| < r < 0, BN_nnmod will output rem + |m| instead). 5386109998Smarkm 5387109998Smarkm BN_mod_XXX_quick(r, a, [b,] m) generates the same result as 5388109998Smarkm BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b] 5389109998Smarkm be reduced modulo m. 5390109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 5391109998Smarkm 5392111147Snectar#if 0 5393111147Snectar The following entry accidentily appeared in the CHANGES file 5394111147Snectar distributed with OpenSSL 0.9.7. The modifications described in 5395111147Snectar it do *not* apply to OpenSSL 0.9.7. 5396111147Snectar 5397109998Smarkm *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 5398109998Smarkm was actually never needed) and in BN_mul(). The removal in BN_mul() 5399109998Smarkm required a small change in bn_mul_part_recursive() and the addition 5400109998Smarkm of the functions bn_cmp_part_words(), bn_sub_part_words() and 5401109998Smarkm bn_add_part_words(), which do the same thing as bn_cmp_words(), 5402109998Smarkm bn_sub_words() and bn_add_words() except they take arrays with 5403109998Smarkm differing sizes. 5404109998Smarkm [Richard Levitte] 5405111147Snectar#endif 5406109998Smarkm 5407109998Smarkm *) In 'openssl passwd', verify passwords read from the terminal 5408109998Smarkm unless the '-salt' option is used (which usually means that 5409109998Smarkm verification would just waste user's time since the resulting 5410109998Smarkm hash is going to be compared with some given password hash) 5411109998Smarkm or the new '-noverify' option is used. 5412109998Smarkm 5413109998Smarkm This is an incompatible change, but it does not affect 5414109998Smarkm non-interactive use of 'openssl passwd' (passwords on the command 5415109998Smarkm line, '-stdin' option, '-in ...' option) and thus should not 5416109998Smarkm cause any problems. 5417109998Smarkm [Bodo Moeller] 5418109998Smarkm 5419109998Smarkm *) Remove all references to RSAref, since there's no more need for it. 5420109998Smarkm [Richard Levitte] 5421109998Smarkm 5422109998Smarkm *) Make DSO load along a path given through an environment variable 5423109998Smarkm (SHLIB_PATH) with shl_load(). 5424109998Smarkm [Richard Levitte] 5425109998Smarkm 5426109998Smarkm *) Constify the ENGINE code as a result of BIGNUM constification. 5427109998Smarkm Also constify the RSA code and most things related to it. In a 5428109998Smarkm few places, most notable in the depth of the ASN.1 code, ugly 5429109998Smarkm casts back to non-const were required (to be solved at a later 5430109998Smarkm time) 5431109998Smarkm [Richard Levitte] 5432109998Smarkm 5433109998Smarkm *) Make it so the openssl application has all engines loaded by default. 5434109998Smarkm [Richard Levitte] 5435109998Smarkm 5436109998Smarkm *) Constify the BIGNUM routines a little more. 5437109998Smarkm [Richard Levitte] 5438109998Smarkm 5439109998Smarkm *) Add the following functions: 5440109998Smarkm 5441109998Smarkm ENGINE_load_cswift() 5442109998Smarkm ENGINE_load_chil() 5443109998Smarkm ENGINE_load_atalla() 5444109998Smarkm ENGINE_load_nuron() 5445109998Smarkm ENGINE_load_builtin_engines() 5446109998Smarkm 5447109998Smarkm That way, an application can itself choose if external engines that 5448109998Smarkm are built-in in OpenSSL shall ever be used or not. The benefit is 5449109998Smarkm that applications won't have to be linked with libdl or other dso 5450109998Smarkm libraries unless it's really needed. 5451109998Smarkm 5452109998Smarkm Changed 'openssl engine' to load all engines on demand. 5453109998Smarkm Changed the engine header files to avoid the duplication of some 5454109998Smarkm declarations (they differed!). 5455109998Smarkm [Richard Levitte] 5456109998Smarkm 5457109998Smarkm *) 'openssl engine' can now list capabilities. 5458109998Smarkm [Richard Levitte] 5459109998Smarkm 5460109998Smarkm *) Better error reporting in 'openssl engine'. 5461109998Smarkm [Richard Levitte] 5462109998Smarkm 5463109998Smarkm *) Never call load_dh_param(NULL) in s_server. 5464109998Smarkm [Bodo Moeller] 5465109998Smarkm 5466109998Smarkm *) Add engine application. It can currently list engines by name and 5467109998Smarkm identity, and test if they are actually available. 5468109998Smarkm [Richard Levitte] 5469109998Smarkm 5470109998Smarkm *) Improve RPM specification file by forcing symbolic linking and making 5471109998Smarkm sure the installed documentation is also owned by root.root. 5472109998Smarkm [Damien Miller <djm@mindrot.org>] 5473109998Smarkm 5474109998Smarkm *) Give the OpenSSL applications more possibilities to make use of 5475109998Smarkm keys (public as well as private) handled by engines. 5476109998Smarkm [Richard Levitte] 5477109998Smarkm 5478109998Smarkm *) Add OCSP code that comes from CertCo. 5479109998Smarkm [Richard Levitte] 5480109998Smarkm 5481109998Smarkm *) Add VMS support for the Rijndael code. 5482109998Smarkm [Richard Levitte] 5483109998Smarkm 5484109998Smarkm *) Added untested support for Nuron crypto accelerator. 5485109998Smarkm [Ben Laurie] 5486109998Smarkm 5487109998Smarkm *) Add support for external cryptographic devices. This code was 5488109998Smarkm previously distributed separately as the "engine" branch. 5489109998Smarkm [Geoff Thorpe, Richard Levitte] 5490109998Smarkm 5491109998Smarkm *) Rework the filename-translation in the DSO code. It is now possible to 5492109998Smarkm have far greater control over how a "name" is turned into a filename 5493109998Smarkm depending on the operating environment and any oddities about the 5494109998Smarkm different shared library filenames on each system. 5495109998Smarkm [Geoff Thorpe] 5496109998Smarkm 5497109998Smarkm *) Support threads on FreeBSD-elf in Configure. 5498109998Smarkm [Richard Levitte] 5499109998Smarkm 5500109998Smarkm *) Fix for SHA1 assembly problem with MASM: it produces 5501109998Smarkm warnings about corrupt line number information when assembling 5502109998Smarkm with debugging information. This is caused by the overlapping 5503109998Smarkm of two sections. 5504109998Smarkm [Bernd Matthes <mainbug@celocom.de>, Steve Henson] 5505109998Smarkm 5506109998Smarkm *) NCONF changes. 5507109998Smarkm NCONF_get_number() has no error checking at all. As a replacement, 5508109998Smarkm NCONF_get_number_e() is defined (_e for "error checking") and is 5509109998Smarkm promoted strongly. The old NCONF_get_number is kept around for 5510109998Smarkm binary backward compatibility. 5511109998Smarkm Make it possible for methods to load from something other than a BIO, 5512109998Smarkm by providing a function pointer that is given a name instead of a BIO. 5513109998Smarkm For example, this could be used to load configuration data from an 5514109998Smarkm LDAP server. 5515109998Smarkm [Richard Levitte] 5516109998Smarkm 5517109998Smarkm *) Fix for non blocking accept BIOs. Added new I/O special reason 5518109998Smarkm BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs 5519109998Smarkm with non blocking I/O was not possible because no retry code was 5520109998Smarkm implemented. Also added new SSL code SSL_WANT_ACCEPT to cover 5521109998Smarkm this case. 5522109998Smarkm [Steve Henson] 5523109998Smarkm 5524109998Smarkm *) Added the beginnings of Rijndael support. 5525109998Smarkm [Ben Laurie] 5526109998Smarkm 5527109998Smarkm *) Fix for bug in DirectoryString mask setting. Add support for 5528109998Smarkm X509_NAME_print_ex() in 'req' and X509_print_ex() function 5529109998Smarkm to allow certificate printing to more controllable, additional 5530109998Smarkm 'certopt' option to 'x509' to allow new printing options to be 5531109998Smarkm set. 5532109998Smarkm [Steve Henson] 5533109998Smarkm 5534109998Smarkm *) Clean old EAY MD5 hack from e_os.h. 5535109998Smarkm [Richard Levitte] 5536109998Smarkm 5537142425Snectar Changes between 0.9.6l and 0.9.6m [17 Mar 2004] 5538142425Snectar 5539142425Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 5540160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 5541142425Snectar [Joe Orton, Steve Henson] 5542142425Snectar 5543142425Snectar Changes between 0.9.6k and 0.9.6l [04 Nov 2003] 5544142425Snectar 5545142425Snectar *) Fix additional bug revealed by the NISCC test suite: 5546142425Snectar 5547142425Snectar Stop bug triggering large recursion when presented with 5548160814Ssimon certain ASN.1 tags (CVE-2003-0851) 5549142425Snectar [Steve Henson] 5550142425Snectar 5551120631Snectar Changes between 0.9.6j and 0.9.6k [30 Sep 2003] 5552120631Snectar 5553120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 5554120631Snectar 5555120631Snectar Stop out of bounds reads in the ASN1 code when presented with 5556160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 5557120631Snectar 5558120631Snectar If verify callback ignores invalid public key errors don't try to check 5559120631Snectar certificate signature with the NULL public key. 5560120631Snectar 5561120631Snectar [Steve Henson] 5562120631Snectar 5563120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 5564120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 5565120631Snectar specifications. 5566120631Snectar [Steve Henson] 5567120631Snectar 5568120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 5569120631Snectar extra data after the compression methods not only for TLS 1.0 5570120631Snectar but also for SSL 3.0 (as required by the specification). 5571120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 5572120631Snectar 5573120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 5574120631Snectar when it's 512 *bits* long, not 512 bytes. 5575120631Snectar [Richard Levitte] 5576120631Snectar 5577120631Snectar Changes between 0.9.6i and 0.9.6j [10 Apr 2003] 5578120631Snectar 5579120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 5580120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 5581120631Snectar a protocol version number mismatch like a decryption error 5582120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 5583120631Snectar [Bodo Moeller] 5584120631Snectar 5585120631Snectar *) Turn on RSA blinding by default in the default implementation 5586120631Snectar to avoid a timing attack. Applications that don't want it can call 5587120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 5588120631Snectar They would be ill-advised to do so in most cases. 5589120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 5590120631Snectar 5591120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 5592120631Snectar seeded (in this case, the secret RSA exponent is abused as 5593120631Snectar an unpredictable seed -- if it is not unpredictable, there 5594120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 5595120631Snectar by remembering the creator's thread ID in rsa->blinding and 5596120631Snectar having all other threads use local one-time blinding factors 5597120631Snectar (this requires more computation than sharing rsa->blinding, but 5598120631Snectar avoids excessive locking; and if an RSA object is not shared 5599120631Snectar between threads, blinding will still be very fast). 5600120631Snectar [Bodo Moeller] 5601120631Snectar 5602111147Snectar Changes between 0.9.6h and 0.9.6i [19 Feb 2003] 5603111147Snectar 5604111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 5605111147Snectar via timing by performing a MAC computation even if incorrrect 5606111147Snectar block cipher padding has been found. This is a countermeasure 5607111147Snectar against active attacks where the attacker has to distinguish 5608160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 5609111147Snectar 5610111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 5611111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 5612111147Snectar Martin Vuagnoux (EPFL, Ilion)] 5613111147Snectar 5614109998Smarkm Changes between 0.9.6g and 0.9.6h [5 Dec 2002] 5615109998Smarkm 5616109998Smarkm *) New function OPENSSL_cleanse(), which is used to cleanse a section of 5617109998Smarkm memory from it's contents. This is done with a counter that will 5618109998Smarkm place alternating values in each byte. This can be used to solve 5619109998Smarkm two issues: 1) the removal of calls to memset() by highly optimizing 5620109998Smarkm compilers, and 2) cleansing with other values than 0, since those can 5621109998Smarkm be read through on certain media, for example a swap space on disk. 5622109998Smarkm [Geoff Thorpe] 5623109998Smarkm 5624109998Smarkm *) Bugfix: client side session caching did not work with external caching, 5625109998Smarkm because the session->cipher setting was not restored when reloading 5626109998Smarkm from the external cache. This problem was masked, when 5627109998Smarkm SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. 5628109998Smarkm (Found by Steve Haslam <steve@araqnid.ddts.net>.) 5629109998Smarkm [Lutz Jaenicke] 5630109998Smarkm 5631109998Smarkm *) Fix client_certificate (ssl/s2_clnt.c): The permissible total 5632109998Smarkm length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. 5633109998Smarkm [Zeev Lieber <zeev-l@yahoo.com>] 5634109998Smarkm 5635109998Smarkm *) Undo an undocumented change introduced in 0.9.6e which caused 5636109998Smarkm repeated calls to OpenSSL_add_all_ciphers() and 5637109998Smarkm OpenSSL_add_all_digests() to be ignored, even after calling 5638109998Smarkm EVP_cleanup(). 5639109998Smarkm [Richard Levitte] 5640109998Smarkm 5641109998Smarkm *) Change the default configuration reader to deal with last line not 5642109998Smarkm being properly terminated. 5643109998Smarkm [Richard Levitte] 5644109998Smarkm 5645109998Smarkm *) Change X509_NAME_cmp() so it applies the special rules on handling 5646109998Smarkm DN values that are of type PrintableString, as well as RDNs of type 5647109998Smarkm emailAddress where the value has the type ia5String. 5648109998Smarkm [stefank@valicert.com via Richard Levitte] 5649109998Smarkm 5650109998Smarkm *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half 5651109998Smarkm the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently 5652109998Smarkm doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be 5653109998Smarkm the bitwise-OR of the two for use by the majority of applications 5654109998Smarkm wanting this behaviour, and update the docs. The documented 5655109998Smarkm behaviour and actual behaviour were inconsistent and had been 5656109998Smarkm changing anyway, so this is more a bug-fix than a behavioural 5657109998Smarkm change. 5658109998Smarkm [Geoff Thorpe, diagnosed by Nadav Har'El] 5659109998Smarkm 5660109998Smarkm *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c 5661109998Smarkm (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). 5662109998Smarkm [Bodo Moeller] 5663109998Smarkm 5664109998Smarkm *) Fix initialization code race conditions in 5665109998Smarkm SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), 5666109998Smarkm SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), 5667109998Smarkm SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), 5668109998Smarkm TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), 5669109998Smarkm ssl2_get_cipher_by_char(), 5670109998Smarkm ssl3_get_cipher_by_char(). 5671109998Smarkm [Patrick McCormick <patrick@tellme.com>, Bodo Moeller] 5672109998Smarkm 5673109998Smarkm *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after 5674109998Smarkm the cached sessions are flushed, as the remove_cb() might use ex_data 5675109998Smarkm contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> 5676109998Smarkm (see [openssl.org #212]). 5677109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 5678109998Smarkm 5679109998Smarkm *) Fix typo in OBJ_txt2obj which incorrectly passed the content 5680109998Smarkm length, instead of the encoding length to d2i_ASN1_OBJECT. 5681109998Smarkm [Steve Henson] 5682109998Smarkm 5683101618Snectar Changes between 0.9.6f and 0.9.6g [9 Aug 2002] 5684101618Snectar 5685101618Snectar *) [In 0.9.6g-engine release:] 5686101618Snectar Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall'). 5687101618Snectar [Lynn Gazis <lgazis@rainbow.com>] 5688101618Snectar 5689101613Snectar Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 5690101613Snectar 5691101613Snectar *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 5692101613Snectar and get fix the header length calculation. 5693101613Snectar [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 5694101613Snectar Alon Kantor <alonk@checkpoint.com> (and others), 5695101613Snectar Steve Henson] 5696101613Snectar 5697101613Snectar *) Use proper error handling instead of 'assertions' in buffer 5698101613Snectar overflow checks added in 0.9.6e. This prevents DoS (the 5699101613Snectar assertions could call abort()). 5700101613Snectar [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] 5701101613Snectar 5702100936Snectar Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 5703100936Snectar 5704109998Smarkm *) Add various sanity checks to asn1_get_length() to reject 5705109998Smarkm the ASN1 length bytes if they exceed sizeof(long), will appear 5706109998Smarkm negative or the content length exceeds the length of the 5707109998Smarkm supplied buffer. 5708109998Smarkm [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 5709109998Smarkm 5710100936Snectar *) Fix cipher selection routines: ciphers without encryption had no flags 5711100936Snectar for the cipher strength set and where therefore not handled correctly 5712100936Snectar by the selection routines (PR #130). 5713100936Snectar [Lutz Jaenicke] 5714100936Snectar 5715100936Snectar *) Fix EVP_dsa_sha macro. 5716100936Snectar [Nils Larsch] 5717100936Snectar 5718100936Snectar *) New option 5719100936Snectar SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 5720100936Snectar for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 5721100936Snectar that was added in OpenSSL 0.9.6d. 5722100936Snectar 5723100936Snectar As the countermeasure turned out to be incompatible with some 5724100936Snectar broken SSL implementations, the new option is part of SSL_OP_ALL. 5725100936Snectar SSL_OP_ALL is usually employed when compatibility with weird SSL 5726100936Snectar implementations is desired (e.g. '-bugs' option to 's_client' and 5727100936Snectar 's_server'), so the new option is automatically set in many 5728100936Snectar applications. 5729100936Snectar [Bodo Moeller] 5730100936Snectar 5731100936Snectar *) Changes in security patch: 5732100936Snectar 5733100936Snectar Changes marked "(CHATS)" were sponsored by the Defense Advanced 5734100936Snectar Research Projects Agency (DARPA) and Air Force Research Laboratory, 5735100936Snectar Air Force Materiel Command, USAF, under agreement number 5736100936Snectar F30602-01-2-0537. 5737100936Snectar 5738100936Snectar *) Add various sanity checks to asn1_get_length() to reject 5739100936Snectar the ASN1 length bytes if they exceed sizeof(long), will appear 5740100936Snectar negative or the content length exceeds the length of the 5741160814Ssimon supplied buffer. (CVE-2002-0659) 5742100936Snectar [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 5743100936Snectar 5744100936Snectar *) Assertions for various potential buffer overflows, not known to 5745100936Snectar happen in practice. 5746100936Snectar [Ben Laurie (CHATS)] 5747100936Snectar 5748100936Snectar *) Various temporary buffers to hold ASCII versions of integers were 5749160814Ssimon too small for 64 bit platforms. (CVE-2002-0655) 5750100936Snectar [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 5751100936Snectar 5752100936Snectar *) Remote buffer overflow in SSL3 protocol - an attacker could 5753160814Ssimon supply an oversized session ID to a client. (CVE-2002-0656) 5754100936Snectar [Ben Laurie (CHATS)] 5755100936Snectar 5756100936Snectar *) Remote buffer overflow in SSL2 protocol - an attacker could 5757160814Ssimon supply an oversized client master key. (CVE-2002-0656) 5758100936Snectar [Ben Laurie (CHATS)] 5759100936Snectar 5760100928Snectar Changes between 0.9.6c and 0.9.6d [9 May 2002] 5761100928Snectar 5762100928Snectar *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 5763100928Snectar encoded as NULL) with id-dsa-with-sha1. 5764100928Snectar [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 5765100928Snectar 5766100928Snectar *) Check various X509_...() return values in apps/req.c. 5767100928Snectar [Nils Larsch <nla@trustcenter.de>] 5768100928Snectar 5769100928Snectar *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 5770100928Snectar an end-of-file condition would erronously be flagged, when the CRLF 5771100928Snectar was just at the end of a processed block. The bug was discovered when 5772100928Snectar processing data through a buffering memory BIO handing the data to a 5773100928Snectar BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 5774100928Snectar <ptsekov@syntrex.com> and Nedelcho Stanev. 5775100928Snectar [Lutz Jaenicke] 5776100928Snectar 5777100928Snectar *) Implement a countermeasure against a vulnerability recently found 5778100928Snectar in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 5779100928Snectar before application data chunks to avoid the use of known IVs 5780100928Snectar with data potentially chosen by the attacker. 5781100928Snectar [Bodo Moeller] 5782100928Snectar 5783100928Snectar *) Fix length checks in ssl3_get_client_hello(). 5784100928Snectar [Bodo Moeller] 5785100928Snectar 5786100928Snectar *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 5787100928Snectar to prevent ssl3_read_internal() from incorrectly assuming that 5788100928Snectar ssl3_read_bytes() found application data while handshake 5789100928Snectar processing was enabled when in fact s->s3->in_read_app_data was 5790100928Snectar merely automatically cleared during the initial handshake. 5791100928Snectar [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 5792100928Snectar 5793100928Snectar *) Fix object definitions for Private and Enterprise: they were not 5794100928Snectar recognized in their shortname (=lowercase) representation. Extend 5795100928Snectar obj_dat.pl to issue an error when using undefined keywords instead 5796100928Snectar of silently ignoring the problem (Svenning Sorensen 5797100928Snectar <sss@sss.dnsalias.net>). 5798100928Snectar [Lutz Jaenicke] 5799100928Snectar 5800100928Snectar *) Fix DH_generate_parameters() so that it works for 'non-standard' 5801100928Snectar generators, i.e. generators other than 2 and 5. (Previously, the 5802100928Snectar code did not properly initialise the 'add' and 'rem' values to 5803100928Snectar BN_generate_prime().) 5804100928Snectar 5805100928Snectar In the new general case, we do not insist that 'generator' is 5806100928Snectar actually a primitive root: This requirement is rather pointless; 5807100928Snectar a generator of the order-q subgroup is just as good, if not 5808100928Snectar better. 5809100928Snectar [Bodo Moeller] 5810100928Snectar 5811100928Snectar *) Map new X509 verification errors to alerts. Discovered and submitted by 5812100928Snectar Tom Wu <tom@arcot.com>. 5813100928Snectar [Lutz Jaenicke] 5814100928Snectar 5815100928Snectar *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 5816100928Snectar returning non-zero before the data has been completely received 5817100928Snectar when using non-blocking I/O. 5818100928Snectar [Bodo Moeller; problem pointed out by John Hughes] 5819100928Snectar 5820100928Snectar *) Some of the ciphers missed the strength entry (SSL_LOW etc). 5821100928Snectar [Ben Laurie, Lutz Jaenicke] 5822100928Snectar 5823100928Snectar *) Fix bug in SSL_clear(): bad sessions were not removed (found by 5824100928Snectar Yoram Zahavi <YoramZ@gilian.com>). 5825100928Snectar [Lutz Jaenicke] 5826100928Snectar 5827100928Snectar *) Add information about CygWin 1.3 and on, and preserve proper 5828100928Snectar configuration for the versions before that. 5829100928Snectar [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 5830100928Snectar 5831100928Snectar *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 5832100928Snectar check whether we deal with a copy of a session and do not delete from 5833100928Snectar the cache in this case. Problem reported by "Izhar Shoshani Levi" 5834100928Snectar <izhar@checkpoint.com>. 5835100928Snectar [Lutz Jaenicke] 5836100928Snectar 5837100928Snectar *) Do not store session data into the internal session cache, if it 5838100928Snectar is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 5839100928Snectar flag is set). Proposed by Aslam <aslam@funk.com>. 5840100928Snectar [Lutz Jaenicke] 5841100928Snectar 5842100928Snectar *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 5843100928Snectar value is 0. 5844100928Snectar [Richard Levitte] 5845100928Snectar 5846109998Smarkm *) [In 0.9.6d-engine release:] 5847109998Smarkm Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 5848100928Snectar [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 5849100928Snectar 5850100928Snectar *) Add the configuration target linux-s390x. 5851100928Snectar [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 5852100928Snectar 5853100928Snectar *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 5854100928Snectar ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 5855100928Snectar variable as an indication that a ClientHello message has been 5856100928Snectar received. As the flag value will be lost between multiple 5857100928Snectar invocations of ssl3_accept when using non-blocking I/O, the 5858100928Snectar function may not be aware that a handshake has actually taken 5859100928Snectar place, thus preventing a new session from being added to the 5860100928Snectar session cache. 5861100928Snectar 5862100928Snectar To avoid this problem, we now set s->new_session to 2 instead of 5863100928Snectar using a local variable. 5864100928Snectar [Lutz Jaenicke, Bodo Moeller] 5865100928Snectar 5866100928Snectar *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 5867100928Snectar if the SSL_R_LENGTH_MISMATCH error is detected. 5868100928Snectar [Geoff Thorpe, Bodo Moeller] 5869100928Snectar 5870100928Snectar *) New 'shared_ldflag' column in Configure platform table. 5871100928Snectar [Richard Levitte] 5872100928Snectar 5873100928Snectar *) Fix EVP_CIPHER_mode macro. 5874100928Snectar ["Dan S. Camper" <dan@bti.net>] 5875100928Snectar 5876100928Snectar *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 5877100928Snectar type, we must throw them away by setting rr->length to 0. 5878100928Snectar [D P Chang <dpc@qualys.com>] 5879100928Snectar 588089837Skris Changes between 0.9.6b and 0.9.6c [21 dec 2001] 588189837Skris 588289837Skris *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 588389837Skris <Dominikus.Scherkl@biodata.com>. (The previous implementation 588489837Skris worked incorrectly for those cases where range = 10..._2 and 588589837Skris 3*range is two bits longer than range.) 588689837Skris [Bodo Moeller] 588789837Skris 588889837Skris *) Only add signing time to PKCS7 structures if it is not already 588989837Skris present. 589089837Skris [Steve Henson] 589189837Skris 589289837Skris *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 589389837Skris OBJ_ld_ce should be OBJ_id_ce. 589489837Skris Also some ip-pda OIDs in crypto/objects/objects.txt were 589589837Skris incorrect (cf. RFC 3039). 589689837Skris [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 589789837Skris 589889837Skris *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 589989837Skris returns early because it has nothing to do. 590089837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 590189837Skris 590289837Skris *) [In 0.9.6c-engine release:] 590389837Skris Fix mutex callback return values in crypto/engine/hw_ncipher.c. 590489837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 590589837Skris 590689837Skris *) [In 0.9.6c-engine release:] 590789837Skris Add support for Cryptographic Appliance's keyserver technology. 590889837Skris (Use engine 'keyclient') 590989837Skris [Cryptographic Appliances and Geoff Thorpe] 591089837Skris 591189837Skris *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 591289837Skris is called via tools/c89.sh because arguments have to be 591389837Skris rearranged (all '-L' options must appear before the first object 591489837Skris modules). 591589837Skris [Richard Shapiro <rshapiro@abinitio.com>] 591689837Skris 591789837Skris *) [In 0.9.6c-engine release:] 591889837Skris Add support for Broadcom crypto accelerator cards, backported 591989837Skris from 0.9.7. 592089837Skris [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 592189837Skris 592289837Skris *) [In 0.9.6c-engine release:] 592389837Skris Add support for SureWare crypto accelerator cards from 592489837Skris Baltimore Technologies. (Use engine 'sureware') 592589837Skris [Baltimore Technologies and Mark Cox] 592689837Skris 592789837Skris *) [In 0.9.6c-engine release:] 592889837Skris Add support for crypto accelerator cards from Accelerated 592989837Skris Encryption Processing, www.aep.ie. (Use engine 'aep') 593089837Skris [AEP Inc. and Mark Cox] 593189837Skris 593289837Skris *) Add a configuration entry for gcc on UnixWare. 593389837Skris [Gary Benson <gbenson@redhat.com>] 593489837Skris 593589837Skris *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 593689837Skris messages are stored in a single piece (fixed-length part and 593789837Skris variable-length part combined) and fix various bugs found on the way. 593889837Skris [Bodo Moeller] 593989837Skris 594089837Skris *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 594189837Skris instead. BIO_gethostbyname() does not know what timeouts are 594289837Skris appropriate, so entries would stay in cache even when they have 594389837Skris become invalid. 594489837Skris [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 594589837Skris 594689837Skris *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 594789837Skris faced with a pathologically small ClientHello fragment that does 594889837Skris not contain client_version: Instead of aborting with an error, 594989837Skris simply choose the highest available protocol version (i.e., 595089837Skris TLS 1.0 unless it is disabled). In practice, ClientHello 595189837Skris messages are never sent like this, but this change gives us 595289837Skris strictly correct behaviour at least for TLS. 595389837Skris [Bodo Moeller] 595489837Skris 595589837Skris *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 595689837Skris never resets s->method to s->ctx->method when called from within 595789837Skris one of the SSL handshake functions. 595889837Skris [Bodo Moeller; problem pointed out by Niko Baric] 595989837Skris 596089837Skris *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 596189837Skris (sent using the client's version number) if client_version is 596289837Skris smaller than the protocol version in use. Also change 596389837Skris ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 596489837Skris the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 596589837Skris the client will at least see that alert. 596689837Skris [Bodo Moeller] 596789837Skris 596889837Skris *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 596989837Skris correctly. 597089837Skris [Bodo Moeller] 597189837Skris 597289837Skris *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 597389837Skris client receives HelloRequest while in a handshake. 597489837Skris [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 597589837Skris 597689837Skris *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 597789837Skris should end in 'break', not 'goto end' which circuments various 597889837Skris cleanups done in state SSL_ST_OK. But session related stuff 597989837Skris must be disabled for SSL_ST_OK in the case that we just sent a 598089837Skris HelloRequest. 598189837Skris 598289837Skris Also avoid some overhead by not calling ssl_init_wbio_buffer() 598389837Skris before just sending a HelloRequest. 598489837Skris [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 598589837Skris 598689837Skris *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 598789837Skris reveal whether illegal block cipher padding was found or a MAC 598889837Skris verification error occured. (Neither SSLerr() codes nor alerts 598989837Skris are directly visible to potential attackers, but the information 599089837Skris may leak via logfiles.) 599189837Skris 599289837Skris Similar changes are not required for the SSL 2.0 implementation 599389837Skris because the number of padding bytes is sent in clear for SSL 2.0, 599489837Skris and the extra bytes are just ignored. However ssl/s2_pkt.c 599589837Skris failed to verify that the purported number of padding bytes is in 599689837Skris the legal range. 599789837Skris [Bodo Moeller] 599889837Skris 599989837Skris *) Add OpenUNIX-8 support including shared libraries 600089837Skris (Boyd Lynn Gerber <gerberb@zenez.com>). 600189837Skris [Lutz Jaenicke] 600289837Skris 600389837Skris *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 600489837Skris 'wristwatch attack' using huge encoding parameters (cf. 600589837Skris James H. Manger's CRYPTO 2001 paper). Note that the 600689837Skris RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 600789837Skris encoding parameters and hence was not vulnerable. 600889837Skris [Bodo Moeller] 600989837Skris 601089837Skris *) BN_sqr() bug fix. 601189837Skris [Ulf M�ller, reported by Jim Ellis <jim.ellis@cavium.com>] 601289837Skris 601389837Skris *) Rabin-Miller test analyses assume uniformly distributed witnesses, 601489837Skris so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 601589837Skris followed by modular reduction. 601689837Skris [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 601789837Skris 601889837Skris *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 601989837Skris equivalent based on BN_pseudo_rand() instead of BN_rand(). 602089837Skris [Bodo Moeller] 602189837Skris 602289837Skris *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 602389837Skris This function was broken, as the check for a new client hello message 602489837Skris to handle SGC did not allow these large messages. 602589837Skris (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 602689837Skris [Lutz Jaenicke] 602789837Skris 602889837Skris *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 602989837Skris [Lutz Jaenicke] 603089837Skris 603189837Skris *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 603289837Skris for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 603389837Skris [Lutz Jaenicke] 603489837Skris 603589837Skris *) Rework the configuration and shared library support for Tru64 Unix. 603689837Skris The configuration part makes use of modern compiler features and 603789837Skris still retains old compiler behavior for those that run older versions 603889837Skris of the OS. The shared library support part includes a variant that 603989837Skris uses the RPATH feature, and is available through the special 604089837Skris configuration target "alpha-cc-rpath", which will never be selected 604189837Skris automatically. 604289837Skris [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 604389837Skris 604489837Skris *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 604589837Skris with the same message size as in ssl3_get_certificate_request(). 604689837Skris Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 604789837Skris messages might inadvertently be reject as too long. 604889837Skris [Petr Lampa <lampa@fee.vutbr.cz>] 604989837Skris 605089837Skris *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 605189837Skris [Andy Polyakov] 605289837Skris 605389837Skris *) Modified SSL library such that the verify_callback that has been set 605489837Skris specificly for an SSL object with SSL_set_verify() is actually being 605589837Skris used. Before the change, a verify_callback set with this function was 605689837Skris ignored and the verify_callback() set in the SSL_CTX at the time of 605789837Skris the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 605889837Skris to allow the necessary settings. 605989837Skris [Lutz Jaenicke] 606089837Skris 606189837Skris *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 606289837Skris explicitly to NULL, as at least on Solaris 8 this seems not always to be 606389837Skris done automatically (in contradiction to the requirements of the C 606489837Skris standard). This made problems when used from OpenSSH. 606589837Skris [Lutz Jaenicke] 606689837Skris 606789837Skris *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 606889837Skris dh->length and always used 606989837Skris 607089837Skris BN_rand_range(priv_key, dh->p). 607189837Skris 607289837Skris BN_rand_range() is not necessary for Diffie-Hellman, and this 607389837Skris specific range makes Diffie-Hellman unnecessarily inefficient if 607489837Skris dh->length (recommended exponent length) is much smaller than the 607589837Skris length of dh->p. We could use BN_rand_range() if the order of 607689837Skris the subgroup was stored in the DH structure, but we only have 607789837Skris dh->length. 607889837Skris 607989837Skris So switch back to 608089837Skris 608189837Skris BN_rand(priv_key, l, ...) 608289837Skris 608389837Skris where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 608489837Skris otherwise. 608589837Skris [Bodo Moeller] 608689837Skris 608789837Skris *) In 608889837Skris 608989837Skris RSA_eay_public_encrypt 609089837Skris RSA_eay_private_decrypt 609189837Skris RSA_eay_private_encrypt (signing) 609289837Skris RSA_eay_public_decrypt (signature verification) 609389837Skris 609489837Skris (default implementations for RSA_public_encrypt, 609589837Skris RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 609689837Skris always reject numbers >= n. 609789837Skris [Bodo Moeller] 609889837Skris 609989837Skris *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 610089837Skris to synchronize access to 'locking_thread'. This is necessary on 610189837Skris systems where access to 'locking_thread' (an 'unsigned long' 610289837Skris variable) is not atomic. 610389837Skris [Bodo Moeller] 610489837Skris 610589837Skris *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 610689837Skris *before* setting the 'crypto_lock_rand' flag. The previous code had 610789837Skris a race condition if 0 is a valid thread ID. 610889837Skris [Travis Vitek <vitek@roguewave.com>] 610989837Skris 611089837Skris *) Add support for shared libraries under Irix. 611189837Skris [Albert Chin-A-Young <china@thewrittenword.com>] 611289837Skris 611389837Skris *) Add configuration option to build on Linux on both big-endian and 611489837Skris little-endian MIPS. 611589837Skris [Ralf Baechle <ralf@uni-koblenz.de>] 611689837Skris 611789837Skris *) Add the possibility to create shared libraries on HP-UX. 611889837Skris [Richard Levitte] 611989837Skris 612079998Skris Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 612179998Skris 612279998Skris *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 612379998Skris to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 612479998Skris Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 612579998Skris PRNG state recovery was possible based on the output of 612679998Skris one PRNG request appropriately sized to gain knowledge on 612779998Skris 'md' followed by enough consecutive 1-byte PRNG requests 612879998Skris to traverse all of 'state'. 612979998Skris 613079998Skris 1. When updating 'md_local' (the current thread's copy of 'md') 613179998Skris during PRNG output generation, hash all of the previous 613279998Skris 'md_local' value, not just the half used for PRNG output. 613379998Skris 613479998Skris 2. Make the number of bytes from 'state' included into the hash 613579998Skris independent from the number of PRNG bytes requested. 613679998Skris 613779998Skris The first measure alone would be sufficient to avoid 613879998Skris Markku-Juhani's attack. (Actually it had never occurred 613979998Skris to me that the half of 'md_local' used for chaining was the 614079998Skris half from which PRNG output bytes were taken -- I had always 614179998Skris assumed that the secret half would be used.) The second 614279998Skris measure makes sure that additional data from 'state' is never 614379998Skris mixed into 'md_local' in small portions; this heuristically 614479998Skris further strengthens the PRNG. 614579998Skris [Bodo Moeller] 614679998Skris 614779998Skris *) Fix crypto/bn/asm/mips3.s. 614879998Skris [Andy Polyakov] 614979998Skris 615079998Skris *) When only the key is given to "enc", the IV is undefined. Print out 615179998Skris an error message in this case. 615279998Skris [Lutz Jaenicke] 615379998Skris 615479998Skris *) Handle special case when X509_NAME is empty in X509 printing routines. 615579998Skris [Steve Henson] 615679998Skris 615779998Skris *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 615879998Skris positive and less than q. 615979998Skris [Bodo Moeller] 616079998Skris 616179998Skris *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 616279998Skris used: it isn't thread safe and the add_lock_callback should handle 616379998Skris that itself. 616479998Skris [Paul Rose <Paul.Rose@bridge.com>] 616579998Skris 616679998Skris *) Verify that incoming data obeys the block size in 616779998Skris ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 616879998Skris [Bodo Moeller] 616979998Skris 617079998Skris *) Fix OAEP check. 617179998Skris [Ulf M�ller, Bodo M�ller] 617279998Skris 617379998Skris *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 617489837Skris RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 617579998Skris when fixing the server behaviour for backwards-compatible 'client 617679998Skris hello' messages. (Note that the attack is impractical against 617779998Skris SSL 3.0 and TLS 1.0 anyway because length and version checking 617879998Skris means that the probability of guessing a valid ciphertext is 617979998Skris around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 618079998Skris paper.) 618179998Skris 618279998Skris Before 0.9.5, the countermeasure (hide the error by generating a 618379998Skris random 'decryption result') did not work properly because 618479998Skris ERR_clear_error() was missing, meaning that SSL_get_error() would 618579998Skris detect the supposedly ignored error. 618679998Skris 618779998Skris Both problems are now fixed. 618879998Skris [Bodo Moeller] 618979998Skris 619079998Skris *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 619179998Skris (previously it was 1024). 619279998Skris [Bodo Moeller] 619379998Skris 619479998Skris *) Fix for compatibility mode trust settings: ignore trust settings 619579998Skris unless some valid trust or reject settings are present. 619679998Skris [Steve Henson] 619779998Skris 619879998Skris *) Fix for blowfish EVP: its a variable length cipher. 619979998Skris [Steve Henson] 620079998Skris 620179998Skris *) Fix various bugs related to DSA S/MIME verification. Handle missing 620279998Skris parameters in DSA public key structures and return an error in the 620379998Skris DSA routines if parameters are absent. 620479998Skris [Steve Henson] 620579998Skris 620679998Skris *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 620779998Skris in the current directory if neither $RANDFILE nor $HOME was set. 620879998Skris RAND_file_name() in 0.9.6a returned NULL in this case. This has 620979998Skris caused some confusion to Windows users who haven't defined $HOME. 621079998Skris Thus RAND_file_name() is changed again: e_os.h can define a 621179998Skris DEFAULT_HOME, which will be used if $HOME is not set. 621279998Skris For Windows, we use "C:"; on other platforms, we still require 621379998Skris environment variables. 621479998Skris 621579998Skris *) Move 'if (!initialized) RAND_poll()' into regions protected by 621679998Skris CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 621779998Skris having multiple threads call RAND_poll() concurrently. 621879998Skris [Bodo Moeller] 621979998Skris 622079998Skris *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 622179998Skris combination of a flag and a thread ID variable. 622279998Skris Otherwise while one thread is in ssleay_rand_bytes (which sets the 622379998Skris flag), *other* threads can enter ssleay_add_bytes without obeying 622489837Skris the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 622579998Skris that they do not hold after the first thread unsets add_do_not_lock). 622679998Skris [Bodo Moeller] 622779998Skris 622879998Skris *) Change bctest again: '-x' expressions are not available in all 622979998Skris versions of 'test'. 623079998Skris [Bodo Moeller] 623179998Skris 623276866Skris Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 623372613Skris 623476866Skris *) Fix a couple of memory leaks in PKCS7_dataDecode() 623576866Skris [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 623676866Skris 623776866Skris *) Change Configure and Makefiles to provide EXE_EXT, which will contain 623876866Skris the default extension for executables, if any. Also, make the perl 623976866Skris scripts that use symlink() to test if it really exists and use "cp" 624076866Skris if it doesn't. All this made OpenSSL compilable and installable in 624176866Skris CygWin. 624276866Skris [Richard Levitte] 624376866Skris 624476866Skris *) Fix for asn1_GetSequence() for indefinite length constructed data. 624576866Skris If SEQUENCE is length is indefinite just set c->slen to the total 624676866Skris amount of data available. 624776866Skris [Steve Henson, reported by shige@FreeBSD.org] 624876866Skris [This change does not apply to 0.9.7.] 624976866Skris 625076866Skris *) Change bctest to avoid here-documents inside command substitution 625176866Skris (workaround for FreeBSD /bin/sh bug). 625276866Skris For compatibility with Ultrix, avoid shell functions (introduced 625376866Skris in the bctest version that searches along $PATH). 625476866Skris [Bodo Moeller] 625576866Skris 625676866Skris *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 625776866Skris with des_encrypt() defined on some operating systems, like Solaris 625876866Skris and UnixWare. 625976866Skris [Richard Levitte] 626076866Skris 626176866Skris *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 626276866Skris On the Importance of Eliminating Errors in Cryptographic 626376866Skris Computations, J. Cryptology 14 (2001) 2, 101-119, 626476866Skris http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 626576866Skris [Ulf Moeller] 626676866Skris 626776866Skris *) MIPS assembler BIGNUM division bug fix. 626876866Skris [Andy Polyakov] 626976866Skris 627076866Skris *) Disabled incorrect Alpha assembler code. 627176866Skris [Richard Levitte] 627276866Skris 627376866Skris *) Fix PKCS#7 decode routines so they correctly update the length 627476866Skris after reading an EOC for the EXPLICIT tag. 627576866Skris [Steve Henson] 627676866Skris [This change does not apply to 0.9.7.] 627776866Skris 627876866Skris *) Fix bug in PKCS#12 key generation routines. This was triggered 627976866Skris if a 3DES key was generated with a 0 initial byte. Include 628076866Skris PKCS12_BROKEN_KEYGEN compilation option to retain the old 628176866Skris (but broken) behaviour. 628276866Skris [Steve Henson] 628376866Skris 628476866Skris *) Enhance bctest to search for a working bc along $PATH and print 628576866Skris it when found. 628676866Skris [Tim Rice <tim@multitalents.net> via Richard Levitte] 628776866Skris 628876866Skris *) Fix memory leaks in err.c: free err_data string if necessary; 628976866Skris don't write to the wrong index in ERR_set_error_data. 629076866Skris [Bodo Moeller] 629176866Skris 629276866Skris *) Implement ssl23_peek (analogous to ssl23_read), which previously 629376866Skris did not exist. 629476866Skris [Bodo Moeller] 629576866Skris 629676866Skris *) Replace rdtsc with _emit statements for VC++ version 5. 629776866Skris [Jeremy Cooper <jeremy@baymoo.org>] 629876866Skris 629976866Skris *) Make it possible to reuse SSLv2 sessions. 630076866Skris [Richard Levitte] 630176866Skris 630276866Skris *) In copy_email() check for >= 0 as a return value for 630376866Skris X509_NAME_get_index_by_NID() since 0 is a valid index. 630476866Skris [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 630576866Skris 630676866Skris *) Avoid coredump with unsupported or invalid public keys by checking if 630776866Skris X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 630876866Skris PKCS7_verify() fails with non detached data. 630976866Skris [Steve Henson] 631076866Skris 631176866Skris *) Don't use getenv in library functions when run as setuid/setgid. 631276866Skris New function OPENSSL_issetugid(). 631376866Skris [Ulf Moeller] 631476866Skris 631576866Skris *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 631676866Skris due to incorrect handling of multi-threading: 631776866Skris 631876866Skris 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 631976866Skris 632076866Skris 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 632176866Skris 632276866Skris 3. Count how many times MemCheck_off() has been called so that 632376866Skris nested use can be treated correctly. This also avoids 632476866Skris inband-signalling in the previous code (which relied on the 632576866Skris assumption that thread ID 0 is impossible). 632676866Skris [Bodo Moeller] 632776866Skris 632876866Skris *) Add "-rand" option also to s_client and s_server. 632976866Skris [Lutz Jaenicke] 633076866Skris 633176866Skris *) Fix CPU detection on Irix 6.x. 633276866Skris [Kurt Hockenbury <khockenb@stevens-tech.edu> and 633376866Skris "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 633476866Skris 633576866Skris *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 633676866Skris was empty. 633776866Skris [Steve Henson] 633876866Skris [This change does not apply to 0.9.7.] 633976866Skris 634076866Skris *) Use the cached encoding of an X509_NAME structure rather than 634176866Skris copying it. This is apparently the reason for the libsafe "errors" 634276866Skris but the code is actually correct. 634376866Skris [Steve Henson] 634476866Skris 634572613Skris *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 634672613Skris Bleichenbacher's DSA attack. 634776866Skris Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 634876866Skris to be set and top=0 forces the highest bit to be set; top=-1 is new 634976866Skris and leaves the highest bit random. 635076866Skris [Ulf Moeller, Bodo Moeller] 635172613Skris 635272613Skris *) In the NCONF_...-based implementations for CONF_... queries 635372613Skris (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 635472613Skris a temporary CONF structure with the data component set to NULL 635572613Skris (which gives segmentation faults in lh_retrieve). 635672613Skris Instead, use NULL for the CONF pointer in CONF_get_string and 635772613Skris CONF_get_number (which may use environment variables) and directly 635872613Skris return NULL from CONF_get_section. 635972613Skris [Bodo Moeller] 636072613Skris 636172613Skris *) Fix potential buffer overrun for EBCDIC. 636272613Skris [Ulf Moeller] 636372613Skris 636472613Skris *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 636572613Skris keyUsage if basicConstraints absent for a CA. 636672613Skris [Steve Henson] 636772613Skris 636872613Skris *) Make SMIME_write_PKCS7() write mail header values with a format that 636972613Skris is more generally accepted (no spaces before the semicolon), since 637072613Skris some programs can't parse those values properly otherwise. Also make 637172613Skris sure BIO's that break lines after each write do not create invalid 637272613Skris headers. 637372613Skris [Richard Levitte] 637472613Skris 637572613Skris *) Make the CRL encoding routines work with empty SEQUENCE OF. The 637672613Skris macros previously used would not encode an empty SEQUENCE OF 637772613Skris and break the signature. 637872613Skris [Steve Henson] 637976866Skris [This change does not apply to 0.9.7.] 638072613Skris 638172613Skris *) Zero the premaster secret after deriving the master secret in 638272613Skris DH ciphersuites. 638372613Skris [Steve Henson] 638472613Skris 638572613Skris *) Add some EVP_add_digest_alias registrations (as found in 638672613Skris OpenSSL_add_all_digests()) to SSL_library_init() 638772613Skris aka OpenSSL_add_ssl_algorithms(). This provides improved 638872613Skris compatibility with peers using X.509 certificates 638972613Skris with unconventional AlgorithmIdentifier OIDs. 639072613Skris [Bodo Moeller] 639172613Skris 639272613Skris *) Fix for Irix with NO_ASM. 639372613Skris ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 639472613Skris 639572613Skris *) ./config script fixes. 639672613Skris [Ulf Moeller, Richard Levitte] 639772613Skris 639872613Skris *) Fix 'openssl passwd -1'. 639972613Skris [Bodo Moeller] 640072613Skris 640172613Skris *) Change PKCS12_key_gen_asc() so it can cope with non null 640272613Skris terminated strings whose length is passed in the passlen 640372613Skris parameter, for example from PEM callbacks. This was done 640472613Skris by adding an extra length parameter to asc2uni(). 640572613Skris [Steve Henson, reported by <oddissey@samsung.co.kr>] 640672613Skris 640772613Skris *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 640872613Skris call failed, free the DSA structure. 640972613Skris [Bodo Moeller] 641072613Skris 641172613Skris *) Fix to uni2asc() to cope with zero length Unicode strings. 641272613Skris These are present in some PKCS#12 files. 641372613Skris [Steve Henson] 641472613Skris 641572613Skris *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 641672613Skris Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 641772613Skris when writing a 32767 byte record. 641872613Skris [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 641972613Skris 642072613Skris *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 642172613Skris obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 642272613Skris 642372613Skris (RSA objects have a reference count access to which is protected 642472613Skris by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 642572613Skris so they are meant to be shared between threads.) 642672613Skris [Bodo Moeller, Geoff Thorpe; original patch submitted by 642772613Skris "Reddie, Steven" <Steven.Reddie@ca.com>] 642872613Skris 642972613Skris *) Fix a deadlock in CRYPTO_mem_leaks(). 643072613Skris [Bodo Moeller] 643172613Skris 643276866Skris *) Use better test patterns in bntest. 643376866Skris [Ulf M�ller] 643476866Skris 643572613Skris *) rand_win.c fix for Borland C. 643672613Skris [Ulf M�ller] 643772613Skris 643872613Skris *) BN_rshift bugfix for n == 0. 643972613Skris [Bodo Moeller] 644072613Skris 644176866Skris *) Add a 'bctest' script that checks for some known 'bc' bugs 644276866Skris so that 'make test' does not abort just because 'bc' is broken. 644376866Skris [Bodo Moeller] 644476866Skris 644572613Skris *) Store verify_result within SSL_SESSION also for client side to 644672613Skris avoid potential security hole. (Re-used sessions on the client side 644772613Skris always resulted in verify_result==X509_V_OK, not using the original 644872613Skris result of the server certificate verification.) 644972613Skris [Lutz Jaenicke] 645072613Skris 645172613Skris *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 645272613Skris SSL3_RT_APPLICATION_DATA, return 0. 645372613Skris Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 645472613Skris [Bodo Moeller] 645572613Skris 645672613Skris *) Fix SSL_peek: 645772613Skris Both ssl2_peek and ssl3_peek, which were totally broken in earlier 645872613Skris releases, have been re-implemented by renaming the previous 645972613Skris implementations of ssl2_read and ssl3_read to ssl2_read_internal 646072613Skris and ssl3_read_internal, respectively, and adding 'peek' parameters 646172613Skris to them. The new ssl[23]_{read,peek} functions are calls to 646272613Skris ssl[23]_read_internal with the 'peek' flag set appropriately. 646372613Skris A 'peek' parameter has also been added to ssl3_read_bytes, which 646472613Skris does the actual work for ssl3_read_internal. 646572613Skris [Bodo Moeller] 646672613Skris 646776866Skris *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 646876866Skris the method-specific "init()" handler. Also clean up ex_data after 646976866Skris calling the method-specific "finish()" handler. Previously, this was 647076866Skris happening the other way round. 647176866Skris [Geoff Thorpe] 647276866Skris 647372613Skris *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 647472613Skris The previous value, 12, was not always sufficient for BN_mod_exp(). 647572613Skris [Bodo Moeller] 647672613Skris 647776866Skris *) Make sure that shared libraries get the internal name engine with 647876866Skris the full version number and not just 0. This should mark the 647976866Skris shared libraries as not backward compatible. Of course, this should 648076866Skris be changed again when we can guarantee backward binary compatibility. 648176866Skris [Richard Levitte] 648276866Skris 648372613Skris *) Fix typo in get_cert_by_subject() in by_dir.c 648472613Skris [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 648572613Skris 648676866Skris *) Rework the system to generate shared libraries: 648776866Skris 648876866Skris - Make note of the expected extension for the shared libraries and 648976866Skris if there is a need for symbolic links from for example libcrypto.so.0 649076866Skris to libcrypto.so.0.9.7. There is extended info in Configure for 649176866Skris that. 649276866Skris 649376866Skris - Make as few rebuilds of the shared libraries as possible. 649476866Skris 649576866Skris - Still avoid linking the OpenSSL programs with the shared libraries. 649676866Skris 649776866Skris - When installing, install the shared libraries separately from the 649876866Skris static ones. 649976866Skris [Richard Levitte] 650076866Skris 650172613Skris *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 650272613Skris 650372613Skris Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 650472613Skris and not in SSL_clear because the latter is also used by the 650572613Skris accept/connect functions; previously, the settings made by 650672613Skris SSL_set_read_ahead would be lost during the handshake. 650772613Skris [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 650872613Skris 650972613Skris *) Correct util/mkdef.pl to be selective about disabled algorithms. 651072613Skris Previously, it would create entries for disableed algorithms no 651172613Skris matter what. 651272613Skris [Richard Levitte] 651372613Skris 651472613Skris *) Added several new manual pages for SSL_* function. 651572613Skris [Lutz Jaenicke] 651672613Skris 651768651Skris Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 651868651Skris 651968651Skris *) In ssl23_get_client_hello, generate an error message when faced 652068651Skris with an initial SSL 3.0/TLS record that is too small to contain the 652168651Skris first two bytes of the ClientHello message, i.e. client_version. 652268651Skris (Note that this is a pathologic case that probably has never happened 652368651Skris in real life.) The previous approach was to use the version number 652468651Skris from the record header as a substitute; but our protocol choice 652568651Skris should not depend on that one because it is not authenticated 652668651Skris by the Finished messages. 652768651Skris [Bodo Moeller] 652868651Skris 652968651Skris *) More robust randomness gathering functions for Windows. 653068651Skris [Jeffrey Altman <jaltman@columbia.edu>] 653168651Skris 653268651Skris *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 653368651Skris not set then we don't setup the error code for issuer check errors 653468651Skris to avoid possibly overwriting other errors which the callback does 653568651Skris handle. If an application does set the flag then we assume it knows 653668651Skris what it is doing and can handle the new informational codes 653768651Skris appropriately. 653868651Skris [Steve Henson] 653968651Skris 654068651Skris *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 654168651Skris a general "ANY" type, as such it should be able to decode anything 654268651Skris including tagged types. However it didn't check the class so it would 654368651Skris wrongly interpret tagged types in the same way as their universal 654468651Skris counterpart and unknown types were just rejected. Changed so that the 654568651Skris tagged and unknown types are handled in the same way as a SEQUENCE: 654668651Skris that is the encoding is stored intact. There is also a new type 654768651Skris "V_ASN1_OTHER" which is used when the class is not universal, in this 654868651Skris case we have no idea what the actual type is so we just lump them all 654968651Skris together. 655068651Skris [Steve Henson] 655168651Skris 655268651Skris *) On VMS, stdout may very well lead to a file that is written to 655368651Skris in a record-oriented fashion. That means that every write() will 655468651Skris write a separate record, which will be read separately by the 655568651Skris programs trying to read from it. This can be very confusing. 655668651Skris 655768651Skris The solution is to put a BIO filter in the way that will buffer 655868651Skris text until a linefeed is reached, and then write everything a 655968651Skris line at a time, so every record written will be an actual line, 656068651Skris not chunks of lines and not (usually doesn't happen, but I've 656168651Skris seen it once) several lines in one record. BIO_f_linebuffer() is 656268651Skris the answer. 656368651Skris 656468651Skris Currently, it's a VMS-only method, because that's where it has 656568651Skris been tested well enough. 656668651Skris [Richard Levitte] 656768651Skris 656868651Skris *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 656968651Skris it can return incorrect results. 657068651Skris (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 657168651Skris but it was in 0.9.6-beta[12].) 657268651Skris [Bodo Moeller] 657368651Skris 657468651Skris *) Disable the check for content being present when verifying detached 657568651Skris signatures in pk7_smime.c. Some versions of Netscape (wrongly) 657668651Skris include zero length content when signing messages. 657768651Skris [Steve Henson] 657868651Skris 657968651Skris *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 658068651Skris BIO_ctrl (for BIO pairs). 658168651Skris [Bodo M�ller] 658268651Skris 658368651Skris *) Add DSO method for VMS. 658468651Skris [Richard Levitte] 658568651Skris 658668651Skris *) Bug fix: Montgomery multiplication could produce results with the 658768651Skris wrong sign. 658868651Skris [Ulf M�ller] 658968651Skris 659068651Skris *) Add RPM specification openssl.spec and modify it to build three 659168651Skris packages. The default package contains applications, application 659268651Skris documentation and run-time libraries. The devel package contains 659368651Skris include files, static libraries and function documentation. The 659468651Skris doc package contains the contents of the doc directory. The original 659568651Skris openssl.spec was provided by Damien Miller <djm@mindrot.org>. 659668651Skris [Richard Levitte] 659768651Skris 659868651Skris *) Add a large number of documentation files for many SSL routines. 659968651Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 660068651Skris 660168651Skris *) Add a configuration entry for Sony News 4. 660268651Skris [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 660368651Skris 660468651Skris *) Don't set the two most significant bits to one when generating a 660568651Skris random number < q in the DSA library. 660668651Skris [Ulf M�ller] 660768651Skris 660868651Skris *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 660968651Skris behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 661068651Skris the underlying transport is blocking) if a handshake took place. 661168651Skris (The default behaviour is needed by applications such as s_client 661268651Skris and s_server that use select() to determine when to use SSL_read; 661368651Skris but for applications that know in advance when to expect data, it 661468651Skris just makes things more complicated.) 661568651Skris [Bodo Moeller] 661668651Skris 661768651Skris *) Add RAND_egd_bytes(), which gives control over the number of bytes read 661868651Skris from EGD. 661968651Skris [Ben Laurie] 662068651Skris 662168651Skris *) Add a few more EBCDIC conditionals that make `req' and `x509' 662268651Skris work better on such systems. 662368651Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 662468651Skris 662568651Skris *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 662668651Skris Update PKCS12_parse() so it copies the friendlyName and the 662768651Skris keyid to the certificates aux info. 662868651Skris [Steve Henson] 662968651Skris 663068651Skris *) Fix bug in PKCS7_verify() which caused an infinite loop 663168651Skris if there was more than one signature. 663268651Skris [Sven Uszpelkat <su@celocom.de>] 663368651Skris 663468651Skris *) Major change in util/mkdef.pl to include extra information 663568651Skris about each symbol, as well as presentig variables as well 663668651Skris as functions. This change means that there's n more need 663768651Skris to rebuild the .num files when some algorithms are excluded. 663868651Skris [Richard Levitte] 663968651Skris 664068651Skris *) Allow the verify time to be set by an application, 664168651Skris rather than always using the current time. 664268651Skris [Steve Henson] 664368651Skris 664468651Skris *) Phase 2 verify code reorganisation. The certificate 664568651Skris verify code now looks up an issuer certificate by a 664668651Skris number of criteria: subject name, authority key id 664768651Skris and key usage. It also verifies self signed certificates 664868651Skris by the same criteria. The main comparison function is 664968651Skris X509_check_issued() which performs these checks. 665068651Skris 665168651Skris Lot of changes were necessary in order to support this 665268651Skris without completely rewriting the lookup code. 665368651Skris 665468651Skris Authority and subject key identifier are now cached. 665568651Skris 665668651Skris The LHASH 'certs' is X509_STORE has now been replaced 665768651Skris by a STACK_OF(X509_OBJECT). This is mainly because an 665868651Skris LHASH can't store or retrieve multiple objects with 665968651Skris the same hash value. 666068651Skris 666168651Skris As a result various functions (which were all internal 666268651Skris use only) have changed to handle the new X509_STORE 666368651Skris structure. This will break anything that messed round 666468651Skris with X509_STORE internally. 666568651Skris 666668651Skris The functions X509_STORE_add_cert() now checks for an 666768651Skris exact match, rather than just subject name. 666868651Skris 666968651Skris The X509_STORE API doesn't directly support the retrieval 667068651Skris of multiple certificates matching a given criteria, however 667168651Skris this can be worked round by performing a lookup first 667268651Skris (which will fill the cache with candidate certificates) 667368651Skris and then examining the cache for matches. This is probably 667468651Skris the best we can do without throwing out X509_LOOKUP 667568651Skris entirely (maybe later...). 667668651Skris 667768651Skris The X509_VERIFY_CTX structure has been enhanced considerably. 667868651Skris 667968651Skris All certificate lookup operations now go via a get_issuer() 668068651Skris callback. Although this currently uses an X509_STORE it 668168651Skris can be replaced by custom lookups. This is a simple way 668268651Skris to bypass the X509_STORE hackery necessary to make this 668368651Skris work and makes it possible to use more efficient techniques 668468651Skris in future. A very simple version which uses a simple 668568651Skris STACK for its trusted certificate store is also provided 668668651Skris using X509_STORE_CTX_trusted_stack(). 668768651Skris 668868651Skris The verify_cb() and verify() callbacks now have equivalents 668968651Skris in the X509_STORE_CTX structure. 669068651Skris 669168651Skris X509_STORE_CTX also has a 'flags' field which can be used 669268651Skris to customise the verify behaviour. 669368651Skris [Steve Henson] 669468651Skris 669568651Skris *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 669668651Skris excludes S/MIME capabilities. 669768651Skris [Steve Henson] 669868651Skris 669968651Skris *) When a certificate request is read in keep a copy of the 670068651Skris original encoding of the signed data and use it when outputing 670168651Skris again. Signatures then use the original encoding rather than 670268651Skris a decoded, encoded version which may cause problems if the 670368651Skris request is improperly encoded. 670468651Skris [Steve Henson] 670568651Skris 670668651Skris *) For consistency with other BIO_puts implementations, call 670768651Skris buffer_write(b, ...) directly in buffer_puts instead of calling 670868651Skris BIO_write(b, ...). 670968651Skris 671068651Skris In BIO_puts, increment b->num_write as in BIO_write. 671168651Skris [Peter.Sylvester@EdelWeb.fr] 671268651Skris 671368651Skris *) Fix BN_mul_word for the case where the word is 0. (We have to use 671468651Skris BN_zero, we may not return a BIGNUM with an array consisting of 671568651Skris words set to zero.) 671668651Skris [Bodo Moeller] 671768651Skris 671868651Skris *) Avoid calling abort() from within the library when problems are 671968651Skris detected, except if preprocessor symbols have been defined 672068651Skris (such as REF_CHECK, BN_DEBUG etc.). 672168651Skris [Bodo Moeller] 672268651Skris 672368651Skris *) New openssl application 'rsautl'. This utility can be 672468651Skris used for low level RSA operations. DER public key 672568651Skris BIO/fp routines also added. 672668651Skris [Steve Henson] 672768651Skris 672868651Skris *) New Configure entry and patches for compiling on QNX 4. 672968651Skris [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 673068651Skris 673168651Skris *) A demo state-machine implementation was sponsored by 673268651Skris Nuron (http://www.nuron.com/) and is now available in 673368651Skris demos/state_machine. 673468651Skris [Ben Laurie] 673568651Skris 673668651Skris *) New options added to the 'dgst' utility for signature 673768651Skris generation and verification. 673868651Skris [Steve Henson] 673968651Skris 674068651Skris *) Unrecognized PKCS#7 content types are now handled via a 674168651Skris catch all ASN1_TYPE structure. This allows unsupported 674268651Skris types to be stored as a "blob" and an application can 674368651Skris encode and decode it manually. 674468651Skris [Steve Henson] 674568651Skris 674668651Skris *) Fix various signed/unsigned issues to make a_strex.c 674768651Skris compile under VC++. 674868651Skris [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 674968651Skris 675068651Skris *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 675168651Skris length if passed a buffer. ASN1_INTEGER_to_BN failed 675268651Skris if passed a NULL BN and its argument was negative. 675368651Skris [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 675468651Skris 675568651Skris *) Modification to PKCS#7 encoding routines to output definite 675668651Skris length encoding. Since currently the whole structures are in 675768651Skris memory there's not real point in using indefinite length 675868651Skris constructed encoding. However if OpenSSL is compiled with 675968651Skris the flag PKCS7_INDEFINITE_ENCODING the old form is used. 676068651Skris [Steve Henson] 676168651Skris 676268651Skris *) Added BIO_vprintf() and BIO_vsnprintf(). 676368651Skris [Richard Levitte] 676468651Skris 676568651Skris *) Added more prefixes to parse for in the the strings written 676668651Skris through a logging bio, to cover all the levels that are available 676768651Skris through syslog. The prefixes are now: 676868651Skris 676968651Skris PANIC, EMERG, EMR => LOG_EMERG 677068651Skris ALERT, ALR => LOG_ALERT 677168651Skris CRIT, CRI => LOG_CRIT 677268651Skris ERROR, ERR => LOG_ERR 677368651Skris WARNING, WARN, WAR => LOG_WARNING 677468651Skris NOTICE, NOTE, NOT => LOG_NOTICE 677568651Skris INFO, INF => LOG_INFO 677668651Skris DEBUG, DBG => LOG_DEBUG 677768651Skris 677868651Skris and as before, if none of those prefixes are present at the 677968651Skris beginning of the string, LOG_ERR is chosen. 678068651Skris 678168651Skris On Win32, the LOG_* levels are mapped according to this: 678268651Skris 678368651Skris LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 678468651Skris LOG_WARNING => EVENTLOG_WARNING_TYPE 678568651Skris LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 678668651Skris 678768651Skris [Richard Levitte] 678868651Skris 678968651Skris *) Made it possible to reconfigure with just the configuration 679068651Skris argument "reconf" or "reconfigure". The command line arguments 679168651Skris are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 679268651Skris and are retrieved from there when reconfiguring. 679368651Skris [Richard Levitte] 679468651Skris 679568651Skris *) MD4 implemented. 679668651Skris [Assar Westerlund <assar@sics.se>, Richard Levitte] 679768651Skris 679868651Skris *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 679968651Skris [Richard Levitte] 680068651Skris 680168651Skris *) The obj_dat.pl script was messing up the sorting of object 680268651Skris names. The reason was that it compared the quoted version 680368651Skris of strings as a result "OCSP" > "OCSP Signing" because 680468651Skris " > SPACE. Changed script to store unquoted versions of 680568651Skris names and add quotes on output. It was also omitting some 680668651Skris names from the lookup table if they were given a default 680768651Skris value (that is if SN is missing it is given the same 680868651Skris value as LN and vice versa), these are now added on the 680968651Skris grounds that if an object has a name we should be able to 681068651Skris look it up. Finally added warning output when duplicate 681168651Skris short or long names are found. 681268651Skris [Steve Henson] 681368651Skris 681468651Skris *) Changes needed for Tandem NSK. 681568651Skris [Scott Uroff <scott@xypro.com>] 681668651Skris 681768651Skris *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 681868651Skris RSA_padding_check_SSLv23(), special padding was never detected 681968651Skris and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 682068651Skris version rollback attacks was not effective. 682168651Skris 682268651Skris In s23_clnt.c, don't use special rollback-attack detection padding 682368651Skris (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 682468651Skris client; similarly, in s23_srvr.c, don't do the rollback check if 682568651Skris SSL 2.0 is the only protocol enabled in the server. 682668651Skris [Bodo Moeller] 682768651Skris 682868651Skris *) Make it possible to get hexdumps of unprintable data with 'openssl 682968651Skris asn1parse'. By implication, the functions ASN1_parse_dump() and 683068651Skris BIO_dump_indent() are added. 683168651Skris [Richard Levitte] 683268651Skris 683368651Skris *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 683468651Skris these print out strings and name structures based on various 683568651Skris flags including RFC2253 support and proper handling of 683668651Skris multibyte characters. Added options to the 'x509' utility 683768651Skris to allow the various flags to be set. 683868651Skris [Steve Henson] 683968651Skris 684068651Skris *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 684168651Skris Also change the functions X509_cmp_current_time() and 684268651Skris X509_gmtime_adj() work with an ASN1_TIME structure, 684368651Skris this will enable certificates using GeneralizedTime in validity 684468651Skris dates to be checked. 684568651Skris [Steve Henson] 684668651Skris 684768651Skris *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 684868651Skris negative public key encodings) on by default, 684968651Skris NO_NEG_PUBKEY_BUG can be set to disable it. 685068651Skris [Steve Henson] 685168651Skris 685268651Skris *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 685368651Skris content octets. An i2c_ASN1_OBJECT is unnecessary because 685468651Skris the encoding can be trivially obtained from the structure. 685568651Skris [Steve Henson] 685668651Skris 685768651Skris *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 685868651Skris not read locks (CRYPTO_r_[un]lock). 685968651Skris [Bodo Moeller] 686068651Skris 686168651Skris *) A first attempt at creating official support for shared 686268651Skris libraries through configuration. I've kept it so the 686368651Skris default is static libraries only, and the OpenSSL programs 686468651Skris are always statically linked for now, but there are 686568651Skris preparations for dynamic linking in place. 686689837Skris This has been tested on Linux and Tru64. 686768651Skris [Richard Levitte] 686868651Skris 686968651Skris *) Randomness polling function for Win9x, as described in: 687068651Skris Peter Gutmann, Software Generation of Practically Strong 687168651Skris Random Numbers. 687268651Skris [Ulf M�ller] 687368651Skris 687468651Skris *) Fix so PRNG is seeded in req if using an already existing 687568651Skris DSA key. 687668651Skris [Steve Henson] 687768651Skris 687868651Skris *) New options to smime application. -inform and -outform 687968651Skris allow alternative formats for the S/MIME message including 688068651Skris PEM and DER. The -content option allows the content to be 688168651Skris specified separately. This should allow things like Netscape 688268651Skris form signing output easier to verify. 688368651Skris [Steve Henson] 688468651Skris 688568651Skris *) Fix the ASN1 encoding of tags using the 'long form'. 688668651Skris [Steve Henson] 688768651Skris 688868651Skris *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 688968651Skris STRING types. These convert content octets to and from the 689068651Skris underlying type. The actual tag and length octets are 689168651Skris already assumed to have been read in and checked. These 689268651Skris are needed because all other string types have virtually 689368651Skris identical handling apart from the tag. By having versions 689468651Skris of the ASN1 functions that just operate on content octets 689568651Skris IMPLICIT tagging can be handled properly. It also allows 689668651Skris the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 689768651Skris and ASN1_INTEGER are identical apart from the tag. 689868651Skris [Steve Henson] 689968651Skris 690068651Skris *) Change the handling of OID objects as follows: 690168651Skris 690268651Skris - New object identifiers are inserted in objects.txt, following 690368651Skris the syntax given in objects.README. 690468651Skris - objects.pl is used to process obj_mac.num and create a new 690568651Skris obj_mac.h. 690668651Skris - obj_dat.pl is used to create a new obj_dat.h, using the data in 690768651Skris obj_mac.h. 690868651Skris 690968651Skris This is currently kind of a hack, and the perl code in objects.pl 691068651Skris isn't very elegant, but it works as I intended. The simplest way 691168651Skris to check that it worked correctly is to look in obj_dat.h and 691268651Skris check the array nid_objs and make sure the objects haven't moved 691368651Skris around (this is important!). Additions are OK, as well as 691468651Skris consistent name changes. 691568651Skris [Richard Levitte] 691668651Skris 691768651Skris *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 691868651Skris [Bodo Moeller] 691968651Skris 692068651Skris *) Addition of the command line parameter '-rand file' to 'openssl req'. 692168651Skris The given file adds to whatever has already been seeded into the 692268651Skris random pool through the RANDFILE configuration file option or 692368651Skris environment variable, or the default random state file. 692468651Skris [Richard Levitte] 692568651Skris 692668651Skris *) mkstack.pl now sorts each macro group into lexical order. 692768651Skris Previously the output order depended on the order the files 692868651Skris appeared in the directory, resulting in needless rewriting 692968651Skris of safestack.h . 693068651Skris [Steve Henson] 693168651Skris 693268651Skris *) Patches to make OpenSSL compile under Win32 again. Mostly 693368651Skris work arounds for the VC++ problem that it treats func() as 693468651Skris func(void). Also stripped out the parts of mkdef.pl that 693568651Skris added extra typesafe functions: these no longer exist. 693668651Skris [Steve Henson] 693768651Skris 693868651Skris *) Reorganisation of the stack code. The macros are now all 693968651Skris collected in safestack.h . Each macro is defined in terms of 694068651Skris a "stack macro" of the form SKM_<name>(type, a, b). The 694168651Skris DEBUG_SAFESTACK is now handled in terms of function casts, 694268651Skris this has the advantage of retaining type safety without the 694368651Skris use of additional functions. If DEBUG_SAFESTACK is not defined 694468651Skris then the non typesafe macros are used instead. Also modified the 694568651Skris mkstack.pl script to handle the new form. Needs testing to see 694668651Skris if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 694768651Skris the default if no major problems. Similar behaviour for ASN1_SET_OF 694868651Skris and PKCS12_STACK_OF. 694968651Skris [Steve Henson] 695068651Skris 695168651Skris *) When some versions of IIS use the 'NET' form of private key the 695268651Skris key derivation algorithm is different. Normally MD5(password) is 695368651Skris used as a 128 bit RC4 key. In the modified case 695468651Skris MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 695568651Skris new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 695668651Skris as the old Netscape_RSA functions except they have an additional 695768651Skris 'sgckey' parameter which uses the modified algorithm. Also added 695868651Skris an -sgckey command line option to the rsa utility. Thanks to 695968651Skris Adrian Peck <bertie@ncipher.com> for posting details of the modified 696068651Skris algorithm to openssl-dev. 696168651Skris [Steve Henson] 696268651Skris 696368651Skris *) The evp_local.h macros were using 'c.##kname' which resulted in 696468651Skris invalid expansion on some systems (SCO 5.0.5 for example). 696568651Skris Corrected to 'c.kname'. 696668651Skris [Phillip Porch <root@theporch.com>] 696768651Skris 696868651Skris *) New X509_get1_email() and X509_REQ_get1_email() functions that return 696968651Skris a STACK of email addresses from a certificate or request, these look 697068651Skris in the subject name and the subject alternative name extensions and 697168651Skris omit any duplicate addresses. 697268651Skris [Steve Henson] 697368651Skris 697468651Skris *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 697568651Skris This makes DSA verification about 2 % faster. 697668651Skris [Bodo Moeller] 697768651Skris 697868651Skris *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 697968651Skris (meaning that now 2^5 values will be precomputed, which is only 4 KB 698068651Skris plus overhead for 1024 bit moduli). 698168651Skris This makes exponentiations about 0.5 % faster for 1024 bit 698268651Skris exponents (as measured by "openssl speed rsa2048"). 698368651Skris [Bodo Moeller] 698468651Skris 698568651Skris *) Rename memory handling macros to avoid conflicts with other 698668651Skris software: 698768651Skris Malloc => OPENSSL_malloc 698868651Skris Malloc_locked => OPENSSL_malloc_locked 698968651Skris Realloc => OPENSSL_realloc 699068651Skris Free => OPENSSL_free 699168651Skris [Richard Levitte] 699268651Skris 699368651Skris *) New function BN_mod_exp_mont_word for small bases (roughly 15% 699468651Skris faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 699568651Skris [Bodo Moeller] 699668651Skris 699768651Skris *) CygWin32 support. 699868651Skris [John Jarvie <jjarvie@newsguy.com>] 699968651Skris 700068651Skris *) The type-safe stack code has been rejigged. It is now only compiled 700168651Skris in when OpenSSL is configured with the DEBUG_SAFESTACK option and 700268651Skris by default all type-specific stack functions are "#define"d back to 700368651Skris standard stack functions. This results in more streamlined output 700468651Skris but retains the type-safety checking possibilities of the original 700568651Skris approach. 700668651Skris [Geoff Thorpe] 700768651Skris 700868651Skris *) The STACK code has been cleaned up, and certain type declarations 700968651Skris that didn't make a lot of sense have been brought in line. This has 701068651Skris also involved a cleanup of sorts in safestack.h to more correctly 701168651Skris map type-safe stack functions onto their plain stack counterparts. 701268651Skris This work has also resulted in a variety of "const"ifications of 701368651Skris lots of the code, especially "_cmp" operations which should normally 701468651Skris be prototyped with "const" parameters anyway. 701568651Skris [Geoff Thorpe] 701668651Skris 701768651Skris *) When generating bytes for the first time in md_rand.c, 'stir the pool' 701868651Skris by seeding with STATE_SIZE dummy bytes (with zero entropy count). 701968651Skris (The PRNG state consists of two parts, the large pool 'state' and 'md', 702068651Skris where all of 'md' is used each time the PRNG is used, but 'state' 702168651Skris is used only indexed by a cyclic counter. As entropy may not be 702268651Skris well distributed from the beginning, 'md' is important as a 702368651Skris chaining variable. However, the output function chains only half 702468651Skris of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 702568651Skris all of 'md', and seeding with STATE_SIZE dummy bytes will result 702668651Skris in all of 'state' being rewritten, with the new values depending 702768651Skris on virtually all of 'md'. This overcomes the 80 bit limitation.) 702868651Skris [Bodo Moeller] 702968651Skris 703068651Skris *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 703168651Skris the handshake is continued after ssl_verify_cert_chain(); 703268651Skris otherwise, if SSL_VERIFY_NONE is set, remaining error codes 703368651Skris can lead to 'unexplainable' connection aborts later. 703468651Skris [Bodo Moeller; problem tracked down by Lutz Jaenicke] 703568651Skris 703668651Skris *) Major EVP API cipher revision. 703768651Skris Add hooks for extra EVP features. This allows various cipher 703868651Skris parameters to be set in the EVP interface. Support added for variable 703968651Skris key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 704068651Skris setting of RC2 and RC5 parameters. 704168651Skris 704268651Skris Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 704368651Skris ciphers. 704468651Skris 704568651Skris Remove lots of duplicated code from the EVP library. For example *every* 704668651Skris cipher init() function handles the 'iv' in the same way according to the 704768651Skris cipher mode. They also all do nothing if the 'key' parameter is NULL and 704868651Skris for CFB and OFB modes they zero ctx->num. 704968651Skris 705068651Skris New functionality allows removal of S/MIME code RC2 hack. 705168651Skris 705268651Skris Most of the routines have the same form and so can be declared in terms 705368651Skris of macros. 705468651Skris 705568651Skris By shifting this to the top level EVP_CipherInit() it can be removed from 705668651Skris all individual ciphers. If the cipher wants to handle IVs or keys 705768651Skris differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 705868651Skris flags. 705968651Skris 706068651Skris Change lots of functions like EVP_EncryptUpdate() to now return a 706168651Skris value: although software versions of the algorithms cannot fail 706268651Skris any installed hardware versions can. 706368651Skris [Steve Henson] 706468651Skris 706568651Skris *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 706668651Skris this option is set, tolerate broken clients that send the negotiated 706768651Skris protocol version number instead of the requested protocol version 706868651Skris number. 706968651Skris [Bodo Moeller] 707068651Skris 707168651Skris *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 707268651Skris i.e. non-zero for export ciphersuites, zero otherwise. 707368651Skris Previous versions had this flag inverted, inconsistent with 707468651Skris rsa_tmp_cb (..._TMP_RSA_CB). 707568651Skris [Bodo Moeller; problem reported by Amit Chopra] 707668651Skris 707768651Skris *) Add missing DSA library text string. Work around for some IIS 707868651Skris key files with invalid SEQUENCE encoding. 707968651Skris [Steve Henson] 708068651Skris 708168651Skris *) Add a document (doc/standards.txt) that list all kinds of standards 708268651Skris and so on that are implemented in OpenSSL. 708368651Skris [Richard Levitte] 708468651Skris 708568651Skris *) Enhance c_rehash script. Old version would mishandle certificates 708668651Skris with the same subject name hash and wouldn't handle CRLs at all. 708768651Skris Added -fingerprint option to crl utility, to support new c_rehash 708868651Skris features. 708968651Skris [Steve Henson] 709068651Skris 709168651Skris *) Eliminate non-ANSI declarations in crypto.h and stack.h. 709268651Skris [Ulf M�ller] 709368651Skris 709468651Skris *) Fix for SSL server purpose checking. Server checking was 709568651Skris rejecting certificates which had extended key usage present 709668651Skris but no ssl client purpose. 709768651Skris [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 709868651Skris 709968651Skris *) Make PKCS#12 code work with no password. The PKCS#12 spec 710068651Skris is a little unclear about how a blank password is handled. 710168651Skris Since the password in encoded as a BMPString with terminating 710268651Skris double NULL a zero length password would end up as just the 710368651Skris double NULL. However no password at all is different and is 710468651Skris handled differently in the PKCS#12 key generation code. NS 710568651Skris treats a blank password as zero length. MSIE treats it as no 710668651Skris password on export: but it will try both on import. We now do 710768651Skris the same: PKCS12_parse() tries zero length and no password if 710868651Skris the password is set to "" or NULL (NULL is now a valid password: 710968651Skris it wasn't before) as does the pkcs12 application. 711068651Skris [Steve Henson] 711168651Skris 711268651Skris *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 711368651Skris perror when PEM_read_bio_X509_REQ fails, the error message must 711468651Skris be obtained from the error queue. 711568651Skris [Bodo Moeller] 711668651Skris 711768651Skris *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 711868651Skris it in ERR_remove_state if appropriate, and change ERR_get_state 711968651Skris accordingly to avoid race conditions (this is necessary because 712068651Skris thread_hash is no longer constant once set). 712168651Skris [Bodo Moeller] 712268651Skris 712368651Skris *) Bugfix for linux-elf makefile.one. 712468651Skris [Ulf M�ller] 712568651Skris 712668651Skris *) RSA_get_default_method() will now cause a default 712768651Skris RSA_METHOD to be chosen if one doesn't exist already. 712868651Skris Previously this was only set during a call to RSA_new() 712968651Skris or RSA_new_method(NULL) meaning it was possible for 713068651Skris RSA_get_default_method() to return NULL. 713168651Skris [Geoff Thorpe] 713268651Skris 713368651Skris *) Added native name translation to the existing DSO code 713468651Skris that will convert (if the flag to do so is set) filenames 713568651Skris that are sufficiently small and have no path information 713668651Skris into a canonical native form. Eg. "blah" converted to 713768651Skris "libblah.so" or "blah.dll" etc. 713868651Skris [Geoff Thorpe] 713968651Skris 714068651Skris *) New function ERR_error_string_n(e, buf, len) which is like 714168651Skris ERR_error_string(e, buf), but writes at most 'len' bytes 714268651Skris including the 0 terminator. For ERR_error_string_n, 'buf' 714368651Skris may not be NULL. 714468651Skris [Damien Miller <djm@mindrot.org>, Bodo Moeller] 714568651Skris 714668651Skris *) CONF library reworked to become more general. A new CONF 714768651Skris configuration file reader "class" is implemented as well as a 714868651Skris new functions (NCONF_*, for "New CONF") to handle it. The now 714968651Skris old CONF_* functions are still there, but are reimplemented to 715068651Skris work in terms of the new functions. Also, a set of functions 715168651Skris to handle the internal storage of the configuration data is 715268651Skris provided to make it easier to write new configuration file 715368651Skris reader "classes" (I can definitely see something reading a 715468651Skris configuration file in XML format, for example), called _CONF_*, 715568651Skris or "the configuration storage API"... 715668651Skris 715768651Skris The new configuration file reading functions are: 715868651Skris 715968651Skris NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 716068651Skris NCONF_get_section, NCONF_get_string, NCONF_get_numbre 716168651Skris 716268651Skris NCONF_default, NCONF_WIN32 716368651Skris 716468651Skris NCONF_dump_fp, NCONF_dump_bio 716568651Skris 716668651Skris NCONF_default and NCONF_WIN32 are method (or "class") choosers, 716768651Skris NCONF_new creates a new CONF object. This works in the same way 716868651Skris as other interfaces in OpenSSL, like the BIO interface. 716968651Skris NCONF_dump_* dump the internal storage of the configuration file, 717068651Skris which is useful for debugging. All other functions take the same 717168651Skris arguments as the old CONF_* functions wth the exception of the 717268651Skris first that must be a `CONF *' instead of a `LHASH *'. 717368651Skris 717468651Skris To make it easer to use the new classes with the old CONF_* functions, 717568651Skris the function CONF_set_default_method is provided. 717668651Skris [Richard Levitte] 717768651Skris 717868651Skris *) Add '-tls1' option to 'openssl ciphers', which was already 717968651Skris mentioned in the documentation but had not been implemented. 718068651Skris (This option is not yet really useful because even the additional 718168651Skris experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 718268651Skris [Bodo Moeller] 718368651Skris 718468651Skris *) Initial DSO code added into libcrypto for letting OpenSSL (and 718568651Skris OpenSSL-based applications) load shared libraries and bind to 718668651Skris them in a portable way. 718768651Skris [Geoff Thorpe, with contributions from Richard Levitte] 718868651Skris 718959191Skris Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 719059191Skris 719159191Skris *) Make sure _lrotl and _lrotr are only used with MSVC. 719259191Skris 719359191Skris *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 719459191Skris (the default implementation of RAND_status). 719559191Skris 719659191Skris *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 719759191Skris to '-clrext' (= clear extensions), as intended and documented. 719859191Skris [Bodo Moeller; inconsistency pointed out by Michael Attili 719959191Skris <attili@amaxo.com>] 720059191Skris 720159191Skris *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 720259191Skris was larger than the MD block size. 720359191Skris [Steve Henson, pointed out by Yost William <YostW@tce.com>] 720459191Skris 720559191Skris *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 720659191Skris fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 720759191Skris using the passed key: if the passed key was a private key the result 720859191Skris of X509_print(), for example, would be to print out all the private key 720959191Skris components. 721059191Skris [Steve Henson] 721159191Skris 721259191Skris *) des_quad_cksum() byte order bug fix. 721359191Skris [Ulf M�ller, using the problem description in krb4-0.9.7, where 721459191Skris the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 721559191Skris 721659191Skris *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 721759191Skris discouraged. 721859191Skris [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 721959191Skris 722059191Skris *) For easily testing in shell scripts whether some command 722159191Skris 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 722259191Skris returns with exit code 0 iff no command of the given name is available. 722359191Skris 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 722459191Skris the output goes to stdout and nothing is printed to stderr. 722559191Skris Additional arguments are always ignored. 722659191Skris 722759191Skris Since for each cipher there is a command of the same name, 722859191Skris the 'no-cipher' compilation switches can be tested this way. 722959191Skris 723059191Skris ('openssl no-XXX' is not able to detect pseudo-commands such 723159191Skris as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 723259191Skris [Bodo Moeller] 723359191Skris 723459191Skris *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 723559191Skris [Bodo Moeller] 723659191Skris 723759191Skris *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 723859191Skris is set; it will be thrown away anyway because each handshake creates 723959191Skris its own key. 724059191Skris ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 724159191Skris to parameters -- in previous versions (since OpenSSL 0.9.3) the 724259191Skris 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 724359191Skris you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 724459191Skris [Bodo Moeller] 724559191Skris 724659191Skris *) New s_client option -ign_eof: EOF at stdin is ignored, and 724759191Skris 'Q' and 'R' lose their special meanings (quit/renegotiate). 724859191Skris This is part of what -quiet does; unlike -quiet, -ign_eof 724959191Skris does not suppress any output. 725059191Skris [Richard Levitte] 725159191Skris 725259191Skris *) Add compatibility options to the purpose and trust code. The 725359191Skris purpose X509_PURPOSE_ANY is "any purpose" which automatically 725459191Skris accepts a certificate or CA, this was the previous behaviour, 725559191Skris with all the associated security issues. 725659191Skris 725759191Skris X509_TRUST_COMPAT is the old trust behaviour: only and 725859191Skris automatically trust self signed roots in certificate store. A 725959191Skris new trust setting X509_TRUST_DEFAULT is used to specify that 726059191Skris a purpose has no associated trust setting and it should instead 726159191Skris use the value in the default purpose. 726259191Skris [Steve Henson] 726359191Skris 726459191Skris *) Fix the PKCS#8 DSA private key code so it decodes keys again 726559191Skris and fix a memory leak. 726659191Skris [Steve Henson] 726759191Skris 726859191Skris *) In util/mkerr.pl (which implements 'make errors'), preserve 726959191Skris reason strings from the previous version of the .c file, as 727059191Skris the default to have only downcase letters (and digits) in 727159191Skris automatically generated reasons codes is not always appropriate. 727259191Skris [Bodo Moeller] 727359191Skris 727459191Skris *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 727559191Skris using strerror. Previously, ERR_reason_error_string() returned 727659191Skris library names as reason strings for SYSerr; but SYSerr is a special 727759191Skris case where small numbers are errno values, not library numbers. 727859191Skris [Bodo Moeller] 727959191Skris 728059191Skris *) Add '-dsaparam' option to 'openssl dhparam' application. This 728159191Skris converts DSA parameters into DH parameters. (When creating parameters, 728259191Skris DSA_generate_parameters is used.) 728359191Skris [Bodo Moeller] 728459191Skris 728559191Skris *) Include 'length' (recommended exponent length) in C code generated 728659191Skris by 'openssl dhparam -C'. 728759191Skris [Bodo Moeller] 728859191Skris 728959191Skris *) The second argument to set_label in perlasm was already being used 729059191Skris so couldn't be used as a "file scope" flag. Moved to third argument 729159191Skris which was free. 729259191Skris [Steve Henson] 729359191Skris 729459191Skris *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 729559191Skris instead of RAND_bytes for encryption IVs and salts. 729659191Skris [Bodo Moeller] 729759191Skris 729859191Skris *) Include RAND_status() into RAND_METHOD instead of implementing 729959191Skris it only for md_rand.c Otherwise replacing the PRNG by calling 730059191Skris RAND_set_rand_method would be impossible. 730159191Skris [Bodo Moeller] 730259191Skris 730359191Skris *) Don't let DSA_generate_key() enter an infinite loop if the random 730459191Skris number generation fails. 730559191Skris [Bodo Moeller] 730659191Skris 730759191Skris *) New 'rand' application for creating pseudo-random output. 730859191Skris [Bodo Moeller] 730959191Skris 731059191Skris *) Added configuration support for Linux/IA64 731159191Skris [Rolf Haberrecker <rolf@suse.de>] 731259191Skris 731359191Skris *) Assembler module support for Mingw32. 731459191Skris [Ulf M�ller] 731559191Skris 731659191Skris *) Shared library support for HPUX (in shlib/). 731759191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 731859191Skris 731959191Skris *) Shared library support for Solaris gcc. 732059191Skris [Lutz Behnke <behnke@trustcenter.de>] 732159191Skris 732259191Skris Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 732359191Skris 732459191Skris *) PKCS7_encrypt() was adding text MIME headers twice because they 732559191Skris were added manually and by SMIME_crlf_copy(). 732659191Skris [Steve Henson] 732759191Skris 732859191Skris *) In bntest.c don't call BN_rand with zero bits argument. 732959191Skris [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 733059191Skris 733159191Skris *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 733259191Skris case was implemented. This caused BN_div_recp() to fail occasionally. 733359191Skris [Ulf M�ller] 733459191Skris 733559191Skris *) Add an optional second argument to the set_label() in the perl 733659191Skris assembly language builder. If this argument exists and is set 733759191Skris to 1 it signals that the assembler should use a symbol whose 733859191Skris scope is the entire file, not just the current function. This 733959191Skris is needed with MASM which uses the format label:: for this scope. 734059191Skris [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 734159191Skris 734259191Skris *) Change the ASN1 types so they are typedefs by default. Before 734359191Skris almost all types were #define'd to ASN1_STRING which was causing 734459191Skris STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 734559191Skris for example. 734659191Skris [Steve Henson] 734759191Skris 734859191Skris *) Change names of new functions to the new get1/get0 naming 734959191Skris convention: After 'get1', the caller owns a reference count 735059191Skris and has to call ..._free; 'get0' returns a pointer to some 735159191Skris data structure without incrementing reference counters. 735259191Skris (Some of the existing 'get' functions increment a reference 735359191Skris counter, some don't.) 735459191Skris Similarly, 'set1' and 'add1' functions increase reference 735559191Skris counters or duplicate objects. 735659191Skris [Steve Henson] 735759191Skris 735859191Skris *) Allow for the possibility of temp RSA key generation failure: 735959191Skris the code used to assume it always worked and crashed on failure. 736059191Skris [Steve Henson] 736159191Skris 736259191Skris *) Fix potential buffer overrun problem in BIO_printf(). 736359191Skris [Ulf M�ller, using public domain code by Patrick Powell; problem 736459191Skris pointed out by David Sacerdote <das33@cornell.edu>] 736559191Skris 736659191Skris *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 736759191Skris RAND_egd() and RAND_status(). In the command line application, 736859191Skris the EGD socket can be specified like a seed file using RANDFILE 736959191Skris or -rand. 737059191Skris [Ulf M�ller] 737159191Skris 737259191Skris *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 737359191Skris Some CAs (e.g. Verisign) distribute certificates in this form. 737459191Skris [Steve Henson] 737559191Skris 737659191Skris *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 737759191Skris list to exclude them. This means that no special compilation option 737859191Skris is needed to use anonymous DH: it just needs to be included in the 737959191Skris cipher list. 738059191Skris [Steve Henson] 738159191Skris 738259191Skris *) Change the EVP_MD_CTX_type macro so its meaning consistent with 738359191Skris EVP_MD_type. The old functionality is available in a new macro called 738459191Skris EVP_MD_md(). Change code that uses it and update docs. 738559191Skris [Steve Henson] 738659191Skris 738759191Skris *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 738859191Skris where the 'void *' argument is replaced by a function pointer argument. 738959191Skris Previously 'void *' was abused to point to functions, which works on 739059191Skris many platforms, but is not correct. As these functions are usually 739159191Skris called by macros defined in OpenSSL header files, most source code 739259191Skris should work without changes. 739359191Skris [Richard Levitte] 739459191Skris 739559191Skris *) <openssl/opensslconf.h> (which is created by Configure) now contains 739659191Skris sections with information on -D... compiler switches used for 739759191Skris compiling the library so that applications can see them. To enable 739859191Skris one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 739959191Skris must be defined. E.g., 740059191Skris #define OPENSSL_ALGORITHM_DEFINES 740159191Skris #include <openssl/opensslconf.h> 740259191Skris defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 740359191Skris [Richard Levitte, Ulf and Bodo M�ller] 740459191Skris 740559191Skris *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 740659191Skris record layer. 740759191Skris [Bodo Moeller] 740859191Skris 740959191Skris *) Change the 'other' type in certificate aux info to a STACK_OF 741059191Skris X509_ALGOR. Although not an AlgorithmIdentifier as such it has 741159191Skris the required ASN1 format: arbitrary types determined by an OID. 741259191Skris [Steve Henson] 741359191Skris 741459191Skris *) Add some PEM_write_X509_REQ_NEW() functions and a command line 741559191Skris argument to 'req'. This is not because the function is newer or 741659191Skris better than others it just uses the work 'NEW' in the certificate 741759191Skris request header lines. Some software needs this. 741859191Skris [Steve Henson] 741959191Skris 742059191Skris *) Reorganise password command line arguments: now passwords can be 742159191Skris obtained from various sources. Delete the PEM_cb function and make 742259191Skris it the default behaviour: i.e. if the callback is NULL and the 742359191Skris usrdata argument is not NULL interpret it as a null terminated pass 742459191Skris phrase. If usrdata and the callback are NULL then the pass phrase 742559191Skris is prompted for as usual. 742659191Skris [Steve Henson] 742759191Skris 742859191Skris *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 742959191Skris the support is automatically enabled. The resulting binaries will 743059191Skris autodetect the card and use it if present. 743159191Skris [Ben Laurie and Compaq Inc.] 743259191Skris 743359191Skris *) Work around for Netscape hang bug. This sends certificate request 743459191Skris and server done in one record. Since this is perfectly legal in the 743559191Skris SSL/TLS protocol it isn't a "bug" option and is on by default. See 743659191Skris the bugs/SSLv3 entry for more info. 743759191Skris [Steve Henson] 743859191Skris 743959191Skris *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 744059191Skris [Andy Polyakov] 744159191Skris 744259191Skris *) Add -rand argument to smime and pkcs12 applications and read/write 744359191Skris of seed file. 744459191Skris [Steve Henson] 744559191Skris 744659191Skris *) New 'passwd' tool for crypt(3) and apr1 password hashes. 744759191Skris [Bodo Moeller] 744859191Skris 744959191Skris *) Add command line password options to the remaining applications. 745059191Skris [Steve Henson] 745159191Skris 745259191Skris *) Bug fix for BN_div_recp() for numerators with an even number of 745359191Skris bits. 745459191Skris [Ulf M�ller] 745559191Skris 745659191Skris *) More tests in bntest.c, and changed test_bn output. 745759191Skris [Ulf M�ller] 745859191Skris 745959191Skris *) ./config recognizes MacOS X now. 746059191Skris [Andy Polyakov] 746159191Skris 746259191Skris *) Bug fix for BN_div() when the first words of num and divsor are 746359191Skris equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 746459191Skris [Ulf M�ller] 746559191Skris 746659191Skris *) Add support for various broken PKCS#8 formats, and command line 746759191Skris options to produce them. 746859191Skris [Steve Henson] 746959191Skris 747059191Skris *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 747159191Skris get temporary BIGNUMs from a BN_CTX. 747259191Skris [Ulf M�ller] 747359191Skris 747459191Skris *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 747559191Skris for p == 0. 747659191Skris [Ulf M�ller] 747759191Skris 747859191Skris *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 747959191Skris include a #define from the old name to the new. The original intent 748059191Skris was that statically linked binaries could for example just call 748159191Skris SSLeay_add_all_ciphers() to just add ciphers to the table and not 748259191Skris link with digests. This never worked becayse SSLeay_add_all_digests() 748359191Skris and SSLeay_add_all_ciphers() were in the same source file so calling 748459191Skris one would link with the other. They are now in separate source files. 748559191Skris [Steve Henson] 748659191Skris 748759191Skris *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 748859191Skris [Steve Henson] 748959191Skris 749059191Skris *) Use a less unusual form of the Miller-Rabin primality test (it used 749159191Skris a binary algorithm for exponentiation integrated into the Miller-Rabin 749259191Skris loop, our standard modexp algorithms are faster). 749359191Skris [Bodo Moeller] 749459191Skris 749559191Skris *) Support for the EBCDIC character set completed. 749659191Skris [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 749759191Skris 749859191Skris *) Source code cleanups: use const where appropriate, eliminate casts, 749959191Skris use void * instead of char * in lhash. 750059191Skris [Ulf M�ller] 750159191Skris 750259191Skris *) Bugfix: ssl3_send_server_key_exchange was not restartable 750359191Skris (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 750459191Skris this the server could overwrite ephemeral keys that the client 750559191Skris has already seen). 750659191Skris [Bodo Moeller] 750759191Skris 750859191Skris *) Turn DSA_is_prime into a macro that calls BN_is_prime, 750959191Skris using 50 iterations of the Rabin-Miller test. 751059191Skris 751159191Skris DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 751259191Skris iterations of the Rabin-Miller test as required by the appendix 751359191Skris to FIPS PUB 186[-1]) instead of DSA_is_prime. 751459191Skris As BN_is_prime_fasttest includes trial division, DSA parameter 751559191Skris generation becomes much faster. 751659191Skris 751759191Skris This implies a change for the callback functions in DSA_is_prime 751859191Skris and DSA_generate_parameters: The callback function is called once 751959191Skris for each positive witness in the Rabin-Miller test, not just 752059191Skris occasionally in the inner loop; and the parameters to the 752159191Skris callback function now provide an iteration count for the outer 752259191Skris loop rather than for the current invocation of the inner loop. 752359191Skris DSA_generate_parameters additionally can call the callback 752459191Skris function with an 'iteration count' of -1, meaning that a 752559191Skris candidate has passed the trial division test (when q is generated 752659191Skris from an application-provided seed, trial division is skipped). 752759191Skris [Bodo Moeller] 752859191Skris 752959191Skris *) New function BN_is_prime_fasttest that optionally does trial 753059191Skris division before starting the Rabin-Miller test and has 753159191Skris an additional BN_CTX * argument (whereas BN_is_prime always 753259191Skris has to allocate at least one BN_CTX). 753359191Skris 'callback(1, -1, cb_arg)' is called when a number has passed the 753459191Skris trial division stage. 753559191Skris [Bodo Moeller] 753659191Skris 753759191Skris *) Fix for bug in CRL encoding. The validity dates weren't being handled 753859191Skris as ASN1_TIME. 753959191Skris [Steve Henson] 754059191Skris 754159191Skris *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 754259191Skris [Steve Henson] 754359191Skris 754459191Skris *) New function BN_pseudo_rand(). 754559191Skris [Ulf M�ller] 754659191Skris 754759191Skris *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 754859191Skris bignum version of BN_from_montgomery() with the working code from 754959191Skris SSLeay 0.9.0 (the word based version is faster anyway), and clean up 755059191Skris the comments. 755159191Skris [Ulf M�ller] 755259191Skris 755359191Skris *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 755459191Skris made it impossible to use the same SSL_SESSION data structure in 755559191Skris SSL2 clients in multiple threads. 755659191Skris [Bodo Moeller] 755759191Skris 755859191Skris *) The return value of RAND_load_file() no longer counts bytes obtained 755959191Skris by stat(). RAND_load_file(..., -1) is new and uses the complete file 756059191Skris to seed the PRNG (previously an explicit byte count was required). 756159191Skris [Ulf M�ller, Bodo M�ller] 756259191Skris 756359191Skris *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 756459191Skris used (char *) instead of (void *) and had casts all over the place. 756559191Skris [Steve Henson] 756659191Skris 756759191Skris *) Make BN_generate_prime() return NULL on error if ret!=NULL. 756859191Skris [Ulf M�ller] 756959191Skris 757059191Skris *) Retain source code compatibility for BN_prime_checks macro: 757159191Skris BN_is_prime(..., BN_prime_checks, ...) now uses 757259191Skris BN_prime_checks_for_size to determine the appropriate number of 757359191Skris Rabin-Miller iterations. 757459191Skris [Ulf M�ller] 757559191Skris 757659191Skris *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 757759191Skris DH_CHECK_P_NOT_SAFE_PRIME. 757859191Skris (Check if this is true? OpenPGP calls them "strong".) 757959191Skris [Ulf M�ller] 758059191Skris 758159191Skris *) Merge the functionality of "dh" and "gendh" programs into a new program 758259191Skris "dhparam". The old programs are retained for now but will handle DH keys 758359191Skris (instead of parameters) in future. 758459191Skris [Steve Henson] 758559191Skris 758659191Skris *) Make the ciphers, s_server and s_client programs check the return values 758759191Skris when a new cipher list is set. 758859191Skris [Steve Henson] 758959191Skris 759059191Skris *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 759159191Skris ciphers. Before when the 56bit ciphers were enabled the sorting was 759259191Skris wrong. 759359191Skris 759459191Skris The syntax for the cipher sorting has been extended to support sorting by 759559191Skris cipher-strength (using the strength_bits hard coded in the tables). 759659191Skris The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 759759191Skris 759859191Skris Fix a bug in the cipher-command parser: when supplying a cipher command 759959191Skris string with an "undefined" symbol (neither command nor alphanumeric 760059191Skris [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 760159191Skris an error is flagged. 760259191Skris 760359191Skris Due to the strength-sorting extension, the code of the 760459191Skris ssl_create_cipher_list() function was completely rearranged. I hope that 760559191Skris the readability was also increased :-) 760659191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 760759191Skris 760859191Skris *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 760959191Skris for the first serial number and places 2 in the serial number file. This 761059191Skris avoids problems when the root CA is created with serial number zero and 761159191Skris the first user certificate has the same issuer name and serial number 761259191Skris as the root CA. 761359191Skris [Steve Henson] 761459191Skris 761559191Skris *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 761659191Skris the new code. Add documentation for this stuff. 761759191Skris [Steve Henson] 761859191Skris 761959191Skris *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 762059191Skris X509_*() to X509at_*() on the grounds that they don't handle X509 762159191Skris structures and behave in an analagous way to the X509v3 functions: 762259191Skris they shouldn't be called directly but wrapper functions should be used 762359191Skris instead. 762459191Skris 762559191Skris So we also now have some wrapper functions that call the X509at functions 762659191Skris when passed certificate requests. (TO DO: similar things can be done with 762759191Skris PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 762859191Skris things. Some of these need some d2i or i2d and print functionality 762959191Skris because they handle more complex structures.) 763059191Skris [Steve Henson] 763159191Skris 763259191Skris *) Add missing #ifndefs that caused missing symbols when building libssl 763359191Skris as a shared library without RSA. Use #ifndef NO_SSL2 instead of 763459191Skris NO_RSA in ssl/s2*.c. 763559191Skris [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 763659191Skris 763759191Skris *) Precautions against using the PRNG uninitialized: RAND_bytes() now 763859191Skris has a return value which indicates the quality of the random data 763959191Skris (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 764059191Skris error queue. New function RAND_pseudo_bytes() generates output that is 764159191Skris guaranteed to be unique but not unpredictable. RAND_add is like 764259191Skris RAND_seed, but takes an extra argument for an entropy estimate 764359191Skris (RAND_seed always assumes full entropy). 764459191Skris [Ulf M�ller] 764559191Skris 764659191Skris *) Do more iterations of Rabin-Miller probable prime test (specifically, 764759191Skris 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 764859191Skris instead of only 2 for all lengths; see BN_prime_checks_for_size definition 764959191Skris in crypto/bn/bn_prime.c for the complete table). This guarantees a 765059191Skris false-positive rate of at most 2^-80 for random input. 765159191Skris [Bodo Moeller] 765259191Skris 765359191Skris *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 765459191Skris [Bodo Moeller] 765559191Skris 765659191Skris *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 765759191Skris in the 0.9.5 release), this returns the chain 765859191Skris from an X509_CTX structure with a dup of the stack and all 765959191Skris the X509 reference counts upped: so the stack will exist 766059191Skris after X509_CTX_cleanup() has been called. Modify pkcs12.c 766159191Skris to use this. 766259191Skris 766359191Skris Also make SSL_SESSION_print() print out the verify return 766459191Skris code. 766559191Skris [Steve Henson] 766659191Skris 766759191Skris *) Add manpage for the pkcs12 command. Also change the default 766859191Skris behaviour so MAC iteration counts are used unless the new 766959191Skris -nomaciter option is used. This improves file security and 767059191Skris only older versions of MSIE (4.0 for example) need it. 767159191Skris [Steve Henson] 767259191Skris 767359191Skris *) Honor the no-xxx Configure options when creating .DEF files. 767459191Skris [Ulf M�ller] 767559191Skris 767659191Skris *) Add PKCS#10 attributes to field table: challengePassword, 767759191Skris unstructuredName and unstructuredAddress. These are taken from 767859191Skris draft PKCS#9 v2.0 but are compatible with v1.2 provided no 767959191Skris international characters are used. 768059191Skris 768159191Skris More changes to X509_ATTRIBUTE code: allow the setting of types 768259191Skris based on strings. Remove the 'loc' parameter when adding 768359191Skris attributes because these will be a SET OF encoding which is sorted 768459191Skris in ASN1 order. 768559191Skris [Steve Henson] 768659191Skris 768759191Skris *) Initial changes to the 'req' utility to allow request generation 768859191Skris automation. This will allow an application to just generate a template 768959191Skris file containing all the field values and have req construct the 769059191Skris request. 769159191Skris 769259191Skris Initial support for X509_ATTRIBUTE handling. Stacks of these are 769359191Skris used all over the place including certificate requests and PKCS#7 769459191Skris structures. They are currently handled manually where necessary with 769559191Skris some primitive wrappers for PKCS#7. The new functions behave in a 769659191Skris manner analogous to the X509 extension functions: they allow 769759191Skris attributes to be looked up by NID and added. 769859191Skris 769959191Skris Later something similar to the X509V3 code would be desirable to 770059191Skris automatically handle the encoding, decoding and printing of the 770159191Skris more complex types. The string types like challengePassword can 770259191Skris be handled by the string table functions. 770359191Skris 770459191Skris Also modified the multi byte string table handling. Now there is 770559191Skris a 'global mask' which masks out certain types. The table itself 770659191Skris can use the flag STABLE_NO_MASK to ignore the mask setting: this 770759191Skris is useful when for example there is only one permissible type 770859191Skris (as in countryName) and using the mask might result in no valid 770959191Skris types at all. 771059191Skris [Steve Henson] 771159191Skris 771259191Skris *) Clean up 'Finished' handling, and add functions SSL_get_finished and 771359191Skris SSL_get_peer_finished to allow applications to obtain the latest 771459191Skris Finished messages sent to the peer or expected from the peer, 771559191Skris respectively. (SSL_get_peer_finished is usually the Finished message 771659191Skris actually received from the peer, otherwise the protocol will be aborted.) 771759191Skris 771859191Skris As the Finished message are message digests of the complete handshake 771959191Skris (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 772059191Skris be used for external authentication procedures when the authentication 772159191Skris provided by SSL/TLS is not desired or is not enough. 772259191Skris [Bodo Moeller] 772359191Skris 772459191Skris *) Enhanced support for Alpha Linux is added. Now ./config checks if 772559191Skris the host supports BWX extension and if Compaq C is present on the 772659191Skris $PATH. Just exploiting of the BWX extension results in 20-30% 772759191Skris performance kick for some algorithms, e.g. DES and RC4 to mention 772859191Skris a couple. Compaq C in turn generates ~20% faster code for MD5 and 772959191Skris SHA1. 773059191Skris [Andy Polyakov] 773159191Skris 773259191Skris *) Add support for MS "fast SGC". This is arguably a violation of the 773359191Skris SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 773459191Skris weak crypto and after checking the certificate is SGC a second one 773559191Skris with strong crypto. MS SGC stops the first handshake after receiving 773659191Skris the server certificate message and sends a second client hello. Since 773759191Skris a server will typically do all the time consuming operations before 773859191Skris expecting any further messages from the client (server key exchange 773959191Skris is the most expensive) there is little difference between the two. 774059191Skris 774159191Skris To get OpenSSL to support MS SGC we have to permit a second client 774259191Skris hello message after we have sent server done. In addition we have to 774359191Skris reset the MAC if we do get this second client hello. 774459191Skris [Steve Henson] 774559191Skris 774659191Skris *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 774759191Skris if a DER encoded private key is RSA or DSA traditional format. Changed 774859191Skris d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 774959191Skris format DER encoded private key. Newer code should use PKCS#8 format which 775059191Skris has the key type encoded in the ASN1 structure. Added DER private key 775159191Skris support to pkcs8 application. 775259191Skris [Steve Henson] 775359191Skris 775459191Skris *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 775559191Skris ciphersuites has been selected (as required by the SSL 3/TLS 1 775659191Skris specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 775759191Skris is set, we interpret this as a request to violate the specification 775859191Skris (the worst that can happen is a handshake failure, and 'correct' 775959191Skris behaviour would result in a handshake failure anyway). 776059191Skris [Bodo Moeller] 776159191Skris 776259191Skris *) In SSL_CTX_add_session, take into account that there might be multiple 776359191Skris SSL_SESSION structures with the same session ID (e.g. when two threads 776459191Skris concurrently obtain them from an external cache). 776559191Skris The internal cache can handle only one SSL_SESSION with a given ID, 776659191Skris so if there's a conflict, we now throw out the old one to achieve 776759191Skris consistency. 776859191Skris [Bodo Moeller] 776959191Skris 777059191Skris *) Add OIDs for idea and blowfish in CBC mode. This will allow both 777159191Skris to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 777259191Skris some routines that use cipher OIDs: some ciphers do not have OIDs 777359191Skris defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 777459191Skris example. 777559191Skris [Steve Henson] 777659191Skris 777759191Skris *) Simplify the trust setting structure and code. Now we just have 777859191Skris two sequences of OIDs for trusted and rejected settings. These will 777959191Skris typically have values the same as the extended key usage extension 778059191Skris and any application specific purposes. 778159191Skris 778259191Skris The trust checking code now has a default behaviour: it will just 778359191Skris check for an object with the same NID as the passed id. Functions can 778459191Skris be provided to override either the default behaviour or the behaviour 778559191Skris for a given id. SSL client, server and email already have functions 778659191Skris in place for compatibility: they check the NID and also return "trusted" 778759191Skris if the certificate is self signed. 778859191Skris [Steve Henson] 778959191Skris 779059191Skris *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 779159191Skris traditional format into an EVP_PKEY structure. 779259191Skris [Steve Henson] 779359191Skris 779459191Skris *) Add a password callback function PEM_cb() which either prompts for 779559191Skris a password if usr_data is NULL or otherwise assumes it is a null 779659191Skris terminated password. Allow passwords to be passed on command line 779759191Skris environment or config files in a few more utilities. 779859191Skris [Steve Henson] 779959191Skris 780059191Skris *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 780159191Skris keys. Add some short names for PKCS#8 PBE algorithms and allow them 780259191Skris to be specified on the command line for the pkcs8 and pkcs12 utilities. 780359191Skris Update documentation. 780459191Skris [Steve Henson] 780559191Skris 780659191Skris *) Support for ASN1 "NULL" type. This could be handled before by using 780759191Skris ASN1_TYPE but there wasn't any function that would try to read a NULL 780859191Skris and produce an error if it couldn't. For compatibility we also have 780959191Skris ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 781059191Skris don't allocate anything because they don't need to. 781159191Skris [Steve Henson] 781259191Skris 781359191Skris *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 781459191Skris for details. 781559191Skris [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 781659191Skris 781759191Skris *) Rebuild of the memory allocation routines used by OpenSSL code and 781859191Skris possibly others as well. The purpose is to make an interface that 781959191Skris provide hooks so anyone can build a separate set of allocation and 782059191Skris deallocation routines to be used by OpenSSL, for example memory 782159191Skris pool implementations, or something else, which was previously hard 782259191Skris since Malloc(), Realloc() and Free() were defined as macros having 782359191Skris the values malloc, realloc and free, respectively (except for Win32 782459191Skris compilations). The same is provided for memory debugging code. 782559191Skris OpenSSL already comes with functionality to find memory leaks, but 782659191Skris this gives people a chance to debug other memory problems. 782759191Skris 782859191Skris With these changes, a new set of functions and macros have appeared: 782959191Skris 783068651Skris CRYPTO_set_mem_debug_functions() [F] 783168651Skris CRYPTO_get_mem_debug_functions() [F] 783268651Skris CRYPTO_dbg_set_options() [F] 783368651Skris CRYPTO_dbg_get_options() [F] 783468651Skris CRYPTO_malloc_debug_init() [M] 783559191Skris 783659191Skris The memory debug functions are NULL by default, unless the library 783759191Skris is compiled with CRYPTO_MDEBUG or friends is defined. If someone 783859191Skris wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 783959191Skris gives the standard debugging functions that come with OpenSSL) or 784059191Skris CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 784159191Skris provided by the library user) must be used. When the standard 784259191Skris debugging functions are used, CRYPTO_dbg_set_options can be used to 784359191Skris request additional information: 784459191Skris CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 784559191Skris the CRYPTO_MDEBUG_xxx macro when compiling the library. 784659191Skris 784759191Skris Also, things like CRYPTO_set_mem_functions will always give the 784859191Skris expected result (the new set of functions is used for allocation 784959191Skris and deallocation) at all times, regardless of platform and compiler 785059191Skris options. 785159191Skris 785259191Skris To finish it up, some functions that were never use in any other 785359191Skris way than through macros have a new API and new semantic: 785459191Skris 785559191Skris CRYPTO_dbg_malloc() 785659191Skris CRYPTO_dbg_realloc() 785759191Skris CRYPTO_dbg_free() 785859191Skris 785959191Skris All macros of value have retained their old syntax. 786059191Skris [Richard Levitte and Bodo Moeller] 786159191Skris 786259191Skris *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 786359191Skris ordering of SMIMECapabilities wasn't in "strength order" and there 786459191Skris was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 786559191Skris algorithm. 786659191Skris [Steve Henson] 786759191Skris 786859191Skris *) Some ASN1 types with illegal zero length encoding (INTEGER, 786959191Skris ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 787059191Skris [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 787159191Skris 787259191Skris *) Merge in my S/MIME library for OpenSSL. This provides a simple 787359191Skris S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 787459191Skris functionality to handle multipart/signed properly) and a utility 787559191Skris called 'smime' to call all this stuff. This is based on code I 787659191Skris originally wrote for Celo who have kindly allowed it to be 787759191Skris included in OpenSSL. 787859191Skris [Steve Henson] 787959191Skris 788059191Skris *) Add variants des_set_key_checked and des_set_key_unchecked of 788159191Skris des_set_key (aka des_key_sched). Global variable des_check_key 788259191Skris decides which of these is called by des_set_key; this way 788359191Skris des_check_key behaves as it always did, but applications and 788459191Skris the library itself, which was buggy for des_check_key == 1, 788559191Skris have a cleaner way to pick the version they need. 788659191Skris [Bodo Moeller] 788759191Skris 788859191Skris *) New function PKCS12_newpass() which changes the password of a 788959191Skris PKCS12 structure. 789059191Skris [Steve Henson] 789159191Skris 789259191Skris *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 789359191Skris dynamic mix. In both cases the ids can be used as an index into the 789459191Skris table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 789559191Skris functions so they accept a list of the field values and the 789659191Skris application doesn't need to directly manipulate the X509_TRUST 789759191Skris structure. 789859191Skris [Steve Henson] 789959191Skris 790059191Skris *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 790159191Skris need initialising. 790259191Skris [Steve Henson] 790359191Skris 790459191Skris *) Modify the way the V3 extension code looks up extensions. This now 790559191Skris works in a similar way to the object code: we have some "standard" 790659191Skris extensions in a static table which is searched with OBJ_bsearch() 790759191Skris and the application can add dynamic ones if needed. The file 790859191Skris crypto/x509v3/ext_dat.h now has the info: this file needs to be 790959191Skris updated whenever a new extension is added to the core code and kept 791059191Skris in ext_nid order. There is a simple program 'tabtest.c' which checks 791159191Skris this. New extensions are not added too often so this file can readily 791259191Skris be maintained manually. 791359191Skris 791459191Skris There are two big advantages in doing things this way. The extensions 791559191Skris can be looked up immediately and no longer need to be "added" using 791659191Skris X509V3_add_standard_extensions(): this function now does nothing. 791759191Skris [Side note: I get *lots* of email saying the extension code doesn't 791859191Skris work because people forget to call this function] 791959191Skris Also no dynamic allocation is done unless new extensions are added: 792059191Skris so if we don't add custom extensions there is no need to call 792159191Skris X509V3_EXT_cleanup(). 792259191Skris [Steve Henson] 792359191Skris 792459191Skris *) Modify enc utility's salting as follows: make salting the default. Add a 792559191Skris magic header, so unsalted files fail gracefully instead of just decrypting 792659191Skris to garbage. This is because not salting is a big security hole, so people 792759191Skris should be discouraged from doing it. 792859191Skris [Ben Laurie] 792959191Skris 793059191Skris *) Fixes and enhancements to the 'x509' utility. It allowed a message 793159191Skris digest to be passed on the command line but it only used this 793259191Skris parameter when signing a certificate. Modified so all relevant 793359191Skris operations are affected by the digest parameter including the 793459191Skris -fingerprint and -x509toreq options. Also -x509toreq choked if a 793559191Skris DSA key was used because it didn't fix the digest. 793659191Skris [Steve Henson] 793759191Skris 793859191Skris *) Initial certificate chain verify code. Currently tests the untrusted 793959191Skris certificates for consistency with the verify purpose (which is set 794059191Skris when the X509_STORE_CTX structure is set up) and checks the pathlength. 794159191Skris 794259191Skris There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 794359191Skris this is because it will reject chains with invalid extensions whereas 794459191Skris every previous version of OpenSSL and SSLeay made no checks at all. 794559191Skris 794659191Skris Trust code: checks the root CA for the relevant trust settings. Trust 794759191Skris settings have an initial value consistent with the verify purpose: e.g. 794859191Skris if the verify purpose is for SSL client use it expects the CA to be 794959191Skris trusted for SSL client use. However the default value can be changed to 795059191Skris permit custom trust settings: one example of this would be to only trust 795159191Skris certificates from a specific "secure" set of CAs. 795259191Skris 795359191Skris Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 795459191Skris which should be used for version portability: especially since the 795559191Skris verify structure is likely to change more often now. 795659191Skris 795759191Skris SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 795859191Skris to set them. If not set then assume SSL clients will verify SSL servers 795959191Skris and vice versa. 796059191Skris 796159191Skris Two new options to the verify program: -untrusted allows a set of 796259191Skris untrusted certificates to be passed in and -purpose which sets the 796359191Skris intended purpose of the certificate. If a purpose is set then the 796459191Skris new chain verify code is used to check extension consistency. 796559191Skris [Steve Henson] 796659191Skris 796759191Skris *) Support for the authority information access extension. 796859191Skris [Steve Henson] 796959191Skris 797059191Skris *) Modify RSA and DSA PEM read routines to transparently handle 797159191Skris PKCS#8 format private keys. New *_PUBKEY_* functions that handle 797259191Skris public keys in a format compatible with certificate 797359191Skris SubjectPublicKeyInfo structures. Unfortunately there were already 797459191Skris functions called *_PublicKey_* which used various odd formats so 797559191Skris these are retained for compatibility: however the DSA variants were 797659191Skris never in a public release so they have been deleted. Changed dsa/rsa 797759191Skris utilities to handle the new format: note no releases ever handled public 797859191Skris keys so we should be OK. 797959191Skris 798059191Skris The primary motivation for this change is to avoid the same fiasco 798159191Skris that dogs private keys: there are several incompatible private key 798259191Skris formats some of which are standard and some OpenSSL specific and 798359191Skris require various evil hacks to allow partial transparent handling and 798459191Skris even then it doesn't work with DER formats. Given the option anything 798559191Skris other than PKCS#8 should be dumped: but the other formats have to 798659191Skris stay in the name of compatibility. 798759191Skris 798859191Skris With public keys and the benefit of hindsight one standard format 798959191Skris is used which works with EVP_PKEY, RSA or DSA structures: though 799059191Skris it clearly returns an error if you try to read the wrong kind of key. 799159191Skris 799259191Skris Added a -pubkey option to the 'x509' utility to output the public key. 799359191Skris Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 799459191Skris (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 799559191Skris EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 799659191Skris that do the same as the EVP_PKEY_assign_*() except they up the 799759191Skris reference count of the added key (they don't "swallow" the 799859191Skris supplied key). 799959191Skris [Steve Henson] 800059191Skris 800159191Skris *) Fixes to crypto/x509/by_file.c the code to read in certificates and 800259191Skris CRLs would fail if the file contained no certificates or no CRLs: 800359191Skris added a new function to read in both types and return the number 800459191Skris read: this means that if none are read it will be an error. The 800559191Skris DER versions of the certificate and CRL reader would always fail 800659191Skris because it isn't possible to mix certificates and CRLs in DER format 800759191Skris without choking one or the other routine. Changed this to just read 800859191Skris a certificate: this is the best we can do. Also modified the code 800959191Skris in apps/verify.c to take notice of return codes: it was previously 801059191Skris attempting to read in certificates from NULL pointers and ignoring 801159191Skris any errors: this is one reason why the cert and CRL reader seemed 801259191Skris to work. It doesn't check return codes from the default certificate 801359191Skris routines: these may well fail if the certificates aren't installed. 801459191Skris [Steve Henson] 801559191Skris 801659191Skris *) Code to support otherName option in GeneralName. 801759191Skris [Steve Henson] 801859191Skris 801959191Skris *) First update to verify code. Change the verify utility 802059191Skris so it warns if it is passed a self signed certificate: 802159191Skris for consistency with the normal behaviour. X509_verify 802259191Skris has been modified to it will now verify a self signed 802359191Skris certificate if *exactly* the same certificate appears 802459191Skris in the store: it was previously impossible to trust a 802559191Skris single self signed certificate. This means that: 802659191Skris openssl verify ss.pem 802759191Skris now gives a warning about a self signed certificate but 802859191Skris openssl verify -CAfile ss.pem ss.pem 802959191Skris is OK. 803059191Skris [Steve Henson] 803159191Skris 803259191Skris *) For servers, store verify_result in SSL_SESSION data structure 803359191Skris (and add it to external session representation). 803459191Skris This is needed when client certificate verifications fails, 803559191Skris but an application-provided verification callback (set by 803659191Skris SSL_CTX_set_cert_verify_callback) allows accepting the session 803759191Skris anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 803859191Skris but returns 1): When the session is reused, we have to set 803959191Skris ssl->verify_result to the appropriate error code to avoid 804059191Skris security holes. 804159191Skris [Bodo Moeller, problem pointed out by Lutz Jaenicke] 804259191Skris 804359191Skris *) Fix a bug in the new PKCS#7 code: it didn't consider the 804459191Skris case in PKCS7_dataInit() where the signed PKCS7 structure 804559191Skris didn't contain any existing data because it was being created. 804659191Skris [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 804759191Skris 804859191Skris *) Add a salt to the key derivation routines in enc.c. This 804959191Skris forms the first 8 bytes of the encrypted file. Also add a 805059191Skris -S option to allow a salt to be input on the command line. 805159191Skris [Steve Henson] 805259191Skris 805359191Skris *) New function X509_cmp(). Oddly enough there wasn't a function 805459191Skris to compare two certificates. We do this by working out the SHA1 805559191Skris hash and comparing that. X509_cmp() will be needed by the trust 805659191Skris code. 805759191Skris [Steve Henson] 805859191Skris 805959191Skris *) SSL_get1_session() is like SSL_get_session(), but increments 806059191Skris the reference count in the SSL_SESSION returned. 806159191Skris [Geoff Thorpe <geoff@eu.c2.net>] 806259191Skris 806359191Skris *) Fix for 'req': it was adding a null to request attributes. 806459191Skris Also change the X509_LOOKUP and X509_INFO code to handle 806559191Skris certificate auxiliary information. 806659191Skris [Steve Henson] 806759191Skris 806859191Skris *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 806959191Skris the 'enc' command. 807059191Skris [Steve Henson] 807159191Skris 807259191Skris *) Add the possibility to add extra information to the memory leak 807359191Skris detecting output, to form tracebacks, showing from where each 807459191Skris allocation was originated: CRYPTO_push_info("constant string") adds 807559191Skris the string plus current file name and line number to a per-thread 807659191Skris stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 807759191Skris is like calling CYRPTO_pop_info() until the stack is empty. 807859191Skris Also updated memory leak detection code to be multi-thread-safe. 807959191Skris [Richard Levitte] 808059191Skris 808159191Skris *) Add options -text and -noout to pkcs7 utility and delete the 808259191Skris encryption options which never did anything. Update docs. 808359191Skris [Steve Henson] 808459191Skris 808559191Skris *) Add options to some of the utilities to allow the pass phrase 808659191Skris to be included on either the command line (not recommended on 808759191Skris OSes like Unix) or read from the environment. Update the 808859191Skris manpages and fix a few bugs. 808959191Skris [Steve Henson] 809059191Skris 809159191Skris *) Add a few manpages for some of the openssl commands. 809259191Skris [Steve Henson] 809359191Skris 809459191Skris *) Fix the -revoke option in ca. It was freeing up memory twice, 809559191Skris leaking and not finding already revoked certificates. 809659191Skris [Steve Henson] 809759191Skris 809859191Skris *) Extensive changes to support certificate auxiliary information. 809959191Skris This involves the use of X509_CERT_AUX structure and X509_AUX 810059191Skris functions. An X509_AUX function such as PEM_read_X509_AUX() 810159191Skris can still read in a certificate file in the usual way but it 810259191Skris will also read in any additional "auxiliary information". By 810359191Skris doing things this way a fair degree of compatibility can be 810459191Skris retained: existing certificates can have this information added 810559191Skris using the new 'x509' options. 810659191Skris 810759191Skris Current auxiliary information includes an "alias" and some trust 810859191Skris settings. The trust settings will ultimately be used in enhanced 810959191Skris certificate chain verification routines: currently a certificate 811059191Skris can only be trusted if it is self signed and then it is trusted 811159191Skris for all purposes. 811259191Skris [Steve Henson] 811359191Skris 811459191Skris *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 811559191Skris The problem was that one of the replacement routines had not been working 811659191Skris since SSLeay releases. For now the offending routine has been replaced 811759191Skris with non-optimised assembler. Even so, this now gives around 95% 811859191Skris performance improvement for 1024 bit RSA signs. 811959191Skris [Mark Cox] 812059191Skris 812159191Skris *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 812259191Skris handling. Most clients have the effective key size in bits equal to 812359191Skris the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 812459191Skris A few however don't do this and instead use the size of the decrypted key 812559191Skris to determine the RC2 key length and the AlgorithmIdentifier to determine 812659191Skris the effective key length. In this case the effective key length can still 812759191Skris be 40 bits but the key length can be 168 bits for example. This is fixed 812859191Skris by manually forcing an RC2 key into the EVP_PKEY structure because the 812959191Skris EVP code can't currently handle unusual RC2 key sizes: it always assumes 813059191Skris the key length and effective key length are equal. 813159191Skris [Steve Henson] 813259191Skris 813359191Skris *) Add a bunch of functions that should simplify the creation of 813459191Skris X509_NAME structures. Now you should be able to do: 813559191Skris X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 813659191Skris and have it automatically work out the correct field type and fill in 813759191Skris the structures. The more adventurous can try: 813859191Skris X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 813959191Skris and it will (hopefully) work out the correct multibyte encoding. 814059191Skris [Steve Henson] 814159191Skris 814259191Skris *) Change the 'req' utility to use the new field handling and multibyte 814359191Skris copy routines. Before the DN field creation was handled in an ad hoc 814459191Skris way in req, ca, and x509 which was rather broken and didn't support 814559191Skris BMPStrings or UTF8Strings. Since some software doesn't implement 814659191Skris BMPStrings or UTF8Strings yet, they can be enabled using the config file 814759191Skris using the dirstring_type option. See the new comment in the default 814859191Skris openssl.cnf for more info. 814959191Skris [Steve Henson] 815059191Skris 815159191Skris *) Make crypto/rand/md_rand.c more robust: 815259191Skris - Assure unique random numbers after fork(). 815359191Skris - Make sure that concurrent threads access the global counter and 815459191Skris md serializably so that we never lose entropy in them 815559191Skris or use exactly the same state in multiple threads. 815659191Skris Access to the large state is not always serializable because 815759191Skris the additional locking could be a performance killer, and 815859191Skris md should be large enough anyway. 815959191Skris [Bodo Moeller] 816059191Skris 816159191Skris *) New file apps/app_rand.c with commonly needed functionality 816259191Skris for handling the random seed file. 816359191Skris 816459191Skris Use the random seed file in some applications that previously did not: 816559191Skris ca, 816659191Skris dsaparam -genkey (which also ignored its '-rand' option), 816759191Skris s_client, 816859191Skris s_server, 816959191Skris x509 (when signing). 817059191Skris Except on systems with /dev/urandom, it is crucial to have a random 817159191Skris seed file at least for key creation, DSA signing, and for DH exchanges; 817259191Skris for RSA signatures we could do without one. 817359191Skris 817459191Skris gendh and gendsa (unlike genrsa) used to read only the first byte 817559191Skris of each file listed in the '-rand' option. The function as previously 817659191Skris found in genrsa is now in app_rand.c and is used by all programs 817759191Skris that support '-rand'. 817859191Skris [Bodo Moeller] 817959191Skris 818059191Skris *) In RAND_write_file, use mode 0600 for creating files; 818159191Skris don't just chmod when it may be too late. 818259191Skris [Bodo Moeller] 818359191Skris 818459191Skris *) Report an error from X509_STORE_load_locations 818559191Skris when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 818659191Skris [Bill Perry] 818759191Skris 818859191Skris *) New function ASN1_mbstring_copy() this copies a string in either 818959191Skris ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 819059191Skris into an ASN1_STRING type. A mask of permissible types is passed 819159191Skris and it chooses the "minimal" type to use or an error if not type 819259191Skris is suitable. 819359191Skris [Steve Henson] 819459191Skris 819559191Skris *) Add function equivalents to the various macros in asn1.h. The old 819659191Skris macros are retained with an M_ prefix. Code inside the library can 819759191Skris use the M_ macros. External code (including the openssl utility) 819859191Skris should *NOT* in order to be "shared library friendly". 819959191Skris [Steve Henson] 820059191Skris 820159191Skris *) Add various functions that can check a certificate's extensions 820259191Skris to see if it usable for various purposes such as SSL client, 820359191Skris server or S/MIME and CAs of these types. This is currently 820459191Skris VERY EXPERIMENTAL but will ultimately be used for certificate chain 820559191Skris verification. Also added a -purpose flag to x509 utility to 820659191Skris print out all the purposes. 820759191Skris [Steve Henson] 820859191Skris 820959191Skris *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 821059191Skris functions. 821159191Skris [Steve Henson] 821259191Skris 821359191Skris *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 821459191Skris for, obtain and decode and extension and obtain its critical flag. 821559191Skris This allows all the necessary extension code to be handled in a 821659191Skris single function call. 821759191Skris [Steve Henson] 821859191Skris 821959191Skris *) RC4 tune-up featuring 30-40% performance improvement on most RISC 822059191Skris platforms. See crypto/rc4/rc4_enc.c for further details. 822159191Skris [Andy Polyakov] 822259191Skris 822359191Skris *) New -noout option to asn1parse. This causes no output to be produced 822459191Skris its main use is when combined with -strparse and -out to extract data 822559191Skris from a file (which may not be in ASN.1 format). 822659191Skris [Steve Henson] 822759191Skris 822859191Skris *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 822959191Skris when producing the local key id. 823059191Skris [Richard Levitte <levitte@stacken.kth.se>] 823159191Skris 823259191Skris *) New option -dhparam in s_server. This allows a DH parameter file to be 823359191Skris stated explicitly. If it is not stated then it tries the first server 823459191Skris certificate file. The previous behaviour hard coded the filename 823559191Skris "server.pem". 823659191Skris [Steve Henson] 823759191Skris 823859191Skris *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 823959191Skris a public key to be input or output. For example: 824059191Skris openssl rsa -in key.pem -pubout -out pubkey.pem 824159191Skris Also added necessary DSA public key functions to handle this. 824259191Skris [Steve Henson] 824359191Skris 824459191Skris *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 824559191Skris in the message. This was handled by allowing 824659191Skris X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 824759191Skris [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 824859191Skris 824959191Skris *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 825059191Skris to the end of the strings whereas this didn't. This would cause problems 825159191Skris if strings read with d2i_ASN1_bytes() were later modified. 825259191Skris [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 825359191Skris 825459191Skris *) Fix for base64 decode bug. When a base64 bio reads only one line of 825559191Skris data and it contains EOF it will end up returning an error. This is 825659191Skris caused by input 46 bytes long. The cause is due to the way base64 825759191Skris BIOs find the start of base64 encoded data. They do this by trying a 825859191Skris trial decode on each line until they find one that works. When they 825959191Skris do a flag is set and it starts again knowing it can pass all the 826059191Skris data directly through the decoder. Unfortunately it doesn't reset 826159191Skris the context it uses. This means that if EOF is reached an attempt 826259191Skris is made to pass two EOFs through the context and this causes the 826359191Skris resulting error. This can also cause other problems as well. As is 826459191Skris usual with these problems it takes *ages* to find and the fix is 826559191Skris trivial: move one line. 826659191Skris [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 826759191Skris 826859191Skris *) Ugly workaround to get s_client and s_server working under Windows. The 826959191Skris old code wouldn't work because it needed to select() on sockets and the 827059191Skris tty (for keypresses and to see if data could be written). Win32 only 827159191Skris supports select() on sockets so we select() with a 1s timeout on the 827259191Skris sockets and then see if any characters are waiting to be read, if none 827359191Skris are present then we retry, we also assume we can always write data to 827459191Skris the tty. This isn't nice because the code then blocks until we've 827559191Skris received a complete line of data and it is effectively polling the 827659191Skris keyboard at 1s intervals: however it's quite a bit better than not 827759191Skris working at all :-) A dedicated Windows application might handle this 827859191Skris with an event loop for example. 827959191Skris [Steve Henson] 828059191Skris 828159191Skris *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 828259191Skris and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 828359191Skris will be called when RSA_sign() and RSA_verify() are used. This is useful 828459191Skris if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 828559191Skris For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 828659191Skris should *not* be used: RSA_sign() and RSA_verify() must be used instead. 828759191Skris This necessitated the support of an extra signature type NID_md5_sha1 828859191Skris for SSL signatures and modifications to the SSL library to use it instead 828959191Skris of calling RSA_public_decrypt() and RSA_private_encrypt(). 829059191Skris [Steve Henson] 829159191Skris 829259191Skris *) Add new -verify -CAfile and -CApath options to the crl program, these 829359191Skris will lookup a CRL issuers certificate and verify the signature in a 829459191Skris similar way to the verify program. Tidy up the crl program so it 829559191Skris no longer accesses structures directly. Make the ASN1 CRL parsing a bit 829659191Skris less strict. It will now permit CRL extensions even if it is not 829759191Skris a V2 CRL: this will allow it to tolerate some broken CRLs. 829859191Skris [Steve Henson] 829959191Skris 830059191Skris *) Initialize all non-automatic variables each time one of the openssl 830159191Skris sub-programs is started (this is necessary as they may be started 830259191Skris multiple times from the "OpenSSL>" prompt). 830359191Skris [Lennart Bang, Bodo Moeller] 830459191Skris 830559191Skris *) Preliminary compilation option RSA_NULL which disables RSA crypto without 830659191Skris removing all other RSA functionality (this is what NO_RSA does). This 830759191Skris is so (for example) those in the US can disable those operations covered 830859191Skris by the RSA patent while allowing storage and parsing of RSA keys and RSA 830959191Skris key generation. 831059191Skris [Steve Henson] 831159191Skris 831259191Skris *) Non-copying interface to BIO pairs. 831359191Skris (still largely untested) 831459191Skris [Bodo Moeller] 831559191Skris 831659191Skris *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 831759191Skris ASCII string. This was handled independently in various places before. 831859191Skris [Steve Henson] 831959191Skris 832059191Skris *) New functions UTF8_getc() and UTF8_putc() that parse and generate 832159191Skris UTF8 strings a character at a time. 832259191Skris [Steve Henson] 832359191Skris 832459191Skris *) Use client_version from client hello to select the protocol 832559191Skris (s23_srvr.c) and for RSA client key exchange verification 832659191Skris (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 832759191Skris [Bodo Moeller] 832859191Skris 832959191Skris *) Add various utility functions to handle SPKACs, these were previously 833059191Skris handled by poking round in the structure internals. Added new function 833159191Skris NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 833259191Skris print, verify and generate SPKACs. Based on an original idea from 833359191Skris Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 833459191Skris [Steve Henson] 833559191Skris 833659191Skris *) RIPEMD160 is operational on all platforms and is back in 'make test'. 833759191Skris [Andy Polyakov] 833859191Skris 833959191Skris *) Allow the config file extension section to be overwritten on the 834059191Skris command line. Based on an original idea from Massimiliano Pala 834159191Skris <madwolf@comune.modena.it>. The new option is called -extensions 834259191Skris and can be applied to ca, req and x509. Also -reqexts to override 834359191Skris the request extensions in req and -crlexts to override the crl extensions 834459191Skris in ca. 834559191Skris [Steve Henson] 834659191Skris 834759191Skris *) Add new feature to the SPKAC handling in ca. Now you can include 834859191Skris the same field multiple times by preceding it by "XXXX." for example: 834959191Skris 1.OU="Unit name 1" 835059191Skris 2.OU="Unit name 2" 835159191Skris this is the same syntax as used in the req config file. 835259191Skris [Steve Henson] 835359191Skris 835459191Skris *) Allow certificate extensions to be added to certificate requests. These 835559191Skris are specified in a 'req_extensions' option of the req section of the 835659191Skris config file. They can be printed out with the -text option to req but 835759191Skris are otherwise ignored at present. 835859191Skris [Steve Henson] 835959191Skris 836059191Skris *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 836159191Skris data read consists of only the final block it would not decrypted because 836259191Skris EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 836359191Skris A misplaced 'break' also meant the decrypted final block might not be 836459191Skris copied until the next read. 836559191Skris [Steve Henson] 836659191Skris 836759191Skris *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 836859191Skris a few extra parameters to the DH structure: these will be useful if 836959191Skris for example we want the value of 'q' or implement X9.42 DH. 837059191Skris [Steve Henson] 837159191Skris 837259191Skris *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 837359191Skris provides hooks that allow the default DSA functions or functions on a 837459191Skris "per key" basis to be replaced. This allows hardware acceleration and 837559191Skris hardware key storage to be handled without major modification to the 837659191Skris library. Also added low level modexp hooks and CRYPTO_EX structure and 837759191Skris associated functions. 837859191Skris [Steve Henson] 837959191Skris 838059191Skris *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 838159191Skris as "read only": it can't be written to and the buffer it points to will 838259191Skris not be freed. Reading from a read only BIO is much more efficient than 838359191Skris a normal memory BIO. This was added because there are several times when 838459191Skris an area of memory needs to be read from a BIO. The previous method was 838559191Skris to create a memory BIO and write the data to it, this results in two 838659191Skris copies of the data and an O(n^2) reading algorithm. There is a new 838759191Skris function BIO_new_mem_buf() which creates a read only memory BIO from 838859191Skris an area of memory. Also modified the PKCS#7 routines to use read only 838959191Skris memory BIOs. 839059191Skris [Steve Henson] 839159191Skris 839259191Skris *) Bugfix: ssl23_get_client_hello did not work properly when called in 839359191Skris state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 839459191Skris a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 839559191Skris but a retry condition occured while trying to read the rest. 839659191Skris [Bodo Moeller] 839759191Skris 839859191Skris *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 839959191Skris NID_pkcs7_encrypted by default: this was wrong since this should almost 840059191Skris always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 840159191Skris the encrypted data type: this is a more sensible place to put it and it 840259191Skris allows the PKCS#12 code to be tidied up that duplicated this 840359191Skris functionality. 840459191Skris [Steve Henson] 840559191Skris 840659191Skris *) Changed obj_dat.pl script so it takes its input and output files on 840759191Skris the command line. This should avoid shell escape redirection problems 840859191Skris under Win32. 840959191Skris [Steve Henson] 841059191Skris 841159191Skris *) Initial support for certificate extension requests, these are included 841259191Skris in things like Xenroll certificate requests. Included functions to allow 841359191Skris extensions to be obtained and added. 841459191Skris [Steve Henson] 841559191Skris 841659191Skris *) -crlf option to s_client and s_server for sending newlines as 841759191Skris CRLF (as required by many protocols). 841859191Skris [Bodo Moeller] 841959191Skris 842055714Skris Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 842155714Skris 842255714Skris *) Install libRSAglue.a when OpenSSL is built with RSAref. 842355714Skris [Ralf S. Engelschall] 842455714Skris 842555714Skris *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 842655714Skris [Andrija Antonijevic <TheAntony2@bigfoot.com>] 842755714Skris 842855714Skris *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 842955714Skris program. 843055714Skris [Steve Henson] 843155714Skris 843255714Skris *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 843355714Skris DH parameters/keys (q is lost during that conversion, but the resulting 843455714Skris DH parameters contain its length). 843555714Skris 843655714Skris For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 843755714Skris much faster than DH_generate_parameters (which creates parameters 843855714Skris where p = 2*q + 1), and also the smaller q makes DH computations 843955714Skris much more efficient (160-bit exponentiation instead of 1024-bit 844055714Skris exponentiation); so this provides a convenient way to support DHE 844155714Skris ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 844255714Skris utter importance to use 844355714Skris SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 844455714Skris or 844555714Skris SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 844655714Skris when such DH parameters are used, because otherwise small subgroup 844755714Skris attacks may become possible! 844855714Skris [Bodo Moeller] 844955714Skris 845055714Skris *) Avoid memory leak in i2d_DHparams. 845155714Skris [Bodo Moeller] 845255714Skris 845355714Skris *) Allow the -k option to be used more than once in the enc program: 845455714Skris this allows the same encrypted message to be read by multiple recipients. 845555714Skris [Steve Henson] 845655714Skris 845755714Skris *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 845855714Skris an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 845955714Skris it will always use the numerical form of the OID, even if it has a short 846055714Skris or long name. 846155714Skris [Steve Henson] 846255714Skris 846355714Skris *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 846455714Skris method only got called if p,q,dmp1,dmq1,iqmp components were present, 846555714Skris otherwise bn_mod_exp was called. In the case of hardware keys for example 846655714Skris no private key components need be present and it might store extra data 846759191Skris in the RSA structure, which cannot be accessed from bn_mod_exp. 846859191Skris By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 846959191Skris private key operations. 847055714Skris [Steve Henson] 847155714Skris 847255714Skris *) Added support for SPARC Linux. 847355714Skris [Andy Polyakov] 847455714Skris 847555714Skris *) pem_password_cb function type incompatibly changed from 847655714Skris typedef int pem_password_cb(char *buf, int size, int rwflag); 847755714Skris to 847855714Skris ....(char *buf, int size, int rwflag, void *userdata); 847955714Skris so that applications can pass data to their callbacks: 848055714Skris The PEM[_ASN1]_{read,write}... functions and macros now take an 848155714Skris additional void * argument, which is just handed through whenever 848255714Skris the password callback is called. 848359191Skris [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 848455714Skris 848555714Skris New function SSL_CTX_set_default_passwd_cb_userdata. 848655714Skris 848755714Skris Compatibility note: As many C implementations push function arguments 848855714Skris onto the stack in reverse order, the new library version is likely to 848955714Skris interoperate with programs that have been compiled with the old 849055714Skris pem_password_cb definition (PEM_whatever takes some data that 849155714Skris happens to be on the stack as its last argument, and the callback 849255714Skris just ignores this garbage); but there is no guarantee whatsoever that 849355714Skris this will work. 849455714Skris 849555714Skris *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 849655714Skris (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 849755714Skris problems not only on Windows, but also on some Unix platforms. 849855714Skris To avoid problematic command lines, these definitions are now in an 849955714Skris auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 850055714Skris for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 850155714Skris [Bodo Moeller] 850255714Skris 850355714Skris *) MIPS III/IV assembler module is reimplemented. 850455714Skris [Andy Polyakov] 850555714Skris 850655714Skris *) More DES library cleanups: remove references to srand/rand and 850755714Skris delete an unused file. 850855714Skris [Ulf M�ller] 850955714Skris 851055714Skris *) Add support for the the free Netwide assembler (NASM) under Win32, 851155714Skris since not many people have MASM (ml) and it can be hard to obtain. 851255714Skris This is currently experimental but it seems to work OK and pass all 851355714Skris the tests. Check out INSTALL.W32 for info. 851455714Skris [Steve Henson] 851555714Skris 851655714Skris *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 851755714Skris without temporary keys kept an extra copy of the server key, 851855714Skris and connections with temporary keys did not free everything in case 851955714Skris of an error. 852055714Skris [Bodo Moeller] 852155714Skris 852255714Skris *) New function RSA_check_key and new openssl rsa option -check 852355714Skris for verifying the consistency of RSA keys. 852455714Skris [Ulf Moeller, Bodo Moeller] 852555714Skris 852655714Skris *) Various changes to make Win32 compile work: 852755714Skris 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 852855714Skris 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 852955714Skris comparison" warnings. 853055714Skris 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 853155714Skris [Steve Henson] 853255714Skris 853355714Skris *) Add a debugging option to PKCS#5 v2 key generation function: when 853455714Skris you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 853555714Skris derived keys are printed to stderr. 853655714Skris [Steve Henson] 853755714Skris 853855714Skris *) Copy the flags in ASN1_STRING_dup(). 853955714Skris [Roman E. Pavlov <pre@mo.msk.ru>] 854055714Skris 854155714Skris *) The x509 application mishandled signing requests containing DSA 854255714Skris keys when the signing key was also DSA and the parameters didn't match. 854355714Skris 854455714Skris It was supposed to omit the parameters when they matched the signing key: 854555714Skris the verifying software was then supposed to automatically use the CA's 854655714Skris parameters if they were absent from the end user certificate. 854755714Skris 854855714Skris Omitting parameters is no longer recommended. The test was also 854955714Skris the wrong way round! This was probably due to unusual behaviour in 855055714Skris EVP_cmp_parameters() which returns 1 if the parameters match. 855155714Skris This meant that parameters were omitted when they *didn't* match and 855255714Skris the certificate was useless. Certificates signed with 'ca' didn't have 855355714Skris this bug. 855455714Skris [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 855555714Skris 855655714Skris *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 855755714Skris The interface is as follows: 855855714Skris Applications can use 855955714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 856055714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 856155714Skris "off" is now the default. 856255714Skris The library internally uses 856355714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 856455714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 856555714Skris to disable memory-checking temporarily. 856655714Skris 856755714Skris Some inconsistent states that previously were possible (and were 856855714Skris even the default) are now avoided. 856955714Skris 857055714Skris -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 857155714Skris with each memory chunk allocated; this is occasionally more helpful 857255714Skris than just having a counter. 857355714Skris 857455714Skris -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 857555714Skris 857655714Skris -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 857755714Skris extensions. 857855714Skris [Bodo Moeller] 857955714Skris 858055714Skris *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 858155714Skris which largely parallels "options", but is for changing API behaviour, 858255714Skris whereas "options" are about protocol behaviour. 858355714Skris Initial "mode" flags are: 858455714Skris 858555714Skris SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 858655714Skris a single record has been written. 858755714Skris SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 858855714Skris retries use the same buffer location. 858955714Skris (But all of the contents must be 859055714Skris copied!) 859155714Skris [Bodo Moeller] 859255714Skris 859379998Skris *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 859455714Skris worked. 859555714Skris 859655714Skris *) Fix problems with no-hmac etc. 859755714Skris [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 859855714Skris 859955714Skris *) New functions RSA_get_default_method(), RSA_set_method() and 860055714Skris RSA_get_method(). These allows replacement of RSA_METHODs without having 860155714Skris to mess around with the internals of an RSA structure. 860255714Skris [Steve Henson] 860355714Skris 860455714Skris *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 860555714Skris Also really enable memory leak checks in openssl.c and in some 860655714Skris test programs. 860755714Skris [Chad C. Mulligan, Bodo Moeller] 860855714Skris 860955714Skris *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 861055714Skris up the length of negative integers. This has now been simplified to just 861155714Skris store the length when it is first determined and use it later, rather 861255714Skris than trying to keep track of where data is copied and updating it to 861355714Skris point to the end. 861455714Skris [Steve Henson, reported by Brien Wheeler 861555714Skris <bwheeler@authentica-security.com>] 861655714Skris 861755714Skris *) Add a new function PKCS7_signatureVerify. This allows the verification 861855714Skris of a PKCS#7 signature but with the signing certificate passed to the 861955714Skris function itself. This contrasts with PKCS7_dataVerify which assumes the 862055714Skris certificate is present in the PKCS#7 structure. This isn't always the 862155714Skris case: certificates can be omitted from a PKCS#7 structure and be 862255714Skris distributed by "out of band" means (such as a certificate database). 862355714Skris [Steve Henson] 862455714Skris 862555714Skris *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 862655714Skris function prototypes in pem.h, also change util/mkdef.pl to add the 862755714Skris necessary function names. 862855714Skris [Steve Henson] 862955714Skris 863055714Skris *) mk1mf.pl (used by Windows builds) did not properly read the 863155714Skris options set by Configure in the top level Makefile, and Configure 863255714Skris was not even able to write more than one option correctly. 863355714Skris Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 863455714Skris [Bodo Moeller] 863555714Skris 863655714Skris *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 863755714Skris file to be loaded from a BIO or FILE pointer. The BIO version will 863855714Skris for example allow memory BIOs to contain config info. 863955714Skris [Steve Henson] 864055714Skris 864155714Skris *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 864255714Skris Whoever hopes to achieve shared-library compatibility across versions 864355714Skris must use this, not the compile-time macro. 864455714Skris (Exercise 0.9.4: Which is the minimum library version required by 864555714Skris such programs?) 864655714Skris Note: All this applies only to multi-threaded programs, others don't 864755714Skris need locks. 864855714Skris [Bodo Moeller] 864955714Skris 865055714Skris *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 865155714Skris through a BIO pair triggered the default case, i.e. 865255714Skris SSLerr(...,SSL_R_UNKNOWN_STATE). 865355714Skris [Bodo Moeller] 865455714Skris 865555714Skris *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 865655714Skris can use the SSL library even if none of the specific BIOs is 865755714Skris appropriate. 865855714Skris [Bodo Moeller] 865955714Skris 866055714Skris *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 866155714Skris for the encoded length. 866255714Skris [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 866355714Skris 866455714Skris *) Add initial documentation of the X509V3 functions. 866555714Skris [Steve Henson] 866655714Skris 866755714Skris *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 866855714Skris PEM_write_bio_PKCS8PrivateKey() that are equivalent to 866955714Skris PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 867055714Skris secure PKCS#8 private key format with a high iteration count. 867155714Skris [Steve Henson] 867255714Skris 867355714Skris *) Fix determination of Perl interpreter: A perl or perl5 867455714Skris _directory_ in $PATH was also accepted as the interpreter. 867555714Skris [Ralf S. Engelschall] 867655714Skris 867755714Skris *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 867855714Skris wrong with it but it was very old and did things like calling 867955714Skris PEM_ASN1_read() directly and used MD5 for the hash not to mention some 868055714Skris unusual formatting. 868155714Skris [Steve Henson] 868255714Skris 868355714Skris *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 868455714Skris to use the new extension code. 868555714Skris [Steve Henson] 868655714Skris 868755714Skris *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 868855714Skris with macros. This should make it easier to change their form, add extra 868955714Skris arguments etc. Fix a few PEM prototypes which didn't have cipher as a 869055714Skris constant. 869155714Skris [Steve Henson] 869255714Skris 869355714Skris *) Add to configuration table a new entry that can specify an alternative 869455714Skris name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 869555714Skris according to Mark Crispin <MRC@Panda.COM>. 869655714Skris [Bodo Moeller] 869755714Skris 869855714Skris#if 0 869955714Skris *) DES CBC did not update the IV. Weird. 870055714Skris [Ben Laurie] 870155714Skris#else 870255714Skris des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 870355714Skris Changing the behaviour of the former might break existing programs -- 870455714Skris where IV updating is needed, des_ncbc_encrypt can be used. 870555714Skris#endif 870655714Skris 870755714Skris *) When bntest is run from "make test" it drives bc to check its 870855714Skris calculations, as well as internally checking them. If an internal check 870955714Skris fails, it needs to cause bc to give a non-zero result or make test carries 871055714Skris on without noticing the failure. Fixed. 871155714Skris [Ben Laurie] 871255714Skris 871355714Skris *) DES library cleanups. 871455714Skris [Ulf M�ller] 871555714Skris 871655714Skris *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 871755714Skris used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 871855714Skris ciphers. NOTE: although the key derivation function has been verified 871955714Skris against some published test vectors it has not been extensively tested 872055714Skris yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 872155714Skris of v2.0. 872255714Skris [Steve Henson] 872355714Skris 872455714Skris *) Instead of "mkdir -p", which is not fully portable, use new 872555714Skris Perl script "util/mkdir-p.pl". 872655714Skris [Bodo Moeller] 872755714Skris 872855714Skris *) Rewrite the way password based encryption (PBE) is handled. It used to 872955714Skris assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 873055714Skris structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 873155714Skris but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 873255714Skris the 'parameter' field of the AlgorithmIdentifier is passed to the 873355714Skris underlying key generation function so it must do its own ASN1 parsing. 873455714Skris This has also changed the EVP_PBE_CipherInit() function which now has a 873555714Skris 'parameter' argument instead of literal salt and iteration count values 873655714Skris and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 873755714Skris [Steve Henson] 873855714Skris 873955714Skris *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 874055714Skris and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 874155714Skris Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 874255714Skris KEY" because this clashed with PKCS#8 unencrypted string. Since this 874355714Skris value was just used as a "magic string" and not used directly its 874455714Skris value doesn't matter. 874555714Skris [Steve Henson] 874655714Skris 874755714Skris *) Introduce some semblance of const correctness to BN. Shame C doesn't 874855714Skris support mutable. 874955714Skris [Ben Laurie] 875055714Skris 875155714Skris *) "linux-sparc64" configuration (ultrapenguin). 875255714Skris [Ray Miller <ray.miller@oucs.ox.ac.uk>] 875355714Skris "linux-sparc" configuration. 875455714Skris [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 875555714Skris 875655714Skris *) config now generates no-xxx options for missing ciphers. 875755714Skris [Ulf M�ller] 875855714Skris 875955714Skris *) Support the EBCDIC character set (work in progress). 876055714Skris File ebcdic.c not yet included because it has a different license. 876155714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 876255714Skris 876355714Skris *) Support BS2000/OSD-POSIX. 876455714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 876555714Skris 876655714Skris *) Make callbacks for key generation use void * instead of char *. 876755714Skris [Ben Laurie] 876855714Skris 876955714Skris *) Make S/MIME samples compile (not yet tested). 877055714Skris [Ben Laurie] 877155714Skris 877255714Skris *) Additional typesafe stacks. 877355714Skris [Ben Laurie] 877455714Skris 877555714Skris *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 877655714Skris [Bodo Moeller] 877755714Skris 877855714Skris 877955714Skris Changes between 0.9.3 and 0.9.3a [29 May 1999] 878055714Skris 878155714Skris *) New configuration variant "sco5-gcc". 878255714Skris 878355714Skris *) Updated some demos. 878455714Skris [Sean O Riordain, Wade Scholine] 878555714Skris 878655714Skris *) Add missing BIO_free at exit of pkcs12 application. 878755714Skris [Wu Zhigang] 878855714Skris 878955714Skris *) Fix memory leak in conf.c. 879055714Skris [Steve Henson] 879155714Skris 879255714Skris *) Updates for Win32 to assembler version of MD5. 879355714Skris [Steve Henson] 879455714Skris 879555714Skris *) Set #! path to perl in apps/der_chop to where we found it 879655714Skris instead of using a fixed path. 879755714Skris [Bodo Moeller] 879855714Skris 879955714Skris *) SHA library changes for irix64-mips4-cc. 880055714Skris [Andy Polyakov] 880155714Skris 880255714Skris *) Improvements for VMS support. 880355714Skris [Richard Levitte] 880455714Skris 880555714Skris 880655714Skris Changes between 0.9.2b and 0.9.3 [24 May 1999] 880755714Skris 880855714Skris *) Bignum library bug fix. IRIX 6 passes "make test" now! 880955714Skris This also avoids the problems with SC4.2 and unpatched SC5. 881055714Skris [Andy Polyakov <appro@fy.chalmers.se>] 881155714Skris 881255714Skris *) New functions sk_num, sk_value and sk_set to replace the previous macros. 881355714Skris These are required because of the typesafe stack would otherwise break 881455714Skris existing code. If old code used a structure member which used to be STACK 881555714Skris and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 881655714Skris sk_num or sk_value it would produce an error because the num, data members 881755714Skris are not present in STACK_OF. Now it just produces a warning. sk_set 881855714Skris replaces the old method of assigning a value to sk_value 881955714Skris (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 882055714Skris that does this will no longer work (and should use sk_set instead) but 882155714Skris this could be regarded as a "questionable" behaviour anyway. 882255714Skris [Steve Henson] 882355714Skris 882455714Skris *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 882555714Skris correctly handle encrypted S/MIME data. 882655714Skris [Steve Henson] 882755714Skris 882855714Skris *) Change type of various DES function arguments from des_cblock 882955714Skris (which means, in function argument declarations, pointer to char) 883055714Skris to des_cblock * (meaning pointer to array with 8 char elements), 883155714Skris which allows the compiler to do more typechecking; it was like 883255714Skris that back in SSLeay, but with lots of ugly casts. 883355714Skris 883455714Skris Introduce new type const_des_cblock. 883555714Skris [Bodo Moeller] 883655714Skris 883755714Skris *) Reorganise the PKCS#7 library and get rid of some of the more obvious 883855714Skris problems: find RecipientInfo structure that matches recipient certificate 883955714Skris and initialise the ASN1 structures properly based on passed cipher. 884055714Skris [Steve Henson] 884155714Skris 884255714Skris *) Belatedly make the BN tests actually check the results. 884355714Skris [Ben Laurie] 884455714Skris 884555714Skris *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 884655714Skris to and from BNs: it was completely broken. New compilation option 884755714Skris NEG_PUBKEY_BUG to allow for some broken certificates that encode public 884855714Skris key elements as negative integers. 884955714Skris [Steve Henson] 885055714Skris 885155714Skris *) Reorganize and speed up MD5. 885255714Skris [Andy Polyakov <appro@fy.chalmers.se>] 885355714Skris 885455714Skris *) VMS support. 885555714Skris [Richard Levitte <richard@levitte.org>] 885655714Skris 885755714Skris *) New option -out to asn1parse to allow the parsed structure to be 885855714Skris output to a file. This is most useful when combined with the -strparse 885955714Skris option to examine the output of things like OCTET STRINGS. 886055714Skris [Steve Henson] 886155714Skris 886255714Skris *) Make SSL library a little more fool-proof by not requiring any longer 886355714Skris that SSL_set_{accept,connect}_state be called before 886455714Skris SSL_{accept,connect} may be used (SSL_set_..._state is omitted 886555714Skris in many applications because usually everything *appeared* to work as 886655714Skris intended anyway -- now it really works as intended). 886755714Skris [Bodo Moeller] 886855714Skris 886955714Skris *) Move openssl.cnf out of lib/. 887055714Skris [Ulf M�ller] 887155714Skris 887255714Skris *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 887355714Skris -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 887455714Skris -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 887555714Skris [Ralf S. Engelschall] 887655714Skris 887755714Skris *) Various fixes to the EVP and PKCS#7 code. It may now be able to 887855714Skris handle PKCS#7 enveloped data properly. 887955714Skris [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 888055714Skris 888155714Skris *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 888255714Skris copying pointers. The cert_st handling is changed by this in 888355714Skris various ways (and thus what used to be known as ctx->default_cert 888455714Skris is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 888555714Skris any longer when s->cert does not give us what we need). 888655714Skris ssl_cert_instantiate becomes obsolete by this change. 888755714Skris As soon as we've got the new code right (possibly it already is?), 888855714Skris we have solved a couple of bugs of the earlier code where s->cert 888955714Skris was used as if it could not have been shared with other SSL structures. 889055714Skris 889155714Skris Note that using the SSL API in certain dirty ways now will result 889255714Skris in different behaviour than observed with earlier library versions: 889355714Skris Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 889455714Skris does not influence s as it used to. 889555714Skris 889655714Skris In order to clean up things more thoroughly, inside SSL_SESSION 889755714Skris we don't use CERT any longer, but a new structure SESS_CERT 889855714Skris that holds per-session data (if available); currently, this is 889955714Skris the peer's certificate chain and, for clients, the server's certificate 890055714Skris and temporary key. CERT holds only those values that can have 890155714Skris meaningful defaults in an SSL_CTX. 890255714Skris [Bodo Moeller] 890355714Skris 890455714Skris *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 890555714Skris from the internal representation. Various PKCS#7 fixes: remove some 890655714Skris evil casts and set the enc_dig_alg field properly based on the signing 890755714Skris key type. 890855714Skris [Steve Henson] 890955714Skris 891055714Skris *) Allow PKCS#12 password to be set from the command line or the 891155714Skris environment. Let 'ca' get its config file name from the environment 891255714Skris variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 891355714Skris and 'x509'). 891455714Skris [Steve Henson] 891555714Skris 891655714Skris *) Allow certificate policies extension to use an IA5STRING for the 891755714Skris organization field. This is contrary to the PKIX definition but 891855714Skris VeriSign uses it and IE5 only recognises this form. Document 'x509' 891955714Skris extension option. 892055714Skris [Steve Henson] 892155714Skris 892255714Skris *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 892355714Skris without disallowing inline assembler and the like for non-pedantic builds. 892455714Skris [Ben Laurie] 892555714Skris 892655714Skris *) Support Borland C++ builder. 892755714Skris [Janez Jere <jj@void.si>, modified by Ulf M�ller] 892855714Skris 892955714Skris *) Support Mingw32. 893055714Skris [Ulf M�ller] 893155714Skris 893255714Skris *) SHA-1 cleanups and performance enhancements. 893355714Skris [Andy Polyakov <appro@fy.chalmers.se>] 893455714Skris 893555714Skris *) Sparc v8plus assembler for the bignum library. 893655714Skris [Andy Polyakov <appro@fy.chalmers.se>] 893755714Skris 893855714Skris *) Accept any -xxx and +xxx compiler options in Configure. 893955714Skris [Ulf M�ller] 894055714Skris 894155714Skris *) Update HPUX configuration. 894255714Skris [Anonymous] 894355714Skris 894455714Skris *) Add missing sk_<type>_unshift() function to safestack.h 894555714Skris [Ralf S. Engelschall] 894655714Skris 894755714Skris *) New function SSL_CTX_use_certificate_chain_file that sets the 894855714Skris "extra_cert"s in addition to the certificate. (This makes sense 894955714Skris only for "PEM" format files, as chains as a whole are not 895055714Skris DER-encoded.) 895155714Skris [Bodo Moeller] 895255714Skris 895355714Skris *) Support verify_depth from the SSL API. 895455714Skris x509_vfy.c had what can be considered an off-by-one-error: 895555714Skris Its depth (which was not part of the external interface) 895655714Skris was actually counting the number of certificates in a chain; 895755714Skris now it really counts the depth. 895855714Skris [Bodo Moeller] 895955714Skris 896055714Skris *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 896155714Skris instead of X509err, which often resulted in confusing error 896255714Skris messages since the error codes are not globally unique 896355714Skris (e.g. an alleged error in ssl3_accept when a certificate 896455714Skris didn't match the private key). 896555714Skris 896655714Skris *) New function SSL_CTX_set_session_id_context that allows to set a default 896755714Skris value (so that you don't need SSL_set_session_id_context for each 896855714Skris connection using the SSL_CTX). 896955714Skris [Bodo Moeller] 897055714Skris 897155714Skris *) OAEP decoding bug fix. 897255714Skris [Ulf M�ller] 897355714Skris 897455714Skris *) Support INSTALL_PREFIX for package builders, as proposed by 897555714Skris David Harris. 897655714Skris [Bodo Moeller] 897755714Skris 897855714Skris *) New Configure options "threads" and "no-threads". For systems 897955714Skris where the proper compiler options are known (currently Solaris 898055714Skris and Linux), "threads" is the default. 898155714Skris [Bodo Moeller] 898255714Skris 898355714Skris *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 898455714Skris [Bodo Moeller] 898555714Skris 898655714Skris *) Install various scripts to $(OPENSSLDIR)/misc, not to 898755714Skris $(INSTALLTOP)/bin -- they shouldn't clutter directories 898855714Skris such as /usr/local/bin. 898955714Skris [Bodo Moeller] 899055714Skris 899155714Skris *) "make linux-shared" to build shared libraries. 899255714Skris [Niels Poppe <niels@netbox.org>] 899355714Skris 899455714Skris *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 899555714Skris [Ulf M�ller] 899655714Skris 899755714Skris *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 899855714Skris extension adding in x509 utility. 899955714Skris [Steve Henson] 900055714Skris 900155714Skris *) Remove NOPROTO sections and error code comments. 900255714Skris [Ulf M�ller] 900355714Skris 900455714Skris *) Partial rewrite of the DEF file generator to now parse the ANSI 900555714Skris prototypes. 900655714Skris [Steve Henson] 900755714Skris 900855714Skris *) New Configure options --prefix=DIR and --openssldir=DIR. 900955714Skris [Ulf M�ller] 901055714Skris 901155714Skris *) Complete rewrite of the error code script(s). It is all now handled 901255714Skris by one script at the top level which handles error code gathering, 901355714Skris header rewriting and C source file generation. It should be much better 901455714Skris than the old method: it now uses a modified version of Ulf's parser to 901555714Skris read the ANSI prototypes in all header files (thus the old K&R definitions 901655714Skris aren't needed for error creation any more) and do a better job of 901755714Skris translating function codes into names. The old 'ASN1 error code imbedded 901855714Skris in a comment' is no longer necessary and it doesn't use .err files which 901955714Skris have now been deleted. Also the error code call doesn't have to appear all 902055714Skris on one line (which resulted in some large lines...). 902155714Skris [Steve Henson] 902255714Skris 902355714Skris *) Change #include filenames from <foo.h> to <openssl/foo.h>. 902455714Skris [Bodo Moeller] 902555714Skris 902655714Skris *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 902755714Skris 0 (which usually indicates a closed connection), but continue reading. 902855714Skris [Bodo Moeller] 902955714Skris 903055714Skris *) Fix some race conditions. 903155714Skris [Bodo Moeller] 903255714Skris 903355714Skris *) Add support for CRL distribution points extension. Add Certificate 903455714Skris Policies and CRL distribution points documentation. 903555714Skris [Steve Henson] 903655714Skris 903755714Skris *) Move the autogenerated header file parts to crypto/opensslconf.h. 903855714Skris [Ulf M�ller] 903955714Skris 904055714Skris *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 904155714Skris 8 of keying material. Merlin has also confirmed interop with this fix 904255714Skris between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 904355714Skris [Merlin Hughes <merlin@baltimore.ie>] 904455714Skris 904555714Skris *) Fix lots of warnings. 904655714Skris [Richard Levitte <levitte@stacken.kth.se>] 904755714Skris 904855714Skris *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 904955714Skris the directory spec didn't end with a LIST_SEPARATOR_CHAR. 905055714Skris [Richard Levitte <levitte@stacken.kth.se>] 905155714Skris 905255714Skris *) Fix problems with sizeof(long) == 8. 905355714Skris [Andy Polyakov <appro@fy.chalmers.se>] 905455714Skris 905555714Skris *) Change functions to ANSI C. 905655714Skris [Ulf M�ller] 905755714Skris 905855714Skris *) Fix typos in error codes. 905955714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 906055714Skris 906155714Skris *) Remove defunct assembler files from Configure. 906255714Skris [Ulf M�ller] 906355714Skris 906455714Skris *) SPARC v8 assembler BIGNUM implementation. 906555714Skris [Andy Polyakov <appro@fy.chalmers.se>] 906655714Skris 906755714Skris *) Support for Certificate Policies extension: both print and set. 906855714Skris Various additions to support the r2i method this uses. 906955714Skris [Steve Henson] 907055714Skris 907155714Skris *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 907255714Skris return a const string when you are expecting an allocated buffer. 907355714Skris [Ben Laurie] 907455714Skris 907555714Skris *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 907655714Skris types DirectoryString and DisplayText. 907755714Skris [Steve Henson] 907855714Skris 907955714Skris *) Add code to allow r2i extensions to access the configuration database, 908055714Skris add an LHASH database driver and add several ctx helper functions. 908155714Skris [Steve Henson] 908255714Skris 908355714Skris *) Fix an evil bug in bn_expand2() which caused various BN functions to 908455714Skris fail when they extended the size of a BIGNUM. 908555714Skris [Steve Henson] 908655714Skris 908755714Skris *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 908855714Skris support typesafe stack. 908955714Skris [Steve Henson] 909055714Skris 909155714Skris *) Fix typo in SSL_[gs]et_options(). 909255714Skris [Nils Frostberg <nils@medcom.se>] 909355714Skris 909455714Skris *) Delete various functions and files that belonged to the (now obsolete) 909555714Skris old X509V3 handling code. 909655714Skris [Steve Henson] 909755714Skris 909855714Skris *) New Configure option "rsaref". 909955714Skris [Ulf M�ller] 910055714Skris 910155714Skris *) Don't auto-generate pem.h. 910255714Skris [Bodo Moeller] 910355714Skris 910455714Skris *) Introduce type-safe ASN.1 SETs. 910555714Skris [Ben Laurie] 910655714Skris 910755714Skris *) Convert various additional casted stacks to type-safe STACK_OF() variants. 910855714Skris [Ben Laurie, Ralf S. Engelschall, Steve Henson] 910955714Skris 911055714Skris *) Introduce type-safe STACKs. This will almost certainly break lots of code 911155714Skris that links with OpenSSL (well at least cause lots of warnings), but fear 911255714Skris not: the conversion is trivial, and it eliminates loads of evil casts. A 911355714Skris few STACKed things have been converted already. Feel free to convert more. 911455714Skris In the fullness of time, I'll do away with the STACK type altogether. 911555714Skris [Ben Laurie] 911655714Skris 911755714Skris *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 911855714Skris specified in <certfile> by updating the entry in the index.txt file. 911955714Skris This way one no longer has to edit the index.txt file manually for 912055714Skris revoking a certificate. The -revoke option does the gory details now. 912155714Skris [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 912255714Skris 912355714Skris *) Fix `openssl crl -noout -text' combination where `-noout' killed the 912455714Skris `-text' option at all and this way the `-noout -text' combination was 912555714Skris inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 912655714Skris [Ralf S. Engelschall] 912755714Skris 912855714Skris *) Make sure a corresponding plain text error message exists for the 912955714Skris X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 913055714Skris verify callback function determined that a certificate was revoked. 913155714Skris [Ralf S. Engelschall] 913255714Skris 913355714Skris *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 913455714Skris ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 913555714Skris all available cipers including rc5, which was forgotten until now. 913655714Skris In order to let the testing shell script know which algorithms 913755714Skris are available, a new (up to now undocumented) command 913855714Skris "openssl list-cipher-commands" is used. 913955714Skris [Bodo Moeller] 914055714Skris 914155714Skris *) Bugfix: s_client occasionally would sleep in select() when 914255714Skris it should have checked SSL_pending() first. 914355714Skris [Bodo Moeller] 914455714Skris 914555714Skris *) New functions DSA_do_sign and DSA_do_verify to provide access to 914655714Skris the raw DSA values prior to ASN.1 encoding. 914755714Skris [Ulf M�ller] 914855714Skris 914955714Skris *) Tweaks to Configure 915055714Skris [Niels Poppe <niels@netbox.org>] 915155714Skris 915255714Skris *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 915355714Skris yet... 915455714Skris [Steve Henson] 915555714Skris 915655714Skris *) New variables $(RANLIB) and $(PERL) in the Makefiles. 915755714Skris [Ulf M�ller] 915855714Skris 915955714Skris *) New config option to avoid instructions that are illegal on the 80386. 916055714Skris The default code is faster, but requires at least a 486. 916155714Skris [Ulf M�ller] 916255714Skris 916355714Skris *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 916455714Skris SSL2_SERVER_VERSION (not used at all) macros, which are now the 916555714Skris same as SSL2_VERSION anyway. 916655714Skris [Bodo Moeller] 916755714Skris 916855714Skris *) New "-showcerts" option for s_client. 916955714Skris [Bodo Moeller] 917055714Skris 917155714Skris *) Still more PKCS#12 integration. Add pkcs12 application to openssl 917255714Skris application. Various cleanups and fixes. 917355714Skris [Steve Henson] 917455714Skris 917555714Skris *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 917655714Skris modify error routines to work internally. Add error codes and PBE init 917755714Skris to library startup routines. 917855714Skris [Steve Henson] 917955714Skris 918055714Skris *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 918155714Skris packing functions to asn1 and evp. Changed function names and error 918255714Skris codes along the way. 918355714Skris [Steve Henson] 918455714Skris 918555714Skris *) PKCS12 integration: and so it begins... First of several patches to 918655714Skris slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 918755714Skris objects to objects.h 918855714Skris [Steve Henson] 918955714Skris 919055714Skris *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 919155714Skris and display support for Thawte strong extranet extension. 919255714Skris [Steve Henson] 919355714Skris 919455714Skris *) Add LinuxPPC support. 919555714Skris [Jeff Dubrule <igor@pobox.org>] 919655714Skris 919755714Skris *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 919855714Skris bn_div_words in alpha.s. 919955714Skris [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 920055714Skris 920155714Skris *) Make sure the RSA OAEP test is skipped under -DRSAref because 920255714Skris OAEP isn't supported when OpenSSL is built with RSAref. 920355714Skris [Ulf Moeller <ulf@fitug.de>] 920455714Skris 920555714Skris *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 920655714Skris so they no longer are missing under -DNOPROTO. 920755714Skris [Soren S. Jorvang <soren@t.dk>] 920855714Skris 920955714Skris 921055714Skris Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 921155714Skris 921255714Skris *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 921355714Skris doesn't work when the session is reused. Coming soon! 921455714Skris [Ben Laurie] 921555714Skris 921655714Skris *) Fix a security hole, that allows sessions to be reused in the wrong 921755714Skris context thus bypassing client cert protection! All software that uses 921855714Skris client certs and session caches in multiple contexts NEEDS PATCHING to 921955714Skris allow session reuse! A fuller solution is in the works. 922055714Skris [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 922155714Skris 922255714Skris *) Some more source tree cleanups (removed obsolete files 922355714Skris crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 922455714Skris permission on "config" script to be executable) and a fix for the INSTALL 922555714Skris document. 922655714Skris [Ulf Moeller <ulf@fitug.de>] 922755714Skris 922855714Skris *) Remove some legacy and erroneous uses of malloc, free instead of 922955714Skris Malloc, Free. 923055714Skris [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 923155714Skris 923255714Skris *) Make rsa_oaep_test return non-zero on error. 923355714Skris [Ulf Moeller <ulf@fitug.de>] 923455714Skris 923555714Skris *) Add support for native Solaris shared libraries. Configure 923655714Skris solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 923755714Skris if someone would make that last step automatic. 923855714Skris [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 923955714Skris 924055714Skris *) ctx_size was not built with the right compiler during "make links". Fixed. 924155714Skris [Ben Laurie] 924255714Skris 924355714Skris *) Change the meaning of 'ALL' in the cipher list. It now means "everything 924455714Skris except NULL ciphers". This means the default cipher list will no longer 924555714Skris enable NULL ciphers. They need to be specifically enabled e.g. with 924655714Skris the string "DEFAULT:eNULL". 924755714Skris [Steve Henson] 924855714Skris 924955714Skris *) Fix to RSA private encryption routines: if p < q then it would 925055714Skris occasionally produce an invalid result. This will only happen with 925155714Skris externally generated keys because OpenSSL (and SSLeay) ensure p > q. 925255714Skris [Steve Henson] 925355714Skris 925455714Skris *) Be less restrictive and allow also `perl util/perlpath.pl 925555714Skris /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 925655714Skris because this way one can also use an interpreter named `perl5' (which is 925755714Skris usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 925855714Skris installed as `perl'). 925955714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 926055714Skris 926155714Skris *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 926255714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 926355714Skris 926455714Skris *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 926555714Skris advapi32.lib to Win32 build and change the pem test comparision 926655714Skris to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 926755714Skris suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 926855714Skris and crypto/des/ede_cbcm_enc.c. 926955714Skris [Steve Henson] 927055714Skris 927155714Skris *) DES quad checksum was broken on big-endian architectures. Fixed. 927255714Skris [Ben Laurie] 927355714Skris 927455714Skris *) Comment out two functions in bio.h that aren't implemented. Fix up the 927555714Skris Win32 test batch file so it (might) work again. The Win32 test batch file 927655714Skris is horrible: I feel ill.... 927755714Skris [Steve Henson] 927855714Skris 927955714Skris *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 928055714Skris in e_os.h. Audit of header files to check ANSI and non ANSI 928155714Skris sections: 10 functions were absent from non ANSI section and not exported 928255714Skris from Windows DLLs. Fixed up libeay.num for new functions. 928355714Skris [Steve Henson] 928455714Skris 928555714Skris *) Make `openssl version' output lines consistent. 928655714Skris [Ralf S. Engelschall] 928755714Skris 928855714Skris *) Fix Win32 symbol export lists for BIO functions: Added 928955714Skris BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 929055714Skris to ms/libeay{16,32}.def. 929155714Skris [Ralf S. Engelschall] 929255714Skris 929355714Skris *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 929455714Skris fine under Unix and passes some trivial tests I've now added. But the 929555714Skris whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 929655714Skris added to make sure no one expects that this stuff really works in the 929755714Skris OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 929855714Skris up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 929955714Skris openssl_bio.xs. 930055714Skris [Ralf S. Engelschall] 930155714Skris 930255714Skris *) Fix the generation of two part addresses in perl. 930355714Skris [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 930455714Skris 930555714Skris *) Add config entry for Linux on MIPS. 930655714Skris [John Tobey <jtobey@channel1.com>] 930755714Skris 930855714Skris *) Make links whenever Configure is run, unless we are on Windoze. 930955714Skris [Ben Laurie] 931055714Skris 931155714Skris *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 931255714Skris Currently only issuerAltName and AuthorityKeyIdentifier make any sense 931355714Skris in CRLs. 931455714Skris [Steve Henson] 931555714Skris 931655714Skris *) Add a useful kludge to allow package maintainers to specify compiler and 931755714Skris other platforms details on the command line without having to patch the 931855714Skris Configure script everytime: One now can use ``perl Configure 931955714Skris <id>:<details>'', i.e. platform ids are allowed to have details appended 932055714Skris to them (seperated by colons). This is treated as there would be a static 932155714Skris pre-configured entry in Configure's %table under key <id> with value 932255714Skris <details> and ``perl Configure <id>'' is called. So, when you want to 932355714Skris perform a quick test-compile under FreeBSD 3.1 with pgcc and without 932455714Skris assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 932555714Skris now, which overrides the FreeBSD-elf entry on-the-fly. 932655714Skris [Ralf S. Engelschall] 932755714Skris 932855714Skris *) Disable new TLS1 ciphersuites by default: they aren't official yet. 932955714Skris [Ben Laurie] 933055714Skris 933155714Skris *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 933255714Skris on the `perl Configure ...' command line. This way one can compile 933355714Skris OpenSSL libraries with Position Independent Code (PIC) which is needed 933455714Skris for linking it into DSOs. 933555714Skris [Ralf S. Engelschall] 933655714Skris 933755714Skris *) Remarkably, export ciphers were totally broken and no-one had noticed! 933855714Skris Fixed. 933955714Skris [Ben Laurie] 934055714Skris 934155714Skris *) Cleaned up the LICENSE document: The official contact for any license 934255714Skris questions now is the OpenSSL core team under openssl-core@openssl.org. 934355714Skris And add a paragraph about the dual-license situation to make sure people 934455714Skris recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 934555714Skris to the OpenSSL toolkit. 934655714Skris [Ralf S. Engelschall] 934755714Skris 934855714Skris *) General source tree makefile cleanups: Made `making xxx in yyy...' 934955714Skris display consistent in the source tree and replaced `/bin/rm' by `rm'. 935055714Skris Additonally cleaned up the `make links' target: Remove unnecessary 935155714Skris semicolons, subsequent redundant removes, inline point.sh into mklink.sh 935255714Skris to speed processing and no longer clutter the display with confusing 935355714Skris stuff. Instead only the actually done links are displayed. 935455714Skris [Ralf S. Engelschall] 935555714Skris 935655714Skris *) Permit null encryption ciphersuites, used for authentication only. It used 935755714Skris to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 935855714Skris It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 935955714Skris encryption. 936055714Skris [Ben Laurie] 936155714Skris 936255714Skris *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 936355714Skris signed attributes when verifying signatures (this would break them), 936455714Skris the detached data encoding was wrong and public keys obtained using 936555714Skris X509_get_pubkey() weren't freed. 936655714Skris [Steve Henson] 936755714Skris 936855714Skris *) Add text documentation for the BUFFER functions. Also added a work around 936955714Skris to a Win95 console bug. This was triggered by the password read stuff: the 937055714Skris last character typed gets carried over to the next fread(). If you were 937155714Skris generating a new cert request using 'req' for example then the last 937255714Skris character of the passphrase would be CR which would then enter the first 937355714Skris field as blank. 937455714Skris [Steve Henson] 937555714Skris 937655714Skris *) Added the new `Includes OpenSSL Cryptography Software' button as 937755714Skris doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 937855714Skris button and can be used by applications based on OpenSSL to show the 937955714Skris relationship to the OpenSSL project. 938055714Skris [Ralf S. Engelschall] 938155714Skris 938255714Skris *) Remove confusing variables in function signatures in files 938355714Skris ssl/ssl_lib.c and ssl/ssl.h. 938455714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 938555714Skris 938655714Skris *) Don't install bss_file.c under PREFIX/include/ 938755714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 938855714Skris 938955714Skris *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 939055714Skris functions that return function pointers and has support for NT specific 939155714Skris stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 939255714Skris #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 939355714Skris unsigned to signed types: this was killing the Win32 compile. 939455714Skris [Steve Henson] 939555714Skris 939655714Skris *) Add new certificate file to stack functions, 939755714Skris SSL_add_dir_cert_subjects_to_stack() and 939855714Skris SSL_add_file_cert_subjects_to_stack(). These largely supplant 939955714Skris SSL_load_client_CA_file(), and can be used to add multiple certs easily 940055714Skris to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 940155714Skris This means that Apache-SSL and similar packages don't have to mess around 940255714Skris to add as many CAs as they want to the preferred list. 940355714Skris [Ben Laurie] 940455714Skris 940555714Skris *) Experiment with doxygen documentation. Currently only partially applied to 940655714Skris ssl/ssl_lib.c. 940755714Skris See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 940855714Skris openssl.doxy as the configuration file. 940955714Skris [Ben Laurie] 941055714Skris 941155714Skris *) Get rid of remaining C++-style comments which strict C compilers hate. 941255714Skris [Ralf S. Engelschall, pointed out by Carlos Amengual] 941355714Skris 941455714Skris *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 941555714Skris compiled in by default: it has problems with large keys. 941655714Skris [Steve Henson] 941755714Skris 941855714Skris *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 941955714Skris DH private keys and/or callback functions which directly correspond to 942055714Skris their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 942155714Skris is needed for applications which have to configure certificates on a 942255714Skris per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 942355714Skris (e.g. s_server). 942455714Skris For the RSA certificate situation is makes no difference, but 942555714Skris for the DSA certificate situation this fixes the "no shared cipher" 942655714Skris problem where the OpenSSL cipher selection procedure failed because the 942755714Skris temporary keys were not overtaken from the context and the API provided 942855714Skris no way to reconfigure them. 942955714Skris The new functions now let applications reconfigure the stuff and they 943055714Skris are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 943155714Skris SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 943255714Skris non-public-API function ssl_cert_instantiate() is used as a helper 943355714Skris function and also to reduce code redundancy inside ssl_rsa.c. 943455714Skris [Ralf S. Engelschall] 943555714Skris 943655714Skris *) Move s_server -dcert and -dkey options out of the undocumented feature 943755714Skris area because they are useful for the DSA situation and should be 943855714Skris recognized by the users. 943955714Skris [Ralf S. Engelschall] 944055714Skris 944155714Skris *) Fix the cipher decision scheme for export ciphers: the export bits are 944255714Skris *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 944355714Skris SSL_EXP_MASK. So, the original variable has to be used instead of the 944455714Skris already masked variable. 944555714Skris [Richard Levitte <levitte@stacken.kth.se>] 944655714Skris 944755714Skris *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 944855714Skris [Richard Levitte <levitte@stacken.kth.se>] 944955714Skris 945055714Skris *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 945155714Skris from `int' to `unsigned int' because it's a length and initialized by 945255714Skris EVP_DigestFinal() which expects an `unsigned int *'. 945355714Skris [Richard Levitte <levitte@stacken.kth.se>] 945455714Skris 945555714Skris *) Don't hard-code path to Perl interpreter on shebang line of Configure 945655714Skris script. Instead use the usual Shell->Perl transition trick. 945755714Skris [Ralf S. Engelschall] 945855714Skris 945955714Skris *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 946055714Skris (in addition to RSA certificates) to match the behaviour of `openssl dsa 946155714Skris -noout -modulus' as it's already the case for `openssl rsa -noout 946255714Skris -modulus'. For RSA the -modulus is the real "modulus" while for DSA 946355714Skris currently the public key is printed (a decision which was already done by 946455714Skris `openssl dsa -modulus' in the past) which serves a similar purpose. 946555714Skris Additionally the NO_RSA no longer completely removes the whole -modulus 946655714Skris option; it now only avoids using the RSA stuff. Same applies to NO_DSA 946755714Skris now, too. 946855714Skris [Ralf S. Engelschall] 946955714Skris 947055714Skris *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 947155714Skris BIO. See the source (crypto/evp/bio_ok.c) for more info. 947255714Skris [Arne Ansper <arne@ats.cyber.ee>] 947355714Skris 947455714Skris *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 947555714Skris to be added. Now both 'req' and 'ca' can use new objects defined in the 947655714Skris config file. 947755714Skris [Steve Henson] 947855714Skris 947955714Skris *) Add cool BIO that does syslog (or event log on NT). 948055714Skris [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 948155714Skris 948255714Skris *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 948355714Skris TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 948455714Skris TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 948555714Skris Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 948655714Skris [Ben Laurie] 948755714Skris 948855714Skris *) Add preliminary config info for new extension code. 948955714Skris [Steve Henson] 949055714Skris 949155714Skris *) Make RSA_NO_PADDING really use no padding. 949255714Skris [Ulf Moeller <ulf@fitug.de>] 949355714Skris 949455714Skris *) Generate errors when private/public key check is done. 949555714Skris [Ben Laurie] 949655714Skris 949755714Skris *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 949855714Skris for some CRL extensions and new objects added. 949955714Skris [Steve Henson] 950055714Skris 950155714Skris *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 950255714Skris key usage extension and fuller support for authority key id. 950355714Skris [Steve Henson] 950455714Skris 950555714Skris *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 950655714Skris padding method for RSA, which is recommended for new applications in PKCS 950755714Skris #1 v2.0 (RFC 2437, October 1998). 950855714Skris OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 950955714Skris foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 951055714Skris against Bleichbacher's attack on RSA. 951155714Skris [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 951255714Skris Ben Laurie] 951355714Skris 951455714Skris *) Updates to the new SSL compression code 951555714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 951655714Skris 951755714Skris *) Fix so that the version number in the master secret, when passed 951855714Skris via RSA, checks that if TLS was proposed, but we roll back to SSLv3 951955714Skris (because the server will not accept higher), that the version number 952055714Skris is 0x03,0x01, not 0x03,0x00 952155714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 952255714Skris 952355714Skris *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 952455714Skris leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 952555714Skris in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 952655714Skris [Steve Henson] 952755714Skris 952855714Skris *) Support for RAW extensions where an arbitrary extension can be 952955714Skris created by including its DER encoding. See apps/openssl.cnf for 953055714Skris an example. 953155714Skris [Steve Henson] 953255714Skris 953355714Skris *) Make sure latest Perl versions don't interpret some generated C array 953455714Skris code as Perl array code in the crypto/err/err_genc.pl script. 953555714Skris [Lars Weber <3weber@informatik.uni-hamburg.de>] 953655714Skris 953755714Skris *) Modify ms/do_ms.bat to not generate assembly language makefiles since 953855714Skris not many people have the assembler. Various Win32 compilation fixes and 953955714Skris update to the INSTALL.W32 file with (hopefully) more accurate Win32 954055714Skris build instructions. 954155714Skris [Steve Henson] 954255714Skris 954355714Skris *) Modify configure script 'Configure' to automatically create crypto/date.h 954455714Skris file under Win32 and also build pem.h from pem.org. New script 954555714Skris util/mkfiles.pl to create the MINFO file on environments that can't do a 954655714Skris 'make files': perl util/mkfiles.pl >MINFO should work. 954755714Skris [Steve Henson] 954855714Skris 954955714Skris *) Major rework of DES function declarations, in the pursuit of correctness 955055714Skris and purity. As a result, many evil casts evaporated, and some weirdness, 955155714Skris too. You may find this causes warnings in your code. Zapping your evil 955255714Skris casts will probably fix them. Mostly. 955355714Skris [Ben Laurie] 955455714Skris 955555714Skris *) Fix for a typo in asn1.h. Bug fix to object creation script 955655714Skris obj_dat.pl. It considered a zero in an object definition to mean 955755714Skris "end of object": none of the objects in objects.h have any zeros 955855714Skris so it wasn't spotted. 955955714Skris [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 956055714Skris 956155714Skris *) Add support for Triple DES Cipher Block Chaining with Output Feedback 956255714Skris Masking (CBCM). In the absence of test vectors, the best I have been able 956355714Skris to do is check that the decrypt undoes the encrypt, so far. Send me test 956455714Skris vectors if you have them. 956555714Skris [Ben Laurie] 956655714Skris 956755714Skris *) Correct calculation of key length for export ciphers (too much space was 956855714Skris allocated for null ciphers). This has not been tested! 956955714Skris [Ben Laurie] 957055714Skris 957155714Skris *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 957255714Skris message is now correct (it understands "crypto" and "ssl" on its 957355714Skris command line). There is also now an "update" option. This will update 957455714Skris the util/ssleay.num and util/libeay.num files with any new functions. 957555714Skris If you do a: 957655714Skris perl util/mkdef.pl crypto ssl update 957755714Skris it will update them. 957855714Skris [Steve Henson] 957955714Skris 958055714Skris *) Overhauled the Perl interface (perl/*): 958155714Skris - ported BN stuff to OpenSSL's different BN library 958255714Skris - made the perl/ source tree CVS-aware 958355714Skris - renamed the package from SSLeay to OpenSSL (the files still contain 958455714Skris their history because I've copied them in the repository) 958555714Skris - removed obsolete files (the test scripts will be replaced 958655714Skris by better Test::Harness variants in the future) 958755714Skris [Ralf S. Engelschall] 958855714Skris 958955714Skris *) First cut for a very conservative source tree cleanup: 959055714Skris 1. merge various obsolete readme texts into doc/ssleay.txt 959155714Skris where we collect the old documents and readme texts. 959255714Skris 2. remove the first part of files where I'm already sure that we no 959355714Skris longer need them because of three reasons: either they are just temporary 959455714Skris files which were left by Eric or they are preserved original files where 959555714Skris I've verified that the diff is also available in the CVS via "cvs diff 959655714Skris -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 959755714Skris the crypto/md/ stuff). 959855714Skris [Ralf S. Engelschall] 959955714Skris 960055714Skris *) More extension code. Incomplete support for subject and issuer alt 960155714Skris name, issuer and authority key id. Change the i2v function parameters 960255714Skris and add an extra 'crl' parameter in the X509V3_CTX structure: guess 960355714Skris what that's for :-) Fix to ASN1 macro which messed up 960455714Skris IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 960555714Skris [Steve Henson] 960655714Skris 960755714Skris *) Preliminary support for ENUMERATED type. This is largely copied from the 960855714Skris INTEGER code. 960955714Skris [Steve Henson] 961055714Skris 961155714Skris *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 961255714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 961355714Skris 961455714Skris *) Make sure `make rehash' target really finds the `openssl' program. 961555714Skris [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 961655714Skris 961755714Skris *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 961855714Skris like to hear about it if this slows down other processors. 961955714Skris [Ben Laurie] 962055714Skris 962155714Skris *) Add CygWin32 platform information to Configure script. 962255714Skris [Alan Batie <batie@aahz.jf.intel.com>] 962355714Skris 962455714Skris *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 962555714Skris [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 962655714Skris 962755714Skris *) New program nseq to manipulate netscape certificate sequences 962855714Skris [Steve Henson] 962955714Skris 963055714Skris *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 963155714Skris few typos. 963255714Skris [Steve Henson] 963355714Skris 963455714Skris *) Fixes to BN code. Previously the default was to define BN_RECURSION 963555714Skris but the BN code had some problems that would cause failures when 963655714Skris doing certificate verification and some other functions. 963755714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 963855714Skris 963955714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 964055714Skris [Steve Henson] 964155714Skris 964255714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 964355714Skris [Steve Henson] 964455714Skris 964555714Skris *) Add several PKIX and private extended key usage OIDs. 964655714Skris [Steve Henson] 964755714Skris 964855714Skris *) Modify the 'ca' program to handle the new extension code. Modify 964955714Skris openssl.cnf for new extension format, add comments. 965055714Skris [Steve Henson] 965155714Skris 965255714Skris *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 965355714Skris and add a sample to openssl.cnf so req -x509 now adds appropriate 965455714Skris CA extensions. 965555714Skris [Steve Henson] 965655714Skris 965755714Skris *) Continued X509 V3 changes. Add to other makefiles, integrate with the 965855714Skris error code, add initial support to X509_print() and x509 application. 965955714Skris [Steve Henson] 966055714Skris 966155714Skris *) Takes a deep breath and start addding X509 V3 extension support code. Add 966255714Skris files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 966355714Skris stuff is currently isolated and isn't even compiled yet. 966455714Skris [Steve Henson] 966555714Skris 966655714Skris *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 966755714Skris ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 966855714Skris Removed the versions check from X509 routines when loading extensions: 966955714Skris this allows certain broken certificates that don't set the version 967055714Skris properly to be processed. 967155714Skris [Steve Henson] 967255714Skris 967355714Skris *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 967455714Skris Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 967555714Skris can still be regenerated with "make depend". 967655714Skris [Ben Laurie] 967755714Skris 967855714Skris *) Spelling mistake in C version of CAST-128. 967955714Skris [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 968055714Skris 968155714Skris *) Changes to the error generation code. The perl script err-code.pl 968255714Skris now reads in the old error codes and retains the old numbers, only 968355714Skris adding new ones if necessary. It also only changes the .err files if new 968455714Skris codes are added. The makefiles have been modified to only insert errors 968555714Skris when needed (to avoid needlessly modifying header files). This is done 968655714Skris by only inserting errors if the .err file is newer than the auto generated 968755714Skris C file. To rebuild all the error codes from scratch (the old behaviour) 968855714Skris either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 968955714Skris or delete all the .err files. 969055714Skris [Steve Henson] 969155714Skris 969255714Skris *) CAST-128 was incorrectly implemented for short keys. The C version has 969355714Skris been fixed, but is untested. The assembler versions are also fixed, but 969455714Skris new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 969555714Skris to regenerate it if needed. 969655714Skris [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 969755714Skris Hagino <itojun@kame.net>] 969855714Skris 969955714Skris *) File was opened incorrectly in randfile.c. 970055714Skris [Ulf M�ller <ulf@fitug.de>] 970155714Skris 970255714Skris *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 970355714Skris functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 970455714Skris GeneralizedTime. ASN1_TIME is the proper type used in certificates et 970555714Skris al: it's just almost always a UTCTime. Note this patch adds new error 970655714Skris codes so do a "make errors" if there are problems. 970755714Skris [Steve Henson] 970855714Skris 970955714Skris *) Correct Linux 1 recognition in config. 971055714Skris [Ulf M�ller <ulf@fitug.de>] 971155714Skris 971255714Skris *) Remove pointless MD5 hash when using DSA keys in ca. 971355714Skris [Anonymous <nobody@replay.com>] 971455714Skris 971555714Skris *) Generate an error if given an empty string as a cert directory. Also 971655714Skris generate an error if handed NULL (previously returned 0 to indicate an 971755714Skris error, but didn't set one). 971855714Skris [Ben Laurie, reported by Anonymous <nobody@replay.com>] 971955714Skris 972055714Skris *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 972155714Skris [Ben Laurie] 972255714Skris 972355714Skris *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 972455714Skris parameters. This was causing a warning which killed off the Win32 compile. 972555714Skris [Steve Henson] 972655714Skris 972755714Skris *) Remove C++ style comments from crypto/bn/bn_local.h. 972855714Skris [Neil Costigan <neil.costigan@celocom.com>] 972955714Skris 973055714Skris *) The function OBJ_txt2nid was broken. It was supposed to return a nid 973155714Skris based on a text string, looking up short and long names and finally 973255714Skris "dot" format. The "dot" format stuff didn't work. Added new function 973355714Skris OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 973455714Skris OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 973555714Skris OID is not part of the table. 973655714Skris [Steve Henson] 973755714Skris 973855714Skris *) Add prototypes to X509 lookup/verify methods, fixing a bug in 973955714Skris X509_LOOKUP_by_alias(). 974055714Skris [Ben Laurie] 974155714Skris 974255714Skris *) Sort openssl functions by name. 974355714Skris [Ben Laurie] 974455714Skris 974555714Skris *) Get the gendsa program working (hopefully) and add it to app list. Remove 974655714Skris encryption from sample DSA keys (in case anyone is interested the password 974755714Skris was "1234"). 974855714Skris [Steve Henson] 974955714Skris 975055714Skris *) Make _all_ *_free functions accept a NULL pointer. 975155714Skris [Frans Heymans <fheymans@isaserver.be>] 975255714Skris 975355714Skris *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 975455714Skris NULL pointers. 975555714Skris [Anonymous <nobody@replay.com>] 975655714Skris 975755714Skris *) s_server should send the CAfile as acceptable CAs, not its own cert. 975855714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 975955714Skris 976055714Skris *) Don't blow it for numeric -newkey arguments to apps/req. 976155714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 976255714Skris 976355714Skris *) Temp key "for export" tests were wrong in s3_srvr.c. 976455714Skris [Anonymous <nobody@replay.com>] 976555714Skris 976655714Skris *) Add prototype for temp key callback functions 976755714Skris SSL_CTX_set_tmp_{rsa,dh}_callback(). 976855714Skris [Ben Laurie] 976955714Skris 977055714Skris *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 977155714Skris DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 977255714Skris [Steve Henson] 977355714Skris 977455714Skris *) X509_name_add_entry() freed the wrong thing after an error. 977555714Skris [Arne Ansper <arne@ats.cyber.ee>] 977655714Skris 977755714Skris *) rsa_eay.c would attempt to free a NULL context. 977855714Skris [Arne Ansper <arne@ats.cyber.ee>] 977955714Skris 978055714Skris *) BIO_s_socket() had a broken should_retry() on Windoze. 978155714Skris [Arne Ansper <arne@ats.cyber.ee>] 978255714Skris 978355714Skris *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 978455714Skris [Arne Ansper <arne@ats.cyber.ee>] 978555714Skris 978655714Skris *) Make sure the already existing X509_STORE->depth variable is initialized 978755714Skris in X509_STORE_new(), but document the fact that this variable is still 978855714Skris unused in the certificate verification process. 978955714Skris [Ralf S. Engelschall] 979055714Skris 979155714Skris *) Fix the various library and apps files to free up pkeys obtained from 979255714Skris X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 979355714Skris [Steve Henson] 979455714Skris 979555714Skris *) Fix reference counting in X509_PUBKEY_get(). This makes 979655714Skris demos/maurice/example2.c work, amongst others, probably. 979755714Skris [Steve Henson and Ben Laurie] 979855714Skris 979955714Skris *) First cut of a cleanup for apps/. First the `ssleay' program is now named 980055714Skris `openssl' and second, the shortcut symlinks for the `openssl <command>' 980155714Skris are no longer created. This way we have a single and consistent command 980255714Skris line interface `openssl <command>', similar to `cvs <command>'. 980355714Skris [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 980455714Skris 980555714Skris *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 980655714Skris BIT STRING wrapper always have zero unused bits. 980755714Skris [Steve Henson] 980855714Skris 980955714Skris *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 981055714Skris [Steve Henson] 981155714Skris 981255714Skris *) Make the top-level INSTALL documentation easier to understand. 981355714Skris [Paul Sutton] 981455714Skris 981555714Skris *) Makefiles updated to exit if an error occurs in a sub-directory 981655714Skris make (including if user presses ^C) [Paul Sutton] 981755714Skris 981855714Skris *) Make Montgomery context stuff explicit in RSA data structure. 981955714Skris [Ben Laurie] 982055714Skris 982155714Skris *) Fix build order of pem and err to allow for generated pem.h. 982255714Skris [Ben Laurie] 982355714Skris 982455714Skris *) Fix renumbering bug in X509_NAME_delete_entry(). 982555714Skris [Ben Laurie] 982655714Skris 982755714Skris *) Enhanced the err-ins.pl script so it makes the error library number 982855714Skris global and can add a library name. This is needed for external ASN1 and 982955714Skris other error libraries. 983055714Skris [Steve Henson] 983155714Skris 983255714Skris *) Fixed sk_insert which never worked properly. 983355714Skris [Steve Henson] 983455714Skris 983555714Skris *) Fix ASN1 macros so they can handle indefinite length construted 983655714Skris EXPLICIT tags. Some non standard certificates use these: they can now 983755714Skris be read in. 983855714Skris [Steve Henson] 983955714Skris 984055714Skris *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 984155714Skris into a single doc/ssleay.txt bundle. This way the information is still 984255714Skris preserved but no longer messes up this directory. Now it's new room for 984355714Skris the new set of documenation files. 984455714Skris [Ralf S. Engelschall] 984555714Skris 984655714Skris *) SETs were incorrectly DER encoded. This was a major pain, because they 984755714Skris shared code with SEQUENCEs, which aren't coded the same. This means that 984855714Skris almost everything to do with SETs or SEQUENCEs has either changed name or 984955714Skris number of arguments. 985055714Skris [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 985155714Skris 985255714Skris *) Fix test data to work with the above. 985355714Skris [Ben Laurie] 985455714Skris 985555714Skris *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 985655714Skris was already fixed by Eric for 0.9.1 it seems. 985755714Skris [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 985855714Skris 985955714Skris *) Autodetect FreeBSD3. 986055714Skris [Ben Laurie] 986155714Skris 986255714Skris *) Fix various bugs in Configure. This affects the following platforms: 986355714Skris nextstep 986455714Skris ncr-scde 986555714Skris unixware-2.0 986655714Skris unixware-2.0-pentium 986755714Skris sco5-cc. 986855714Skris [Ben Laurie] 986955714Skris 987055714Skris *) Eliminate generated files from CVS. Reorder tests to regenerate files 987155714Skris before they are needed. 987255714Skris [Ben Laurie] 987355714Skris 987455714Skris *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 987555714Skris [Ben Laurie] 987655714Skris 987755714Skris 987855714Skris Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 987955714Skris 988055714Skris *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 988155714Skris changed SSLeay to OpenSSL in version strings. 988255714Skris [Ralf S. Engelschall] 988355714Skris 988455714Skris *) Some fixups to the top-level documents. 988555714Skris [Paul Sutton] 988655714Skris 988755714Skris *) Fixed the nasty bug where rsaref.h was not found under compile-time 988855714Skris because the symlink to include/ was missing. 988955714Skris [Ralf S. Engelschall] 989055714Skris 989155714Skris *) Incorporated the popular no-RSA/DSA-only patches 989255714Skris which allow to compile a RSA-free SSLeay. 989355714Skris [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 989455714Skris 989555714Skris *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 989655714Skris when "ssleay" is still not found. 989755714Skris [Ralf S. Engelschall] 989855714Skris 989955714Skris *) Added more platforms to Configure: Cray T3E, HPUX 11, 990055714Skris [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 990155714Skris 990255714Skris *) Updated the README file. 990355714Skris [Ralf S. Engelschall] 990455714Skris 990555714Skris *) Added various .cvsignore files in the CVS repository subdirs 990655714Skris to make a "cvs update" really silent. 990755714Skris [Ralf S. Engelschall] 990855714Skris 990955714Skris *) Recompiled the error-definition header files and added 991055714Skris missing symbols to the Win32 linker tables. 991155714Skris [Ralf S. Engelschall] 991255714Skris 991355714Skris *) Cleaned up the top-level documents; 991455714Skris o new files: CHANGES and LICENSE 991555714Skris o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 991655714Skris o merged COPYRIGHT into LICENSE 991755714Skris o removed obsolete TODO file 991855714Skris o renamed MICROSOFT to INSTALL.W32 991955714Skris [Ralf S. Engelschall] 992055714Skris 992155714Skris *) Removed dummy files from the 0.9.1b source tree: 992255714Skris crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 992355714Skris crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 992455714Skris crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 992555714Skris crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 992655714Skris util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 992755714Skris [Ralf S. Engelschall] 992855714Skris 992955714Skris *) Added various platform portability fixes. 993055714Skris [Mark J. Cox] 993155714Skris 993255714Skris *) The Genesis of the OpenSSL rpject: 993355714Skris We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 993455714Skris Young and Tim J. Hudson created while they were working for C2Net until 993555714Skris summer 1998. 993655714Skris [The OpenSSL Project] 993755714Skris 993855714Skris 993955714Skris Changes between 0.9.0b and 0.9.1b [not released] 994055714Skris 994155714Skris *) Updated a few CA certificates under certs/ 994255714Skris [Eric A. Young] 994355714Skris 994455714Skris *) Changed some BIGNUM api stuff. 994555714Skris [Eric A. Young] 994655714Skris 994755714Skris *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 994855714Skris DGUX x86, Linux Alpha, etc. 994955714Skris [Eric A. Young] 995055714Skris 995155714Skris *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 995255714Skris RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 995355714Skris available). 995455714Skris [Eric A. Young] 995555714Skris 995655714Skris *) Add -strparse option to asn1pars program which parses nested 995755714Skris binary structures 995855714Skris [Dr Stephen Henson <shenson@bigfoot.com>] 995955714Skris 996055714Skris *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 996155714Skris [Eric A. Young] 996255714Skris 996355714Skris *) DSA fix for "ca" program. 996455714Skris [Eric A. Young] 996555714Skris 996655714Skris *) Added "-genkey" option to "dsaparam" program. 996755714Skris [Eric A. Young] 996855714Skris 996955714Skris *) Added RIPE MD160 (rmd160) message digest. 997055714Skris [Eric A. Young] 997155714Skris 997255714Skris *) Added -a (all) option to "ssleay version" command. 997355714Skris [Eric A. Young] 997455714Skris 997555714Skris *) Added PLATFORM define which is the id given to Configure. 997655714Skris [Eric A. Young] 997755714Skris 997855714Skris *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 997955714Skris [Eric A. Young] 998055714Skris 998155714Skris *) Extended the ASN.1 parser routines. 998255714Skris [Eric A. Young] 998355714Skris 998455714Skris *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 998555714Skris [Eric A. Young] 998655714Skris 998755714Skris *) Added a BN_CTX to the BN library. 998855714Skris [Eric A. Young] 998955714Skris 999055714Skris *) Fixed the weak key values in DES library 999155714Skris [Eric A. Young] 999255714Skris 999355714Skris *) Changed API in EVP library for cipher aliases. 999455714Skris [Eric A. Young] 999555714Skris 999655714Skris *) Added support for RC2/64bit cipher. 999755714Skris [Eric A. Young] 999855714Skris 999955714Skris *) Converted the lhash library to the crypto/mem.c functions. 1000055714Skris [Eric A. Young] 1000155714Skris 1000255714Skris *) Added more recognized ASN.1 object ids. 1000355714Skris [Eric A. Young] 1000455714Skris 1000555714Skris *) Added more RSA padding checks for SSL/TLS. 1000655714Skris [Eric A. Young] 1000755714Skris 1000855714Skris *) Added BIO proxy/filter functionality. 1000955714Skris [Eric A. Young] 1001055714Skris 1001155714Skris *) Added extra_certs to SSL_CTX which can be used 1001255714Skris send extra CA certificates to the client in the CA cert chain sending 1001355714Skris process. It can be configured with SSL_CTX_add_extra_chain_cert(). 1001455714Skris [Eric A. Young] 1001555714Skris 1001655714Skris *) Now Fortezza is denied in the authentication phase because 1001755714Skris this is key exchange mechanism is not supported by SSLeay at all. 1001855714Skris [Eric A. Young] 1001955714Skris 1002055714Skris *) Additional PKCS1 checks. 1002155714Skris [Eric A. Young] 1002255714Skris 1002355714Skris *) Support the string "TLSv1" for all TLS v1 ciphers. 1002455714Skris [Eric A. Young] 1002555714Skris 1002655714Skris *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 1002755714Skris ex_data index of the SSL context in the X509_STORE_CTX ex_data. 1002855714Skris [Eric A. Young] 1002955714Skris 1003055714Skris *) Fixed a few memory leaks. 1003155714Skris [Eric A. Young] 1003255714Skris 1003355714Skris *) Fixed various code and comment typos. 1003455714Skris [Eric A. Young] 1003555714Skris 1003655714Skris *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 1003755714Skris bytes sent in the client random. 1003855714Skris [Edward Bishop <ebishop@spyglass.com>] 1003955714Skris 10040