CHANGES revision 194206
155714Skris 255714Skris OpenSSL CHANGES 355714Skris _______________ 455714Skris 5194206Ssimon Changes between 0.9.8j and 0.9.8k [25 Mar 2009] 6167612Ssimon 7194206Ssimon *) Don't set val to NULL when freeing up structures, it is freed up by 8194206Ssimon underlying code. If sizeof(void *) > sizeof(long) this can result in 9194206Ssimon zeroing past the valid field. (CVE-2009-0789) 10194206Ssimon [Paolo Ganci <Paolo.Ganci@AdNovum.CH>] 11194206Ssimon 12194206Ssimon *) Fix bug where return value of CMS_SignerInfo_verify_content() was not 13194206Ssimon checked correctly. This would allow some invalid signed attributes to 14194206Ssimon appear to verify correctly. (CVE-2009-0591) 15194206Ssimon [Ivan Nestlerode <inestlerode@us.ibm.com>] 16194206Ssimon 17194206Ssimon *) Reject UniversalString and BMPString types with invalid lengths. This 18194206Ssimon prevents a crash in ASN1_STRING_print_ex() which assumes the strings have 19194206Ssimon a legal length. (CVE-2009-0590) 20194206Ssimon [Steve Henson] 21194206Ssimon 22194206Ssimon *) Set S/MIME signing as the default purpose rather than setting it 23194206Ssimon unconditionally. This allows applications to override it at the store 24194206Ssimon level. 25194206Ssimon [Steve Henson] 26194206Ssimon 27194206Ssimon *) Permit restricted recursion of ASN1 strings. This is needed in practice 28194206Ssimon to handle some structures. 29194206Ssimon [Steve Henson] 30194206Ssimon 31194206Ssimon *) Improve efficiency of mem_gets: don't search whole buffer each time 32194206Ssimon for a '\n' 33194206Ssimon [Jeremy Shapiro <jnshapir@us.ibm.com>] 34194206Ssimon 35194206Ssimon *) New -hex option for openssl rand. 36194206Ssimon [Matthieu Herrb] 37194206Ssimon 38194206Ssimon *) Print out UTF8String and NumericString when parsing ASN1. 39194206Ssimon [Steve Henson] 40194206Ssimon 41194206Ssimon *) Support NumericString type for name components. 42194206Ssimon [Steve Henson] 43194206Ssimon 44194206Ssimon *) Allow CC in the environment to override the automatically chosen 45194206Ssimon compiler. Note that nothing is done to ensure flags work with the 46194206Ssimon chosen compiler. 47194206Ssimon [Ben Laurie] 48194206Ssimon 49194206Ssimon Changes between 0.9.8i and 0.9.8j [07 Jan 2009] 50194206Ssimon 51194206Ssimon *) Properly check EVP_VerifyFinal() and similar return values 52194206Ssimon (CVE-2008-5077). 53194206Ssimon [Ben Laurie, Bodo Moeller, Google Security Team] 54194206Ssimon 55194206Ssimon *) Enable TLS extensions by default. 56194206Ssimon [Ben Laurie] 57194206Ssimon 58194206Ssimon *) Allow the CHIL engine to be loaded, whether the application is 59194206Ssimon multithreaded or not. (This does not release the developer from the 60194206Ssimon obligation to set up the dynamic locking callbacks.) 61194206Ssimon [Sander Temme <sander@temme.net>] 62194206Ssimon 63194206Ssimon *) Use correct exit code if there is an error in dgst command. 64194206Ssimon [Steve Henson; problem pointed out by Roland Dirlewanger] 65194206Ssimon 66194206Ssimon *) Tweak Configure so that you need to say "experimental-jpake" to enable 67194206Ssimon JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications. 68194206Ssimon [Bodo Moeller] 69194206Ssimon 70194206Ssimon *) Add experimental JPAKE support, including demo authentication in 71194206Ssimon s_client and s_server. 72194206Ssimon [Ben Laurie] 73194206Ssimon 74194206Ssimon *) Set the comparison function in v3_addr_canonize(). 75194206Ssimon [Rob Austein <sra@hactrn.net>] 76194206Ssimon 77194206Ssimon *) Add support for XMPP STARTTLS in s_client. 78194206Ssimon [Philip Paeps <philip@freebsd.org>] 79194206Ssimon 80194206Ssimon *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior 81194206Ssimon to ensure that even with this option, only ciphersuites in the 82194206Ssimon server's preference list will be accepted. (Note that the option 83194206Ssimon applies only when resuming a session, so the earlier behavior was 84194206Ssimon just about the algorithm choice for symmetric cryptography.) 85194206Ssimon [Bodo Moeller] 86194206Ssimon 87194206Ssimon Changes between 0.9.8h and 0.9.8i [15 Sep 2008] 88194206Ssimon 89194206Ssimon *) Fix a state transitition in s3_srvr.c and d1_srvr.c 90194206Ssimon (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...). 91194206Ssimon [Nagendra Modadugu] 92194206Ssimon 93194206Ssimon *) The fix in 0.9.8c that supposedly got rid of unsafe 94194206Ssimon double-checked locking was incomplete for RSA blinding, 95194206Ssimon addressing just one layer of what turns out to have been 96194206Ssimon doubly unsafe triple-checked locking. 97194206Ssimon 98194206Ssimon So now fix this for real by retiring the MONT_HELPER macro 99194206Ssimon in crypto/rsa/rsa_eay.c. 100194206Ssimon 101194206Ssimon [Bodo Moeller; problem pointed out by Marius Schilder] 102194206Ssimon 103194206Ssimon *) Various precautionary measures: 104194206Ssimon 105194206Ssimon - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h). 106194206Ssimon 107194206Ssimon - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c). 108194206Ssimon (NB: This would require knowledge of the secret session ticket key 109194206Ssimon to exploit, in which case you'd be SOL either way.) 110194206Ssimon 111194206Ssimon - Change bn_nist.c so that it will properly handle input BIGNUMs 112194206Ssimon outside the expected range. 113194206Ssimon 114194206Ssimon - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG 115194206Ssimon builds. 116194206Ssimon 117194206Ssimon [Neel Mehta, Bodo Moeller] 118194206Ssimon 119194206Ssimon *) Allow engines to be "soft loaded" - i.e. optionally don't die if 120194206Ssimon the load fails. Useful for distros. 121194206Ssimon [Ben Laurie and the FreeBSD team] 122194206Ssimon 123194206Ssimon *) Add support for Local Machine Keyset attribute in PKCS#12 files. 124194206Ssimon [Steve Henson] 125194206Ssimon 126194206Ssimon *) Fix BN_GF2m_mod_arr() top-bit cleanup code. 127194206Ssimon [Huang Ying] 128194206Ssimon 129194206Ssimon *) Expand ENGINE to support engine supplied SSL client certificate functions. 130194206Ssimon 131194206Ssimon This work was sponsored by Logica. 132194206Ssimon [Steve Henson] 133194206Ssimon 134194206Ssimon *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows 135194206Ssimon keystores. Support for SSL/TLS client authentication too. 136194206Ssimon Not compiled unless enable-capieng specified to Configure. 137194206Ssimon 138194206Ssimon This work was sponsored by Logica. 139194206Ssimon [Steve Henson] 140194206Ssimon 141194206Ssimon *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using 142194206Ssimon ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain 143194206Ssimon attribute creation routines such as certifcate requests and PKCS#12 144194206Ssimon files. 145194206Ssimon [Steve Henson] 146194206Ssimon 147194206Ssimon Changes between 0.9.8g and 0.9.8h [28 May 2008] 148194206Ssimon 149194206Ssimon *) Fix flaw if 'Server Key exchange message' is omitted from a TLS 150194206Ssimon handshake which could lead to a cilent crash as found using the 151194206Ssimon Codenomicon TLS test suite (CVE-2008-1672) 152194206Ssimon [Steve Henson, Mark Cox] 153194206Ssimon 154194206Ssimon *) Fix double free in TLS server name extensions which could lead to 155194206Ssimon a remote crash found by Codenomicon TLS test suite (CVE-2008-0891) 156194206Ssimon [Joe Orton] 157194206Ssimon 158194206Ssimon *) Clear error queue in SSL_CTX_use_certificate_chain_file() 159194206Ssimon 160194206Ssimon Clear the error queue to ensure that error entries left from 161194206Ssimon older function calls do not interfere with the correct operation. 162194206Ssimon [Lutz Jaenicke, Erik de Castro Lopo] 163194206Ssimon 164194206Ssimon *) Remove root CA certificates of commercial CAs: 165194206Ssimon 166194206Ssimon The OpenSSL project does not recommend any specific CA and does not 167194206Ssimon have any policy with respect to including or excluding any CA. 168194206Ssimon Therefore it does not make any sense to ship an arbitrary selection 169194206Ssimon of root CA certificates with the OpenSSL software. 170194206Ssimon [Lutz Jaenicke] 171194206Ssimon 172194206Ssimon *) RSA OAEP patches to fix two separate invalid memory reads. 173194206Ssimon The first one involves inputs when 'lzero' is greater than 174194206Ssimon 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes 175194206Ssimon before the beginning of from). The second one involves inputs where 176194206Ssimon the 'db' section contains nothing but zeroes (there is a one-byte 177194206Ssimon invalid read after the end of 'db'). 178194206Ssimon [Ivan Nestlerode <inestlerode@us.ibm.com>] 179194206Ssimon 180194206Ssimon *) Partial backport from 0.9.9-dev: 181194206Ssimon 182194206Ssimon Introduce bn_mul_mont (dedicated Montgomery multiplication 183194206Ssimon procedure) as a candidate for BIGNUM assembler implementation. 184194206Ssimon While 0.9.9-dev uses assembler for various architectures, only 185194206Ssimon x86_64 is available by default here in the 0.9.8 branch, and 186194206Ssimon 32-bit x86 is available through a compile-time setting. 187194206Ssimon 188194206Ssimon To try the 32-bit x86 assembler implementation, use Configure 189194206Ssimon option "enable-montasm" (which exists only for this backport). 190194206Ssimon 191194206Ssimon As "enable-montasm" for 32-bit x86 disclaims code stability 192194206Ssimon anyway, in this constellation we activate additional code 193194206Ssimon backported from 0.9.9-dev for further performance improvements, 194194206Ssimon namely BN_from_montgomery_word. (To enable this otherwise, 195194206Ssimon e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".) 196194206Ssimon 197194206Ssimon [Andy Polyakov (backport partially by Bodo Moeller)] 198194206Ssimon 199194206Ssimon *) Add TLS session ticket callback. This allows an application to set 200194206Ssimon TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed 201194206Ssimon values. This is useful for key rollover for example where several key 202194206Ssimon sets may exist with different names. 203194206Ssimon [Steve Henson] 204194206Ssimon 205194206Ssimon *) Reverse ENGINE-internal logic for caching default ENGINE handles. 206194206Ssimon This was broken until now in 0.9.8 releases, such that the only way 207194206Ssimon a registered ENGINE could be used (assuming it initialises 208194206Ssimon successfully on the host) was to explicitly set it as the default 209194206Ssimon for the relevant algorithms. This is in contradiction with 0.9.7 210194206Ssimon behaviour and the documentation. With this fix, when an ENGINE is 211194206Ssimon registered into a given algorithm's table of implementations, the 212194206Ssimon 'uptodate' flag is reset so that auto-discovery will be used next 213194206Ssimon time a new context for that algorithm attempts to select an 214194206Ssimon implementation. 215194206Ssimon [Ian Lister (tweaked by Geoff Thorpe)] 216194206Ssimon 217194206Ssimon *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9 218194206Ssimon implemention in the following ways: 219194206Ssimon 220194206Ssimon Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be 221194206Ssimon hard coded. 222194206Ssimon 223194206Ssimon Lack of BER streaming support means one pass streaming processing is 224194206Ssimon only supported if data is detached: setting the streaming flag is 225194206Ssimon ignored for embedded content. 226194206Ssimon 227194206Ssimon CMS support is disabled by default and must be explicitly enabled 228194206Ssimon with the enable-cms configuration option. 229194206Ssimon [Steve Henson] 230194206Ssimon 231194206Ssimon *) Update the GMP engine glue to do direct copies between BIGNUM and 232194206Ssimon mpz_t when openssl and GMP use the same limb size. Otherwise the 233194206Ssimon existing "conversion via a text string export" trick is still used. 234194206Ssimon [Paul Sheer <paulsheer@gmail.com>] 235194206Ssimon 236194206Ssimon *) Zlib compression BIO. This is a filter BIO which compressed and 237194206Ssimon uncompresses any data passed through it. 238194206Ssimon [Steve Henson] 239194206Ssimon 240194206Ssimon *) Add AES_wrap_key() and AES_unwrap_key() functions to implement 241194206Ssimon RFC3394 compatible AES key wrapping. 242194206Ssimon [Steve Henson] 243194206Ssimon 244194206Ssimon *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0(): 245194206Ssimon sets string data without copying. X509_ALGOR_set0() and 246194206Ssimon X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier) 247194206Ssimon data. Attribute function X509at_get0_data_by_OBJ(): retrieves data 248194206Ssimon from an X509_ATTRIBUTE structure optionally checking it occurs only 249194206Ssimon once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied 250194206Ssimon data. 251194206Ssimon [Steve Henson] 252194206Ssimon 253194206Ssimon *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set() 254194206Ssimon to get the expected BN_FLG_CONSTTIME behavior. 255194206Ssimon [Bodo Moeller (Google)] 256194206Ssimon 257194206Ssimon *) Netware support: 258194206Ssimon 259194206Ssimon - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets 260194206Ssimon - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT) 261194206Ssimon - added some more tests to do_tests.pl 262194206Ssimon - fixed RunningProcess usage so that it works with newer LIBC NDKs too 263194206Ssimon - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency 264194206Ssimon - added new Configure targets netware-clib-bsdsock, netware-clib-gcc, 265194206Ssimon netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc 266194206Ssimon - various changes to netware.pl to enable gcc-cross builds on Win32 267194206Ssimon platform 268194206Ssimon - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD) 269194206Ssimon - various changes to fix missing prototype warnings 270194206Ssimon - fixed x86nasm.pl to create correct asm files for NASM COFF output 271194206Ssimon - added AES, WHIRLPOOL and CPUID assembler code to build files 272194206Ssimon - added missing AES assembler make rules to mk1mf.pl 273194206Ssimon - fixed order of includes in apps/ocsp.c so that e_os.h settings apply 274194206Ssimon [Guenter Knauf <eflash@gmx.net>] 275194206Ssimon 276194206Ssimon *) Implement certificate status request TLS extension defined in RFC3546. 277194206Ssimon A client can set the appropriate parameters and receive the encoded 278194206Ssimon OCSP response via a callback. A server can query the supplied parameters 279194206Ssimon and set the encoded OCSP response in the callback. Add simplified examples 280194206Ssimon to s_client and s_server. 281194206Ssimon [Steve Henson] 282194206Ssimon 283194206Ssimon Changes between 0.9.8f and 0.9.8g [19 Oct 2007] 284194206Ssimon 285194206Ssimon *) Fix various bugs: 286194206Ssimon + Binary incompatibility of ssl_ctx_st structure 287194206Ssimon + DTLS interoperation with non-compliant servers 288194206Ssimon + Don't call get_session_cb() without proposed session 289194206Ssimon + Fix ia64 assembler code 290194206Ssimon [Andy Polyakov, Steve Henson] 291194206Ssimon 292194206Ssimon Changes between 0.9.8e and 0.9.8f [11 Oct 2007] 293194206Ssimon 294194206Ssimon *) DTLS Handshake overhaul. There were longstanding issues with 295194206Ssimon OpenSSL DTLS implementation, which were making it impossible for 296194206Ssimon RFC 4347 compliant client to communicate with OpenSSL server. 297194206Ssimon Unfortunately just fixing these incompatibilities would "cut off" 298194206Ssimon pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e 299194206Ssimon server keeps tolerating non RFC compliant syntax. The opposite is 300194206Ssimon not true, 0.9.8f client can not communicate with earlier server. 301194206Ssimon This update even addresses CVE-2007-4995. 302194206Ssimon [Andy Polyakov] 303194206Ssimon 304194206Ssimon *) Changes to avoid need for function casts in OpenSSL: some compilers 305194206Ssimon (gcc 4.2 and later) reject their use. 306194206Ssimon [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>, 307194206Ssimon Steve Henson] 308194206Ssimon 309194206Ssimon *) Add RFC4507 support to OpenSSL. This includes the corrections in 310194206Ssimon RFC4507bis. The encrypted ticket format is an encrypted encoded 311194206Ssimon SSL_SESSION structure, that way new session features are automatically 312194206Ssimon supported. 313194206Ssimon 314194206Ssimon If a client application caches session in an SSL_SESSION structure 315194206Ssimon support is transparent because tickets are now stored in the encoded 316194206Ssimon SSL_SESSION. 317194206Ssimon 318194206Ssimon The SSL_CTX structure automatically generates keys for ticket 319194206Ssimon protection in servers so again support should be possible 320194206Ssimon with no application modification. 321194206Ssimon 322194206Ssimon If a client or server wishes to disable RFC4507 support then the option 323194206Ssimon SSL_OP_NO_TICKET can be set. 324194206Ssimon 325194206Ssimon Add a TLS extension debugging callback to allow the contents of any client 326194206Ssimon or server extensions to be examined. 327194206Ssimon 328194206Ssimon This work was sponsored by Google. 329194206Ssimon [Steve Henson] 330194206Ssimon 331194206Ssimon *) Add initial support for TLS extensions, specifically for the server_name 332194206Ssimon extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 333194206Ssimon have new members for a host name. The SSL data structure has an 334194206Ssimon additional member SSL_CTX *initial_ctx so that new sessions can be 335194206Ssimon stored in that context to allow for session resumption, even after the 336194206Ssimon SSL has been switched to a new SSL_CTX in reaction to a client's 337194206Ssimon server_name extension. 338194206Ssimon 339194206Ssimon New functions (subject to change): 340194206Ssimon 341194206Ssimon SSL_get_servername() 342194206Ssimon SSL_get_servername_type() 343194206Ssimon SSL_set_SSL_CTX() 344194206Ssimon 345194206Ssimon New CTRL codes and macros (subject to change): 346194206Ssimon 347194206Ssimon SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 348194206Ssimon - SSL_CTX_set_tlsext_servername_callback() 349194206Ssimon SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 350194206Ssimon - SSL_CTX_set_tlsext_servername_arg() 351194206Ssimon SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 352194206Ssimon 353194206Ssimon openssl s_client has a new '-servername ...' option. 354194206Ssimon 355194206Ssimon openssl s_server has new options '-servername_host ...', '-cert2 ...', 356194206Ssimon '-key2 ...', '-servername_fatal' (subject to change). This allows 357194206Ssimon testing the HostName extension for a specific single host name ('-cert' 358194206Ssimon and '-key' remain fallbacks for handshakes without HostName 359194206Ssimon negotiation). If the unrecogninzed_name alert has to be sent, this by 360194206Ssimon default is a warning; it becomes fatal with the '-servername_fatal' 361194206Ssimon option. 362194206Ssimon 363194206Ssimon [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson] 364194206Ssimon 365194206Ssimon *) Add AES and SSE2 assembly language support to VC++ build. 366194206Ssimon [Steve Henson] 367194206Ssimon 368194206Ssimon *) Mitigate attack on final subtraction in Montgomery reduction. 369194206Ssimon [Andy Polyakov] 370194206Ssimon 371194206Ssimon *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0 372194206Ssimon (which previously caused an internal error). 373194206Ssimon [Bodo Moeller] 374194206Ssimon 375194206Ssimon *) Squeeze another 10% out of IGE mode when in != out. 376194206Ssimon [Ben Laurie] 377194206Ssimon 378194206Ssimon *) AES IGE mode speedup. 379194206Ssimon [Dean Gaudet (Google)] 380194206Ssimon 381194206Ssimon *) Add the Korean symmetric 128-bit cipher SEED (see 382194206Ssimon http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and 383194206Ssimon add SEED ciphersuites from RFC 4162: 384194206Ssimon 385194206Ssimon TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA" 386194206Ssimon TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA" 387194206Ssimon TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA" 388194206Ssimon TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA" 389194206Ssimon 390194206Ssimon To minimize changes between patchlevels in the OpenSSL 0.9.8 391194206Ssimon series, SEED remains excluded from compilation unless OpenSSL 392194206Ssimon is configured with 'enable-seed'. 393194206Ssimon [KISA, Bodo Moeller] 394194206Ssimon 395194206Ssimon *) Mitigate branch prediction attacks, which can be practical if a 396194206Ssimon single processor is shared, allowing a spy process to extract 397194206Ssimon information. For detailed background information, see 398194206Ssimon http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron, 399194206Ssimon J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL 400194206Ssimon and Necessary Software Countermeasures"). The core of the change 401194206Ssimon are new versions BN_div_no_branch() and 402194206Ssimon BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(), 403194206Ssimon respectively, which are slower, but avoid the security-relevant 404194206Ssimon conditional branches. These are automatically called by BN_div() 405194206Ssimon and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one 406194206Ssimon of the input BIGNUMs. Also, BN_is_bit_set() has been changed to 407194206Ssimon remove a conditional branch. 408194206Ssimon 409194206Ssimon BN_FLG_CONSTTIME is the new name for the previous 410194206Ssimon BN_FLG_EXP_CONSTTIME flag, since it now affects more than just 411194206Ssimon modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag 412194206Ssimon in the exponent causes BN_mod_exp_mont() to use the alternative 413194206Ssimon implementation in BN_mod_exp_mont_consttime().) The old name 414194206Ssimon remains as a deprecated alias. 415194206Ssimon 416194206Ssimon Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general 417194206Ssimon RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses 418194206Ssimon constant-time implementations for more than just exponentiation. 419194206Ssimon Here too the old name is kept as a deprecated alias. 420194206Ssimon 421194206Ssimon BN_BLINDING_new() will now use BN_dup() for the modulus so that 422194206Ssimon the BN_BLINDING structure gets an independent copy of the 423194206Ssimon modulus. This means that the previous "BIGNUM *m" argument to 424194206Ssimon BN_BLINDING_new() and to BN_BLINDING_create_param() now 425194206Ssimon essentially becomes "const BIGNUM *m", although we can't actually 426194206Ssimon change this in the header file before 0.9.9. It allows 427194206Ssimon RSA_setup_blinding() to use BN_with_flags() on the modulus to 428194206Ssimon enable BN_FLG_CONSTTIME. 429194206Ssimon 430194206Ssimon [Matthew D Wood (Intel Corp)] 431194206Ssimon 432194206Ssimon *) In the SSL/TLS server implementation, be strict about session ID 433194206Ssimon context matching (which matters if an application uses a single 434194206Ssimon external cache for different purposes). Previously, 435194206Ssimon out-of-context reuse was forbidden only if SSL_VERIFY_PEER was 436194206Ssimon set. This did ensure strict client verification, but meant that, 437194206Ssimon with applications using a single external cache for quite 438194206Ssimon different requirements, clients could circumvent ciphersuite 439194206Ssimon restrictions for a given session ID context by starting a session 440194206Ssimon in a different context. 441194206Ssimon [Bodo Moeller] 442194206Ssimon 443167612Ssimon *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 444167612Ssimon a ciphersuite string such as "DEFAULT:RSA" cannot enable 445167612Ssimon authentication-only ciphersuites. 446167612Ssimon [Bodo Moeller] 447167612Ssimon 448194206Ssimon *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was 449194206Ssimon not complete and could lead to a possible single byte overflow 450194206Ssimon (CVE-2007-5135) [Ben Laurie] 451194206Ssimon 452194206Ssimon Changes between 0.9.8d and 0.9.8e [23 Feb 2007] 453194206Ssimon 454167612Ssimon *) Since AES128 and AES256 (and similarly Camellia128 and 455167612Ssimon Camellia256) share a single mask bit in the logic of 456167612Ssimon ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 457167612Ssimon kludge to work properly if AES128 is available and AES256 isn't 458167612Ssimon (or if Camellia128 is available and Camellia256 isn't). 459167612Ssimon [Victor Duchovni] 460167612Ssimon 461167612Ssimon *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c 462167612Ssimon (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters): 463167612Ssimon When a point or a seed is encoded in a BIT STRING, we need to 464167612Ssimon prevent the removal of trailing zero bits to get the proper DER 465167612Ssimon encoding. (By default, crypto/asn1/a_bitstr.c assumes the case 466167612Ssimon of a NamedBitList, for which trailing 0 bits need to be removed.) 467167612Ssimon [Bodo Moeller] 468167612Ssimon 469167612Ssimon *) Have SSL/TLS server implementation tolerate "mismatched" record 470167612Ssimon protocol version while receiving ClientHello even if the 471167612Ssimon ClientHello is fragmented. (The server can't insist on the 472167612Ssimon particular protocol version it has chosen before the ServerHello 473167612Ssimon message has informed the client about his choice.) 474167612Ssimon [Bodo Moeller] 475167612Ssimon 476167612Ssimon *) Add RFC 3779 support. 477167612Ssimon [Rob Austein for ARIN, Ben Laurie] 478167612Ssimon 479167612Ssimon *) Load error codes if they are not already present instead of using a 480167612Ssimon static variable. This allows them to be cleanly unloaded and reloaded. 481167612Ssimon Improve header file function name parsing. 482167612Ssimon [Steve Henson] 483167612Ssimon 484167612Ssimon *) extend SMTP and IMAP protocol emulation in s_client to use EHLO 485167612Ssimon or CAPABILITY handshake as required by RFCs. 486167612Ssimon [Goetz Babin-Ebell] 487167612Ssimon 488162911Ssimon Changes between 0.9.8c and 0.9.8d [28 Sep 2006] 489162911Ssimon 490162911Ssimon *) Introduce limits to prevent malicious keys being able to 491162911Ssimon cause a denial of service. (CVE-2006-2940) 492162911Ssimon [Steve Henson, Bodo Moeller] 493162911Ssimon 494162911Ssimon *) Fix ASN.1 parsing of certain invalid structures that can result 495162911Ssimon in a denial of service. (CVE-2006-2937) [Steve Henson] 496162911Ssimon 497162911Ssimon *) Fix buffer overflow in SSL_get_shared_ciphers() function. 498162911Ssimon (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 499162911Ssimon 500162911Ssimon *) Fix SSL client code which could crash if connecting to a 501162911Ssimon malicious SSLv2 server. (CVE-2006-4343) 502162911Ssimon [Tavis Ormandy and Will Drewry, Google Security Team] 503162911Ssimon 504162911Ssimon *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites 505162911Ssimon match only those. Before that, "AES256-SHA" would be interpreted 506162911Ssimon as a pattern and match "AES128-SHA" too (since AES128-SHA got 507162911Ssimon the same strength classification in 0.9.7h) as we currently only 508162911Ssimon have a single AES bit in the ciphersuite description bitmap. 509162911Ssimon That change, however, also applied to ciphersuite strings such as 510162911Ssimon "RC4-MD5" that intentionally matched multiple ciphersuites -- 511162911Ssimon namely, SSL 2.0 ciphersuites in addition to the more common ones 512162911Ssimon from SSL 3.0/TLS 1.0. 513162911Ssimon 514162911Ssimon So we change the selection algorithm again: Naming an explicit 515162911Ssimon ciphersuite selects this one ciphersuite, and any other similar 516162911Ssimon ciphersuite (same bitmap) from *other* protocol versions. 517162911Ssimon Thus, "RC4-MD5" again will properly select both the SSL 2.0 518162911Ssimon ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite. 519162911Ssimon 520162911Ssimon Since SSL 2.0 does not have any ciphersuites for which the 521162911Ssimon 128/256 bit distinction would be relevant, this works for now. 522162911Ssimon The proper fix will be to use different bits for AES128 and 523162911Ssimon AES256, which would have avoided the problems from the beginning; 524162911Ssimon however, bits are scarce, so we can only do this in a new release 525162911Ssimon (not just a patchlevel) when we can change the SSL_CIPHER 526162911Ssimon definition to split the single 'unsigned long mask' bitmap into 527162911Ssimon multiple values to extend the available space. 528162911Ssimon 529162911Ssimon [Bodo Moeller] 530162911Ssimon 531162911Ssimon Changes between 0.9.8b and 0.9.8c [05 Sep 2006] 532162911Ssimon 533162911Ssimon *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 534162911Ssimon (CVE-2006-4339) [Ben Laurie and Google Security Team] 535162911Ssimon 536162911Ssimon *) Add AES IGE and biIGE modes. 537162911Ssimon [Ben Laurie] 538162911Ssimon 539162911Ssimon *) Change the Unix randomness entropy gathering to use poll() when 540162911Ssimon possible instead of select(), since the latter has some 541162911Ssimon undesirable limitations. 542162911Ssimon [Darryl Miles via Richard Levitte and Bodo Moeller] 543162911Ssimon 544162911Ssimon *) Disable "ECCdraft" ciphersuites more thoroughly. Now special 545162911Ssimon treatment in ssl/ssl_ciph.s makes sure that these ciphersuites 546162911Ssimon cannot be implicitly activated as part of, e.g., the "AES" alias. 547162911Ssimon However, please upgrade to OpenSSL 0.9.9[-dev] for 548162911Ssimon non-experimental use of the ECC ciphersuites to get TLS extension 549162911Ssimon support, which is required for curve and point format negotiation 550162911Ssimon to avoid potential handshake problems. 551162911Ssimon [Bodo Moeller] 552162911Ssimon 553162911Ssimon *) Disable rogue ciphersuites: 554162911Ssimon 555162911Ssimon - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 556162911Ssimon - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 557162911Ssimon - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 558162911Ssimon 559162911Ssimon The latter two were purportedly from 560162911Ssimon draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 561162911Ssimon appear there. 562162911Ssimon 563167612Ssimon Also deactivate the remaining ciphersuites from 564162911Ssimon draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 565162911Ssimon unofficial, and the ID has long expired. 566162911Ssimon [Bodo Moeller] 567162911Ssimon 568162911Ssimon *) Fix RSA blinding Heisenbug (problems sometimes occured on 569162911Ssimon dual-core machines) and other potential thread-safety issues. 570162911Ssimon [Bodo Moeller] 571162911Ssimon 572162911Ssimon *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key 573162911Ssimon versions), which is now available for royalty-free use 574162911Ssimon (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html). 575162911Ssimon Also, add Camellia TLS ciphersuites from RFC 4132. 576162911Ssimon 577162911Ssimon To minimize changes between patchlevels in the OpenSSL 0.9.8 578162911Ssimon series, Camellia remains excluded from compilation unless OpenSSL 579162911Ssimon is configured with 'enable-camellia'. 580162911Ssimon [NTT] 581162911Ssimon 582162911Ssimon *) Disable the padding bug check when compression is in use. The padding 583162911Ssimon bug check assumes the first packet is of even length, this is not 584162911Ssimon necessarily true if compresssion is enabled and can result in false 585162911Ssimon positives causing handshake failure. The actual bug test is ancient 586162911Ssimon code so it is hoped that implementations will either have fixed it by 587162911Ssimon now or any which still have the bug do not support compression. 588162911Ssimon [Steve Henson] 589162911Ssimon 590160814Ssimon Changes between 0.9.8a and 0.9.8b [04 May 2006] 591160814Ssimon 592160814Ssimon *) When applying a cipher rule check to see if string match is an explicit 593160814Ssimon cipher suite and only match that one cipher suite if it is. 594160814Ssimon [Steve Henson] 595160814Ssimon 596160814Ssimon *) Link in manifests for VC++ if needed. 597160814Ssimon [Austin Ziegler <halostatue@gmail.com>] 598160814Ssimon 599160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 600160814Ssimon draft-ietf-tls-ecc-12.txt with proposed changes (but without 601160814Ssimon TLS extensions, which are supported starting with the 0.9.9 602160814Ssimon branch, not in the OpenSSL 0.9.8 branch). 603160814Ssimon [Douglas Stebila] 604160814Ssimon 605160814Ssimon *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support 606160814Ssimon opaque EVP_CIPHER_CTX handling. 607160814Ssimon [Steve Henson] 608160814Ssimon 609160814Ssimon *) Fixes and enhancements to zlib compression code. We now only use 610160814Ssimon "zlib1.dll" and use the default __cdecl calling convention on Win32 611160814Ssimon to conform with the standards mentioned here: 612160814Ssimon http://www.zlib.net/DLL_FAQ.txt 613160814Ssimon Static zlib linking now works on Windows and the new --with-zlib-include 614160814Ssimon --with-zlib-lib options to Configure can be used to supply the location 615160814Ssimon of the headers and library. Gracefully handle case where zlib library 616160814Ssimon can't be loaded. 617160814Ssimon [Steve Henson] 618160814Ssimon 619160814Ssimon *) Several fixes and enhancements to the OID generation code. The old code 620160814Ssimon sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't 621160814Ssimon handle numbers larger than ULONG_MAX, truncated printing and had a 622160814Ssimon non standard OBJ_obj2txt() behaviour. 623160814Ssimon [Steve Henson] 624160814Ssimon 625160814Ssimon *) Add support for building of engines under engine/ as shared libraries 626160814Ssimon under VC++ build system. 627160814Ssimon [Steve Henson] 628160814Ssimon 629160814Ssimon *) Corrected the numerous bugs in the Win32 path splitter in DSO. 630160814Ssimon Hopefully, we will not see any false combination of paths any more. 631160814Ssimon [Richard Levitte] 632160814Ssimon 633160814Ssimon Changes between 0.9.8 and 0.9.8a [11 Oct 2005] 634160814Ssimon 635160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 636160814Ssimon (part of SSL_OP_ALL). This option used to disable the 637160814Ssimon countermeasure against man-in-the-middle protocol-version 638160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 639160814Ssimon idea. (CVE-2005-2969) 640160814Ssimon 641160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 642160814Ssimon for Information Security, National Institute of Advanced Industrial 643160814Ssimon Science and Technology [AIST], Japan)] 644160814Ssimon 645160814Ssimon *) Add two function to clear and return the verify parameter flags. 646160814Ssimon [Steve Henson] 647160814Ssimon 648160814Ssimon *) Keep cipherlists sorted in the source instead of sorting them at 649160814Ssimon runtime, thus removing the need for a lock. 650160814Ssimon [Nils Larsch] 651160814Ssimon 652160814Ssimon *) Avoid some small subgroup attacks in Diffie-Hellman. 653160814Ssimon [Nick Mathewson and Ben Laurie] 654160814Ssimon 655160814Ssimon *) Add functions for well-known primes. 656160814Ssimon [Nick Mathewson] 657160814Ssimon 658160814Ssimon *) Extended Windows CE support. 659160814Ssimon [Satoshi Nakamura and Andy Polyakov] 660160814Ssimon 661160814Ssimon *) Initialize SSL_METHOD structures at compile time instead of during 662160814Ssimon runtime, thus removing the need for a lock. 663160814Ssimon [Steve Henson] 664160814Ssimon 665160814Ssimon *) Make PKCS7_decrypt() work even if no certificate is supplied by 666160814Ssimon attempting to decrypt each encrypted key in turn. Add support to 667160814Ssimon smime utility. 668160814Ssimon [Steve Henson] 669160814Ssimon 670160814Ssimon Changes between 0.9.7h and 0.9.8 [05 Jul 2005] 671160814Ssimon 672162911Ssimon [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after 673162911Ssimon OpenSSL 0.9.8.] 674162911Ssimon 675160814Ssimon *) Add libcrypto.pc and libssl.pc for those who feel they need them. 676160814Ssimon [Richard Levitte] 677160814Ssimon 678160814Ssimon *) Change CA.sh and CA.pl so they don't bundle the CSR and the private 679160814Ssimon key into the same file any more. 680160814Ssimon [Richard Levitte] 681160814Ssimon 682160814Ssimon *) Add initial support for Win64, both IA64 and AMD64/x64 flavors. 683160814Ssimon [Andy Polyakov] 684160814Ssimon 685160814Ssimon *) Add -utf8 command line and config file option to 'ca'. 686160814Ssimon [Stefan <stf@udoma.org] 687160814Ssimon 688160814Ssimon *) Removed the macro des_crypt(), as it seems to conflict with some 689160814Ssimon libraries. Use DES_crypt(). 690160814Ssimon [Richard Levitte] 691160814Ssimon 692160814Ssimon *) Correct naming of the 'chil' and '4758cca' ENGINEs. This 693160814Ssimon involves renaming the source and generated shared-libs for 694160814Ssimon both. The engines will accept the corrected or legacy ids 695160814Ssimon ('ncipher' and '4758_cca' respectively) when binding. NB, 696160814Ssimon this only applies when building 'shared'. 697160814Ssimon [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe] 698160814Ssimon 699160814Ssimon *) Add attribute functions to EVP_PKEY structure. Modify 700160814Ssimon PKCS12_create() to recognize a CSP name attribute and 701160814Ssimon use it. Make -CSP option work again in pkcs12 utility. 702160814Ssimon [Steve Henson] 703160814Ssimon 704160814Ssimon *) Add new functionality to the bn blinding code: 705160814Ssimon - automatic re-creation of the BN_BLINDING parameters after 706160814Ssimon a fixed number of uses (currently 32) 707160814Ssimon - add new function for parameter creation 708160814Ssimon - introduce flags to control the update behaviour of the 709160814Ssimon BN_BLINDING parameters 710160814Ssimon - hide BN_BLINDING structure 711160814Ssimon Add a second BN_BLINDING slot to the RSA structure to improve 712160814Ssimon performance when a single RSA object is shared among several 713160814Ssimon threads. 714160814Ssimon [Nils Larsch] 715160814Ssimon 716160814Ssimon *) Add support for DTLS. 717160814Ssimon [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie] 718160814Ssimon 719160814Ssimon *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1) 720160814Ssimon to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file() 721160814Ssimon [Walter Goulet] 722160814Ssimon 723160814Ssimon *) Remove buggy and incompletet DH cert support from 724160814Ssimon ssl/ssl_rsa.c and ssl/s3_both.c 725160814Ssimon [Nils Larsch] 726160814Ssimon 727160814Ssimon *) Use SHA-1 instead of MD5 as the default digest algorithm for 728160814Ssimon the apps/openssl applications. 729160814Ssimon [Nils Larsch] 730160814Ssimon 731160814Ssimon *) Compile clean with "-Wall -Wmissing-prototypes 732160814Ssimon -Wstrict-prototypes -Wmissing-declarations -Werror". Currently 733160814Ssimon DEBUG_SAFESTACK must also be set. 734160814Ssimon [Ben Laurie] 735160814Ssimon 736160814Ssimon *) Change ./Configure so that certain algorithms can be disabled by default. 737160814Ssimon The new counterpiece to "no-xxx" is "enable-xxx". 738160814Ssimon 739160814Ssimon The patented RC5 and MDC2 algorithms will now be disabled unless 740160814Ssimon "enable-rc5" and "enable-mdc2", respectively, are specified. 741160814Ssimon 742160814Ssimon (IDEA remains enabled despite being patented. This is because IDEA 743160814Ssimon is frequently required for interoperability, and there is no license 744160814Ssimon fee for non-commercial use. As before, "no-idea" can be used to 745160814Ssimon avoid this algorithm.) 746160814Ssimon 747160814Ssimon [Bodo Moeller] 748160814Ssimon 749160814Ssimon *) Add processing of proxy certificates (see RFC 3820). This work was 750160814Ssimon sponsored by KTH (The Royal Institute of Technology in Stockholm) and 751160814Ssimon EGEE (Enabling Grids for E-science in Europe). 752160814Ssimon [Richard Levitte] 753160814Ssimon 754160814Ssimon *) RC4 performance overhaul on modern architectures/implementations, such 755160814Ssimon as Intel P4, IA-64 and AMD64. 756160814Ssimon [Andy Polyakov] 757160814Ssimon 758160814Ssimon *) New utility extract-section.pl. This can be used specify an alternative 759160814Ssimon section number in a pod file instead of having to treat each file as 760160814Ssimon a separate case in Makefile. This can be done by adding two lines to the 761160814Ssimon pod file: 762160814Ssimon 763160814Ssimon =for comment openssl_section:XXX 764160814Ssimon 765160814Ssimon The blank line is mandatory. 766160814Ssimon 767160814Ssimon [Steve Henson] 768160814Ssimon 769160814Ssimon *) New arguments -certform, -keyform and -pass for s_client and s_server 770160814Ssimon to allow alternative format key and certificate files and passphrase 771160814Ssimon sources. 772160814Ssimon [Steve Henson] 773160814Ssimon 774160814Ssimon *) New structure X509_VERIFY_PARAM which combines current verify parameters, 775160814Ssimon update associated structures and add various utility functions. 776160814Ssimon 777160814Ssimon Add new policy related verify parameters, include policy checking in 778160814Ssimon standard verify code. Enhance 'smime' application with extra parameters 779160814Ssimon to support policy checking and print out. 780160814Ssimon [Steve Henson] 781160814Ssimon 782160814Ssimon *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3 783160814Ssimon Nehemiah processors. These extensions support AES encryption in hardware 784160814Ssimon as well as RNG (though RNG support is currently disabled). 785160814Ssimon [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov] 786160814Ssimon 787160814Ssimon *) Deprecate BN_[get|set]_params() functions (they were ignored internally). 788160814Ssimon [Geoff Thorpe] 789160814Ssimon 790160814Ssimon *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented. 791160814Ssimon [Andy Polyakov and a number of other people] 792160814Ssimon 793160814Ssimon *) Improved PowerPC platform support. Most notably BIGNUM assembler 794160814Ssimon implementation contributed by IBM. 795160814Ssimon [Suresh Chari, Peter Waltenberg, Andy Polyakov] 796160814Ssimon 797160814Ssimon *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public 798160814Ssimon exponent rather than 'unsigned long'. There is a corresponding change to 799160814Ssimon the new 'rsa_keygen' element of the RSA_METHOD structure. 800160814Ssimon [Jelte Jansen, Geoff Thorpe] 801160814Ssimon 802160814Ssimon *) Functionality for creating the initial serial number file is now 803160814Ssimon moved from CA.pl to the 'ca' utility with a new option -create_serial. 804160814Ssimon 805160814Ssimon (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial 806160814Ssimon number file to 1, which is bound to cause problems. To avoid 807160814Ssimon the problems while respecting compatibility between different 0.9.7 808160814Ssimon patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in 809160814Ssimon CA.pl for serial number initialization. With the new release 0.9.8, 810160814Ssimon we can fix the problem directly in the 'ca' utility.) 811160814Ssimon [Steve Henson] 812160814Ssimon 813160814Ssimon *) Reduced header interdepencies by declaring more opaque objects in 814160814Ssimon ossl_typ.h. As a consequence, including some headers (eg. engine.h) will 815160814Ssimon give fewer recursive includes, which could break lazy source code - so 816160814Ssimon this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always, 817160814Ssimon developers should define this symbol when building and using openssl to 818160814Ssimon ensure they track the recommended behaviour, interfaces, [etc], but 819160814Ssimon backwards-compatible behaviour prevails when this isn't defined. 820160814Ssimon [Geoff Thorpe] 821160814Ssimon 822160814Ssimon *) New function X509_POLICY_NODE_print() which prints out policy nodes. 823160814Ssimon [Steve Henson] 824160814Ssimon 825160814Ssimon *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality. 826160814Ssimon This will generate a random key of the appropriate length based on the 827160814Ssimon cipher context. The EVP_CIPHER can provide its own random key generation 828160814Ssimon routine to support keys of a specific form. This is used in the des and 829160814Ssimon 3des routines to generate a key of the correct parity. Update S/MIME 830160814Ssimon code to use new functions and hence generate correct parity DES keys. 831160814Ssimon Add EVP_CHECK_DES_KEY #define to return an error if the key is not 832160814Ssimon valid (weak or incorrect parity). 833160814Ssimon [Steve Henson] 834160814Ssimon 835160814Ssimon *) Add a local set of CRLs that can be used by X509_verify_cert() as well 836160814Ssimon as looking them up. This is useful when the verified structure may contain 837160814Ssimon CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs 838160814Ssimon present unless the new PKCS7_NO_CRL flag is asserted. 839160814Ssimon [Steve Henson] 840160814Ssimon 841160814Ssimon *) Extend ASN1 oid configuration module. It now additionally accepts the 842160814Ssimon syntax: 843160814Ssimon 844160814Ssimon shortName = some long name, 1.2.3.4 845160814Ssimon [Steve Henson] 846160814Ssimon 847160814Ssimon *) Reimplemented the BN_CTX implementation. There is now no more static 848160814Ssimon limitation on the number of variables it can handle nor the depth of the 849160814Ssimon "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack 850160814Ssimon information can now expand as required, and rather than having a single 851160814Ssimon static array of bignums, BN_CTX now uses a linked-list of such arrays 852160814Ssimon allowing it to expand on demand whilst maintaining the usefulness of 853160814Ssimon BN_CTX's "bundling". 854160814Ssimon [Geoff Thorpe] 855160814Ssimon 856160814Ssimon *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD 857160814Ssimon to allow all RSA operations to function using a single BN_CTX. 858160814Ssimon [Geoff Thorpe] 859160814Ssimon 860160814Ssimon *) Preliminary support for certificate policy evaluation and checking. This 861160814Ssimon is initially intended to pass the tests outlined in "Conformance Testing 862160814Ssimon of Relying Party Client Certificate Path Processing Logic" v1.07. 863160814Ssimon [Steve Henson] 864160814Ssimon 865160814Ssimon *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and 866160814Ssimon remained unused and not that useful. A variety of other little bignum 867160814Ssimon tweaks and fixes have also been made continuing on from the audit (see 868160814Ssimon below). 869160814Ssimon [Geoff Thorpe] 870160814Ssimon 871160814Ssimon *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with 872160814Ssimon associated ASN1, EVP and SSL functions and old ASN1 macros. 873160814Ssimon [Richard Levitte] 874160814Ssimon 875160814Ssimon *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results, 876160814Ssimon and this should never fail. So the return value from the use of 877160814Ssimon BN_set_word() (which can fail due to needless expansion) is now deprecated; 878160814Ssimon if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro. 879160814Ssimon [Geoff Thorpe] 880160814Ssimon 881160814Ssimon *) BN_CTX_get() should return zero-valued bignums, providing the same 882160814Ssimon initialised value as BN_new(). 883160814Ssimon [Geoff Thorpe, suggested by Ulf M�ller] 884160814Ssimon 885160814Ssimon *) Support for inhibitAnyPolicy certificate extension. 886160814Ssimon [Steve Henson] 887160814Ssimon 888160814Ssimon *) An audit of the BIGNUM code is underway, for which debugging code is 889160814Ssimon enabled when BN_DEBUG is defined. This makes stricter enforcements on what 890160814Ssimon is considered valid when processing BIGNUMs, and causes execution to 891160814Ssimon assert() when a problem is discovered. If BN_DEBUG_RAND is defined, 892160814Ssimon further steps are taken to deliberately pollute unused data in BIGNUM 893160814Ssimon structures to try and expose faulty code further on. For now, openssl will 894160814Ssimon (in its default mode of operation) continue to tolerate the inconsistent 895160814Ssimon forms that it has tolerated in the past, but authors and packagers should 896160814Ssimon consider trying openssl and their own applications when compiled with 897160814Ssimon these debugging symbols defined. It will help highlight potential bugs in 898160814Ssimon their own code, and will improve the test coverage for OpenSSL itself. At 899160814Ssimon some point, these tighter rules will become openssl's default to improve 900160814Ssimon maintainability, though the assert()s and other overheads will remain only 901160814Ssimon in debugging configurations. See bn.h for more details. 902160814Ssimon [Geoff Thorpe, Nils Larsch, Ulf M�ller] 903160814Ssimon 904160814Ssimon *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure 905160814Ssimon that can only be obtained through BN_CTX_new() (which implicitly 906160814Ssimon initialises it). The presence of this function only made it possible 907160814Ssimon to overwrite an existing structure (and cause memory leaks). 908160814Ssimon [Geoff Thorpe] 909160814Ssimon 910160814Ssimon *) Because of the callback-based approach for implementing LHASH as a 911160814Ssimon template type, lh_insert() adds opaque objects to hash-tables and 912160814Ssimon lh_doall() or lh_doall_arg() are typically used with a destructor callback 913160814Ssimon to clean up those corresponding objects before destroying the hash table 914160814Ssimon (and losing the object pointers). So some over-zealous constifications in 915160814Ssimon LHASH have been relaxed so that lh_insert() does not take (nor store) the 916160814Ssimon objects as "const" and the lh_doall[_arg] callback wrappers are not 917160814Ssimon prototyped to have "const" restrictions on the object pointers they are 918160814Ssimon given (and so aren't required to cast them away any more). 919160814Ssimon [Geoff Thorpe] 920160814Ssimon 921160814Ssimon *) The tmdiff.h API was so ugly and minimal that our own timing utility 922160814Ssimon (speed) prefers to use its own implementation. The two implementations 923160814Ssimon haven't been consolidated as yet (volunteers?) but the tmdiff API has had 924160814Ssimon its object type properly exposed (MS_TM) instead of casting to/from "char 925160814Ssimon *". This may still change yet if someone realises MS_TM and "ms_time_***" 926160814Ssimon aren't necessarily the greatest nomenclatures - but this is what was used 927160814Ssimon internally to the implementation so I've used that for now. 928160814Ssimon [Geoff Thorpe] 929160814Ssimon 930160814Ssimon *) Ensure that deprecated functions do not get compiled when 931160814Ssimon OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of 932160814Ssimon the self-tests were still using deprecated key-generation functions so 933160814Ssimon these have been updated also. 934160814Ssimon [Geoff Thorpe] 935160814Ssimon 936160814Ssimon *) Reorganise PKCS#7 code to separate the digest location functionality 937160814Ssimon into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest(). 938160814Ssimon New function PKCS7_set_digest() to set the digest type for PKCS#7 939160814Ssimon digestedData type. Add additional code to correctly generate the 940160814Ssimon digestedData type and add support for this type in PKCS7 initialization 941160814Ssimon functions. 942160814Ssimon [Steve Henson] 943160814Ssimon 944160814Ssimon *) New function PKCS7_set0_type_other() this initializes a PKCS7 945160814Ssimon structure of type "other". 946160814Ssimon [Steve Henson] 947160814Ssimon 948160814Ssimon *) Fix prime generation loop in crypto/bn/bn_prime.pl by making 949160814Ssimon sure the loop does correctly stop and breaking ("division by zero") 950160814Ssimon modulus operations are not performed. The (pre-generated) prime 951160814Ssimon table crypto/bn/bn_prime.h was already correct, but it could not be 952160814Ssimon re-generated on some platforms because of the "division by zero" 953160814Ssimon situation in the script. 954160814Ssimon [Ralf S. Engelschall] 955160814Ssimon 956160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 957160814Ssimon draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with 958160814Ssimon SHA-1 now is only used for "small" curves (where the 959160814Ssimon representation of a field element takes up to 24 bytes); for 960160814Ssimon larger curves, the field element resulting from ECDH is directly 961160814Ssimon used as premaster secret. 962160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 963160814Ssimon 964160814Ssimon *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2 965160814Ssimon curve secp160r1 to the tests. 966160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 967160814Ssimon 968160814Ssimon *) Add the possibility to load symbols globally with DSO. 969160814Ssimon [G�tz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte] 970160814Ssimon 971160814Ssimon *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better 972160814Ssimon control of the error stack. 973160814Ssimon [Richard Levitte] 974160814Ssimon 975160814Ssimon *) Add support for STORE in ENGINE. 976160814Ssimon [Richard Levitte] 977160814Ssimon 978160814Ssimon *) Add the STORE type. The intention is to provide a common interface 979160814Ssimon to certificate and key stores, be they simple file-based stores, or 980160814Ssimon HSM-type store, or LDAP stores, or... 981160814Ssimon NOTE: The code is currently UNTESTED and isn't really used anywhere. 982160814Ssimon [Richard Levitte] 983160814Ssimon 984160814Ssimon *) Add a generic structure called OPENSSL_ITEM. This can be used to 985160814Ssimon pass a list of arguments to any function as well as provide a way 986160814Ssimon for a function to pass data back to the caller. 987160814Ssimon [Richard Levitte] 988160814Ssimon 989160814Ssimon *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup() 990160814Ssimon works like BUF_strdup() but can be used to duplicate a portion of 991160814Ssimon a string. The copy gets NUL-terminated. BUF_memdup() duplicates 992160814Ssimon a memory area. 993160814Ssimon [Richard Levitte] 994160814Ssimon 995160814Ssimon *) Add the function sk_find_ex() which works like sk_find(), but will 996160814Ssimon return an index to an element even if an exact match couldn't be 997160814Ssimon found. The index is guaranteed to point at the element where the 998160814Ssimon searched-for key would be inserted to preserve sorting order. 999160814Ssimon [Richard Levitte] 1000160814Ssimon 1001160814Ssimon *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but 1002160814Ssimon takes an extra flags argument for optional functionality. Currently, 1003160814Ssimon the following flags are defined: 1004160814Ssimon 1005160814Ssimon OBJ_BSEARCH_VALUE_ON_NOMATCH 1006160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 1007160814Ssimon element where the comparing function returns a negative or zero 1008160814Ssimon number. 1009160814Ssimon 1010160814Ssimon OBJ_BSEARCH_FIRST_VALUE_ON_MATCH 1011160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 1012160814Ssimon element where the comparing function returns zero. This is useful 1013160814Ssimon if there are more than one element where the comparing function 1014160814Ssimon returns zero. 1015160814Ssimon [Richard Levitte] 1016160814Ssimon 1017160814Ssimon *) Make it possible to create self-signed certificates with 'openssl ca' 1018160814Ssimon in such a way that the self-signed certificate becomes part of the 1019160814Ssimon CA database and uses the same mechanisms for serial number generation 1020160814Ssimon as all other certificate signing. The new flag '-selfsign' enables 1021160814Ssimon this functionality. Adapt CA.sh and CA.pl.in. 1022160814Ssimon [Richard Levitte] 1023160814Ssimon 1024160814Ssimon *) Add functionality to check the public key of a certificate request 1025160814Ssimon against a given private. This is useful to check that a certificate 1026160814Ssimon request can be signed by that key (self-signing). 1027160814Ssimon [Richard Levitte] 1028160814Ssimon 1029160814Ssimon *) Make it possible to have multiple active certificates with the same 1030160814Ssimon subject in the CA index file. This is done only if the keyword 1031160814Ssimon 'unique_subject' is set to 'no' in the main CA section (default 1032160814Ssimon if 'CA_default') of the configuration file. The value is saved 1033160814Ssimon with the database itself in a separate index attribute file, 1034160814Ssimon named like the index file with '.attr' appended to the name. 1035160814Ssimon [Richard Levitte] 1036160814Ssimon 1037160814Ssimon *) Generate muti valued AVAs using '+' notation in config files for 1038160814Ssimon req and dirName. 1039160814Ssimon [Steve Henson] 1040160814Ssimon 1041160814Ssimon *) Support for nameConstraints certificate extension. 1042160814Ssimon [Steve Henson] 1043160814Ssimon 1044160814Ssimon *) Support for policyConstraints certificate extension. 1045160814Ssimon [Steve Henson] 1046160814Ssimon 1047160814Ssimon *) Support for policyMappings certificate extension. 1048160814Ssimon [Steve Henson] 1049160814Ssimon 1050160814Ssimon *) Make sure the default DSA_METHOD implementation only uses its 1051160814Ssimon dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL, 1052160814Ssimon and change its own handlers to be NULL so as to remove unnecessary 1053160814Ssimon indirection. This lets alternative implementations fallback to the 1054160814Ssimon default implementation more easily. 1055160814Ssimon [Geoff Thorpe] 1056160814Ssimon 1057160814Ssimon *) Support for directoryName in GeneralName related extensions 1058160814Ssimon in config files. 1059160814Ssimon [Steve Henson] 1060160814Ssimon 1061160814Ssimon *) Make it possible to link applications using Makefile.shared. 1062160814Ssimon Make that possible even when linking against static libraries! 1063160814Ssimon [Richard Levitte] 1064160814Ssimon 1065160814Ssimon *) Support for single pass processing for S/MIME signing. This now 1066160814Ssimon means that S/MIME signing can be done from a pipe, in addition 1067160814Ssimon cleartext signing (multipart/signed type) is effectively streaming 1068160814Ssimon and the signed data does not need to be all held in memory. 1069160814Ssimon 1070160814Ssimon This is done with a new flag PKCS7_STREAM. When this flag is set 1071160814Ssimon PKCS7_sign() only initializes the PKCS7 structure and the actual signing 1072160814Ssimon is done after the data is output (and digests calculated) in 1073160814Ssimon SMIME_write_PKCS7(). 1074160814Ssimon [Steve Henson] 1075160814Ssimon 1076160814Ssimon *) Add full support for -rpath/-R, both in shared libraries and 1077160814Ssimon applications, at least on the platforms where it's known how 1078160814Ssimon to do it. 1079160814Ssimon [Richard Levitte] 1080160814Ssimon 1081160814Ssimon *) In crypto/ec/ec_mult.c, implement fast point multiplication with 1082160814Ssimon precomputation, based on wNAF splitting: EC_GROUP_precompute_mult() 1083160814Ssimon will now compute a table of multiples of the generator that 1084160814Ssimon makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul() 1085160814Ssimon faster (notably in the case of a single point multiplication, 1086160814Ssimon scalar * generator). 1087160814Ssimon [Nils Larsch, Bodo Moeller] 1088160814Ssimon 1089160814Ssimon *) IPv6 support for certificate extensions. The various extensions 1090160814Ssimon which use the IP:a.b.c.d can now take IPv6 addresses using the 1091160814Ssimon formats of RFC1884 2.2 . IPv6 addresses are now also displayed 1092160814Ssimon correctly. 1093160814Ssimon [Steve Henson] 1094160814Ssimon 1095160814Ssimon *) Added an ENGINE that implements RSA by performing private key 1096160814Ssimon exponentiations with the GMP library. The conversions to and from 1097160814Ssimon GMP's mpz_t format aren't optimised nor are any montgomery forms 1098160814Ssimon cached, and on x86 it appears OpenSSL's own performance has caught up. 1099160814Ssimon However there are likely to be other architectures where GMP could 1100160814Ssimon provide a boost. This ENGINE is not built in by default, but it can be 1101160814Ssimon specified at Configure time and should be accompanied by the necessary 1102160814Ssimon linker additions, eg; 1103160814Ssimon ./config -DOPENSSL_USE_GMP -lgmp 1104160814Ssimon [Geoff Thorpe] 1105160814Ssimon 1106160814Ssimon *) "openssl engine" will not display ENGINE/DSO load failure errors when 1107160814Ssimon testing availability of engines with "-t" - the old behaviour is 1108160814Ssimon produced by increasing the feature's verbosity with "-tt". 1109160814Ssimon [Geoff Thorpe] 1110160814Ssimon 1111160814Ssimon *) ECDSA routines: under certain error conditions uninitialized BN objects 1112160814Ssimon could be freed. Solution: make sure initialization is performed early 1113160814Ssimon enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de> 1114160814Ssimon via PR#459) 1115160814Ssimon [Lutz Jaenicke] 1116160814Ssimon 1117160814Ssimon *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD 1118160814Ssimon and DH_METHOD (eg. by ENGINE implementations) to override the normal 1119160814Ssimon software implementations. For DSA and DH, parameter generation can 1120160814Ssimon also be overriden by providing the appropriate method callbacks. 1121160814Ssimon [Geoff Thorpe] 1122160814Ssimon 1123160814Ssimon *) Change the "progress" mechanism used in key-generation and 1124160814Ssimon primality testing to functions that take a new BN_GENCB pointer in 1125160814Ssimon place of callback/argument pairs. The new API functions have "_ex" 1126160814Ssimon postfixes and the older functions are reimplemented as wrappers for 1127160814Ssimon the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide 1128160814Ssimon declarations of the old functions to help (graceful) attempts to 1129160814Ssimon migrate to the new functions. Also, the new key-generation API 1130160814Ssimon functions operate on a caller-supplied key-structure and return 1131160814Ssimon success/failure rather than returning a key or NULL - this is to 1132160814Ssimon help make "keygen" another member function of RSA_METHOD etc. 1133160814Ssimon 1134160814Ssimon Example for using the new callback interface: 1135160814Ssimon 1136160814Ssimon int (*my_callback)(int a, int b, BN_GENCB *cb) = ...; 1137160814Ssimon void *my_arg = ...; 1138160814Ssimon BN_GENCB my_cb; 1139160814Ssimon 1140160814Ssimon BN_GENCB_set(&my_cb, my_callback, my_arg); 1141160814Ssimon 1142160814Ssimon return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb); 1143160814Ssimon /* For the meaning of a, b in calls to my_callback(), see the 1144160814Ssimon * documentation of the function that calls the callback. 1145160814Ssimon * cb will point to my_cb; my_arg can be retrieved as cb->arg. 1146160814Ssimon * my_callback should return 1 if it wants BN_is_prime_ex() 1147160814Ssimon * to continue, or 0 to stop. 1148160814Ssimon */ 1149160814Ssimon 1150160814Ssimon [Geoff Thorpe] 1151160814Ssimon 1152160814Ssimon *) Change the ZLIB compression method to be stateful, and make it 1153160814Ssimon available to TLS with the number defined in 1154160814Ssimon draft-ietf-tls-compression-04.txt. 1155160814Ssimon [Richard Levitte] 1156160814Ssimon 1157160814Ssimon *) Add the ASN.1 structures and functions for CertificatePair, which 1158160814Ssimon is defined as follows (according to X.509_4thEditionDraftV6.pdf): 1159160814Ssimon 1160160814Ssimon CertificatePair ::= SEQUENCE { 1161160814Ssimon forward [0] Certificate OPTIONAL, 1162160814Ssimon reverse [1] Certificate OPTIONAL, 1163160814Ssimon -- at least one of the pair shall be present -- } 1164160814Ssimon 1165160814Ssimon Also implement the PEM functions to read and write certificate 1166160814Ssimon pairs, and defined the PEM tag as "CERTIFICATE PAIR". 1167160814Ssimon 1168160814Ssimon This needed to be defined, mostly for the sake of the LDAP 1169160814Ssimon attribute crossCertificatePair, but may prove useful elsewhere as 1170160814Ssimon well. 1171160814Ssimon [Richard Levitte] 1172160814Ssimon 1173160814Ssimon *) Make it possible to inhibit symlinking of shared libraries in 1174160814Ssimon Makefile.shared, for Cygwin's sake. 1175160814Ssimon [Richard Levitte] 1176160814Ssimon 1177160814Ssimon *) Extend the BIGNUM API by creating a function 1178160814Ssimon void BN_set_negative(BIGNUM *a, int neg); 1179160814Ssimon and a macro that behave like 1180160814Ssimon int BN_is_negative(const BIGNUM *a); 1181160814Ssimon 1182160814Ssimon to avoid the need to access 'a->neg' directly in applications. 1183160814Ssimon [Nils Larsch] 1184160814Ssimon 1185160814Ssimon *) Implement fast modular reduction for pseudo-Mersenne primes 1186160814Ssimon used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c). 1187160814Ssimon EC_GROUP_new_curve_GFp() will now automatically use this 1188160814Ssimon if applicable. 1189160814Ssimon [Nils Larsch <nla@trustcenter.de>] 1190160814Ssimon 1191160814Ssimon *) Add new lock type (CRYPTO_LOCK_BN). 1192160814Ssimon [Bodo Moeller] 1193160814Ssimon 1194160814Ssimon *) Change the ENGINE framework to automatically load engines 1195160814Ssimon dynamically from specific directories unless they could be 1196160814Ssimon found to already be built in or loaded. Move all the 1197160814Ssimon current engines except for the cryptodev one to a new 1198160814Ssimon directory engines/. 1199160814Ssimon The engines in engines/ are built as shared libraries if 1200160814Ssimon the "shared" options was given to ./Configure or ./config. 1201160814Ssimon Otherwise, they are inserted in libcrypto.a. 1202160814Ssimon /usr/local/ssl/engines is the default directory for dynamic 1203160814Ssimon engines, but that can be overriden at configure time through 1204160814Ssimon the usual use of --prefix and/or --openssldir, and at run 1205160814Ssimon time with the environment variable OPENSSL_ENGINES. 1206160814Ssimon [Geoff Thorpe and Richard Levitte] 1207160814Ssimon 1208160814Ssimon *) Add Makefile.shared, a helper makefile to build shared 1209160814Ssimon libraries. Addapt Makefile.org. 1210160814Ssimon [Richard Levitte] 1211160814Ssimon 1212160814Ssimon *) Add version info to Win32 DLLs. 1213160814Ssimon [Peter 'Luna' Runestig" <peter@runestig.com>] 1214160814Ssimon 1215160814Ssimon *) Add new 'medium level' PKCS#12 API. Certificates and keys 1216160814Ssimon can be added using this API to created arbitrary PKCS#12 1217160814Ssimon files while avoiding the low level API. 1218160814Ssimon 1219160814Ssimon New options to PKCS12_create(), key or cert can be NULL and 1220160814Ssimon will then be omitted from the output file. The encryption 1221160814Ssimon algorithm NIDs can be set to -1 for no encryption, the mac 1222160814Ssimon iteration count can be set to 0 to omit the mac. 1223160814Ssimon 1224160814Ssimon Enhance pkcs12 utility by making the -nokeys and -nocerts 1225160814Ssimon options work when creating a PKCS#12 file. New option -nomac 1226160814Ssimon to omit the mac, NONE can be set for an encryption algorithm. 1227160814Ssimon New code is modified to use the enhanced PKCS12_create() 1228160814Ssimon instead of the low level API. 1229160814Ssimon [Steve Henson] 1230160814Ssimon 1231160814Ssimon *) Extend ASN1 encoder to support indefinite length constructed 1232160814Ssimon encoding. This can output sequences tags and octet strings in 1233160814Ssimon this form. Modify pk7_asn1.c to support indefinite length 1234160814Ssimon encoding. This is experimental and needs additional code to 1235160814Ssimon be useful, such as an ASN1 bio and some enhanced streaming 1236160814Ssimon PKCS#7 code. 1237160814Ssimon 1238160814Ssimon Extend template encode functionality so that tagging is passed 1239160814Ssimon down to the template encoder. 1240160814Ssimon [Steve Henson] 1241160814Ssimon 1242160814Ssimon *) Let 'openssl req' fail if an argument to '-newkey' is not 1243160814Ssimon recognized instead of using RSA as a default. 1244160814Ssimon [Bodo Moeller] 1245160814Ssimon 1246160814Ssimon *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt. 1247160814Ssimon As these are not official, they are not included in "ALL"; 1248160814Ssimon the "ECCdraft" ciphersuite group alias can be used to select them. 1249160814Ssimon [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)] 1250160814Ssimon 1251160814Ssimon *) Add ECDH engine support. 1252160814Ssimon [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)] 1253160814Ssimon 1254160814Ssimon *) Add ECDH in new directory crypto/ecdh/. 1255160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 1256160814Ssimon 1257160814Ssimon *) Let BN_rand_range() abort with an error after 100 iterations 1258160814Ssimon without success (which indicates a broken PRNG). 1259160814Ssimon [Bodo Moeller] 1260160814Ssimon 1261160814Ssimon *) Change BN_mod_sqrt() so that it verifies that the input value 1262160814Ssimon is really the square of the return value. (Previously, 1263160814Ssimon BN_mod_sqrt would show GIGO behaviour.) 1264160814Ssimon [Bodo Moeller] 1265160814Ssimon 1266160814Ssimon *) Add named elliptic curves over binary fields from X9.62, SECG, 1267160814Ssimon and WAP/WTLS; add OIDs that were still missing. 1268160814Ssimon 1269160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 1270160814Ssimon (Sun Microsystems Laboratories)] 1271160814Ssimon 1272160814Ssimon *) Extend the EC library for elliptic curves over binary fields 1273160814Ssimon (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/). 1274160814Ssimon New EC_METHOD: 1275160814Ssimon 1276160814Ssimon EC_GF2m_simple_method 1277160814Ssimon 1278160814Ssimon New API functions: 1279160814Ssimon 1280160814Ssimon EC_GROUP_new_curve_GF2m 1281160814Ssimon EC_GROUP_set_curve_GF2m 1282160814Ssimon EC_GROUP_get_curve_GF2m 1283160814Ssimon EC_POINT_set_affine_coordinates_GF2m 1284160814Ssimon EC_POINT_get_affine_coordinates_GF2m 1285160814Ssimon EC_POINT_set_compressed_coordinates_GF2m 1286160814Ssimon 1287160814Ssimon Point compression for binary fields is disabled by default for 1288160814Ssimon patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to 1289160814Ssimon enable it). 1290160814Ssimon 1291160814Ssimon As binary polynomials are represented as BIGNUMs, various members 1292160814Ssimon of the EC_GROUP and EC_POINT data structures can be shared 1293160814Ssimon between the implementations for prime fields and binary fields; 1294160814Ssimon the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m) 1295160814Ssimon are essentially identical to their ..._GFp counterparts. 1296160814Ssimon (For simplicity, the '..._GFp' prefix has been dropped from 1297160814Ssimon various internal method names.) 1298160814Ssimon 1299160814Ssimon An internal 'field_div' method (similar to 'field_mul' and 1300160814Ssimon 'field_sqr') has been added; this is used only for binary fields. 1301160814Ssimon 1302160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 1303160814Ssimon (Sun Microsystems Laboratories)] 1304160814Ssimon 1305160814Ssimon *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult() 1306160814Ssimon through methods ('mul', 'precompute_mult'). 1307160814Ssimon 1308160814Ssimon The generic implementations (now internally called 'ec_wNAF_mul' 1309160814Ssimon and 'ec_wNAF_precomputed_mult') remain the default if these 1310160814Ssimon methods are undefined. 1311160814Ssimon 1312160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 1313160814Ssimon (Sun Microsystems Laboratories)] 1314160814Ssimon 1315160814Ssimon *) New function EC_GROUP_get_degree, which is defined through 1316160814Ssimon EC_METHOD. For curves over prime fields, this returns the bit 1317160814Ssimon length of the modulus. 1318160814Ssimon 1319160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 1320160814Ssimon (Sun Microsystems Laboratories)] 1321160814Ssimon 1322160814Ssimon *) New functions EC_GROUP_dup, EC_POINT_dup. 1323160814Ssimon (These simply call ..._new and ..._copy). 1324160814Ssimon 1325160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 1326160814Ssimon (Sun Microsystems Laboratories)] 1327160814Ssimon 1328160814Ssimon *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c. 1329160814Ssimon Polynomials are represented as BIGNUMs (where the sign bit is not 1330160814Ssimon used) in the following functions [macros]: 1331160814Ssimon 1332160814Ssimon BN_GF2m_add 1333160814Ssimon BN_GF2m_sub [= BN_GF2m_add] 1334160814Ssimon BN_GF2m_mod [wrapper for BN_GF2m_mod_arr] 1335160814Ssimon BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr] 1336160814Ssimon BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr] 1337160814Ssimon BN_GF2m_mod_inv 1338160814Ssimon BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr] 1339160814Ssimon BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr] 1340160814Ssimon BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr] 1341160814Ssimon BN_GF2m_cmp [= BN_ucmp] 1342160814Ssimon 1343160814Ssimon (Note that only the 'mod' functions are actually for fields GF(2^m). 1344160814Ssimon BN_GF2m_add() is misnomer, but this is for the sake of consistency.) 1345160814Ssimon 1346160814Ssimon For some functions, an the irreducible polynomial defining a 1347160814Ssimon field can be given as an 'unsigned int[]' with strictly 1348160814Ssimon decreasing elements giving the indices of those bits that are set; 1349160814Ssimon i.e., p[] represents the polynomial 1350160814Ssimon f(t) = t^p[0] + t^p[1] + ... + t^p[k] 1351160814Ssimon where 1352160814Ssimon p[0] > p[1] > ... > p[k] = 0. 1353160814Ssimon This applies to the following functions: 1354160814Ssimon 1355160814Ssimon BN_GF2m_mod_arr 1356160814Ssimon BN_GF2m_mod_mul_arr 1357160814Ssimon BN_GF2m_mod_sqr_arr 1358160814Ssimon BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv] 1359160814Ssimon BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div] 1360160814Ssimon BN_GF2m_mod_exp_arr 1361160814Ssimon BN_GF2m_mod_sqrt_arr 1362160814Ssimon BN_GF2m_mod_solve_quad_arr 1363160814Ssimon BN_GF2m_poly2arr 1364160814Ssimon BN_GF2m_arr2poly 1365160814Ssimon 1366160814Ssimon Conversion can be performed by the following functions: 1367160814Ssimon 1368160814Ssimon BN_GF2m_poly2arr 1369160814Ssimon BN_GF2m_arr2poly 1370160814Ssimon 1371160814Ssimon bntest.c has additional tests for binary polynomial arithmetic. 1372160814Ssimon 1373160814Ssimon Two implementations for BN_GF2m_mod_div() are available. 1374160814Ssimon The default algorithm simply uses BN_GF2m_mod_inv() and 1375160814Ssimon BN_GF2m_mod_mul(). The alternative algorithm is compiled in only 1376160814Ssimon if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the 1377160814Ssimon copyright notice in crypto/bn/bn_gf2m.c before enabling it). 1378160814Ssimon 1379160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 1380160814Ssimon (Sun Microsystems Laboratories)] 1381160814Ssimon 1382160814Ssimon *) Add new error code 'ERR_R_DISABLED' that can be used when some 1383160814Ssimon functionality is disabled at compile-time. 1384160814Ssimon [Douglas Stebila <douglas.stebila@sun.com>] 1385160814Ssimon 1386160814Ssimon *) Change default behaviour of 'openssl asn1parse' so that more 1387160814Ssimon information is visible when viewing, e.g., a certificate: 1388160814Ssimon 1389160814Ssimon Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump' 1390160814Ssimon mode the content of non-printable OCTET STRINGs is output in a 1391160814Ssimon style similar to INTEGERs, but with '[HEX DUMP]' prepended to 1392160814Ssimon avoid the appearance of a printable string. 1393160814Ssimon [Nils Larsch <nla@trustcenter.de>] 1394160814Ssimon 1395160814Ssimon *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access 1396160814Ssimon functions 1397160814Ssimon EC_GROUP_set_asn1_flag() 1398160814Ssimon EC_GROUP_get_asn1_flag() 1399160814Ssimon EC_GROUP_set_point_conversion_form() 1400160814Ssimon EC_GROUP_get_point_conversion_form() 1401160814Ssimon These control ASN1 encoding details: 1402160814Ssimon - Curves (i.e., groups) are encoded explicitly unless asn1_flag 1403160814Ssimon has been set to OPENSSL_EC_NAMED_CURVE. 1404160814Ssimon - Points are encoded in uncompressed form by default; options for 1405160814Ssimon asn1_for are as for point2oct, namely 1406160814Ssimon POINT_CONVERSION_COMPRESSED 1407160814Ssimon POINT_CONVERSION_UNCOMPRESSED 1408160814Ssimon POINT_CONVERSION_HYBRID 1409160814Ssimon 1410160814Ssimon Also add 'seed' and 'seed_len' members to EC_GROUP with access 1411160814Ssimon functions 1412160814Ssimon EC_GROUP_set_seed() 1413160814Ssimon EC_GROUP_get0_seed() 1414160814Ssimon EC_GROUP_get_seed_len() 1415160814Ssimon This is used only for ASN1 purposes (so far). 1416160814Ssimon [Nils Larsch <nla@trustcenter.de>] 1417160814Ssimon 1418160814Ssimon *) Add 'field_type' member to EC_METHOD, which holds the NID 1419160814Ssimon of the appropriate field type OID. The new function 1420160814Ssimon EC_METHOD_get_field_type() returns this value. 1421160814Ssimon [Nils Larsch <nla@trustcenter.de>] 1422160814Ssimon 1423160814Ssimon *) Add functions 1424160814Ssimon EC_POINT_point2bn() 1425160814Ssimon EC_POINT_bn2point() 1426160814Ssimon EC_POINT_point2hex() 1427160814Ssimon EC_POINT_hex2point() 1428160814Ssimon providing useful interfaces to EC_POINT_point2oct() and 1429160814Ssimon EC_POINT_oct2point(). 1430160814Ssimon [Nils Larsch <nla@trustcenter.de>] 1431160814Ssimon 1432160814Ssimon *) Change internals of the EC library so that the functions 1433160814Ssimon EC_GROUP_set_generator() 1434160814Ssimon EC_GROUP_get_generator() 1435160814Ssimon EC_GROUP_get_order() 1436160814Ssimon EC_GROUP_get_cofactor() 1437160814Ssimon are implemented directly in crypto/ec/ec_lib.c and not dispatched 1438160814Ssimon to methods, which would lead to unnecessary code duplication when 1439160814Ssimon adding different types of curves. 1440160814Ssimon [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller] 1441160814Ssimon 1442160814Ssimon *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM 1443160814Ssimon arithmetic, and such that modified wNAFs are generated 1444160814Ssimon (which avoid length expansion in many cases). 1445160814Ssimon [Bodo Moeller] 1446160814Ssimon 1447160814Ssimon *) Add a function EC_GROUP_check_discriminant() (defined via 1448160814Ssimon EC_METHOD) that verifies that the curve discriminant is non-zero. 1449160814Ssimon 1450160814Ssimon Add a function EC_GROUP_check() that makes some sanity tests 1451160814Ssimon on a EC_GROUP, its generator and order. This includes 1452160814Ssimon EC_GROUP_check_discriminant(). 1453160814Ssimon [Nils Larsch <nla@trustcenter.de>] 1454160814Ssimon 1455160814Ssimon *) Add ECDSA in new directory crypto/ecdsa/. 1456160814Ssimon 1457160814Ssimon Add applications 'openssl ecparam' and 'openssl ecdsa' 1458160814Ssimon (these are based on 'openssl dsaparam' and 'openssl dsa'). 1459160814Ssimon 1460160814Ssimon ECDSA support is also included in various other files across the 1461160814Ssimon library. Most notably, 1462160814Ssimon - 'openssl req' now has a '-newkey ecdsa:file' option; 1463160814Ssimon - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA; 1464160814Ssimon - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and 1465160814Ssimon d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make 1466160814Ssimon them suitable for ECDSA where domain parameters must be 1467160814Ssimon extracted before the specific public key; 1468160814Ssimon - ECDSA engine support has been added. 1469160814Ssimon [Nils Larsch <nla@trustcenter.de>] 1470160814Ssimon 1471160814Ssimon *) Include some named elliptic curves, and add OIDs from X9.62, 1472160814Ssimon SECG, and WAP/WTLS. Each curve can be obtained from the new 1473160814Ssimon function 1474160814Ssimon EC_GROUP_new_by_curve_name(), 1475160814Ssimon and the list of available named curves can be obtained with 1476160814Ssimon EC_get_builtin_curves(). 1477160814Ssimon Also add a 'curve_name' member to EC_GROUP objects, which can be 1478160814Ssimon accessed via 1479160814Ssimon EC_GROUP_set_curve_name() 1480160814Ssimon EC_GROUP_get_curve_name() 1481160814Ssimon [Nils Larsch <larsch@trustcenter.de, Bodo Moeller] 1482160814Ssimon 1483160814Ssimon *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 1484160814Ssimon was actually never needed) and in BN_mul(). The removal in BN_mul() 1485160814Ssimon required a small change in bn_mul_part_recursive() and the addition 1486160814Ssimon of the functions bn_cmp_part_words(), bn_sub_part_words() and 1487160814Ssimon bn_add_part_words(), which do the same thing as bn_cmp_words(), 1488160814Ssimon bn_sub_words() and bn_add_words() except they take arrays with 1489160814Ssimon differing sizes. 1490160814Ssimon [Richard Levitte] 1491160814Ssimon 1492194206Ssimon Changes between 0.9.7m and 0.9.7n [xx XXX xxxx] 1493162911Ssimon 1494194206Ssimon *) In the SSL/TLS server implementation, be strict about session ID 1495194206Ssimon context matching (which matters if an application uses a single 1496194206Ssimon external cache for different purposes). Previously, 1497194206Ssimon out-of-context reuse was forbidden only if SSL_VERIFY_PEER was 1498194206Ssimon set. This did ensure strict client verification, but meant that, 1499194206Ssimon with applications using a single external cache for quite 1500194206Ssimon different requirements, clients could circumvent ciphersuite 1501194206Ssimon restrictions for a given session ID context by starting a session 1502194206Ssimon in a different context. 1503194206Ssimon [Bodo Moeller] 1504194206Ssimon 1505194206Ssimon Changes between 0.9.7l and 0.9.7m [23 Feb 2007] 1506194206Ssimon 1507167612Ssimon *) Cleanse PEM buffers before freeing them since they may contain 1508167612Ssimon sensitive data. 1509167612Ssimon [Benjamin Bennett <ben@psc.edu>] 1510167612Ssimon 1511167612Ssimon *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 1512167612Ssimon a ciphersuite string such as "DEFAULT:RSA" cannot enable 1513167612Ssimon authentication-only ciphersuites. 1514167612Ssimon [Bodo Moeller] 1515167612Ssimon 1516167612Ssimon *) Since AES128 and AES256 share a single mask bit in the logic of 1517167612Ssimon ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 1518167612Ssimon kludge to work properly if AES128 is available and AES256 isn't. 1519167612Ssimon [Victor Duchovni] 1520167612Ssimon 1521194206Ssimon *) Expand security boundary to match 1.1.1 module. 1522194206Ssimon [Steve Henson] 1523194206Ssimon 1524194206Ssimon *) Remove redundant features: hash file source, editing of test vectors 1525194206Ssimon modify fipsld to use external fips_premain.c signature. 1526194206Ssimon [Steve Henson] 1527194206Ssimon 1528194206Ssimon *) New perl script mkfipsscr.pl to create shell scripts or batch files to 1529194206Ssimon run algorithm test programs. 1530194206Ssimon [Steve Henson] 1531194206Ssimon 1532194206Ssimon *) Make algorithm test programs more tolerant of whitespace. 1533194206Ssimon [Steve Henson] 1534194206Ssimon 1535167612Ssimon *) Have SSL/TLS server implementation tolerate "mismatched" record 1536167612Ssimon protocol version while receiving ClientHello even if the 1537167612Ssimon ClientHello is fragmented. (The server can't insist on the 1538167612Ssimon particular protocol version it has chosen before the ServerHello 1539167612Ssimon message has informed the client about his choice.) 1540167612Ssimon [Bodo Moeller] 1541167612Ssimon 1542167612Ssimon *) Load error codes if they are not already present instead of using a 1543167612Ssimon static variable. This allows them to be cleanly unloaded and reloaded. 1544167612Ssimon [Steve Henson] 1545167612Ssimon 1546167612Ssimon Changes between 0.9.7k and 0.9.7l [28 Sep 2006] 1547167612Ssimon 1548167612Ssimon *) Introduce limits to prevent malicious keys being able to 1549167612Ssimon cause a denial of service. (CVE-2006-2940) 1550167612Ssimon [Steve Henson, Bodo Moeller] 1551167612Ssimon 1552167612Ssimon *) Fix ASN.1 parsing of certain invalid structures that can result 1553167612Ssimon in a denial of service. (CVE-2006-2937) [Steve Henson] 1554167612Ssimon 1555167612Ssimon *) Fix buffer overflow in SSL_get_shared_ciphers() function. 1556167612Ssimon (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 1557167612Ssimon 1558167612Ssimon *) Fix SSL client code which could crash if connecting to a 1559167612Ssimon malicious SSLv2 server. (CVE-2006-4343) 1560167612Ssimon [Tavis Ormandy and Will Drewry, Google Security Team] 1561167612Ssimon 1562162911Ssimon *) Change ciphersuite string processing so that an explicit 1563162911Ssimon ciphersuite selects this one ciphersuite (so that "AES256-SHA" 1564162911Ssimon will no longer include "AES128-SHA"), and any other similar 1565162911Ssimon ciphersuite (same bitmap) from *other* protocol versions (so that 1566162911Ssimon "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the 1567162911Ssimon SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining 1568162911Ssimon changes from 0.9.8b and 0.9.8d. 1569162911Ssimon [Bodo Moeller] 1570162911Ssimon 1571162911Ssimon Changes between 0.9.7j and 0.9.7k [05 Sep 2006] 1572162911Ssimon 1573162911Ssimon *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 1574162911Ssimon (CVE-2006-4339) [Ben Laurie and Google Security Team] 1575162911Ssimon 1576162911Ssimon *) Change the Unix randomness entropy gathering to use poll() when 1577162911Ssimon possible instead of select(), since the latter has some 1578162911Ssimon undesirable limitations. 1579162911Ssimon [Darryl Miles via Richard Levitte and Bodo Moeller] 1580162911Ssimon 1581162911Ssimon *) Disable rogue ciphersuites: 1582162911Ssimon 1583162911Ssimon - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 1584162911Ssimon - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 1585162911Ssimon - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 1586162911Ssimon 1587162911Ssimon The latter two were purportedly from 1588162911Ssimon draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 1589162911Ssimon appear there. 1590162911Ssimon 1591162911Ssimon Also deactive the remaining ciphersuites from 1592162911Ssimon draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 1593162911Ssimon unofficial, and the ID has long expired. 1594162911Ssimon [Bodo Moeller] 1595162911Ssimon 1596162911Ssimon *) Fix RSA blinding Heisenbug (problems sometimes occured on 1597162911Ssimon dual-core machines) and other potential thread-safety issues. 1598162911Ssimon [Bodo Moeller] 1599162911Ssimon 1600162911Ssimon Changes between 0.9.7i and 0.9.7j [04 May 2006] 1601162911Ssimon 1602162911Ssimon *) Adapt fipsld and the build system to link against the validated FIPS 1603162911Ssimon module in FIPS mode. 1604162911Ssimon [Steve Henson] 1605162911Ssimon 1606162911Ssimon *) Fixes for VC++ 2005 build under Windows. 1607162911Ssimon [Steve Henson] 1608162911Ssimon 1609162911Ssimon *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make 1610162911Ssimon from a Windows bash shell such as MSYS. It is autodetected from the 1611162911Ssimon "config" script when run from a VC++ environment. Modify standard VC++ 1612162911Ssimon build to use fipscanister.o from the GNU make build. 1613162911Ssimon [Steve Henson] 1614162911Ssimon 1615160814Ssimon Changes between 0.9.7h and 0.9.7i [14 Oct 2005] 1616160814Ssimon 1617160814Ssimon *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS. 1618160814Ssimon The value now differs depending on if you build for FIPS or not. 1619160814Ssimon BEWARE! A program linked with a shared FIPSed libcrypto can't be 1620160814Ssimon safely run with a non-FIPSed libcrypto, as it may crash because of 1621160814Ssimon the difference induced by this change. 1622160814Ssimon [Andy Polyakov] 1623160814Ssimon 1624160814Ssimon Changes between 0.9.7g and 0.9.7h [11 Oct 2005] 1625160814Ssimon 1626160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 1627160814Ssimon (part of SSL_OP_ALL). This option used to disable the 1628160814Ssimon countermeasure against man-in-the-middle protocol-version 1629160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 1630160814Ssimon idea. (CVE-2005-2969) 1631160814Ssimon 1632160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 1633160814Ssimon for Information Security, National Institute of Advanced Industrial 1634160814Ssimon Science and Technology [AIST], Japan)] 1635160814Ssimon 1636160814Ssimon *) Minimal support for X9.31 signatures and PSS padding modes. This is 1637160814Ssimon mainly for FIPS compliance and not fully integrated at this stage. 1638160814Ssimon [Steve Henson] 1639160814Ssimon 1640160814Ssimon *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform 1641160814Ssimon the exponentiation using a fixed-length exponent. (Otherwise, 1642160814Ssimon the information leaked through timing could expose the secret key 1643160814Ssimon after many signatures; cf. Bleichenbacher's attack on DSA with 1644160814Ssimon biased k.) 1645160814Ssimon [Bodo Moeller] 1646160814Ssimon 1647160814Ssimon *) Make a new fixed-window mod_exp implementation the default for 1648160814Ssimon RSA, DSA, and DH private-key operations so that the sequence of 1649160814Ssimon squares and multiplies and the memory access pattern are 1650160814Ssimon independent of the particular secret key. This will mitigate 1651160814Ssimon cache-timing and potential related attacks. 1652160814Ssimon 1653160814Ssimon BN_mod_exp_mont_consttime() is the new exponentiation implementation, 1654160814Ssimon and this is automatically used by BN_mod_exp_mont() if the new flag 1655160814Ssimon BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH 1656160814Ssimon will use this BN flag for private exponents unless the flag 1657160814Ssimon RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or 1658160814Ssimon DH_FLAG_NO_EXP_CONSTTIME, respectively, is set. 1659160814Ssimon 1660160814Ssimon [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller] 1661160814Ssimon 1662160814Ssimon *) Change the client implementation for SSLv23_method() and 1663160814Ssimon SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0 1664160814Ssimon Client Hello message format if the SSL_OP_NO_SSLv2 option is set. 1665160814Ssimon (Previously, the SSL 2.0 backwards compatible Client Hello 1666160814Ssimon message format would be used even with SSL_OP_NO_SSLv2.) 1667160814Ssimon [Bodo Moeller] 1668160814Ssimon 1669160814Ssimon *) Add support for smime-type MIME parameter in S/MIME messages which some 1670160814Ssimon clients need. 1671160814Ssimon [Steve Henson] 1672160814Ssimon 1673160814Ssimon *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in 1674160814Ssimon a threadsafe manner. Modify rsa code to use new function and add calls 1675160814Ssimon to dsa and dh code (which had race conditions before). 1676160814Ssimon [Steve Henson] 1677160814Ssimon 1678160814Ssimon *) Include the fixed error library code in the C error file definitions 1679160814Ssimon instead of fixing them up at runtime. This keeps the error code 1680160814Ssimon structures constant. 1681160814Ssimon [Steve Henson] 1682160814Ssimon 1683160814Ssimon Changes between 0.9.7f and 0.9.7g [11 Apr 2005] 1684160814Ssimon 1685160814Ssimon [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after 1686160814Ssimon OpenSSL 0.9.8.] 1687160814Ssimon 1688160814Ssimon *) Fixes for newer kerberos headers. NB: the casts are needed because 1689160814Ssimon the 'length' field is signed on one version and unsigned on another 1690160814Ssimon with no (?) obvious way to tell the difference, without these VC++ 1691160814Ssimon complains. Also the "definition" of FAR (blank) is no longer included 1692160814Ssimon nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up 1693160814Ssimon some needed definitions. 1694160814Ssimon [Steve Henson] 1695160814Ssimon 1696160814Ssimon *) Undo Cygwin change. 1697160814Ssimon [Ulf M�ller] 1698160814Ssimon 1699160814Ssimon *) Added support for proxy certificates according to RFC 3820. 1700160814Ssimon Because they may be a security thread to unaware applications, 1701160814Ssimon they must be explicitely allowed in run-time. See 1702160814Ssimon docs/HOWTO/proxy_certificates.txt for further information. 1703160814Ssimon [Richard Levitte] 1704160814Ssimon 1705160814Ssimon Changes between 0.9.7e and 0.9.7f [22 Mar 2005] 1706160814Ssimon 1707160814Ssimon *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating 1708160814Ssimon server and client random values. Previously 1709160814Ssimon (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in 1710160814Ssimon less random data when sizeof(time_t) > 4 (some 64 bit platforms). 1711160814Ssimon 1712160814Ssimon This change has negligible security impact because: 1713160814Ssimon 1714160814Ssimon 1. Server and client random values still have 24 bytes of pseudo random 1715160814Ssimon data. 1716160814Ssimon 1717160814Ssimon 2. Server and client random values are sent in the clear in the initial 1718160814Ssimon handshake. 1719160814Ssimon 1720160814Ssimon 3. The master secret is derived using the premaster secret (48 bytes in 1721160814Ssimon size for static RSA ciphersuites) as well as client server and random 1722160814Ssimon values. 1723160814Ssimon 1724160814Ssimon The OpenSSL team would like to thank the UK NISCC for bringing this issue 1725160814Ssimon to our attention. 1726160814Ssimon 1727160814Ssimon [Stephen Henson, reported by UK NISCC] 1728160814Ssimon 1729160814Ssimon *) Use Windows randomness collection on Cygwin. 1730160814Ssimon [Ulf M�ller] 1731160814Ssimon 1732160814Ssimon *) Fix hang in EGD/PRNGD query when communication socket is closed 1733160814Ssimon prematurely by EGD/PRNGD. 1734160814Ssimon [Darren Tucker <dtucker@zip.com.au> via Lutz J�nicke, resolves #1014] 1735160814Ssimon 1736160814Ssimon *) Prompt for pass phrases when appropriate for PKCS12 input format. 1737160814Ssimon [Steve Henson] 1738160814Ssimon 1739160814Ssimon *) Back-port of selected performance improvements from development 1740160814Ssimon branch, as well as improved support for PowerPC platforms. 1741160814Ssimon [Andy Polyakov] 1742160814Ssimon 1743160814Ssimon *) Add lots of checks for memory allocation failure, error codes to indicate 1744160814Ssimon failure and freeing up memory if a failure occurs. 1745160814Ssimon [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson] 1746160814Ssimon 1747160814Ssimon *) Add new -passin argument to dgst. 1748160814Ssimon [Steve Henson] 1749160814Ssimon 1750160814Ssimon *) Perform some character comparisons of different types in X509_NAME_cmp: 1751160814Ssimon this is needed for some certificates that reencode DNs into UTF8Strings 1752160814Ssimon (in violation of RFC3280) and can't or wont issue name rollover 1753160814Ssimon certificates. 1754160814Ssimon [Steve Henson] 1755160814Ssimon 1756160814Ssimon *) Make an explicit check during certificate validation to see that 1757160814Ssimon the CA setting in each certificate on the chain is correct. As a 1758160814Ssimon side effect always do the following basic checks on extensions, 1759160814Ssimon not just when there's an associated purpose to the check: 1760160814Ssimon 1761160814Ssimon - if there is an unhandled critical extension (unless the user 1762160814Ssimon has chosen to ignore this fault) 1763160814Ssimon - if the path length has been exceeded (if one is set at all) 1764160814Ssimon - that certain extensions fit the associated purpose (if one has 1765160814Ssimon been given) 1766160814Ssimon [Richard Levitte] 1767160814Ssimon 1768142425Snectar Changes between 0.9.7d and 0.9.7e [25 Oct 2004] 1769142425Snectar 1770142425Snectar *) Avoid a race condition when CRLs are checked in a multi threaded 1771142425Snectar environment. This would happen due to the reordering of the revoked 1772142425Snectar entries during signature checking and serial number lookup. Now the 1773142425Snectar encoding is cached and the serial number sort performed under a lock. 1774142425Snectar Add new STACK function sk_is_sorted(). 1775142425Snectar [Steve Henson] 1776142425Snectar 1777142425Snectar *) Add Delta CRL to the extension code. 1778142425Snectar [Steve Henson] 1779142425Snectar 1780142425Snectar *) Various fixes to s3_pkt.c so alerts are sent properly. 1781142425Snectar [David Holmes <d.holmes@f5.com>] 1782142425Snectar 1783142425Snectar *) Reduce the chances of duplicate issuer name and serial numbers (in 1784142425Snectar violation of RFC3280) using the OpenSSL certificate creation utilities. 1785142425Snectar This is done by creating a random 64 bit value for the initial serial 1786142425Snectar number when a serial number file is created or when a self signed 1787142425Snectar certificate is created using 'openssl req -x509'. The initial serial 1788142425Snectar number file is created using 'openssl x509 -next_serial' in CA.pl 1789142425Snectar rather than being initialized to 1. 1790142425Snectar [Steve Henson] 1791142425Snectar 1792127128Snectar Changes between 0.9.7c and 0.9.7d [17 Mar 2004] 1793127128Snectar 1794127128Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 1795160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 1796127128Snectar [Joe Orton, Steve Henson] 1797127128Snectar 1798127128Snectar *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites 1799160814Ssimon (CVE-2004-0112) 1800127128Snectar [Joe Orton, Steve Henson] 1801127128Snectar 1802127128Snectar *) Make it possible to have multiple active certificates with the same 1803127128Snectar subject in the CA index file. This is done only if the keyword 1804127128Snectar 'unique_subject' is set to 'no' in the main CA section (default 1805127128Snectar if 'CA_default') of the configuration file. The value is saved 1806127128Snectar with the database itself in a separate index attribute file, 1807127128Snectar named like the index file with '.attr' appended to the name. 1808127128Snectar [Richard Levitte] 1809127128Snectar 1810127128Snectar *) X509 verify fixes. Disable broken certificate workarounds when 1811127128Snectar X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if 1812127128Snectar keyUsage extension present. Don't accept CRLs with unhandled critical 1813127128Snectar extensions: since verify currently doesn't process CRL extensions this 1814127128Snectar rejects a CRL with *any* critical extensions. Add new verify error codes 1815127128Snectar for these cases. 1816127128Snectar [Steve Henson] 1817127128Snectar 1818127128Snectar *) When creating an OCSP nonce use an OCTET STRING inside the extnValue. 1819127128Snectar A clarification of RFC2560 will require the use of OCTET STRINGs and 1820127128Snectar some implementations cannot handle the current raw format. Since OpenSSL 1821127128Snectar copies and compares OCSP nonces as opaque blobs without any attempt at 1822127128Snectar parsing them this should not create any compatibility issues. 1823127128Snectar [Steve Henson] 1824127128Snectar 1825127128Snectar *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when 1826127128Snectar calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without 1827127128Snectar this HMAC (and other) operations are several times slower than OpenSSL 1828127128Snectar < 0.9.7. 1829127128Snectar [Steve Henson] 1830127128Snectar 1831127128Snectar *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex(). 1832127128Snectar [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 1833127128Snectar 1834127128Snectar *) Use the correct content when signing type "other". 1835127128Snectar [Steve Henson] 1836127128Snectar 1837120631Snectar Changes between 0.9.7b and 0.9.7c [30 Sep 2003] 1838120631Snectar 1839120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 1840120631Snectar 1841120631Snectar Stop out of bounds reads in the ASN1 code when presented with 1842160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 1843120631Snectar 1844160814Ssimon Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545). 1845120631Snectar 1846120631Snectar If verify callback ignores invalid public key errors don't try to check 1847120631Snectar certificate signature with the NULL public key. 1848120631Snectar 1849120631Snectar [Steve Henson] 1850120631Snectar 1851120631Snectar *) New -ignore_err option in ocsp application to stop the server 1852120631Snectar exiting on the first error in a request. 1853120631Snectar [Steve Henson] 1854120631Snectar 1855120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 1856120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 1857120631Snectar specifications. 1858120631Snectar [Steve Henson] 1859120631Snectar 1860120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 1861120631Snectar extra data after the compression methods not only for TLS 1.0 1862120631Snectar but also for SSL 3.0 (as required by the specification). 1863120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 1864120631Snectar 1865120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 1866120631Snectar when it's 512 *bits* long, not 512 bytes. 1867120631Snectar [Richard Levitte] 1868120631Snectar 1869120631Snectar *) Change AES_cbc_encrypt() so it outputs exact multiple of 1870120631Snectar blocks during encryption. 1871120631Snectar [Richard Levitte] 1872120631Snectar 1873120631Snectar *) Various fixes to base64 BIO and non blocking I/O. On write 1874120631Snectar flushes were not handled properly if the BIO retried. On read 1875120631Snectar data was not being buffered properly and had various logic bugs. 1876120631Snectar This also affects blocking I/O when the data being decoded is a 1877120631Snectar certain size. 1878120631Snectar [Steve Henson] 1879120631Snectar 1880120631Snectar *) Various S/MIME bugfixes and compatibility changes: 1881120631Snectar output correct application/pkcs7 MIME type if 1882120631Snectar PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures. 1883120631Snectar Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening 1884120631Snectar of files as .eml work). Correctly handle very long lines in MIME 1885120631Snectar parser. 1886120631Snectar [Steve Henson] 1887120631Snectar 1888120631Snectar Changes between 0.9.7a and 0.9.7b [10 Apr 2003] 1889120631Snectar 1890120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 1891120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 1892120631Snectar a protocol version number mismatch like a decryption error 1893120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 1894120631Snectar [Bodo Moeller] 1895120631Snectar 1896120631Snectar *) Turn on RSA blinding by default in the default implementation 1897120631Snectar to avoid a timing attack. Applications that don't want it can call 1898120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 1899120631Snectar They would be ill-advised to do so in most cases. 1900120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 1901120631Snectar 1902120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 1903120631Snectar seeded (in this case, the secret RSA exponent is abused as 1904120631Snectar an unpredictable seed -- if it is not unpredictable, there 1905120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 1906120631Snectar by remembering the creator's thread ID in rsa->blinding and 1907120631Snectar having all other threads use local one-time blinding factors 1908120631Snectar (this requires more computation than sharing rsa->blinding, but 1909120631Snectar avoids excessive locking; and if an RSA object is not shared 1910120631Snectar between threads, blinding will still be very fast). 1911120631Snectar [Bodo Moeller] 1912120631Snectar 1913120631Snectar *) Fixed a typo bug that would cause ENGINE_set_default() to set an 1914120631Snectar ENGINE as defaults for all supported algorithms irrespective of 1915120631Snectar the 'flags' parameter. 'flags' is now honoured, so applications 1916120631Snectar should make sure they are passing it correctly. 1917120631Snectar [Geoff Thorpe] 1918120631Snectar 1919120631Snectar *) Target "mingw" now allows native Windows code to be generated in 1920120631Snectar the Cygwin environment as well as with the MinGW compiler. 1921120631Snectar [Ulf Moeller] 1922120631Snectar 1923111147Snectar Changes between 0.9.7 and 0.9.7a [19 Feb 2003] 1924111147Snectar 1925111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 1926111147Snectar via timing by performing a MAC computation even if incorrrect 1927111147Snectar block cipher padding has been found. This is a countermeasure 1928111147Snectar against active attacks where the attacker has to distinguish 1929160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 1930111147Snectar 1931111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 1932111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 1933111147Snectar Martin Vuagnoux (EPFL, Ilion)] 1934111147Snectar 1935111147Snectar *) Make the no-err option work as intended. The intention with no-err 1936111147Snectar is not to have the whole error stack handling routines removed from 1937111147Snectar libcrypto, it's only intended to remove all the function name and 1938111147Snectar reason texts, thereby removing some of the footprint that may not 1939111147Snectar be interesting if those errors aren't displayed anyway. 1940111147Snectar 1941111147Snectar NOTE: it's still possible for any application or module to have it's 1942111147Snectar own set of error texts inserted. The routines are there, just not 1943111147Snectar used by default when no-err is given. 1944111147Snectar [Richard Levitte] 1945111147Snectar 1946111147Snectar *) Add support for FreeBSD on IA64. 1947111147Snectar [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454] 1948111147Snectar 1949111147Snectar *) Adjust DES_cbc_cksum() so it returns the same value as the MIT 1950111147Snectar Kerberos function mit_des_cbc_cksum(). Before this change, 1951111147Snectar the value returned by DES_cbc_cksum() was like the one from 1952111147Snectar mit_des_cbc_cksum(), except the bytes were swapped. 1953111147Snectar [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte] 1954111147Snectar 1955111147Snectar *) Allow an application to disable the automatic SSL chain building. 1956111147Snectar Before this a rather primitive chain build was always performed in 1957111147Snectar ssl3_output_cert_chain(): an application had no way to send the 1958111147Snectar correct chain if the automatic operation produced an incorrect result. 1959111147Snectar 1960111147Snectar Now the chain builder is disabled if either: 1961111147Snectar 1962111147Snectar 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert(). 1963111147Snectar 1964111147Snectar 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set. 1965111147Snectar 1966111147Snectar The reasoning behind this is that an application would not want the 1967111147Snectar auto chain building to take place if extra chain certificates are 1968111147Snectar present and it might also want a means of sending no additional 1969111147Snectar certificates (for example the chain has two certificates and the 1970111147Snectar root is omitted). 1971111147Snectar [Steve Henson] 1972111147Snectar 1973111147Snectar *) Add the possibility to build without the ENGINE framework. 1974111147Snectar [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 1975111147Snectar 1976111147Snectar *) Under Win32 gmtime() can return NULL: check return value in 1977111147Snectar OPENSSL_gmtime(). Add error code for case where gmtime() fails. 1978111147Snectar [Steve Henson] 1979111147Snectar 1980111147Snectar *) DSA routines: under certain error conditions uninitialized BN objects 1981111147Snectar could be freed. Solution: make sure initialization is performed early 1982111147Snectar enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>, 1983111147Snectar Nils Larsch <nla@trustcenter.de> via PR#459) 1984111147Snectar [Lutz Jaenicke] 1985111147Snectar 1986111147Snectar *) Another fix for SSLv2 session ID handling: the session ID was incorrectly 1987111147Snectar checked on reconnect on the client side, therefore session resumption 1988111147Snectar could still fail with a "ssl session id is different" error. This 1989111147Snectar behaviour is masked when SSL_OP_ALL is used due to 1990111147Snectar SSL_OP_MICROSOFT_SESS_ID_BUG being set. 1991111147Snectar Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 1992111147Snectar followup to PR #377. 1993111147Snectar [Lutz Jaenicke] 1994111147Snectar 1995111147Snectar *) IA-32 assembler support enhancements: unified ELF targets, support 1996111147Snectar for SCO/Caldera platforms, fix for Cygwin shared build. 1997111147Snectar [Andy Polyakov] 1998111147Snectar 1999111147Snectar *) Add support for FreeBSD on sparc64. As a consequence, support for 2000111147Snectar FreeBSD on non-x86 processors is separate from x86 processors on 2001111147Snectar the config script, much like the NetBSD support. 2002111147Snectar [Richard Levitte & Kris Kennaway <kris@obsecurity.org>] 2003111147Snectar 2004109998Smarkm Changes between 0.9.6h and 0.9.7 [31 Dec 2002] 2005109998Smarkm 2006120631Snectar [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after 2007120631Snectar OpenSSL 0.9.7.] 2008120631Snectar 2009109998Smarkm *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED 2010109998Smarkm code (06) was taken as the first octet of the session ID and the last 2011109998Smarkm octet was ignored consequently. As a result SSLv2 client side session 2012109998Smarkm caching could not have worked due to the session ID mismatch between 2013109998Smarkm client and server. 2014109998Smarkm Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 2015109998Smarkm PR #377. 2016109998Smarkm [Lutz Jaenicke] 2017109998Smarkm 2018109998Smarkm *) Change the declaration of needed Kerberos libraries to use EX_LIBS 2019109998Smarkm instead of the special (and badly supported) LIBKRB5. LIBKRB5 is 2020109998Smarkm removed entirely. 2021109998Smarkm [Richard Levitte] 2022109998Smarkm 2023109998Smarkm *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it 2024109998Smarkm seems that in spite of existing for more than a year, many application 2025109998Smarkm author have done nothing to provide the necessary callbacks, which 2026109998Smarkm means that this particular engine will not work properly anywhere. 2027109998Smarkm This is a very unfortunate situation which forces us, in the name 2028109998Smarkm of usability, to give the hw_ncipher.c a static lock, which is part 2029109998Smarkm of libcrypto. 2030109998Smarkm NOTE: This is for the 0.9.7 series ONLY. This hack will never 2031109998Smarkm appear in 0.9.8 or later. We EXPECT application authors to have 2032109998Smarkm dealt properly with this when 0.9.8 is released (unless we actually 2033109998Smarkm make such changes in the libcrypto locking code that changes will 2034109998Smarkm have to be made anyway). 2035109998Smarkm [Richard Levitte] 2036109998Smarkm 2037109998Smarkm *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content 2038109998Smarkm octets have been read, EOF or an error occurs. Without this change 2039109998Smarkm some truncated ASN1 structures will not produce an error. 2040109998Smarkm [Steve Henson] 2041109998Smarkm 2042109998Smarkm *) Disable Heimdal support, since it hasn't been fully implemented. 2043109998Smarkm Still give the possibility to force the use of Heimdal, but with 2044109998Smarkm warnings and a request that patches get sent to openssl-dev. 2045109998Smarkm [Richard Levitte] 2046109998Smarkm 2047109998Smarkm *) Add the VC-CE target, introduce the WINCE sysname, and add 2048109998Smarkm INSTALL.WCE and appropriate conditionals to make it build. 2049109998Smarkm [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 2050109998Smarkm 2051109998Smarkm *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and 2052109998Smarkm cygssl-x.y.z.dll, where x, y and z are the major, minor and 2053109998Smarkm edit numbers of the version. 2054109998Smarkm [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 2055109998Smarkm 2056109998Smarkm *) Introduce safe string copy and catenation functions 2057109998Smarkm (BUF_strlcpy() and BUF_strlcat()). 2058109998Smarkm [Ben Laurie (CHATS) and Richard Levitte] 2059109998Smarkm 2060109998Smarkm *) Avoid using fixed-size buffers for one-line DNs. 2061109998Smarkm [Ben Laurie (CHATS)] 2062109998Smarkm 2063109998Smarkm *) Add BUF_MEM_grow_clean() to avoid information leakage when 2064109998Smarkm resizing buffers containing secrets, and use where appropriate. 2065109998Smarkm [Ben Laurie (CHATS)] 2066109998Smarkm 2067109998Smarkm *) Avoid using fixed size buffers for configuration file location. 2068109998Smarkm [Ben Laurie (CHATS)] 2069109998Smarkm 2070109998Smarkm *) Avoid filename truncation for various CA files. 2071109998Smarkm [Ben Laurie (CHATS)] 2072109998Smarkm 2073109998Smarkm *) Use sizeof in preference to magic numbers. 2074109998Smarkm [Ben Laurie (CHATS)] 2075109998Smarkm 2076109998Smarkm *) Avoid filename truncation in cert requests. 2077109998Smarkm [Ben Laurie (CHATS)] 2078109998Smarkm 2079109998Smarkm *) Add assertions to check for (supposedly impossible) buffer 2080109998Smarkm overflows. 2081109998Smarkm [Ben Laurie (CHATS)] 2082109998Smarkm 2083109998Smarkm *) Don't cache truncated DNS entries in the local cache (this could 2084109998Smarkm potentially lead to a spoofing attack). 2085109998Smarkm [Ben Laurie (CHATS)] 2086109998Smarkm 2087109998Smarkm *) Fix various buffers to be large enough for hex/decimal 2088109998Smarkm representations in a platform independent manner. 2089109998Smarkm [Ben Laurie (CHATS)] 2090109998Smarkm 2091109998Smarkm *) Add CRYPTO_realloc_clean() to avoid information leakage when 2092109998Smarkm resizing buffers containing secrets, and use where appropriate. 2093109998Smarkm [Ben Laurie (CHATS)] 2094109998Smarkm 2095109998Smarkm *) Add BIO_indent() to avoid much slightly worrying code to do 2096109998Smarkm indents. 2097109998Smarkm [Ben Laurie (CHATS)] 2098109998Smarkm 2099109998Smarkm *) Convert sprintf()/BIO_puts() to BIO_printf(). 2100109998Smarkm [Ben Laurie (CHATS)] 2101109998Smarkm 2102109998Smarkm *) buffer_gets() could terminate with the buffer only half 2103109998Smarkm full. Fixed. 2104109998Smarkm [Ben Laurie (CHATS)] 2105109998Smarkm 2106109998Smarkm *) Add assertions to prevent user-supplied crypto functions from 2107109998Smarkm overflowing internal buffers by having large block sizes, etc. 2108109998Smarkm [Ben Laurie (CHATS)] 2109109998Smarkm 2110109998Smarkm *) New OPENSSL_assert() macro (similar to assert(), but enabled 2111109998Smarkm unconditionally). 2112109998Smarkm [Ben Laurie (CHATS)] 2113109998Smarkm 2114109998Smarkm *) Eliminate unused copy of key in RC4. 2115109998Smarkm [Ben Laurie (CHATS)] 2116109998Smarkm 2117109998Smarkm *) Eliminate unused and incorrectly sized buffers for IV in pem.h. 2118109998Smarkm [Ben Laurie (CHATS)] 2119109998Smarkm 2120109998Smarkm *) Fix off-by-one error in EGD path. 2121109998Smarkm [Ben Laurie (CHATS)] 2122109998Smarkm 2123109998Smarkm *) If RANDFILE path is too long, ignore instead of truncating. 2124109998Smarkm [Ben Laurie (CHATS)] 2125109998Smarkm 2126109998Smarkm *) Eliminate unused and incorrectly sized X.509 structure 2127109998Smarkm CBCParameter. 2128109998Smarkm [Ben Laurie (CHATS)] 2129109998Smarkm 2130109998Smarkm *) Eliminate unused and dangerous function knumber(). 2131109998Smarkm [Ben Laurie (CHATS)] 2132109998Smarkm 2133109998Smarkm *) Eliminate unused and dangerous structure, KSSL_ERR. 2134109998Smarkm [Ben Laurie (CHATS)] 2135109998Smarkm 2136109998Smarkm *) Protect against overlong session ID context length in an encoded 2137109998Smarkm session object. Since these are local, this does not appear to be 2138109998Smarkm exploitable. 2139109998Smarkm [Ben Laurie (CHATS)] 2140109998Smarkm 2141109998Smarkm *) Change from security patch (see 0.9.6e below) that did not affect 2142109998Smarkm the 0.9.6 release series: 2143109998Smarkm 2144109998Smarkm Remote buffer overflow in SSL3 protocol - an attacker could 2145109998Smarkm supply an oversized master key in Kerberos-enabled versions. 2146160814Ssimon (CVE-2002-0657) 2147109998Smarkm [Ben Laurie (CHATS)] 2148109998Smarkm 2149109998Smarkm *) Change the SSL kerb5 codes to match RFC 2712. 2150109998Smarkm [Richard Levitte] 2151109998Smarkm 2152109998Smarkm *) Make -nameopt work fully for req and add -reqopt switch. 2153109998Smarkm [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson] 2154109998Smarkm 2155109998Smarkm *) The "block size" for block ciphers in CFB and OFB mode should be 1. 2156109998Smarkm [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>] 2157109998Smarkm 2158109998Smarkm *) Make sure tests can be performed even if the corresponding algorithms 2159109998Smarkm have been removed entirely. This was also the last step to make 2160109998Smarkm OpenSSL compilable with DJGPP under all reasonable conditions. 2161109998Smarkm [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>] 2162109998Smarkm 2163109998Smarkm *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT 2164109998Smarkm to allow version independent disabling of normally unselected ciphers, 2165109998Smarkm which may be activated as a side-effect of selecting a single cipher. 2166109998Smarkm 2167109998Smarkm (E.g., cipher list string "RSA" enables ciphersuites that are left 2168109998Smarkm out of "ALL" because they do not provide symmetric encryption. 2169109998Smarkm "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) 2170109998Smarkm [Lutz Jaenicke, Bodo Moeller] 2171109998Smarkm 2172109998Smarkm *) Add appropriate support for separate platform-dependent build 2173109998Smarkm directories. The recommended way to make a platform-dependent 2174109998Smarkm build directory is the following (tested on Linux), maybe with 2175109998Smarkm some local tweaks: 2176109998Smarkm 2177109998Smarkm # Place yourself outside of the OpenSSL source tree. In 2178109998Smarkm # this example, the environment variable OPENSSL_SOURCE 2179109998Smarkm # is assumed to contain the absolute OpenSSL source directory. 2180109998Smarkm mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" 2181109998Smarkm cd objtree/"`uname -s`-`uname -r`-`uname -m`" 2182111147Snectar (cd $OPENSSL_SOURCE; find . -type f) | while read F; do 2183109998Smarkm mkdir -p `dirname $F` 2184109998Smarkm ln -s $OPENSSL_SOURCE/$F $F 2185109998Smarkm done 2186109998Smarkm 2187109998Smarkm To be absolutely sure not to disturb the source tree, a "make clean" 2188109998Smarkm is a good thing. If it isn't successfull, don't worry about it, 2189109998Smarkm it probably means the source directory is very clean. 2190109998Smarkm [Richard Levitte] 2191109998Smarkm 2192109998Smarkm *) Make sure any ENGINE control commands make local copies of string 2193109998Smarkm pointers passed to them whenever necessary. Otherwise it is possible 2194109998Smarkm the caller may have overwritten (or deallocated) the original string 2195109998Smarkm data when a later ENGINE operation tries to use the stored values. 2196109998Smarkm [G�tz Babin-Ebell <babinebell@trustcenter.de>] 2197109998Smarkm 2198109998Smarkm *) Improve diagnostics in file reading and command-line digests. 2199109998Smarkm [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>] 2200109998Smarkm 2201109998Smarkm *) Add AES modes CFB and OFB to the object database. Correct an 2202109998Smarkm error in AES-CFB decryption. 2203109998Smarkm [Richard Levitte] 2204109998Smarkm 2205109998Smarkm *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this 2206109998Smarkm allows existing EVP_CIPHER_CTX structures to be reused after 2207109998Smarkm calling EVP_*Final(). This behaviour is used by encryption 2208109998Smarkm BIOs and some applications. This has the side effect that 2209109998Smarkm applications must explicitly clean up cipher contexts with 2210109998Smarkm EVP_CIPHER_CTX_cleanup() or they will leak memory. 2211109998Smarkm [Steve Henson] 2212109998Smarkm 2213109998Smarkm *) Check the values of dna and dnb in bn_mul_recursive before calling 2214109998Smarkm bn_mul_comba (a non zero value means the a or b arrays do not contain 2215109998Smarkm n2 elements) and fallback to bn_mul_normal if either is not zero. 2216109998Smarkm [Steve Henson] 2217109998Smarkm 2218109998Smarkm *) Fix escaping of non-ASCII characters when using the -subj option 2219109998Smarkm of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) 2220109998Smarkm [Lutz Jaenicke] 2221109998Smarkm 2222109998Smarkm *) Make object definitions compliant to LDAP (RFC2256): SN is the short 2223109998Smarkm form for "surname", serialNumber has no short form. 2224109998Smarkm Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; 2225109998Smarkm therefore remove "mail" short name for "internet 7". 2226109998Smarkm The OID for unique identifiers in X509 certificates is 2227109998Smarkm x500UniqueIdentifier, not uniqueIdentifier. 2228109998Smarkm Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) 2229109998Smarkm [Lutz Jaenicke] 2230109998Smarkm 2231109998Smarkm *) Add an "init" command to the ENGINE config module and auto initialize 2232109998Smarkm ENGINEs. Without any "init" command the ENGINE will be initialized 2233109998Smarkm after all ctrl commands have been executed on it. If init=1 the 2234109998Smarkm ENGINE is initailized at that point (ctrls before that point are run 2235109998Smarkm on the uninitialized ENGINE and after on the initialized one). If 2236109998Smarkm init=0 then the ENGINE will not be iniatialized at all. 2237109998Smarkm [Steve Henson] 2238109998Smarkm 2239109998Smarkm *) Fix the 'app_verify_callback' interface so that the user-defined 2240109998Smarkm argument is actually passed to the callback: In the 2241109998Smarkm SSL_CTX_set_cert_verify_callback() prototype, the callback 2242109998Smarkm declaration has been changed from 2243109998Smarkm int (*cb)() 2244109998Smarkm into 2245109998Smarkm int (*cb)(X509_STORE_CTX *,void *); 2246109998Smarkm in ssl_verify_cert_chain (ssl/ssl_cert.c), the call 2247109998Smarkm i=s->ctx->app_verify_callback(&ctx) 2248109998Smarkm has been changed into 2249109998Smarkm i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). 2250109998Smarkm 2251109998Smarkm To update applications using SSL_CTX_set_cert_verify_callback(), 2252109998Smarkm a dummy argument can be added to their callback functions. 2253109998Smarkm [D. K. Smetters <smetters@parc.xerox.com>] 2254109998Smarkm 2255109998Smarkm *) Added the '4758cca' ENGINE to support IBM 4758 cards. 2256109998Smarkm [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe] 2257109998Smarkm 2258109998Smarkm *) Add and OPENSSL_LOAD_CONF define which will cause 2259109998Smarkm OpenSSL_add_all_algorithms() to load the openssl.cnf config file. 2260109998Smarkm This allows older applications to transparently support certain 2261109998Smarkm OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. 2262109998Smarkm Two new functions OPENSSL_add_all_algorithms_noconf() which will never 2263109998Smarkm load the config file and OPENSSL_add_all_algorithms_conf() which will 2264109998Smarkm always load it have also been added. 2265109998Smarkm [Steve Henson] 2266109998Smarkm 2267109998Smarkm *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. 2268109998Smarkm Adjust NIDs and EVP layer. 2269109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 2270109998Smarkm 2271109998Smarkm *) Config modules support in openssl utility. 2272109998Smarkm 2273109998Smarkm Most commands now load modules from the config file, 2274109998Smarkm though in a few (such as version) this isn't done 2275109998Smarkm because it couldn't be used for anything. 2276109998Smarkm 2277109998Smarkm In the case of ca and req the config file used is 2278109998Smarkm the same as the utility itself: that is the -config 2279109998Smarkm command line option can be used to specify an 2280109998Smarkm alternative file. 2281109998Smarkm [Steve Henson] 2282109998Smarkm 2283109998Smarkm *) Move default behaviour from OPENSSL_config(). If appname is NULL 2284109998Smarkm use "openssl_conf" if filename is NULL use default openssl config file. 2285109998Smarkm [Steve Henson] 2286109998Smarkm 2287109998Smarkm *) Add an argument to OPENSSL_config() to allow the use of an alternative 2288109998Smarkm config section name. Add a new flag to tolerate a missing config file 2289109998Smarkm and move code to CONF_modules_load_file(). 2290109998Smarkm [Steve Henson] 2291109998Smarkm 2292109998Smarkm *) Support for crypto accelerator cards from Accelerated Encryption 2293109998Smarkm Processing, www.aep.ie. (Use engine 'aep') 2294109998Smarkm The support was copied from 0.9.6c [engine] and adapted/corrected 2295109998Smarkm to work with the new engine framework. 2296109998Smarkm [AEP Inc. and Richard Levitte] 2297109998Smarkm 2298109998Smarkm *) Support for SureWare crypto accelerator cards from Baltimore 2299109998Smarkm Technologies. (Use engine 'sureware') 2300109998Smarkm The support was copied from 0.9.6c [engine] and adapted 2301109998Smarkm to work with the new engine framework. 2302109998Smarkm [Richard Levitte] 2303109998Smarkm 2304109998Smarkm *) Have the CHIL engine fork-safe (as defined by nCipher) and actually 2305109998Smarkm make the newer ENGINE framework commands for the CHIL engine work. 2306109998Smarkm [Toomas Kiisk <vix@cyber.ee> and Richard Levitte] 2307109998Smarkm 2308109998Smarkm *) Make it possible to produce shared libraries on ReliantUNIX. 2309109998Smarkm [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte] 2310109998Smarkm 2311109998Smarkm *) Add the configuration target debug-linux-ppro. 2312109998Smarkm Make 'openssl rsa' use the general key loading routines 2313109998Smarkm implemented in apps.c, and make those routines able to 2314109998Smarkm handle the key format FORMAT_NETSCAPE and the variant 2315109998Smarkm FORMAT_IISSGC. 2316109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 2317109998Smarkm 2318109998Smarkm *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 2319109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 2320109998Smarkm 2321109998Smarkm *) Add -keyform to rsautl, and document -engine. 2322109998Smarkm [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>] 2323109998Smarkm 2324109998Smarkm *) Change BIO_new_file (crypto/bio/bss_file.c) to use new 2325109998Smarkm BIO_R_NO_SUCH_FILE error code rather than the generic 2326109998Smarkm ERR_R_SYS_LIB error code if fopen() fails with ENOENT. 2327109998Smarkm [Ben Laurie] 2328109998Smarkm 2329109998Smarkm *) Add new functions 2330109998Smarkm ERR_peek_last_error 2331109998Smarkm ERR_peek_last_error_line 2332109998Smarkm ERR_peek_last_error_line_data. 2333109998Smarkm These are similar to 2334109998Smarkm ERR_peek_error 2335109998Smarkm ERR_peek_error_line 2336109998Smarkm ERR_peek_error_line_data, 2337109998Smarkm but report on the latest error recorded rather than the first one 2338109998Smarkm still in the error queue. 2339109998Smarkm [Ben Laurie, Bodo Moeller] 2340109998Smarkm 2341109998Smarkm *) default_algorithms option in ENGINE config module. This allows things 2342109998Smarkm like: 2343109998Smarkm default_algorithms = ALL 2344109998Smarkm default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS 2345109998Smarkm [Steve Henson] 2346109998Smarkm 2347109998Smarkm *) Prelminary ENGINE config module. 2348109998Smarkm [Steve Henson] 2349109998Smarkm 2350109998Smarkm *) New experimental application configuration code. 2351109998Smarkm [Steve Henson] 2352109998Smarkm 2353109998Smarkm *) Change the AES code to follow the same name structure as all other 2354109998Smarkm symmetric ciphers, and behave the same way. Move everything to 2355109998Smarkm the directory crypto/aes, thereby obsoleting crypto/rijndael. 2356109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 2357109998Smarkm 2358109998Smarkm *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. 2359109998Smarkm [Ben Laurie and Theo de Raadt] 2360109998Smarkm 2361109998Smarkm *) Add option to output public keys in req command. 2362109998Smarkm [Massimiliano Pala madwolf@openca.org] 2363109998Smarkm 2364109998Smarkm *) Use wNAFs in EC_POINTs_mul() for improved efficiency 2365109998Smarkm (up to about 10% better than before for P-192 and P-224). 2366109998Smarkm [Bodo Moeller] 2367109998Smarkm 2368109998Smarkm *) New functions/macros 2369109998Smarkm 2370109998Smarkm SSL_CTX_set_msg_callback(ctx, cb) 2371109998Smarkm SSL_CTX_set_msg_callback_arg(ctx, arg) 2372109998Smarkm SSL_set_msg_callback(ssl, cb) 2373109998Smarkm SSL_set_msg_callback_arg(ssl, arg) 2374109998Smarkm 2375109998Smarkm to request calling a callback function 2376109998Smarkm 2377109998Smarkm void cb(int write_p, int version, int content_type, 2378109998Smarkm const void *buf, size_t len, SSL *ssl, void *arg) 2379109998Smarkm 2380109998Smarkm whenever a protocol message has been completely received 2381109998Smarkm (write_p == 0) or sent (write_p == 1). Here 'version' is the 2382109998Smarkm protocol version according to which the SSL library interprets 2383109998Smarkm the current protocol message (SSL2_VERSION, SSL3_VERSION, or 2384109998Smarkm TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or 2385109998Smarkm the content type as defined in the SSL 3.0/TLS 1.0 protocol 2386109998Smarkm specification (change_cipher_spec(20), alert(21), handshake(22)). 2387109998Smarkm 'buf' and 'len' point to the actual message, 'ssl' to the 2388109998Smarkm SSL object, and 'arg' is the application-defined value set by 2389109998Smarkm SSL[_CTX]_set_msg_callback_arg(). 2390109998Smarkm 2391109998Smarkm 'openssl s_client' and 'openssl s_server' have new '-msg' options 2392109998Smarkm to enable a callback that displays all protocol messages. 2393109998Smarkm [Bodo Moeller] 2394109998Smarkm 2395109998Smarkm *) Change the shared library support so shared libraries are built as 2396109998Smarkm soon as the corresponding static library is finished, and thereby get 2397109998Smarkm openssl and the test programs linked against the shared library. 2398109998Smarkm This still only happens when the keyword "shard" has been given to 2399109998Smarkm the configuration scripts. 2400109998Smarkm 2401109998Smarkm NOTE: shared library support is still an experimental thing, and 2402109998Smarkm backward binary compatibility is still not guaranteed. 2403109998Smarkm ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte] 2404109998Smarkm 2405109998Smarkm *) Add support for Subject Information Access extension. 2406109998Smarkm [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 2407109998Smarkm 2408109998Smarkm *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero 2409109998Smarkm additional bytes when new memory had to be allocated, not just 2410109998Smarkm when reusing an existing buffer. 2411109998Smarkm [Bodo Moeller] 2412109998Smarkm 2413109998Smarkm *) New command line and configuration option 'utf8' for the req command. 2414109998Smarkm This allows field values to be specified as UTF8 strings. 2415109998Smarkm [Steve Henson] 2416109998Smarkm 2417109998Smarkm *) Add -multi and -mr options to "openssl speed" - giving multiple parallel 2418109998Smarkm runs for the former and machine-readable output for the latter. 2419109998Smarkm [Ben Laurie] 2420109998Smarkm 2421109998Smarkm *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion 2422109998Smarkm of the e-mail address in the DN (i.e., it will go into a certificate 2423109998Smarkm extension only). The new configuration file option 'email_in_dn = no' 2424109998Smarkm has the same effect. 2425109998Smarkm [Massimiliano Pala madwolf@openca.org] 2426109998Smarkm 2427109998Smarkm *) Change all functions with names starting with des_ to be starting 2428109998Smarkm with DES_ instead. Add wrappers that are compatible with libdes, 2429109998Smarkm but are named _ossl_old_des_*. Finally, add macros that map the 2430109998Smarkm des_* symbols to the corresponding _ossl_old_des_* if libdes 2431109998Smarkm compatibility is desired. If OpenSSL 0.9.6c compatibility is 2432109998Smarkm desired, the des_* symbols will be mapped to DES_*, with one 2433109998Smarkm exception. 2434109998Smarkm 2435109998Smarkm Since we provide two compatibility mappings, the user needs to 2436109998Smarkm define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes 2437109998Smarkm compatibility is desired. The default (i.e., when that macro 2438109998Smarkm isn't defined) is OpenSSL 0.9.6c compatibility. 2439109998Smarkm 2440109998Smarkm There are also macros that enable and disable the support of old 2441109998Smarkm des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT 2442109998Smarkm and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those 2443109998Smarkm are defined, the default will apply: to support the old des routines. 2444109998Smarkm 2445109998Smarkm In either case, one must include openssl/des.h to get the correct 2446109998Smarkm definitions. Do not try to just include openssl/des_old.h, that 2447109998Smarkm won't work. 2448109998Smarkm 2449109998Smarkm NOTE: This is a major break of an old API into a new one. Software 2450109998Smarkm authors are encouraged to switch to the DES_ style functions. Some 2451109998Smarkm time in the future, des_old.h and the libdes compatibility functions 2452109998Smarkm will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the 2453109998Smarkm default), and then completely removed. 2454109998Smarkm [Richard Levitte] 2455109998Smarkm 2456109998Smarkm *) Test for certificates which contain unsupported critical extensions. 2457109998Smarkm If such a certificate is found during a verify operation it is 2458109998Smarkm rejected by default: this behaviour can be overridden by either 2459109998Smarkm handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or 2460109998Smarkm by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function 2461109998Smarkm X509_supported_extension() has also been added which returns 1 if a 2462109998Smarkm particular extension is supported. 2463109998Smarkm [Steve Henson] 2464109998Smarkm 2465109998Smarkm *) Modify the behaviour of EVP cipher functions in similar way to digests 2466109998Smarkm to retain compatibility with existing code. 2467109998Smarkm [Steve Henson] 2468109998Smarkm 2469109998Smarkm *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain 2470109998Smarkm compatibility with existing code. In particular the 'ctx' parameter does 2471109998Smarkm not have to be to be initialized before the call to EVP_DigestInit() and 2472109998Smarkm it is tidied up after a call to EVP_DigestFinal(). New function 2473109998Smarkm EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function 2474109998Smarkm EVP_MD_CTX_copy() changed to not require the destination to be 2475109998Smarkm initialized valid and new function EVP_MD_CTX_copy_ex() added which 2476109998Smarkm requires the destination to be valid. 2477109998Smarkm 2478109998Smarkm Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), 2479109998Smarkm EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). 2480109998Smarkm [Steve Henson] 2481109998Smarkm 2482109998Smarkm *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it 2483109998Smarkm so that complete 'Handshake' protocol structures are kept in memory 2484109998Smarkm instead of overwriting 'msg_type' and 'length' with 'body' data. 2485109998Smarkm [Bodo Moeller] 2486109998Smarkm 2487109998Smarkm *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. 2488109998Smarkm [Massimo Santin via Richard Levitte] 2489109998Smarkm 2490109998Smarkm *) Major restructuring to the underlying ENGINE code. This includes 2491109998Smarkm reduction of linker bloat, separation of pure "ENGINE" manipulation 2492109998Smarkm (initialisation, etc) from functionality dealing with implementations 2493109998Smarkm of specific crypto iterfaces. This change also introduces integrated 2494109998Smarkm support for symmetric ciphers and digest implementations - so ENGINEs 2495109998Smarkm can now accelerate these by providing EVP_CIPHER and EVP_MD 2496109998Smarkm implementations of their own. This is detailed in crypto/engine/README 2497109998Smarkm as it couldn't be adequately described here. However, there are a few 2498109998Smarkm API changes worth noting - some RSA, DSA, DH, and RAND functions that 2499109998Smarkm were changed in the original introduction of ENGINE code have now 2500109998Smarkm reverted back - the hooking from this code to ENGINE is now a good 2501109998Smarkm deal more passive and at run-time, operations deal directly with 2502109998Smarkm RSA_METHODs, DSA_METHODs (etc) as they did before, rather than 2503109998Smarkm dereferencing through an ENGINE pointer any more. Also, the ENGINE 2504109998Smarkm functions dealing with BN_MOD_EXP[_CRT] handlers have been removed - 2505109998Smarkm they were not being used by the framework as there is no concept of a 2506109998Smarkm BIGNUM_METHOD and they could not be generalised to the new 2507109998Smarkm 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, 2508109998Smarkm ENGINE_cpy() has been removed as it cannot be consistently defined in 2509109998Smarkm the new code. 2510109998Smarkm [Geoff Thorpe] 2511109998Smarkm 2512109998Smarkm *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. 2513109998Smarkm [Steve Henson] 2514109998Smarkm 2515109998Smarkm *) Change mkdef.pl to sort symbols that get the same entry number, 2516109998Smarkm and make sure the automatically generated functions ERR_load_* 2517109998Smarkm become part of libeay.num as well. 2518109998Smarkm [Richard Levitte] 2519109998Smarkm 2520109998Smarkm *) New function SSL_renegotiate_pending(). This returns true once 2521109998Smarkm renegotiation has been requested (either SSL_renegotiate() call 2522109998Smarkm or HelloRequest/ClientHello receveived from the peer) and becomes 2523109998Smarkm false once a handshake has been completed. 2524109998Smarkm (For servers, SSL_renegotiate() followed by SSL_do_handshake() 2525109998Smarkm sends a HelloRequest, but does not ensure that a handshake takes 2526109998Smarkm place. SSL_renegotiate_pending() is useful for checking if the 2527109998Smarkm client has followed the request.) 2528109998Smarkm [Bodo Moeller] 2529109998Smarkm 2530109998Smarkm *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. 2531109998Smarkm By default, clients may request session resumption even during 2532109998Smarkm renegotiation (if session ID contexts permit); with this option, 2533109998Smarkm session resumption is possible only in the first handshake. 2534109998Smarkm 2535109998Smarkm SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes 2536109998Smarkm more bits available for options that should not be part of 2537109998Smarkm SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). 2538109998Smarkm [Bodo Moeller] 2539109998Smarkm 2540109998Smarkm *) Add some demos for certificate and certificate request creation. 2541109998Smarkm [Steve Henson] 2542109998Smarkm 2543109998Smarkm *) Make maximum certificate chain size accepted from the peer application 2544109998Smarkm settable (SSL*_get/set_max_cert_list()), as proposed by 2545109998Smarkm "Douglas E. Engert" <deengert@anl.gov>. 2546109998Smarkm [Lutz Jaenicke] 2547109998Smarkm 2548109998Smarkm *) Add support for shared libraries for Unixware-7 2549109998Smarkm (Boyd Lynn Gerber <gerberb@zenez.com>). 2550109998Smarkm [Lutz Jaenicke] 2551109998Smarkm 2552109998Smarkm *) Add a "destroy" handler to ENGINEs that allows structural cleanup to 2553109998Smarkm be done prior to destruction. Use this to unload error strings from 2554109998Smarkm ENGINEs that load their own error strings. NB: This adds two new API 2555109998Smarkm functions to "get" and "set" this destroy handler in an ENGINE. 2556109998Smarkm [Geoff Thorpe] 2557109998Smarkm 2558109998Smarkm *) Alter all existing ENGINE implementations (except "openssl" and 2559109998Smarkm "openbsd") to dynamically instantiate their own error strings. This 2560109998Smarkm makes them more flexible to be built both as statically-linked ENGINEs 2561109998Smarkm and self-contained shared-libraries loadable via the "dynamic" ENGINE. 2562109998Smarkm Also, add stub code to each that makes building them as self-contained 2563109998Smarkm shared-libraries easier (see README.ENGINE). 2564109998Smarkm [Geoff Thorpe] 2565109998Smarkm 2566109998Smarkm *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE 2567109998Smarkm implementations into applications that are completely implemented in 2568109998Smarkm self-contained shared-libraries. The "dynamic" ENGINE exposes control 2569109998Smarkm commands that can be used to configure what shared-library to load and 2570109998Smarkm to control aspects of the way it is handled. Also, made an update to 2571109998Smarkm the README.ENGINE file that brings its information up-to-date and 2572109998Smarkm provides some information and instructions on the "dynamic" ENGINE 2573109998Smarkm (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). 2574109998Smarkm [Geoff Thorpe] 2575109998Smarkm 2576109998Smarkm *) Make it possible to unload ranges of ERR strings with a new 2577109998Smarkm "ERR_unload_strings" function. 2578109998Smarkm [Geoff Thorpe] 2579109998Smarkm 2580109998Smarkm *) Add a copy() function to EVP_MD. 2581109998Smarkm [Ben Laurie] 2582109998Smarkm 2583109998Smarkm *) Make EVP_MD routines take a context pointer instead of just the 2584109998Smarkm md_data void pointer. 2585109998Smarkm [Ben Laurie] 2586109998Smarkm 2587109998Smarkm *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates 2588109998Smarkm that the digest can only process a single chunk of data 2589109998Smarkm (typically because it is provided by a piece of 2590109998Smarkm hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application 2591109998Smarkm is only going to provide a single chunk of data, and hence the 2592109998Smarkm framework needn't accumulate the data for oneshot drivers. 2593109998Smarkm [Ben Laurie] 2594109998Smarkm 2595109998Smarkm *) As with "ERR", make it possible to replace the underlying "ex_data" 2596109998Smarkm functions. This change also alters the storage and management of global 2597109998Smarkm ex_data state - it's now all inside ex_data.c and all "class" code (eg. 2598109998Smarkm RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class 2599109998Smarkm index counters. The API functions that use this state have been changed 2600109998Smarkm to take a "class_index" rather than pointers to the class's local STACK 2601109998Smarkm and counter, and there is now an API function to dynamically create new 2602109998Smarkm classes. This centralisation allows us to (a) plug a lot of the 2603109998Smarkm thread-safety problems that existed, and (b) makes it possible to clean 2604109998Smarkm up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) 2605109998Smarkm such data would previously have always leaked in application code and 2606109998Smarkm workarounds were in place to make the memory debugging turn a blind eye 2607109998Smarkm to it. Application code that doesn't use this new function will still 2608109998Smarkm leak as before, but their memory debugging output will announce it now 2609109998Smarkm rather than letting it slide. 2610109998Smarkm 2611109998Smarkm Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change 2612109998Smarkm induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now 2613109998Smarkm has a return value to indicate success or failure. 2614109998Smarkm [Geoff Thorpe] 2615109998Smarkm 2616109998Smarkm *) Make it possible to replace the underlying "ERR" functions such that the 2617109998Smarkm global state (2 LHASH tables and 2 locks) is only used by the "default" 2618109998Smarkm implementation. This change also adds two functions to "get" and "set" 2619109998Smarkm the implementation prior to it being automatically set the first time 2620109998Smarkm any other ERR function takes place. Ie. an application can call "get", 2621109998Smarkm pass the return value to a module it has just loaded, and that module 2622109998Smarkm can call its own "set" function using that value. This means the 2623109998Smarkm module's "ERR" operations will use (and modify) the error state in the 2624109998Smarkm application and not in its own statically linked copy of OpenSSL code. 2625109998Smarkm [Geoff Thorpe] 2626109998Smarkm 2627109998Smarkm *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment 2628109998Smarkm reference counts. This performs normal REF_PRINT/REF_CHECK macros on 2629109998Smarkm the operation, and provides a more encapsulated way for external code 2630109998Smarkm (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code 2631109998Smarkm to use these functions rather than manually incrementing the counts. 2632109998Smarkm 2633109998Smarkm Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". 2634109998Smarkm [Geoff Thorpe] 2635109998Smarkm 2636109998Smarkm *) Add EVP test program. 2637109998Smarkm [Ben Laurie] 2638109998Smarkm 2639109998Smarkm *) Add symmetric cipher support to ENGINE. Expect the API to change! 2640109998Smarkm [Ben Laurie] 2641109998Smarkm 2642109998Smarkm *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() 2643109998Smarkm X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), 2644109998Smarkm X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). 2645109998Smarkm These allow a CRL to be built without having to access X509_CRL fields 2646109998Smarkm directly. Modify 'ca' application to use new functions. 2647109998Smarkm [Steve Henson] 2648109998Smarkm 2649109998Smarkm *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended 2650109998Smarkm bug workarounds. Rollback attack detection is a security feature. 2651109998Smarkm The problem will only arise on OpenSSL servers when TLSv1 is not 2652109998Smarkm available (sslv3_server_method() or SSL_OP_NO_TLSv1). 2653109998Smarkm Software authors not wanting to support TLSv1 will have special reasons 2654109998Smarkm for their choice and can explicitly enable this option. 2655109998Smarkm [Bodo Moeller, Lutz Jaenicke] 2656109998Smarkm 2657109998Smarkm *) Rationalise EVP so it can be extended: don't include a union of 2658109998Smarkm cipher/digest structures, add init/cleanup functions for EVP_MD_CTX 2659109998Smarkm (similar to those existing for EVP_CIPHER_CTX). 2660109998Smarkm Usage example: 2661109998Smarkm 2662109998Smarkm EVP_MD_CTX md; 2663109998Smarkm 2664109998Smarkm EVP_MD_CTX_init(&md); /* new function call */ 2665109998Smarkm EVP_DigestInit(&md, EVP_sha1()); 2666109998Smarkm EVP_DigestUpdate(&md, in, len); 2667109998Smarkm EVP_DigestFinal(&md, out, NULL); 2668109998Smarkm EVP_MD_CTX_cleanup(&md); /* new function call */ 2669109998Smarkm 2670109998Smarkm [Ben Laurie] 2671109998Smarkm 2672109998Smarkm *) Make DES key schedule conform to the usual scheme, as well as 2673109998Smarkm correcting its structure. This means that calls to DES functions 2674109998Smarkm now have to pass a pointer to a des_key_schedule instead of a 2675109998Smarkm plain des_key_schedule (which was actually always a pointer 2676109998Smarkm anyway): E.g., 2677109998Smarkm 2678109998Smarkm des_key_schedule ks; 2679109998Smarkm 2680109998Smarkm des_set_key_checked(..., &ks); 2681109998Smarkm des_ncbc_encrypt(..., &ks, ...); 2682109998Smarkm 2683109998Smarkm (Note that a later change renames 'des_...' into 'DES_...'.) 2684109998Smarkm [Ben Laurie] 2685109998Smarkm 2686109998Smarkm *) Initial reduction of linker bloat: the use of some functions, such as 2687109998Smarkm PEM causes large amounts of unused functions to be linked in due to 2688109998Smarkm poor organisation. For example pem_all.c contains every PEM function 2689109998Smarkm which has a knock on effect of linking in large amounts of (unused) 2690109998Smarkm ASN1 code. Grouping together similar functions and splitting unrelated 2691109998Smarkm functions prevents this. 2692109998Smarkm [Steve Henson] 2693109998Smarkm 2694109998Smarkm *) Cleanup of EVP macros. 2695109998Smarkm [Ben Laurie] 2696109998Smarkm 2697109998Smarkm *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the 2698109998Smarkm correct _ecb suffix. 2699109998Smarkm [Ben Laurie] 2700109998Smarkm 2701109998Smarkm *) Add initial OCSP responder support to ocsp application. The 2702109998Smarkm revocation information is handled using the text based index 2703109998Smarkm use by the ca application. The responder can either handle 2704109998Smarkm requests generated internally, supplied in files (for example 2705109998Smarkm via a CGI script) or using an internal minimal server. 2706109998Smarkm [Steve Henson] 2707109998Smarkm 2708109998Smarkm *) Add configuration choices to get zlib compression for TLS. 2709109998Smarkm [Richard Levitte] 2710109998Smarkm 2711109998Smarkm *) Changes to Kerberos SSL for RFC 2712 compliance: 2712109998Smarkm 1. Implemented real KerberosWrapper, instead of just using 2713109998Smarkm KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] 2714109998Smarkm 2. Implemented optional authenticator field of KerberosWrapper. 2715109998Smarkm 2716109998Smarkm Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, 2717109998Smarkm and authenticator structs; see crypto/krb5/. 2718109998Smarkm 2719109998Smarkm Generalized Kerberos calls to support multiple Kerberos libraries. 2720109998Smarkm [Vern Staats <staatsvr@asc.hpc.mil>, 2721109998Smarkm Jeffrey Altman <jaltman@columbia.edu> 2722109998Smarkm via Richard Levitte] 2723109998Smarkm 2724109998Smarkm *) Cause 'openssl speed' to use fully hard-coded DSA keys as it 2725109998Smarkm already does with RSA. testdsa.h now has 'priv_key/pub_key' 2726109998Smarkm values for each of the key sizes rather than having just 2727109998Smarkm parameters (and 'speed' generating keys each time). 2728109998Smarkm [Geoff Thorpe] 2729109998Smarkm 2730109998Smarkm *) Speed up EVP routines. 2731109998Smarkm Before: 2732109998Smarkmencrypt 2733109998Smarkmtype 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 2734109998Smarkmdes-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k 2735109998Smarkmdes-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k 2736109998Smarkmdes-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k 2737109998Smarkmdecrypt 2738109998Smarkmdes-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k 2739109998Smarkmdes-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k 2740109998Smarkmdes-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k 2741109998Smarkm After: 2742109998Smarkmencrypt 2743109998Smarkmdes-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k 2744109998Smarkmdecrypt 2745109998Smarkmdes-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k 2746109998Smarkm [Ben Laurie] 2747109998Smarkm 2748109998Smarkm *) Added the OS2-EMX target. 2749109998Smarkm ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte] 2750109998Smarkm 2751109998Smarkm *) Rewrite apps to use NCONF routines instead of the old CONF. New functions 2752109998Smarkm to support NCONF routines in extension code. New function CONF_set_nconf() 2753109998Smarkm to allow functions which take an NCONF to also handle the old LHASH 2754109998Smarkm structure: this means that the old CONF compatible routines can be 2755109998Smarkm retained (in particular wrt extensions) without having to duplicate the 2756109998Smarkm code. New function X509V3_add_ext_nconf_sk to add extensions to a stack. 2757109998Smarkm [Steve Henson] 2758109998Smarkm 2759109998Smarkm *) Enhance the general user interface with mechanisms for inner control 2760109998Smarkm and with possibilities to have yes/no kind of prompts. 2761109998Smarkm [Richard Levitte] 2762109998Smarkm 2763109998Smarkm *) Change all calls to low level digest routines in the library and 2764109998Smarkm applications to use EVP. Add missing calls to HMAC_cleanup() and 2765109998Smarkm don't assume HMAC_CTX can be copied using memcpy(). 2766109998Smarkm [Verdon Walker <VWalker@novell.com>, Steve Henson] 2767109998Smarkm 2768109998Smarkm *) Add the possibility to control engines through control names but with 2769109998Smarkm arbitrary arguments instead of just a string. 2770109998Smarkm Change the key loaders to take a UI_METHOD instead of a callback 2771109998Smarkm function pointer. NOTE: this breaks binary compatibility with earlier 2772109998Smarkm versions of OpenSSL [engine]. 2773109998Smarkm Adapt the nCipher code for these new conditions and add a card insertion 2774109998Smarkm callback. 2775109998Smarkm [Richard Levitte] 2776109998Smarkm 2777109998Smarkm *) Enhance the general user interface with mechanisms to better support 2778109998Smarkm dialog box interfaces, application-defined prompts, the possibility 2779109998Smarkm to use defaults (for example default passwords from somewhere else) 2780109998Smarkm and interrupts/cancellations. 2781109998Smarkm [Richard Levitte] 2782109998Smarkm 2783109998Smarkm *) Tidy up PKCS#12 attribute handling. Add support for the CSP name 2784109998Smarkm attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. 2785109998Smarkm [Steve Henson] 2786109998Smarkm 2787109998Smarkm *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also 2788109998Smarkm tidy up some unnecessarily weird code in 'sk_new()'). 2789109998Smarkm [Geoff, reported by Diego Tartara <dtartara@novamens.com>] 2790109998Smarkm 2791109998Smarkm *) Change the key loading routines for ENGINEs to use the same kind 2792109998Smarkm callback (pem_password_cb) as all other routines that need this 2793109998Smarkm kind of callback. 2794109998Smarkm [Richard Levitte] 2795109998Smarkm 2796109998Smarkm *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with 2797109998Smarkm 256 bit (=32 byte) keys. Of course seeding with more entropy bytes 2798109998Smarkm than this minimum value is recommended. 2799109998Smarkm [Lutz Jaenicke] 2800109998Smarkm 2801109998Smarkm *) New random seeder for OpenVMS, using the system process statistics 2802109998Smarkm that are easily reachable. 2803109998Smarkm [Richard Levitte] 2804109998Smarkm 2805109998Smarkm *) Windows apparently can't transparently handle global 2806109998Smarkm variables defined in DLLs. Initialisations such as: 2807109998Smarkm 2808109998Smarkm const ASN1_ITEM *it = &ASN1_INTEGER_it; 2809109998Smarkm 2810109998Smarkm wont compile. This is used by the any applications that need to 2811109998Smarkm declare their own ASN1 modules. This was fixed by adding the option 2812109998Smarkm EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly 2813109998Smarkm needed for static libraries under Win32. 2814109998Smarkm [Steve Henson] 2815109998Smarkm 2816109998Smarkm *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle 2817109998Smarkm setting of purpose and trust fields. New X509_STORE trust and 2818109998Smarkm purpose functions and tidy up setting in other SSL functions. 2819109998Smarkm [Steve Henson] 2820109998Smarkm 2821109998Smarkm *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE 2822109998Smarkm structure. These are inherited by X509_STORE_CTX when it is 2823109998Smarkm initialised. This allows various defaults to be set in the 2824109998Smarkm X509_STORE structure (such as flags for CRL checking and custom 2825109998Smarkm purpose or trust settings) for functions which only use X509_STORE_CTX 2826109998Smarkm internally such as S/MIME. 2827109998Smarkm 2828109998Smarkm Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and 2829109998Smarkm trust settings if they are not set in X509_STORE. This allows X509_STORE 2830109998Smarkm purposes and trust (in S/MIME for example) to override any set by default. 2831109998Smarkm 2832109998Smarkm Add command line options for CRL checking to smime, s_client and s_server 2833109998Smarkm applications. 2834109998Smarkm [Steve Henson] 2835109998Smarkm 2836109998Smarkm *) Initial CRL based revocation checking. If the CRL checking flag(s) 2837109998Smarkm are set then the CRL is looked up in the X509_STORE structure and 2838109998Smarkm its validity and signature checked, then if the certificate is found 2839109998Smarkm in the CRL the verify fails with a revoked error. 2840109998Smarkm 2841109998Smarkm Various new CRL related callbacks added to X509_STORE_CTX structure. 2842109998Smarkm 2843109998Smarkm Command line options added to 'verify' application to support this. 2844109998Smarkm 2845109998Smarkm This needs some additional work, such as being able to handle multiple 2846109998Smarkm CRLs with different times, extension based lookup (rather than just 2847109998Smarkm by subject name) and ultimately more complete V2 CRL extension 2848109998Smarkm handling. 2849109998Smarkm [Steve Henson] 2850109998Smarkm 2851109998Smarkm *) Add a general user interface API (crypto/ui/). This is designed 2852109998Smarkm to replace things like des_read_password and friends (backward 2853109998Smarkm compatibility functions using this new API are provided). 2854109998Smarkm The purpose is to remove prompting functions from the DES code 2855109998Smarkm section as well as provide for prompting through dialog boxes in 2856109998Smarkm a window system and the like. 2857109998Smarkm [Richard Levitte] 2858109998Smarkm 2859109998Smarkm *) Add "ex_data" support to ENGINE so implementations can add state at a 2860109998Smarkm per-structure level rather than having to store it globally. 2861109998Smarkm [Geoff] 2862109998Smarkm 2863109998Smarkm *) Make it possible for ENGINE structures to be copied when retrieved by 2864109998Smarkm ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. 2865109998Smarkm This causes the "original" ENGINE structure to act like a template, 2866109998Smarkm analogous to the RSA vs. RSA_METHOD type of separation. Because of this 2867109998Smarkm operational state can be localised to each ENGINE structure, despite the 2868109998Smarkm fact they all share the same "methods". New ENGINE structures returned in 2869109998Smarkm this case have no functional references and the return value is the single 2870109998Smarkm structural reference. This matches the single structural reference returned 2871109998Smarkm by ENGINE_by_id() normally, when it is incremented on the pre-existing 2872109998Smarkm ENGINE structure. 2873109998Smarkm [Geoff] 2874109998Smarkm 2875109998Smarkm *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this 2876109998Smarkm needs to match any other type at all we need to manually clear the 2877109998Smarkm tag cache. 2878109998Smarkm [Steve Henson] 2879109998Smarkm 2880109998Smarkm *) Changes to the "openssl engine" utility to include; 2881109998Smarkm - verbosity levels ('-v', '-vv', and '-vvv') that provide information 2882109998Smarkm about an ENGINE's available control commands. 2883109998Smarkm - executing control commands from command line arguments using the 2884109998Smarkm '-pre' and '-post' switches. '-post' is only used if '-t' is 2885109998Smarkm specified and the ENGINE is successfully initialised. The syntax for 2886109998Smarkm the individual commands are colon-separated, for example; 2887109998Smarkm openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so 2888109998Smarkm [Geoff] 2889109998Smarkm 2890109998Smarkm *) New dynamic control command support for ENGINEs. ENGINEs can now 2891109998Smarkm declare their own commands (numbers), names (strings), descriptions, 2892109998Smarkm and input types for run-time discovery by calling applications. A 2893109998Smarkm subset of these commands are implicitly classed as "executable" 2894109998Smarkm depending on their input type, and only these can be invoked through 2895109998Smarkm the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this 2896109998Smarkm can be based on user input, config files, etc). The distinction is 2897109998Smarkm that "executable" commands cannot return anything other than a boolean 2898109998Smarkm result and can only support numeric or string input, whereas some 2899109998Smarkm discoverable commands may only be for direct use through 2900109998Smarkm ENGINE_ctrl(), eg. supporting the exchange of binary data, function 2901109998Smarkm pointers, or other custom uses. The "executable" commands are to 2902109998Smarkm support parameterisations of ENGINE behaviour that can be 2903109998Smarkm unambiguously defined by ENGINEs and used consistently across any 2904109998Smarkm OpenSSL-based application. Commands have been added to all the 2905109998Smarkm existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow 2906109998Smarkm control over shared-library paths without source code alterations. 2907109998Smarkm [Geoff] 2908109998Smarkm 2909109998Smarkm *) Changed all ENGINE implementations to dynamically allocate their 2910109998Smarkm ENGINEs rather than declaring them statically. Apart from this being 2911109998Smarkm necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, 2912109998Smarkm this also allows the implementations to compile without using the 2913109998Smarkm internal engine_int.h header. 2914109998Smarkm [Geoff] 2915109998Smarkm 2916109998Smarkm *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a 2917109998Smarkm 'const' value. Any code that should be able to modify a RAND_METHOD 2918109998Smarkm should already have non-const pointers to it (ie. they should only 2919109998Smarkm modify their own ones). 2920109998Smarkm [Geoff] 2921109998Smarkm 2922109998Smarkm *) Made a variety of little tweaks to the ENGINE code. 2923109998Smarkm - "atalla" and "ubsec" string definitions were moved from header files 2924109998Smarkm to C code. "nuron" string definitions were placed in variables 2925109998Smarkm rather than hard-coded - allowing parameterisation of these values 2926109998Smarkm later on via ctrl() commands. 2927109998Smarkm - Removed unused "#if 0"'d code. 2928109998Smarkm - Fixed engine list iteration code so it uses ENGINE_free() to release 2929109998Smarkm structural references. 2930109998Smarkm - Constified the RAND_METHOD element of ENGINE structures. 2931109998Smarkm - Constified various get/set functions as appropriate and added 2932109998Smarkm missing functions (including a catch-all ENGINE_cpy that duplicates 2933109998Smarkm all ENGINE values onto a new ENGINE except reference counts/state). 2934109998Smarkm - Removed NULL parameter checks in get/set functions. Setting a method 2935109998Smarkm or function to NULL is a way of cancelling out a previously set 2936109998Smarkm value. Passing a NULL ENGINE parameter is just plain stupid anyway 2937109998Smarkm and doesn't justify the extra error symbols and code. 2938109998Smarkm - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for 2939109998Smarkm flags from engine_int.h to engine.h. 2940109998Smarkm - Changed prototypes for ENGINE handler functions (init(), finish(), 2941109998Smarkm ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. 2942109998Smarkm [Geoff] 2943109998Smarkm 2944109998Smarkm *) Implement binary inversion algorithm for BN_mod_inverse in addition 2945109998Smarkm to the algorithm using long division. The binary algorithm can be 2946109998Smarkm used only if the modulus is odd. On 32-bit systems, it is faster 2947109998Smarkm only for relatively small moduli (roughly 20-30% for 128-bit moduli, 2948109998Smarkm roughly 5-15% for 256-bit moduli), so we use it only for moduli 2949109998Smarkm up to 450 bits. In 64-bit environments, the binary algorithm 2950109998Smarkm appears to be advantageous for much longer moduli; here we use it 2951109998Smarkm for moduli up to 2048 bits. 2952109998Smarkm [Bodo Moeller] 2953109998Smarkm 2954109998Smarkm *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code 2955109998Smarkm could not support the combine flag in choice fields. 2956109998Smarkm [Steve Henson] 2957109998Smarkm 2958109998Smarkm *) Add a 'copy_extensions' option to the 'ca' utility. This copies 2959109998Smarkm extensions from a certificate request to the certificate. 2960109998Smarkm [Steve Henson] 2961109998Smarkm 2962109998Smarkm *) Allow multiple 'certopt' and 'nameopt' options to be separated 2963109998Smarkm by commas. Add 'namopt' and 'certopt' options to the 'ca' config 2964109998Smarkm file: this allows the display of the certificate about to be 2965109998Smarkm signed to be customised, to allow certain fields to be included 2966109998Smarkm or excluded and extension details. The old system didn't display 2967109998Smarkm multicharacter strings properly, omitted fields not in the policy 2968109998Smarkm and couldn't display additional details such as extensions. 2969109998Smarkm [Steve Henson] 2970109998Smarkm 2971109998Smarkm *) Function EC_POINTs_mul for multiple scalar multiplication 2972109998Smarkm of an arbitrary number of elliptic curve points 2973109998Smarkm \sum scalars[i]*points[i], 2974109998Smarkm optionally including the generator defined for the EC_GROUP: 2975109998Smarkm scalar*generator + \sum scalars[i]*points[i]. 2976109998Smarkm 2977109998Smarkm EC_POINT_mul is a simple wrapper function for the typical case 2978109998Smarkm that the point list has just one item (besides the optional 2979109998Smarkm generator). 2980109998Smarkm [Bodo Moeller] 2981109998Smarkm 2982109998Smarkm *) First EC_METHODs for curves over GF(p): 2983109998Smarkm 2984109998Smarkm EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr 2985109998Smarkm operations and provides various method functions that can also 2986109998Smarkm operate with faster implementations of modular arithmetic. 2987109998Smarkm 2988109998Smarkm EC_GFp_mont_method() reuses most functions that are part of 2989109998Smarkm EC_GFp_simple_method, but uses Montgomery arithmetic. 2990109998Smarkm 2991109998Smarkm [Bodo Moeller; point addition and point doubling 2992109998Smarkm implementation directly derived from source code provided by 2993109998Smarkm Lenka Fibikova <fibikova@exp-math.uni-essen.de>] 2994109998Smarkm 2995109998Smarkm *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, 2996109998Smarkm crypto/ec/ec_lib.c): 2997109998Smarkm 2998109998Smarkm Curves are EC_GROUP objects (with an optional group generator) 2999109998Smarkm based on EC_METHODs that are built into the library. 3000109998Smarkm 3001109998Smarkm Points are EC_POINT objects based on EC_GROUP objects. 3002109998Smarkm 3003109998Smarkm Most of the framework would be able to handle curves over arbitrary 3004109998Smarkm finite fields, but as there are no obvious types for fields other 3005109998Smarkm than GF(p), some functions are limited to that for now. 3006109998Smarkm [Bodo Moeller] 3007109998Smarkm 3008109998Smarkm *) Add the -HTTP option to s_server. It is similar to -WWW, but requires 3009109998Smarkm that the file contains a complete HTTP response. 3010109998Smarkm [Richard Levitte] 3011109998Smarkm 3012109998Smarkm *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl 3013109998Smarkm change the def and num file printf format specifier from "%-40sXXX" 3014109998Smarkm to "%-39s XXX". The latter will always guarantee a space after the 3015109998Smarkm field while the former will cause them to run together if the field 3016109998Smarkm is 40 of more characters long. 3017109998Smarkm [Steve Henson] 3018109998Smarkm 3019109998Smarkm *) Constify the cipher and digest 'method' functions and structures 3020109998Smarkm and modify related functions to take constant EVP_MD and EVP_CIPHER 3021109998Smarkm pointers. 3022109998Smarkm [Steve Henson] 3023109998Smarkm 3024109998Smarkm *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them 3025109998Smarkm in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. 3026109998Smarkm [Bodo Moeller] 3027109998Smarkm 3028109998Smarkm *) Modify EVP_Digest*() routines so they now return values. Although the 3029109998Smarkm internal software routines can never fail additional hardware versions 3030109998Smarkm might. 3031109998Smarkm [Steve Henson] 3032109998Smarkm 3033109998Smarkm *) Clean up crypto/err/err.h and change some error codes to avoid conflicts: 3034109998Smarkm 3035109998Smarkm Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 3036109998Smarkm (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. 3037109998Smarkm 3038109998Smarkm ASN1 error codes 3039109998Smarkm ERR_R_NESTED_ASN1_ERROR 3040109998Smarkm ... 3041109998Smarkm ERR_R_MISSING_ASN1_EOS 3042109998Smarkm were 4 .. 9, conflicting with 3043109998Smarkm ERR_LIB_RSA (= ERR_R_RSA_LIB) 3044109998Smarkm ... 3045109998Smarkm ERR_LIB_PEM (= ERR_R_PEM_LIB). 3046109998Smarkm They are now 58 .. 63 (i.e., just below ERR_R_FATAL). 3047109998Smarkm 3048109998Smarkm Add new error code 'ERR_R_INTERNAL_ERROR'. 3049109998Smarkm [Bodo Moeller] 3050109998Smarkm 3051109998Smarkm *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock 3052109998Smarkm suffices. 3053109998Smarkm [Bodo Moeller] 3054109998Smarkm 3055109998Smarkm *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This 3056109998Smarkm sets the subject name for a new request or supersedes the 3057109998Smarkm subject name in a given request. Formats that can be parsed are 3058109998Smarkm 'CN=Some Name, OU=myOU, C=IT' 3059109998Smarkm and 3060109998Smarkm 'CN=Some Name/OU=myOU/C=IT'. 3061109998Smarkm 3062109998Smarkm Add options '-batch' and '-verbose' to 'openssl req'. 3063109998Smarkm [Massimiliano Pala <madwolf@hackmasters.net>] 3064109998Smarkm 3065109998Smarkm *) Introduce the possibility to access global variables through 3066109998Smarkm functions on platform were that's the best way to handle exporting 3067109998Smarkm global variables in shared libraries. To enable this functionality, 3068109998Smarkm one must configure with "EXPORT_VAR_AS_FN" or defined the C macro 3069109998Smarkm "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter 3070109998Smarkm is normally done by Configure or something similar). 3071109998Smarkm 3072109998Smarkm To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL 3073109998Smarkm in the source file (foo.c) like this: 3074109998Smarkm 3075109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; 3076109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(double,bar); 3077109998Smarkm 3078109998Smarkm To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL 3079109998Smarkm and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: 3080109998Smarkm 3081109998Smarkm OPENSSL_DECLARE_GLOBAL(int,foo); 3082109998Smarkm #define foo OPENSSL_GLOBAL_REF(foo) 3083109998Smarkm OPENSSL_DECLARE_GLOBAL(double,bar); 3084109998Smarkm #define bar OPENSSL_GLOBAL_REF(bar) 3085109998Smarkm 3086109998Smarkm The #defines are very important, and therefore so is including the 3087109998Smarkm header file everywhere where the defined globals are used. 3088109998Smarkm 3089109998Smarkm The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition 3090109998Smarkm of ASN.1 items, but that structure is a bit different. 3091109998Smarkm 3092109998Smarkm The largest change is in util/mkdef.pl which has been enhanced with 3093109998Smarkm better and easier to understand logic to choose which symbols should 3094109998Smarkm go into the Windows .def files as well as a number of fixes and code 3095109998Smarkm cleanup (among others, algorithm keywords are now sorted 3096109998Smarkm lexicographically to avoid constant rewrites). 3097109998Smarkm [Richard Levitte] 3098109998Smarkm 3099109998Smarkm *) In BN_div() keep a copy of the sign of 'num' before writing the 3100109998Smarkm result to 'rm' because if rm==num the value will be overwritten 3101109998Smarkm and produce the wrong result if 'num' is negative: this caused 3102109998Smarkm problems with BN_mod() and BN_nnmod(). 3103109998Smarkm [Steve Henson] 3104109998Smarkm 3105109998Smarkm *) Function OCSP_request_verify(). This checks the signature on an 3106109998Smarkm OCSP request and verifies the signer certificate. The signer 3107109998Smarkm certificate is just checked for a generic purpose and OCSP request 3108109998Smarkm trust settings. 3109109998Smarkm [Steve Henson] 3110109998Smarkm 3111109998Smarkm *) Add OCSP_check_validity() function to check the validity of OCSP 3112109998Smarkm responses. OCSP responses are prepared in real time and may only 3113109998Smarkm be a few seconds old. Simply checking that the current time lies 3114109998Smarkm between thisUpdate and nextUpdate max reject otherwise valid responses 3115109998Smarkm caused by either OCSP responder or client clock inaccuracy. Instead 3116109998Smarkm we allow thisUpdate and nextUpdate to fall within a certain period of 3117109998Smarkm the current time. The age of the response can also optionally be 3118109998Smarkm checked. Two new options -validity_period and -status_age added to 3119109998Smarkm ocsp utility. 3120109998Smarkm [Steve Henson] 3121109998Smarkm 3122109998Smarkm *) If signature or public key algorithm is unrecognized print out its 3123109998Smarkm OID rather that just UNKNOWN. 3124109998Smarkm [Steve Henson] 3125109998Smarkm 3126109998Smarkm *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and 3127109998Smarkm OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate 3128109998Smarkm ID to be generated from the issuer certificate alone which can then be 3129109998Smarkm passed to OCSP_id_issuer_cmp(). 3130109998Smarkm [Steve Henson] 3131109998Smarkm 3132109998Smarkm *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new 3133109998Smarkm ASN1 modules to export functions returning ASN1_ITEM pointers 3134109998Smarkm instead of the ASN1_ITEM structures themselves. This adds several 3135109998Smarkm new macros which allow the underlying ASN1 function/structure to 3136109998Smarkm be accessed transparently. As a result code should not use ASN1_ITEM 3137109998Smarkm references directly (such as &X509_it) but instead use the relevant 3138109998Smarkm macros (such as ASN1_ITEM_rptr(X509)). This option is to allow 3139109998Smarkm use of the new ASN1 code on platforms where exporting structures 3140109998Smarkm is problematical (for example in shared libraries) but exporting 3141109998Smarkm functions returning pointers to structures is not. 3142109998Smarkm [Steve Henson] 3143109998Smarkm 3144109998Smarkm *) Add support for overriding the generation of SSL/TLS session IDs. 3145109998Smarkm These callbacks can be registered either in an SSL_CTX or per SSL. 3146109998Smarkm The purpose of this is to allow applications to control, if they wish, 3147109998Smarkm the arbitrary values chosen for use as session IDs, particularly as it 3148109998Smarkm can be useful for session caching in multiple-server environments. A 3149109998Smarkm command-line switch for testing this (and any client code that wishes 3150109998Smarkm to use such a feature) has been added to "s_server". 3151109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 3152109998Smarkm 3153109998Smarkm *) Modify mkdef.pl to recognise and parse preprocessor conditionals 3154109998Smarkm of the form '#if defined(...) || defined(...) || ...' and 3155109998Smarkm '#if !defined(...) && !defined(...) && ...'. This also avoids 3156109998Smarkm the growing number of special cases it was previously handling. 3157109998Smarkm [Richard Levitte] 3158109998Smarkm 3159109998Smarkm *) Make all configuration macros available for application by making 3160109998Smarkm sure they are available in opensslconf.h, by giving them names starting 3161109998Smarkm with "OPENSSL_" to avoid conflicts with other packages and by making 3162109998Smarkm sure e_os2.h will cover all platform-specific cases together with 3163109998Smarkm opensslconf.h. 3164109998Smarkm Additionally, it is now possible to define configuration/platform- 3165109998Smarkm specific names (called "system identities"). In the C code, these 3166109998Smarkm are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another 3167109998Smarkm macro with the name beginning with "OPENSSL_SYS_", which is determined 3168109998Smarkm from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on 3169109998Smarkm what is available. 3170109998Smarkm [Richard Levitte] 3171109998Smarkm 3172109998Smarkm *) New option -set_serial to 'req' and 'x509' this allows the serial 3173109998Smarkm number to use to be specified on the command line. Previously self 3174109998Smarkm signed certificates were hard coded with serial number 0 and the 3175109998Smarkm CA options of 'x509' had to use a serial number in a file which was 3176109998Smarkm auto incremented. 3177109998Smarkm [Steve Henson] 3178109998Smarkm 3179109998Smarkm *) New options to 'ca' utility to support V2 CRL entry extensions. 3180109998Smarkm Currently CRL reason, invalidity date and hold instruction are 3181109998Smarkm supported. Add new CRL extensions to V3 code and some new objects. 3182109998Smarkm [Steve Henson] 3183109998Smarkm 3184109998Smarkm *) New function EVP_CIPHER_CTX_set_padding() this is used to 3185109998Smarkm disable standard block padding (aka PKCS#5 padding) in the EVP 3186109998Smarkm API, which was previously mandatory. This means that the data is 3187109998Smarkm not padded in any way and so the total length much be a multiple 3188109998Smarkm of the block size, otherwise an error occurs. 3189109998Smarkm [Steve Henson] 3190109998Smarkm 3191109998Smarkm *) Initial (incomplete) OCSP SSL support. 3192109998Smarkm [Steve Henson] 3193109998Smarkm 3194109998Smarkm *) New function OCSP_parse_url(). This splits up a URL into its host, 3195109998Smarkm port and path components: primarily to parse OCSP URLs. New -url 3196109998Smarkm option to ocsp utility. 3197109998Smarkm [Steve Henson] 3198109998Smarkm 3199109998Smarkm *) New nonce behavior. The return value of OCSP_check_nonce() now 3200109998Smarkm reflects the various checks performed. Applications can decide 3201109998Smarkm whether to tolerate certain situations such as an absent nonce 3202109998Smarkm in a response when one was present in a request: the ocsp application 3203109998Smarkm just prints out a warning. New function OCSP_add1_basic_nonce() 3204109998Smarkm this is to allow responders to include a nonce in a response even if 3205109998Smarkm the request is nonce-less. 3206109998Smarkm [Steve Henson] 3207109998Smarkm 3208109998Smarkm *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are 3209109998Smarkm skipped when using openssl x509 multiple times on a single input file, 3210109998Smarkm e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs". 3211109998Smarkm [Bodo Moeller] 3212109998Smarkm 3213109998Smarkm *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() 3214109998Smarkm set string type: to handle setting ASN1_TIME structures. Fix ca 3215109998Smarkm utility to correctly initialize revocation date of CRLs. 3216109998Smarkm [Steve Henson] 3217109998Smarkm 3218109998Smarkm *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override 3219109998Smarkm the clients preferred ciphersuites and rather use its own preferences. 3220109998Smarkm Should help to work around M$ SGC (Server Gated Cryptography) bug in 3221109998Smarkm Internet Explorer by ensuring unchanged hash method during stepup. 3222109998Smarkm (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) 3223109998Smarkm [Lutz Jaenicke] 3224109998Smarkm 3225109998Smarkm *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael 3226109998Smarkm to aes and add a new 'exist' option to print out symbols that don't 3227109998Smarkm appear to exist. 3228109998Smarkm [Steve Henson] 3229109998Smarkm 3230109998Smarkm *) Additional options to ocsp utility to allow flags to be set and 3231109998Smarkm additional certificates supplied. 3232109998Smarkm [Steve Henson] 3233109998Smarkm 3234109998Smarkm *) Add the option -VAfile to 'openssl ocsp', so the user can give the 3235109998Smarkm OCSP client a number of certificate to only verify the response 3236109998Smarkm signature against. 3237109998Smarkm [Richard Levitte] 3238109998Smarkm 3239109998Smarkm *) Update Rijndael code to version 3.0 and change EVP AES ciphers to 3240109998Smarkm handle the new API. Currently only ECB, CBC modes supported. Add new 3241109998Smarkm AES OIDs. 3242109998Smarkm 3243109998Smarkm Add TLS AES ciphersuites as described in RFC3268, "Advanced 3244109998Smarkm Encryption Standard (AES) Ciphersuites for Transport Layer 3245109998Smarkm Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were 3246109998Smarkm not enabled by default and were not part of the "ALL" ciphersuite 3247109998Smarkm alias because they were not yet official; they could be 3248109998Smarkm explicitly requested by specifying the "AESdraft" ciphersuite 3249109998Smarkm group alias. In the final release of OpenSSL 0.9.7, the group 3250109998Smarkm alias is called "AES" and is part of "ALL".) 3251109998Smarkm [Ben Laurie, Steve Henson, Bodo Moeller] 3252109998Smarkm 3253109998Smarkm *) New function OCSP_copy_nonce() to copy nonce value (if present) from 3254109998Smarkm request to response. 3255109998Smarkm [Steve Henson] 3256109998Smarkm 3257109998Smarkm *) Functions for OCSP responders. OCSP_request_onereq_count(), 3258109998Smarkm OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() 3259109998Smarkm extract information from a certificate request. OCSP_response_create() 3260109998Smarkm creates a response and optionally adds a basic response structure. 3261109998Smarkm OCSP_basic_add1_status() adds a complete single response to a basic 3262109998Smarkm response and returns the OCSP_SINGLERESP structure just added (to allow 3263109998Smarkm extensions to be included for example). OCSP_basic_add1_cert() adds a 3264109998Smarkm certificate to a basic response and OCSP_basic_sign() signs a basic 3265109998Smarkm response with various flags. New helper functions ASN1_TIME_check() 3266109998Smarkm (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() 3267109998Smarkm (converts ASN1_TIME to GeneralizedTime). 3268109998Smarkm [Steve Henson] 3269109998Smarkm 3270109998Smarkm *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() 3271109998Smarkm in a single operation. X509_get0_pubkey_bitstr() extracts the public_key 3272109998Smarkm structure from a certificate. X509_pubkey_digest() digests the public_key 3273109998Smarkm contents: this is used in various key identifiers. 3274109998Smarkm [Steve Henson] 3275109998Smarkm 3276109998Smarkm *) Make sk_sort() tolerate a NULL argument. 3277109998Smarkm [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>] 3278109998Smarkm 3279109998Smarkm *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates 3280109998Smarkm passed by the function are trusted implicitly. If any of them signed the 3281109998Smarkm response then it is assumed to be valid and is not verified. 3282109998Smarkm [Steve Henson] 3283109998Smarkm 3284109998Smarkm *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT 3285109998Smarkm to data. This was previously part of the PKCS7 ASN1 code. This 3286109998Smarkm was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. 3287109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 3288109998Smarkm <support@securenetterm.com>] 3289109998Smarkm 3290109998Smarkm *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 3291109998Smarkm routines: without these tracing memory leaks is very painful. 3292109998Smarkm Fix leaks in PKCS12 and PKCS7 routines. 3293109998Smarkm [Steve Henson] 3294109998Smarkm 3295109998Smarkm *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). 3296109998Smarkm Previously it initialised the 'type' argument to V_ASN1_UTCTIME which 3297109998Smarkm effectively meant GeneralizedTime would never be used. Now it 3298109998Smarkm is initialised to -1 but X509_time_adj() now has to check the value 3299109998Smarkm and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or 3300109998Smarkm V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. 3301109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 3302109998Smarkm <support@securenetterm.com>] 3303109998Smarkm 3304109998Smarkm *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously 3305109998Smarkm result in a zero length in the ASN1_INTEGER structure which was 3306109998Smarkm not consistent with the structure when d2i_ASN1_INTEGER() was used 3307109998Smarkm and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() 3308109998Smarkm to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() 3309109998Smarkm where it did not print out a minus for negative ASN1_INTEGER. 3310109998Smarkm [Steve Henson] 3311109998Smarkm 3312109998Smarkm *) Add summary printout to ocsp utility. The various functions which 3313109998Smarkm convert status values to strings have been renamed to: 3314109998Smarkm OCSP_response_status_str(), OCSP_cert_status_str() and 3315109998Smarkm OCSP_crl_reason_str() and are no longer static. New options 3316109998Smarkm to verify nonce values and to disable verification. OCSP response 3317109998Smarkm printout format cleaned up. 3318109998Smarkm [Steve Henson] 3319109998Smarkm 3320109998Smarkm *) Add additional OCSP certificate checks. These are those specified 3321109998Smarkm in RFC2560. This consists of two separate checks: the CA of the 3322109998Smarkm certificate being checked must either be the OCSP signer certificate 3323109998Smarkm or the issuer of the OCSP signer certificate. In the latter case the 3324109998Smarkm OCSP signer certificate must contain the OCSP signing extended key 3325109998Smarkm usage. This check is performed by attempting to match the OCSP 3326109998Smarkm signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash 3327109998Smarkm in the OCSP_CERTID structures of the response. 3328109998Smarkm [Steve Henson] 3329109998Smarkm 3330109998Smarkm *) Initial OCSP certificate verification added to OCSP_basic_verify() 3331109998Smarkm and related routines. This uses the standard OpenSSL certificate 3332109998Smarkm verify routines to perform initial checks (just CA validity) and 3333109998Smarkm to obtain the certificate chain. Then additional checks will be 3334109998Smarkm performed on the chain. Currently the root CA is checked to see 3335109998Smarkm if it is explicitly trusted for OCSP signing. This is used to set 3336109998Smarkm a root CA as a global signing root: that is any certificate that 3337109998Smarkm chains to that CA is an acceptable OCSP signing certificate. 3338109998Smarkm [Steve Henson] 3339109998Smarkm 3340109998Smarkm *) New '-extfile ...' option to 'openssl ca' for reading X.509v3 3341109998Smarkm extensions from a separate configuration file. 3342109998Smarkm As when reading extensions from the main configuration file, 3343109998Smarkm the '-extensions ...' option may be used for specifying the 3344109998Smarkm section to use. 3345109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 3346109998Smarkm 3347109998Smarkm *) New OCSP utility. Allows OCSP requests to be generated or 3348109998Smarkm read. The request can be sent to a responder and the output 3349109998Smarkm parsed, outputed or printed in text form. Not complete yet: 3350109998Smarkm still needs to check the OCSP response validity. 3351109998Smarkm [Steve Henson] 3352109998Smarkm 3353109998Smarkm *) New subcommands for 'openssl ca': 3354109998Smarkm 'openssl ca -status <serial>' prints the status of the cert with 3355109998Smarkm the given serial number (according to the index file). 3356109998Smarkm 'openssl ca -updatedb' updates the expiry status of certificates 3357109998Smarkm in the index file. 3358109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 3359109998Smarkm 3360109998Smarkm *) New '-newreq-nodes' command option to CA.pl. This is like 3361109998Smarkm '-newreq', but calls 'openssl req' with the '-nodes' option 3362109998Smarkm so that the resulting key is not encrypted. 3363109998Smarkm [Damien Miller <djm@mindrot.org>] 3364109998Smarkm 3365109998Smarkm *) New configuration for the GNU Hurd. 3366109998Smarkm [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte] 3367109998Smarkm 3368109998Smarkm *) Initial code to implement OCSP basic response verify. This 3369109998Smarkm is currently incomplete. Currently just finds the signer's 3370109998Smarkm certificate and verifies the signature on the response. 3371109998Smarkm [Steve Henson] 3372109998Smarkm 3373109998Smarkm *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in 3374109998Smarkm value of OPENSSLDIR. This is available via the new '-d' option 3375109998Smarkm to 'openssl version', and is also included in 'openssl version -a'. 3376109998Smarkm [Bodo Moeller] 3377109998Smarkm 3378109998Smarkm *) Allowing defining memory allocation callbacks that will be given 3379109998Smarkm file name and line number information in additional arguments 3380109998Smarkm (a const char* and an int). The basic functionality remains, as 3381109998Smarkm well as the original possibility to just replace malloc(), 3382109998Smarkm realloc() and free() by functions that do not know about these 3383109998Smarkm additional arguments. To register and find out the current 3384109998Smarkm settings for extended allocation functions, the following 3385109998Smarkm functions are provided: 3386109998Smarkm 3387109998Smarkm CRYPTO_set_mem_ex_functions 3388109998Smarkm CRYPTO_set_locked_mem_ex_functions 3389109998Smarkm CRYPTO_get_mem_ex_functions 3390109998Smarkm CRYPTO_get_locked_mem_ex_functions 3391109998Smarkm 3392109998Smarkm These work the same way as CRYPTO_set_mem_functions and friends. 3393109998Smarkm CRYPTO_get_[locked_]mem_functions now writes 0 where such an 3394109998Smarkm extended allocation function is enabled. 3395109998Smarkm Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where 3396109998Smarkm a conventional allocation function is enabled. 3397109998Smarkm [Richard Levitte, Bodo Moeller] 3398109998Smarkm 3399109998Smarkm *) Finish off removing the remaining LHASH function pointer casts. 3400109998Smarkm There should no longer be any prototype-casting required when using 3401109998Smarkm the LHASH abstraction, and any casts that remain are "bugs". See 3402109998Smarkm the callback types and macros at the head of lhash.h for details 3403109998Smarkm (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). 3404109998Smarkm [Geoff Thorpe] 3405109998Smarkm 3406109998Smarkm *) Add automatic query of EGD sockets in RAND_poll() for the unix variant. 3407109998Smarkm If /dev/[u]random devices are not available or do not return enough 3408109998Smarkm entropy, EGD style sockets (served by EGD or PRNGD) will automatically 3409109998Smarkm be queried. 3410109998Smarkm The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and 3411109998Smarkm /etc/entropy will be queried once each in this sequence, quering stops 3412109998Smarkm when enough entropy was collected without querying more sockets. 3413109998Smarkm [Lutz Jaenicke] 3414109998Smarkm 3415109998Smarkm *) Change the Unix RAND_poll() variant to be able to poll several 3416109998Smarkm random devices, as specified by DEVRANDOM, until a sufficient amount 3417109998Smarkm of data has been collected. We spend at most 10 ms on each file 3418109998Smarkm (select timeout) and read in non-blocking mode. DEVRANDOM now 3419109998Smarkm defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" 3420109998Smarkm (previously it was just the string "/dev/urandom"), so on typical 3421109998Smarkm platforms the 10 ms delay will never occur. 3422109998Smarkm Also separate out the Unix variant to its own file, rand_unix.c. 3423109998Smarkm For VMS, there's a currently-empty rand_vms.c. 3424109998Smarkm [Richard Levitte] 3425109998Smarkm 3426109998Smarkm *) Move OCSP client related routines to ocsp_cl.c. These 3427109998Smarkm provide utility functions which an application needing 3428109998Smarkm to issue a request to an OCSP responder and analyse the 3429109998Smarkm response will typically need: as opposed to those which an 3430109998Smarkm OCSP responder itself would need which will be added later. 3431109998Smarkm 3432109998Smarkm OCSP_request_sign() signs an OCSP request with an API similar 3433109998Smarkm to PKCS7_sign(). OCSP_response_status() returns status of OCSP 3434109998Smarkm response. OCSP_response_get1_basic() extracts basic response 3435109998Smarkm from response. OCSP_resp_find_status(): finds and extracts status 3436109998Smarkm information from an OCSP_CERTID structure (which will be created 3437109998Smarkm when the request structure is built). These are built from lower 3438109998Smarkm level functions which work on OCSP_SINGLERESP structures but 3439109998Smarkm wont normally be used unless the application wishes to examine 3440109998Smarkm extensions in the OCSP response for example. 3441109998Smarkm 3442109998Smarkm Replace nonce routines with a pair of functions. 3443109998Smarkm OCSP_request_add1_nonce() adds a nonce value and optionally 3444109998Smarkm generates a random value. OCSP_check_nonce() checks the 3445109998Smarkm validity of the nonce in an OCSP response. 3446109998Smarkm [Steve Henson] 3447109998Smarkm 3448109998Smarkm *) Change function OCSP_request_add() to OCSP_request_add0_id(). 3449109998Smarkm This doesn't copy the supplied OCSP_CERTID and avoids the 3450109998Smarkm need to free up the newly created id. Change return type 3451109998Smarkm to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. 3452109998Smarkm This can then be used to add extensions to the request. 3453109998Smarkm Deleted OCSP_request_new(), since most of its functionality 3454109998Smarkm is now in OCSP_REQUEST_new() (and the case insensitive name 3455109998Smarkm clash) apart from the ability to set the request name which 3456109998Smarkm will be added elsewhere. 3457109998Smarkm [Steve Henson] 3458109998Smarkm 3459109998Smarkm *) Update OCSP API. Remove obsolete extensions argument from 3460109998Smarkm various functions. Extensions are now handled using the new 3461109998Smarkm OCSP extension code. New simple OCSP HTTP function which 3462109998Smarkm can be used to send requests and parse the response. 3463109998Smarkm [Steve Henson] 3464109998Smarkm 3465109998Smarkm *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new 3466109998Smarkm ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN 3467109998Smarkm uses the special reorder version of SET OF to sort the attributes 3468109998Smarkm and reorder them to match the encoded order. This resolves a long 3469109998Smarkm standing problem: a verify on a PKCS7 structure just after signing 3470109998Smarkm it used to fail because the attribute order did not match the 3471109998Smarkm encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: 3472109998Smarkm it uses the received order. This is necessary to tolerate some broken 3473109998Smarkm software that does not order SET OF. This is handled by encoding 3474109998Smarkm as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) 3475109998Smarkm to produce the required SET OF. 3476109998Smarkm [Steve Henson] 3477109998Smarkm 3478109998Smarkm *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and 3479109998Smarkm OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header 3480109998Smarkm files to get correct declarations of the ASN.1 item variables. 3481109998Smarkm [Richard Levitte] 3482109998Smarkm 3483109998Smarkm *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many 3484109998Smarkm PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: 3485109998Smarkm asn1_check_tlen() would sometimes attempt to use 'ctx' when it was 3486109998Smarkm NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). 3487109998Smarkm New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant 3488109998Smarkm ASN1_ITEM and no wrapper functions. 3489109998Smarkm [Steve Henson] 3490109998Smarkm 3491109998Smarkm *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These 3492109998Smarkm replace the old function pointer based I/O routines. Change most of 3493109998Smarkm the *_d2i_bio() and *_d2i_fp() functions to use these. 3494109998Smarkm [Steve Henson] 3495109998Smarkm 3496109998Smarkm *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor 3497109998Smarkm lines, recognice more "algorithms" that can be deselected, and make 3498109998Smarkm it complain about algorithm deselection that isn't recognised. 3499109998Smarkm [Richard Levitte] 3500109998Smarkm 3501109998Smarkm *) New ASN1 functions to handle dup, sign, verify, digest, pack and 3502109998Smarkm unpack operations in terms of ASN1_ITEM. Modify existing wrappers 3503109998Smarkm to use new functions. Add NO_ASN1_OLD which can be set to remove 3504109998Smarkm some old style ASN1 functions: this can be used to determine if old 3505109998Smarkm code will still work when these eventually go away. 3506109998Smarkm [Steve Henson] 3507109998Smarkm 3508109998Smarkm *) New extension functions for OCSP structures, these follow the 3509109998Smarkm same conventions as certificates and CRLs. 3510109998Smarkm [Steve Henson] 3511109998Smarkm 3512109998Smarkm *) New function X509V3_add1_i2d(). This automatically encodes and 3513109998Smarkm adds an extension. Its behaviour can be customised with various 3514109998Smarkm flags to append, replace or delete. Various wrappers added for 3515109998Smarkm certifcates and CRLs. 3516109998Smarkm [Steve Henson] 3517109998Smarkm 3518109998Smarkm *) Fix to avoid calling the underlying ASN1 print routine when 3519109998Smarkm an extension cannot be parsed. Correct a typo in the 3520109998Smarkm OCSP_SERVICELOC extension. Tidy up print OCSP format. 3521109998Smarkm [Steve Henson] 3522109998Smarkm 3523109998Smarkm *) Make mkdef.pl parse some of the ASN1 macros and add apropriate 3524109998Smarkm entries for variables. 3525109998Smarkm [Steve Henson] 3526109998Smarkm 3527109998Smarkm *) Add functionality to apps/openssl.c for detecting locking 3528109998Smarkm problems: As the program is single-threaded, all we have 3529109998Smarkm to do is register a locking callback using an array for 3530109998Smarkm storing which locks are currently held by the program. 3531109998Smarkm [Bodo Moeller] 3532109998Smarkm 3533109998Smarkm *) Use a lock around the call to CRYPTO_get_ex_new_index() in 3534109998Smarkm SSL_get_ex_data_X509_STORE_idx(), which is used in 3535109998Smarkm ssl_verify_cert_chain() and thus can be called at any time 3536109998Smarkm during TLS/SSL handshakes so that thread-safety is essential. 3537109998Smarkm Unfortunately, the ex_data design is not at all suited 3538109998Smarkm for multi-threaded use, so it probably should be abolished. 3539109998Smarkm [Bodo Moeller] 3540109998Smarkm 3541109998Smarkm *) Added Broadcom "ubsec" ENGINE to OpenSSL. 3542109998Smarkm [Broadcom, tweaked and integrated by Geoff Thorpe] 3543109998Smarkm 3544109998Smarkm *) Move common extension printing code to new function 3545109998Smarkm X509V3_print_extensions(). Reorganise OCSP print routines and 3546109998Smarkm implement some needed OCSP ASN1 functions. Add OCSP extensions. 3547109998Smarkm [Steve Henson] 3548109998Smarkm 3549109998Smarkm *) New function X509_signature_print() to remove duplication in some 3550109998Smarkm print routines. 3551109998Smarkm [Steve Henson] 3552109998Smarkm 3553109998Smarkm *) Add a special meaning when SET OF and SEQUENCE OF flags are both 3554109998Smarkm set (this was treated exactly the same as SET OF previously). This 3555109998Smarkm is used to reorder the STACK representing the structure to match the 3556109998Smarkm encoding. This will be used to get round a problem where a PKCS7 3557109998Smarkm structure which was signed could not be verified because the STACK 3558109998Smarkm order did not reflect the encoded order. 3559109998Smarkm [Steve Henson] 3560109998Smarkm 3561109998Smarkm *) Reimplement the OCSP ASN1 module using the new code. 3562109998Smarkm [Steve Henson] 3563109998Smarkm 3564109998Smarkm *) Update the X509V3 code to permit the use of an ASN1_ITEM structure 3565109998Smarkm for its ASN1 operations. The old style function pointers still exist 3566109998Smarkm for now but they will eventually go away. 3567109998Smarkm [Steve Henson] 3568109998Smarkm 3569109998Smarkm *) Merge in replacement ASN1 code from the ASN1 branch. This almost 3570109998Smarkm completely replaces the old ASN1 functionality with a table driven 3571109998Smarkm encoder and decoder which interprets an ASN1_ITEM structure describing 3572109998Smarkm the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is 3573109998Smarkm largely maintained. Almost all of the old asn1_mac.h macro based ASN1 3574109998Smarkm has also been converted to the new form. 3575109998Smarkm [Steve Henson] 3576109998Smarkm 3577109998Smarkm *) Change BN_mod_exp_recp so that negative moduli are tolerated 3578109998Smarkm (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set 3579109998Smarkm so that BN_mod_exp_mont and BN_mod_exp_mont_word work 3580109998Smarkm for negative moduli. 3581109998Smarkm [Bodo Moeller] 3582109998Smarkm 3583109998Smarkm *) Fix BN_uadd and BN_usub: Always return non-negative results instead 3584109998Smarkm of not touching the result's sign bit. 3585109998Smarkm [Bodo Moeller] 3586109998Smarkm 3587109998Smarkm *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be 3588109998Smarkm set. 3589109998Smarkm [Bodo Moeller] 3590109998Smarkm 3591109998Smarkm *) Changed the LHASH code to use prototypes for callbacks, and created 3592109998Smarkm macros to declare and implement thin (optionally static) functions 3593109998Smarkm that provide type-safety and avoid function pointer casting for the 3594109998Smarkm type-specific callbacks. 3595109998Smarkm [Geoff Thorpe] 3596109998Smarkm 3597109998Smarkm *) Added Kerberos Cipher Suites to be used with TLS, as written in 3598109998Smarkm RFC 2712. 3599109998Smarkm [Veers Staats <staatsvr@asc.hpc.mil>, 3600109998Smarkm Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte] 3601109998Smarkm 3602109998Smarkm *) Reformat the FAQ so the different questions and answers can be divided 3603109998Smarkm in sections depending on the subject. 3604109998Smarkm [Richard Levitte] 3605109998Smarkm 3606109998Smarkm *) Have the zlib compression code load ZLIB.DLL dynamically under 3607109998Smarkm Windows. 3608109998Smarkm [Richard Levitte] 3609109998Smarkm 3610109998Smarkm *) New function BN_mod_sqrt for computing square roots modulo a prime 3611109998Smarkm (using the probabilistic Tonelli-Shanks algorithm unless 3612109998Smarkm p == 3 (mod 4) or p == 5 (mod 8), which are cases that can 3613109998Smarkm be handled deterministically). 3614109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 3615109998Smarkm 3616109998Smarkm *) Make BN_mod_inverse faster by explicitly handling small quotients 3617109998Smarkm in the Euclid loop. (Speed gain about 20% for small moduli [256 or 3618109998Smarkm 512 bits], about 30% for larger ones [1024 or 2048 bits].) 3619109998Smarkm [Bodo Moeller] 3620109998Smarkm 3621109998Smarkm *) New function BN_kronecker. 3622109998Smarkm [Bodo Moeller] 3623109998Smarkm 3624109998Smarkm *) Fix BN_gcd so that it works on negative inputs; the result is 3625109998Smarkm positive unless both parameters are zero. 3626109998Smarkm Previously something reasonably close to an infinite loop was 3627109998Smarkm possible because numbers could be growing instead of shrinking 3628109998Smarkm in the implementation of Euclid's algorithm. 3629109998Smarkm [Bodo Moeller] 3630109998Smarkm 3631109998Smarkm *) Fix BN_is_word() and BN_is_one() macros to take into account the 3632109998Smarkm sign of the number in question. 3633109998Smarkm 3634109998Smarkm Fix BN_is_word(a,w) to work correctly for w == 0. 3635109998Smarkm 3636109998Smarkm The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) 3637109998Smarkm because its test if the absolute value of 'a' equals 'w'. 3638109998Smarkm Note that BN_abs_is_word does *not* handle w == 0 reliably; 3639109998Smarkm it exists mostly for use in the implementations of BN_is_zero(), 3640109998Smarkm BN_is_one(), and BN_is_word(). 3641109998Smarkm [Bodo Moeller] 3642109998Smarkm 3643109998Smarkm *) New function BN_swap. 3644109998Smarkm [Bodo Moeller] 3645109998Smarkm 3646109998Smarkm *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that 3647109998Smarkm the exponentiation functions are more likely to produce reasonable 3648109998Smarkm results on negative inputs. 3649109998Smarkm [Bodo Moeller] 3650109998Smarkm 3651109998Smarkm *) Change BN_mod_mul so that the result is always non-negative. 3652109998Smarkm Previously, it could be negative if one of the factors was negative; 3653109998Smarkm I don't think anyone really wanted that behaviour. 3654109998Smarkm [Bodo Moeller] 3655109998Smarkm 3656109998Smarkm *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c 3657109998Smarkm (except for exponentiation, which stays in crypto/bn/bn_exp.c, 3658109998Smarkm and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c) 3659109998Smarkm and add new functions: 3660109998Smarkm 3661109998Smarkm BN_nnmod 3662109998Smarkm BN_mod_sqr 3663109998Smarkm BN_mod_add 3664109998Smarkm BN_mod_add_quick 3665109998Smarkm BN_mod_sub 3666109998Smarkm BN_mod_sub_quick 3667109998Smarkm BN_mod_lshift1 3668109998Smarkm BN_mod_lshift1_quick 3669109998Smarkm BN_mod_lshift 3670109998Smarkm BN_mod_lshift_quick 3671109998Smarkm 3672109998Smarkm These functions always generate non-negative results. 3673109998Smarkm 3674109998Smarkm BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r 3675109998Smarkm such that |m| < r < 0, BN_nnmod will output rem + |m| instead). 3676109998Smarkm 3677109998Smarkm BN_mod_XXX_quick(r, a, [b,] m) generates the same result as 3678109998Smarkm BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b] 3679109998Smarkm be reduced modulo m. 3680109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 3681109998Smarkm 3682111147Snectar#if 0 3683111147Snectar The following entry accidentily appeared in the CHANGES file 3684111147Snectar distributed with OpenSSL 0.9.7. The modifications described in 3685111147Snectar it do *not* apply to OpenSSL 0.9.7. 3686111147Snectar 3687109998Smarkm *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 3688109998Smarkm was actually never needed) and in BN_mul(). The removal in BN_mul() 3689109998Smarkm required a small change in bn_mul_part_recursive() and the addition 3690109998Smarkm of the functions bn_cmp_part_words(), bn_sub_part_words() and 3691109998Smarkm bn_add_part_words(), which do the same thing as bn_cmp_words(), 3692109998Smarkm bn_sub_words() and bn_add_words() except they take arrays with 3693109998Smarkm differing sizes. 3694109998Smarkm [Richard Levitte] 3695111147Snectar#endif 3696109998Smarkm 3697109998Smarkm *) In 'openssl passwd', verify passwords read from the terminal 3698109998Smarkm unless the '-salt' option is used (which usually means that 3699109998Smarkm verification would just waste user's time since the resulting 3700109998Smarkm hash is going to be compared with some given password hash) 3701109998Smarkm or the new '-noverify' option is used. 3702109998Smarkm 3703109998Smarkm This is an incompatible change, but it does not affect 3704109998Smarkm non-interactive use of 'openssl passwd' (passwords on the command 3705109998Smarkm line, '-stdin' option, '-in ...' option) and thus should not 3706109998Smarkm cause any problems. 3707109998Smarkm [Bodo Moeller] 3708109998Smarkm 3709109998Smarkm *) Remove all references to RSAref, since there's no more need for it. 3710109998Smarkm [Richard Levitte] 3711109998Smarkm 3712109998Smarkm *) Make DSO load along a path given through an environment variable 3713109998Smarkm (SHLIB_PATH) with shl_load(). 3714109998Smarkm [Richard Levitte] 3715109998Smarkm 3716109998Smarkm *) Constify the ENGINE code as a result of BIGNUM constification. 3717109998Smarkm Also constify the RSA code and most things related to it. In a 3718109998Smarkm few places, most notable in the depth of the ASN.1 code, ugly 3719109998Smarkm casts back to non-const were required (to be solved at a later 3720109998Smarkm time) 3721109998Smarkm [Richard Levitte] 3722109998Smarkm 3723109998Smarkm *) Make it so the openssl application has all engines loaded by default. 3724109998Smarkm [Richard Levitte] 3725109998Smarkm 3726109998Smarkm *) Constify the BIGNUM routines a little more. 3727109998Smarkm [Richard Levitte] 3728109998Smarkm 3729109998Smarkm *) Add the following functions: 3730109998Smarkm 3731109998Smarkm ENGINE_load_cswift() 3732109998Smarkm ENGINE_load_chil() 3733109998Smarkm ENGINE_load_atalla() 3734109998Smarkm ENGINE_load_nuron() 3735109998Smarkm ENGINE_load_builtin_engines() 3736109998Smarkm 3737109998Smarkm That way, an application can itself choose if external engines that 3738109998Smarkm are built-in in OpenSSL shall ever be used or not. The benefit is 3739109998Smarkm that applications won't have to be linked with libdl or other dso 3740109998Smarkm libraries unless it's really needed. 3741109998Smarkm 3742109998Smarkm Changed 'openssl engine' to load all engines on demand. 3743109998Smarkm Changed the engine header files to avoid the duplication of some 3744109998Smarkm declarations (they differed!). 3745109998Smarkm [Richard Levitte] 3746109998Smarkm 3747109998Smarkm *) 'openssl engine' can now list capabilities. 3748109998Smarkm [Richard Levitte] 3749109998Smarkm 3750109998Smarkm *) Better error reporting in 'openssl engine'. 3751109998Smarkm [Richard Levitte] 3752109998Smarkm 3753109998Smarkm *) Never call load_dh_param(NULL) in s_server. 3754109998Smarkm [Bodo Moeller] 3755109998Smarkm 3756109998Smarkm *) Add engine application. It can currently list engines by name and 3757109998Smarkm identity, and test if they are actually available. 3758109998Smarkm [Richard Levitte] 3759109998Smarkm 3760109998Smarkm *) Improve RPM specification file by forcing symbolic linking and making 3761109998Smarkm sure the installed documentation is also owned by root.root. 3762109998Smarkm [Damien Miller <djm@mindrot.org>] 3763109998Smarkm 3764109998Smarkm *) Give the OpenSSL applications more possibilities to make use of 3765109998Smarkm keys (public as well as private) handled by engines. 3766109998Smarkm [Richard Levitte] 3767109998Smarkm 3768109998Smarkm *) Add OCSP code that comes from CertCo. 3769109998Smarkm [Richard Levitte] 3770109998Smarkm 3771109998Smarkm *) Add VMS support for the Rijndael code. 3772109998Smarkm [Richard Levitte] 3773109998Smarkm 3774109998Smarkm *) Added untested support for Nuron crypto accelerator. 3775109998Smarkm [Ben Laurie] 3776109998Smarkm 3777109998Smarkm *) Add support for external cryptographic devices. This code was 3778109998Smarkm previously distributed separately as the "engine" branch. 3779109998Smarkm [Geoff Thorpe, Richard Levitte] 3780109998Smarkm 3781109998Smarkm *) Rework the filename-translation in the DSO code. It is now possible to 3782109998Smarkm have far greater control over how a "name" is turned into a filename 3783109998Smarkm depending on the operating environment and any oddities about the 3784109998Smarkm different shared library filenames on each system. 3785109998Smarkm [Geoff Thorpe] 3786109998Smarkm 3787109998Smarkm *) Support threads on FreeBSD-elf in Configure. 3788109998Smarkm [Richard Levitte] 3789109998Smarkm 3790109998Smarkm *) Fix for SHA1 assembly problem with MASM: it produces 3791109998Smarkm warnings about corrupt line number information when assembling 3792109998Smarkm with debugging information. This is caused by the overlapping 3793109998Smarkm of two sections. 3794109998Smarkm [Bernd Matthes <mainbug@celocom.de>, Steve Henson] 3795109998Smarkm 3796109998Smarkm *) NCONF changes. 3797109998Smarkm NCONF_get_number() has no error checking at all. As a replacement, 3798109998Smarkm NCONF_get_number_e() is defined (_e for "error checking") and is 3799109998Smarkm promoted strongly. The old NCONF_get_number is kept around for 3800109998Smarkm binary backward compatibility. 3801109998Smarkm Make it possible for methods to load from something other than a BIO, 3802109998Smarkm by providing a function pointer that is given a name instead of a BIO. 3803109998Smarkm For example, this could be used to load configuration data from an 3804109998Smarkm LDAP server. 3805109998Smarkm [Richard Levitte] 3806109998Smarkm 3807109998Smarkm *) Fix for non blocking accept BIOs. Added new I/O special reason 3808109998Smarkm BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs 3809109998Smarkm with non blocking I/O was not possible because no retry code was 3810109998Smarkm implemented. Also added new SSL code SSL_WANT_ACCEPT to cover 3811109998Smarkm this case. 3812109998Smarkm [Steve Henson] 3813109998Smarkm 3814109998Smarkm *) Added the beginnings of Rijndael support. 3815109998Smarkm [Ben Laurie] 3816109998Smarkm 3817109998Smarkm *) Fix for bug in DirectoryString mask setting. Add support for 3818109998Smarkm X509_NAME_print_ex() in 'req' and X509_print_ex() function 3819109998Smarkm to allow certificate printing to more controllable, additional 3820109998Smarkm 'certopt' option to 'x509' to allow new printing options to be 3821109998Smarkm set. 3822109998Smarkm [Steve Henson] 3823109998Smarkm 3824109998Smarkm *) Clean old EAY MD5 hack from e_os.h. 3825109998Smarkm [Richard Levitte] 3826109998Smarkm 3827142425Snectar Changes between 0.9.6l and 0.9.6m [17 Mar 2004] 3828142425Snectar 3829142425Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 3830160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 3831142425Snectar [Joe Orton, Steve Henson] 3832142425Snectar 3833142425Snectar Changes between 0.9.6k and 0.9.6l [04 Nov 2003] 3834142425Snectar 3835142425Snectar *) Fix additional bug revealed by the NISCC test suite: 3836142425Snectar 3837142425Snectar Stop bug triggering large recursion when presented with 3838160814Ssimon certain ASN.1 tags (CVE-2003-0851) 3839142425Snectar [Steve Henson] 3840142425Snectar 3841120631Snectar Changes between 0.9.6j and 0.9.6k [30 Sep 2003] 3842120631Snectar 3843120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 3844120631Snectar 3845120631Snectar Stop out of bounds reads in the ASN1 code when presented with 3846160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 3847120631Snectar 3848120631Snectar If verify callback ignores invalid public key errors don't try to check 3849120631Snectar certificate signature with the NULL public key. 3850120631Snectar 3851120631Snectar [Steve Henson] 3852120631Snectar 3853120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 3854120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 3855120631Snectar specifications. 3856120631Snectar [Steve Henson] 3857120631Snectar 3858120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 3859120631Snectar extra data after the compression methods not only for TLS 1.0 3860120631Snectar but also for SSL 3.0 (as required by the specification). 3861120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 3862120631Snectar 3863120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 3864120631Snectar when it's 512 *bits* long, not 512 bytes. 3865120631Snectar [Richard Levitte] 3866120631Snectar 3867120631Snectar Changes between 0.9.6i and 0.9.6j [10 Apr 2003] 3868120631Snectar 3869120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 3870120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 3871120631Snectar a protocol version number mismatch like a decryption error 3872120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 3873120631Snectar [Bodo Moeller] 3874120631Snectar 3875120631Snectar *) Turn on RSA blinding by default in the default implementation 3876120631Snectar to avoid a timing attack. Applications that don't want it can call 3877120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 3878120631Snectar They would be ill-advised to do so in most cases. 3879120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 3880120631Snectar 3881120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 3882120631Snectar seeded (in this case, the secret RSA exponent is abused as 3883120631Snectar an unpredictable seed -- if it is not unpredictable, there 3884120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 3885120631Snectar by remembering the creator's thread ID in rsa->blinding and 3886120631Snectar having all other threads use local one-time blinding factors 3887120631Snectar (this requires more computation than sharing rsa->blinding, but 3888120631Snectar avoids excessive locking; and if an RSA object is not shared 3889120631Snectar between threads, blinding will still be very fast). 3890120631Snectar [Bodo Moeller] 3891120631Snectar 3892111147Snectar Changes between 0.9.6h and 0.9.6i [19 Feb 2003] 3893111147Snectar 3894111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 3895111147Snectar via timing by performing a MAC computation even if incorrrect 3896111147Snectar block cipher padding has been found. This is a countermeasure 3897111147Snectar against active attacks where the attacker has to distinguish 3898160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 3899111147Snectar 3900111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 3901111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 3902111147Snectar Martin Vuagnoux (EPFL, Ilion)] 3903111147Snectar 3904109998Smarkm Changes between 0.9.6g and 0.9.6h [5 Dec 2002] 3905109998Smarkm 3906109998Smarkm *) New function OPENSSL_cleanse(), which is used to cleanse a section of 3907109998Smarkm memory from it's contents. This is done with a counter that will 3908109998Smarkm place alternating values in each byte. This can be used to solve 3909109998Smarkm two issues: 1) the removal of calls to memset() by highly optimizing 3910109998Smarkm compilers, and 2) cleansing with other values than 0, since those can 3911109998Smarkm be read through on certain media, for example a swap space on disk. 3912109998Smarkm [Geoff Thorpe] 3913109998Smarkm 3914109998Smarkm *) Bugfix: client side session caching did not work with external caching, 3915109998Smarkm because the session->cipher setting was not restored when reloading 3916109998Smarkm from the external cache. This problem was masked, when 3917109998Smarkm SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. 3918109998Smarkm (Found by Steve Haslam <steve@araqnid.ddts.net>.) 3919109998Smarkm [Lutz Jaenicke] 3920109998Smarkm 3921109998Smarkm *) Fix client_certificate (ssl/s2_clnt.c): The permissible total 3922109998Smarkm length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. 3923109998Smarkm [Zeev Lieber <zeev-l@yahoo.com>] 3924109998Smarkm 3925109998Smarkm *) Undo an undocumented change introduced in 0.9.6e which caused 3926109998Smarkm repeated calls to OpenSSL_add_all_ciphers() and 3927109998Smarkm OpenSSL_add_all_digests() to be ignored, even after calling 3928109998Smarkm EVP_cleanup(). 3929109998Smarkm [Richard Levitte] 3930109998Smarkm 3931109998Smarkm *) Change the default configuration reader to deal with last line not 3932109998Smarkm being properly terminated. 3933109998Smarkm [Richard Levitte] 3934109998Smarkm 3935109998Smarkm *) Change X509_NAME_cmp() so it applies the special rules on handling 3936109998Smarkm DN values that are of type PrintableString, as well as RDNs of type 3937109998Smarkm emailAddress where the value has the type ia5String. 3938109998Smarkm [stefank@valicert.com via Richard Levitte] 3939109998Smarkm 3940109998Smarkm *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half 3941109998Smarkm the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently 3942109998Smarkm doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be 3943109998Smarkm the bitwise-OR of the two for use by the majority of applications 3944109998Smarkm wanting this behaviour, and update the docs. The documented 3945109998Smarkm behaviour and actual behaviour were inconsistent and had been 3946109998Smarkm changing anyway, so this is more a bug-fix than a behavioural 3947109998Smarkm change. 3948109998Smarkm [Geoff Thorpe, diagnosed by Nadav Har'El] 3949109998Smarkm 3950109998Smarkm *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c 3951109998Smarkm (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). 3952109998Smarkm [Bodo Moeller] 3953109998Smarkm 3954109998Smarkm *) Fix initialization code race conditions in 3955109998Smarkm SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), 3956109998Smarkm SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), 3957109998Smarkm SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), 3958109998Smarkm TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), 3959109998Smarkm ssl2_get_cipher_by_char(), 3960109998Smarkm ssl3_get_cipher_by_char(). 3961109998Smarkm [Patrick McCormick <patrick@tellme.com>, Bodo Moeller] 3962109998Smarkm 3963109998Smarkm *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after 3964109998Smarkm the cached sessions are flushed, as the remove_cb() might use ex_data 3965109998Smarkm contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> 3966109998Smarkm (see [openssl.org #212]). 3967109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 3968109998Smarkm 3969109998Smarkm *) Fix typo in OBJ_txt2obj which incorrectly passed the content 3970109998Smarkm length, instead of the encoding length to d2i_ASN1_OBJECT. 3971109998Smarkm [Steve Henson] 3972109998Smarkm 3973101618Snectar Changes between 0.9.6f and 0.9.6g [9 Aug 2002] 3974101618Snectar 3975101618Snectar *) [In 0.9.6g-engine release:] 3976101618Snectar Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall'). 3977101618Snectar [Lynn Gazis <lgazis@rainbow.com>] 3978101618Snectar 3979101613Snectar Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 3980101613Snectar 3981101613Snectar *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 3982101613Snectar and get fix the header length calculation. 3983101613Snectar [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 3984101613Snectar Alon Kantor <alonk@checkpoint.com> (and others), 3985101613Snectar Steve Henson] 3986101613Snectar 3987101613Snectar *) Use proper error handling instead of 'assertions' in buffer 3988101613Snectar overflow checks added in 0.9.6e. This prevents DoS (the 3989101613Snectar assertions could call abort()). 3990101613Snectar [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] 3991101613Snectar 3992100936Snectar Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 3993100936Snectar 3994109998Smarkm *) Add various sanity checks to asn1_get_length() to reject 3995109998Smarkm the ASN1 length bytes if they exceed sizeof(long), will appear 3996109998Smarkm negative or the content length exceeds the length of the 3997109998Smarkm supplied buffer. 3998109998Smarkm [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 3999109998Smarkm 4000100936Snectar *) Fix cipher selection routines: ciphers without encryption had no flags 4001100936Snectar for the cipher strength set and where therefore not handled correctly 4002100936Snectar by the selection routines (PR #130). 4003100936Snectar [Lutz Jaenicke] 4004100936Snectar 4005100936Snectar *) Fix EVP_dsa_sha macro. 4006100936Snectar [Nils Larsch] 4007100936Snectar 4008100936Snectar *) New option 4009100936Snectar SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 4010100936Snectar for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 4011100936Snectar that was added in OpenSSL 0.9.6d. 4012100936Snectar 4013100936Snectar As the countermeasure turned out to be incompatible with some 4014100936Snectar broken SSL implementations, the new option is part of SSL_OP_ALL. 4015100936Snectar SSL_OP_ALL is usually employed when compatibility with weird SSL 4016100936Snectar implementations is desired (e.g. '-bugs' option to 's_client' and 4017100936Snectar 's_server'), so the new option is automatically set in many 4018100936Snectar applications. 4019100936Snectar [Bodo Moeller] 4020100936Snectar 4021100936Snectar *) Changes in security patch: 4022100936Snectar 4023100936Snectar Changes marked "(CHATS)" were sponsored by the Defense Advanced 4024100936Snectar Research Projects Agency (DARPA) and Air Force Research Laboratory, 4025100936Snectar Air Force Materiel Command, USAF, under agreement number 4026100936Snectar F30602-01-2-0537. 4027100936Snectar 4028100936Snectar *) Add various sanity checks to asn1_get_length() to reject 4029100936Snectar the ASN1 length bytes if they exceed sizeof(long), will appear 4030100936Snectar negative or the content length exceeds the length of the 4031160814Ssimon supplied buffer. (CVE-2002-0659) 4032100936Snectar [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 4033100936Snectar 4034100936Snectar *) Assertions for various potential buffer overflows, not known to 4035100936Snectar happen in practice. 4036100936Snectar [Ben Laurie (CHATS)] 4037100936Snectar 4038100936Snectar *) Various temporary buffers to hold ASCII versions of integers were 4039160814Ssimon too small for 64 bit platforms. (CVE-2002-0655) 4040100936Snectar [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 4041100936Snectar 4042100936Snectar *) Remote buffer overflow in SSL3 protocol - an attacker could 4043160814Ssimon supply an oversized session ID to a client. (CVE-2002-0656) 4044100936Snectar [Ben Laurie (CHATS)] 4045100936Snectar 4046100936Snectar *) Remote buffer overflow in SSL2 protocol - an attacker could 4047160814Ssimon supply an oversized client master key. (CVE-2002-0656) 4048100936Snectar [Ben Laurie (CHATS)] 4049100936Snectar 4050100928Snectar Changes between 0.9.6c and 0.9.6d [9 May 2002] 4051100928Snectar 4052100928Snectar *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 4053100928Snectar encoded as NULL) with id-dsa-with-sha1. 4054100928Snectar [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 4055100928Snectar 4056100928Snectar *) Check various X509_...() return values in apps/req.c. 4057100928Snectar [Nils Larsch <nla@trustcenter.de>] 4058100928Snectar 4059100928Snectar *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 4060100928Snectar an end-of-file condition would erronously be flagged, when the CRLF 4061100928Snectar was just at the end of a processed block. The bug was discovered when 4062100928Snectar processing data through a buffering memory BIO handing the data to a 4063100928Snectar BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 4064100928Snectar <ptsekov@syntrex.com> and Nedelcho Stanev. 4065100928Snectar [Lutz Jaenicke] 4066100928Snectar 4067100928Snectar *) Implement a countermeasure against a vulnerability recently found 4068100928Snectar in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 4069100928Snectar before application data chunks to avoid the use of known IVs 4070100928Snectar with data potentially chosen by the attacker. 4071100928Snectar [Bodo Moeller] 4072100928Snectar 4073100928Snectar *) Fix length checks in ssl3_get_client_hello(). 4074100928Snectar [Bodo Moeller] 4075100928Snectar 4076100928Snectar *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 4077100928Snectar to prevent ssl3_read_internal() from incorrectly assuming that 4078100928Snectar ssl3_read_bytes() found application data while handshake 4079100928Snectar processing was enabled when in fact s->s3->in_read_app_data was 4080100928Snectar merely automatically cleared during the initial handshake. 4081100928Snectar [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 4082100928Snectar 4083100928Snectar *) Fix object definitions for Private and Enterprise: they were not 4084100928Snectar recognized in their shortname (=lowercase) representation. Extend 4085100928Snectar obj_dat.pl to issue an error when using undefined keywords instead 4086100928Snectar of silently ignoring the problem (Svenning Sorensen 4087100928Snectar <sss@sss.dnsalias.net>). 4088100928Snectar [Lutz Jaenicke] 4089100928Snectar 4090100928Snectar *) Fix DH_generate_parameters() so that it works for 'non-standard' 4091100928Snectar generators, i.e. generators other than 2 and 5. (Previously, the 4092100928Snectar code did not properly initialise the 'add' and 'rem' values to 4093100928Snectar BN_generate_prime().) 4094100928Snectar 4095100928Snectar In the new general case, we do not insist that 'generator' is 4096100928Snectar actually a primitive root: This requirement is rather pointless; 4097100928Snectar a generator of the order-q subgroup is just as good, if not 4098100928Snectar better. 4099100928Snectar [Bodo Moeller] 4100100928Snectar 4101100928Snectar *) Map new X509 verification errors to alerts. Discovered and submitted by 4102100928Snectar Tom Wu <tom@arcot.com>. 4103100928Snectar [Lutz Jaenicke] 4104100928Snectar 4105100928Snectar *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 4106100928Snectar returning non-zero before the data has been completely received 4107100928Snectar when using non-blocking I/O. 4108100928Snectar [Bodo Moeller; problem pointed out by John Hughes] 4109100928Snectar 4110100928Snectar *) Some of the ciphers missed the strength entry (SSL_LOW etc). 4111100928Snectar [Ben Laurie, Lutz Jaenicke] 4112100928Snectar 4113100928Snectar *) Fix bug in SSL_clear(): bad sessions were not removed (found by 4114100928Snectar Yoram Zahavi <YoramZ@gilian.com>). 4115100928Snectar [Lutz Jaenicke] 4116100928Snectar 4117100928Snectar *) Add information about CygWin 1.3 and on, and preserve proper 4118100928Snectar configuration for the versions before that. 4119100928Snectar [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 4120100928Snectar 4121100928Snectar *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 4122100928Snectar check whether we deal with a copy of a session and do not delete from 4123100928Snectar the cache in this case. Problem reported by "Izhar Shoshani Levi" 4124100928Snectar <izhar@checkpoint.com>. 4125100928Snectar [Lutz Jaenicke] 4126100928Snectar 4127100928Snectar *) Do not store session data into the internal session cache, if it 4128100928Snectar is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 4129100928Snectar flag is set). Proposed by Aslam <aslam@funk.com>. 4130100928Snectar [Lutz Jaenicke] 4131100928Snectar 4132100928Snectar *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 4133100928Snectar value is 0. 4134100928Snectar [Richard Levitte] 4135100928Snectar 4136109998Smarkm *) [In 0.9.6d-engine release:] 4137109998Smarkm Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 4138100928Snectar [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 4139100928Snectar 4140100928Snectar *) Add the configuration target linux-s390x. 4141100928Snectar [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 4142100928Snectar 4143100928Snectar *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 4144100928Snectar ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 4145100928Snectar variable as an indication that a ClientHello message has been 4146100928Snectar received. As the flag value will be lost between multiple 4147100928Snectar invocations of ssl3_accept when using non-blocking I/O, the 4148100928Snectar function may not be aware that a handshake has actually taken 4149100928Snectar place, thus preventing a new session from being added to the 4150100928Snectar session cache. 4151100928Snectar 4152100928Snectar To avoid this problem, we now set s->new_session to 2 instead of 4153100928Snectar using a local variable. 4154100928Snectar [Lutz Jaenicke, Bodo Moeller] 4155100928Snectar 4156100928Snectar *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 4157100928Snectar if the SSL_R_LENGTH_MISMATCH error is detected. 4158100928Snectar [Geoff Thorpe, Bodo Moeller] 4159100928Snectar 4160100928Snectar *) New 'shared_ldflag' column in Configure platform table. 4161100928Snectar [Richard Levitte] 4162100928Snectar 4163100928Snectar *) Fix EVP_CIPHER_mode macro. 4164100928Snectar ["Dan S. Camper" <dan@bti.net>] 4165100928Snectar 4166100928Snectar *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 4167100928Snectar type, we must throw them away by setting rr->length to 0. 4168100928Snectar [D P Chang <dpc@qualys.com>] 4169100928Snectar 417089837Skris Changes between 0.9.6b and 0.9.6c [21 dec 2001] 417189837Skris 417289837Skris *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 417389837Skris <Dominikus.Scherkl@biodata.com>. (The previous implementation 417489837Skris worked incorrectly for those cases where range = 10..._2 and 417589837Skris 3*range is two bits longer than range.) 417689837Skris [Bodo Moeller] 417789837Skris 417889837Skris *) Only add signing time to PKCS7 structures if it is not already 417989837Skris present. 418089837Skris [Steve Henson] 418189837Skris 418289837Skris *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 418389837Skris OBJ_ld_ce should be OBJ_id_ce. 418489837Skris Also some ip-pda OIDs in crypto/objects/objects.txt were 418589837Skris incorrect (cf. RFC 3039). 418689837Skris [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 418789837Skris 418889837Skris *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 418989837Skris returns early because it has nothing to do. 419089837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 419189837Skris 419289837Skris *) [In 0.9.6c-engine release:] 419389837Skris Fix mutex callback return values in crypto/engine/hw_ncipher.c. 419489837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 419589837Skris 419689837Skris *) [In 0.9.6c-engine release:] 419789837Skris Add support for Cryptographic Appliance's keyserver technology. 419889837Skris (Use engine 'keyclient') 419989837Skris [Cryptographic Appliances and Geoff Thorpe] 420089837Skris 420189837Skris *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 420289837Skris is called via tools/c89.sh because arguments have to be 420389837Skris rearranged (all '-L' options must appear before the first object 420489837Skris modules). 420589837Skris [Richard Shapiro <rshapiro@abinitio.com>] 420689837Skris 420789837Skris *) [In 0.9.6c-engine release:] 420889837Skris Add support for Broadcom crypto accelerator cards, backported 420989837Skris from 0.9.7. 421089837Skris [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 421189837Skris 421289837Skris *) [In 0.9.6c-engine release:] 421389837Skris Add support for SureWare crypto accelerator cards from 421489837Skris Baltimore Technologies. (Use engine 'sureware') 421589837Skris [Baltimore Technologies and Mark Cox] 421689837Skris 421789837Skris *) [In 0.9.6c-engine release:] 421889837Skris Add support for crypto accelerator cards from Accelerated 421989837Skris Encryption Processing, www.aep.ie. (Use engine 'aep') 422089837Skris [AEP Inc. and Mark Cox] 422189837Skris 422289837Skris *) Add a configuration entry for gcc on UnixWare. 422389837Skris [Gary Benson <gbenson@redhat.com>] 422489837Skris 422589837Skris *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 422689837Skris messages are stored in a single piece (fixed-length part and 422789837Skris variable-length part combined) and fix various bugs found on the way. 422889837Skris [Bodo Moeller] 422989837Skris 423089837Skris *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 423189837Skris instead. BIO_gethostbyname() does not know what timeouts are 423289837Skris appropriate, so entries would stay in cache even when they have 423389837Skris become invalid. 423489837Skris [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 423589837Skris 423689837Skris *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 423789837Skris faced with a pathologically small ClientHello fragment that does 423889837Skris not contain client_version: Instead of aborting with an error, 423989837Skris simply choose the highest available protocol version (i.e., 424089837Skris TLS 1.0 unless it is disabled). In practice, ClientHello 424189837Skris messages are never sent like this, but this change gives us 424289837Skris strictly correct behaviour at least for TLS. 424389837Skris [Bodo Moeller] 424489837Skris 424589837Skris *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 424689837Skris never resets s->method to s->ctx->method when called from within 424789837Skris one of the SSL handshake functions. 424889837Skris [Bodo Moeller; problem pointed out by Niko Baric] 424989837Skris 425089837Skris *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 425189837Skris (sent using the client's version number) if client_version is 425289837Skris smaller than the protocol version in use. Also change 425389837Skris ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 425489837Skris the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 425589837Skris the client will at least see that alert. 425689837Skris [Bodo Moeller] 425789837Skris 425889837Skris *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 425989837Skris correctly. 426089837Skris [Bodo Moeller] 426189837Skris 426289837Skris *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 426389837Skris client receives HelloRequest while in a handshake. 426489837Skris [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 426589837Skris 426689837Skris *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 426789837Skris should end in 'break', not 'goto end' which circuments various 426889837Skris cleanups done in state SSL_ST_OK. But session related stuff 426989837Skris must be disabled for SSL_ST_OK in the case that we just sent a 427089837Skris HelloRequest. 427189837Skris 427289837Skris Also avoid some overhead by not calling ssl_init_wbio_buffer() 427389837Skris before just sending a HelloRequest. 427489837Skris [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 427589837Skris 427689837Skris *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 427789837Skris reveal whether illegal block cipher padding was found or a MAC 427889837Skris verification error occured. (Neither SSLerr() codes nor alerts 427989837Skris are directly visible to potential attackers, but the information 428089837Skris may leak via logfiles.) 428189837Skris 428289837Skris Similar changes are not required for the SSL 2.0 implementation 428389837Skris because the number of padding bytes is sent in clear for SSL 2.0, 428489837Skris and the extra bytes are just ignored. However ssl/s2_pkt.c 428589837Skris failed to verify that the purported number of padding bytes is in 428689837Skris the legal range. 428789837Skris [Bodo Moeller] 428889837Skris 428989837Skris *) Add OpenUNIX-8 support including shared libraries 429089837Skris (Boyd Lynn Gerber <gerberb@zenez.com>). 429189837Skris [Lutz Jaenicke] 429289837Skris 429389837Skris *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 429489837Skris 'wristwatch attack' using huge encoding parameters (cf. 429589837Skris James H. Manger's CRYPTO 2001 paper). Note that the 429689837Skris RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 429789837Skris encoding parameters and hence was not vulnerable. 429889837Skris [Bodo Moeller] 429989837Skris 430089837Skris *) BN_sqr() bug fix. 430189837Skris [Ulf M�ller, reported by Jim Ellis <jim.ellis@cavium.com>] 430289837Skris 430389837Skris *) Rabin-Miller test analyses assume uniformly distributed witnesses, 430489837Skris so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 430589837Skris followed by modular reduction. 430689837Skris [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 430789837Skris 430889837Skris *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 430989837Skris equivalent based on BN_pseudo_rand() instead of BN_rand(). 431089837Skris [Bodo Moeller] 431189837Skris 431289837Skris *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 431389837Skris This function was broken, as the check for a new client hello message 431489837Skris to handle SGC did not allow these large messages. 431589837Skris (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 431689837Skris [Lutz Jaenicke] 431789837Skris 431889837Skris *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 431989837Skris [Lutz Jaenicke] 432089837Skris 432189837Skris *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 432289837Skris for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 432389837Skris [Lutz Jaenicke] 432489837Skris 432589837Skris *) Rework the configuration and shared library support for Tru64 Unix. 432689837Skris The configuration part makes use of modern compiler features and 432789837Skris still retains old compiler behavior for those that run older versions 432889837Skris of the OS. The shared library support part includes a variant that 432989837Skris uses the RPATH feature, and is available through the special 433089837Skris configuration target "alpha-cc-rpath", which will never be selected 433189837Skris automatically. 433289837Skris [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 433389837Skris 433489837Skris *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 433589837Skris with the same message size as in ssl3_get_certificate_request(). 433689837Skris Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 433789837Skris messages might inadvertently be reject as too long. 433889837Skris [Petr Lampa <lampa@fee.vutbr.cz>] 433989837Skris 434089837Skris *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 434189837Skris [Andy Polyakov] 434289837Skris 434389837Skris *) Modified SSL library such that the verify_callback that has been set 434489837Skris specificly for an SSL object with SSL_set_verify() is actually being 434589837Skris used. Before the change, a verify_callback set with this function was 434689837Skris ignored and the verify_callback() set in the SSL_CTX at the time of 434789837Skris the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 434889837Skris to allow the necessary settings. 434989837Skris [Lutz Jaenicke] 435089837Skris 435189837Skris *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 435289837Skris explicitly to NULL, as at least on Solaris 8 this seems not always to be 435389837Skris done automatically (in contradiction to the requirements of the C 435489837Skris standard). This made problems when used from OpenSSH. 435589837Skris [Lutz Jaenicke] 435689837Skris 435789837Skris *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 435889837Skris dh->length and always used 435989837Skris 436089837Skris BN_rand_range(priv_key, dh->p). 436189837Skris 436289837Skris BN_rand_range() is not necessary for Diffie-Hellman, and this 436389837Skris specific range makes Diffie-Hellman unnecessarily inefficient if 436489837Skris dh->length (recommended exponent length) is much smaller than the 436589837Skris length of dh->p. We could use BN_rand_range() if the order of 436689837Skris the subgroup was stored in the DH structure, but we only have 436789837Skris dh->length. 436889837Skris 436989837Skris So switch back to 437089837Skris 437189837Skris BN_rand(priv_key, l, ...) 437289837Skris 437389837Skris where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 437489837Skris otherwise. 437589837Skris [Bodo Moeller] 437689837Skris 437789837Skris *) In 437889837Skris 437989837Skris RSA_eay_public_encrypt 438089837Skris RSA_eay_private_decrypt 438189837Skris RSA_eay_private_encrypt (signing) 438289837Skris RSA_eay_public_decrypt (signature verification) 438389837Skris 438489837Skris (default implementations for RSA_public_encrypt, 438589837Skris RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 438689837Skris always reject numbers >= n. 438789837Skris [Bodo Moeller] 438889837Skris 438989837Skris *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 439089837Skris to synchronize access to 'locking_thread'. This is necessary on 439189837Skris systems where access to 'locking_thread' (an 'unsigned long' 439289837Skris variable) is not atomic. 439389837Skris [Bodo Moeller] 439489837Skris 439589837Skris *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 439689837Skris *before* setting the 'crypto_lock_rand' flag. The previous code had 439789837Skris a race condition if 0 is a valid thread ID. 439889837Skris [Travis Vitek <vitek@roguewave.com>] 439989837Skris 440089837Skris *) Add support for shared libraries under Irix. 440189837Skris [Albert Chin-A-Young <china@thewrittenword.com>] 440289837Skris 440389837Skris *) Add configuration option to build on Linux on both big-endian and 440489837Skris little-endian MIPS. 440589837Skris [Ralf Baechle <ralf@uni-koblenz.de>] 440689837Skris 440789837Skris *) Add the possibility to create shared libraries on HP-UX. 440889837Skris [Richard Levitte] 440989837Skris 441079998Skris Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 441179998Skris 441279998Skris *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 441379998Skris to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 441479998Skris Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 441579998Skris PRNG state recovery was possible based on the output of 441679998Skris one PRNG request appropriately sized to gain knowledge on 441779998Skris 'md' followed by enough consecutive 1-byte PRNG requests 441879998Skris to traverse all of 'state'. 441979998Skris 442079998Skris 1. When updating 'md_local' (the current thread's copy of 'md') 442179998Skris during PRNG output generation, hash all of the previous 442279998Skris 'md_local' value, not just the half used for PRNG output. 442379998Skris 442479998Skris 2. Make the number of bytes from 'state' included into the hash 442579998Skris independent from the number of PRNG bytes requested. 442679998Skris 442779998Skris The first measure alone would be sufficient to avoid 442879998Skris Markku-Juhani's attack. (Actually it had never occurred 442979998Skris to me that the half of 'md_local' used for chaining was the 443079998Skris half from which PRNG output bytes were taken -- I had always 443179998Skris assumed that the secret half would be used.) The second 443279998Skris measure makes sure that additional data from 'state' is never 443379998Skris mixed into 'md_local' in small portions; this heuristically 443479998Skris further strengthens the PRNG. 443579998Skris [Bodo Moeller] 443679998Skris 443779998Skris *) Fix crypto/bn/asm/mips3.s. 443879998Skris [Andy Polyakov] 443979998Skris 444079998Skris *) When only the key is given to "enc", the IV is undefined. Print out 444179998Skris an error message in this case. 444279998Skris [Lutz Jaenicke] 444379998Skris 444479998Skris *) Handle special case when X509_NAME is empty in X509 printing routines. 444579998Skris [Steve Henson] 444679998Skris 444779998Skris *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 444879998Skris positive and less than q. 444979998Skris [Bodo Moeller] 445079998Skris 445179998Skris *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 445279998Skris used: it isn't thread safe and the add_lock_callback should handle 445379998Skris that itself. 445479998Skris [Paul Rose <Paul.Rose@bridge.com>] 445579998Skris 445679998Skris *) Verify that incoming data obeys the block size in 445779998Skris ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 445879998Skris [Bodo Moeller] 445979998Skris 446079998Skris *) Fix OAEP check. 446179998Skris [Ulf M�ller, Bodo M�ller] 446279998Skris 446379998Skris *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 446489837Skris RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 446579998Skris when fixing the server behaviour for backwards-compatible 'client 446679998Skris hello' messages. (Note that the attack is impractical against 446779998Skris SSL 3.0 and TLS 1.0 anyway because length and version checking 446879998Skris means that the probability of guessing a valid ciphertext is 446979998Skris around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 447079998Skris paper.) 447179998Skris 447279998Skris Before 0.9.5, the countermeasure (hide the error by generating a 447379998Skris random 'decryption result') did not work properly because 447479998Skris ERR_clear_error() was missing, meaning that SSL_get_error() would 447579998Skris detect the supposedly ignored error. 447679998Skris 447779998Skris Both problems are now fixed. 447879998Skris [Bodo Moeller] 447979998Skris 448079998Skris *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 448179998Skris (previously it was 1024). 448279998Skris [Bodo Moeller] 448379998Skris 448479998Skris *) Fix for compatibility mode trust settings: ignore trust settings 448579998Skris unless some valid trust or reject settings are present. 448679998Skris [Steve Henson] 448779998Skris 448879998Skris *) Fix for blowfish EVP: its a variable length cipher. 448979998Skris [Steve Henson] 449079998Skris 449179998Skris *) Fix various bugs related to DSA S/MIME verification. Handle missing 449279998Skris parameters in DSA public key structures and return an error in the 449379998Skris DSA routines if parameters are absent. 449479998Skris [Steve Henson] 449579998Skris 449679998Skris *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 449779998Skris in the current directory if neither $RANDFILE nor $HOME was set. 449879998Skris RAND_file_name() in 0.9.6a returned NULL in this case. This has 449979998Skris caused some confusion to Windows users who haven't defined $HOME. 450079998Skris Thus RAND_file_name() is changed again: e_os.h can define a 450179998Skris DEFAULT_HOME, which will be used if $HOME is not set. 450279998Skris For Windows, we use "C:"; on other platforms, we still require 450379998Skris environment variables. 450479998Skris 450579998Skris *) Move 'if (!initialized) RAND_poll()' into regions protected by 450679998Skris CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 450779998Skris having multiple threads call RAND_poll() concurrently. 450879998Skris [Bodo Moeller] 450979998Skris 451079998Skris *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 451179998Skris combination of a flag and a thread ID variable. 451279998Skris Otherwise while one thread is in ssleay_rand_bytes (which sets the 451379998Skris flag), *other* threads can enter ssleay_add_bytes without obeying 451489837Skris the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 451579998Skris that they do not hold after the first thread unsets add_do_not_lock). 451679998Skris [Bodo Moeller] 451779998Skris 451879998Skris *) Change bctest again: '-x' expressions are not available in all 451979998Skris versions of 'test'. 452079998Skris [Bodo Moeller] 452179998Skris 452276866Skris Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 452372613Skris 452476866Skris *) Fix a couple of memory leaks in PKCS7_dataDecode() 452576866Skris [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 452676866Skris 452776866Skris *) Change Configure and Makefiles to provide EXE_EXT, which will contain 452876866Skris the default extension for executables, if any. Also, make the perl 452976866Skris scripts that use symlink() to test if it really exists and use "cp" 453076866Skris if it doesn't. All this made OpenSSL compilable and installable in 453176866Skris CygWin. 453276866Skris [Richard Levitte] 453376866Skris 453476866Skris *) Fix for asn1_GetSequence() for indefinite length constructed data. 453576866Skris If SEQUENCE is length is indefinite just set c->slen to the total 453676866Skris amount of data available. 453776866Skris [Steve Henson, reported by shige@FreeBSD.org] 453876866Skris [This change does not apply to 0.9.7.] 453976866Skris 454076866Skris *) Change bctest to avoid here-documents inside command substitution 454176866Skris (workaround for FreeBSD /bin/sh bug). 454276866Skris For compatibility with Ultrix, avoid shell functions (introduced 454376866Skris in the bctest version that searches along $PATH). 454476866Skris [Bodo Moeller] 454576866Skris 454676866Skris *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 454776866Skris with des_encrypt() defined on some operating systems, like Solaris 454876866Skris and UnixWare. 454976866Skris [Richard Levitte] 455076866Skris 455176866Skris *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 455276866Skris On the Importance of Eliminating Errors in Cryptographic 455376866Skris Computations, J. Cryptology 14 (2001) 2, 101-119, 455476866Skris http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 455576866Skris [Ulf Moeller] 455676866Skris 455776866Skris *) MIPS assembler BIGNUM division bug fix. 455876866Skris [Andy Polyakov] 455976866Skris 456076866Skris *) Disabled incorrect Alpha assembler code. 456176866Skris [Richard Levitte] 456276866Skris 456376866Skris *) Fix PKCS#7 decode routines so they correctly update the length 456476866Skris after reading an EOC for the EXPLICIT tag. 456576866Skris [Steve Henson] 456676866Skris [This change does not apply to 0.9.7.] 456776866Skris 456876866Skris *) Fix bug in PKCS#12 key generation routines. This was triggered 456976866Skris if a 3DES key was generated with a 0 initial byte. Include 457076866Skris PKCS12_BROKEN_KEYGEN compilation option to retain the old 457176866Skris (but broken) behaviour. 457276866Skris [Steve Henson] 457376866Skris 457476866Skris *) Enhance bctest to search for a working bc along $PATH and print 457576866Skris it when found. 457676866Skris [Tim Rice <tim@multitalents.net> via Richard Levitte] 457776866Skris 457876866Skris *) Fix memory leaks in err.c: free err_data string if necessary; 457976866Skris don't write to the wrong index in ERR_set_error_data. 458076866Skris [Bodo Moeller] 458176866Skris 458276866Skris *) Implement ssl23_peek (analogous to ssl23_read), which previously 458376866Skris did not exist. 458476866Skris [Bodo Moeller] 458576866Skris 458676866Skris *) Replace rdtsc with _emit statements for VC++ version 5. 458776866Skris [Jeremy Cooper <jeremy@baymoo.org>] 458876866Skris 458976866Skris *) Make it possible to reuse SSLv2 sessions. 459076866Skris [Richard Levitte] 459176866Skris 459276866Skris *) In copy_email() check for >= 0 as a return value for 459376866Skris X509_NAME_get_index_by_NID() since 0 is a valid index. 459476866Skris [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 459576866Skris 459676866Skris *) Avoid coredump with unsupported or invalid public keys by checking if 459776866Skris X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 459876866Skris PKCS7_verify() fails with non detached data. 459976866Skris [Steve Henson] 460076866Skris 460176866Skris *) Don't use getenv in library functions when run as setuid/setgid. 460276866Skris New function OPENSSL_issetugid(). 460376866Skris [Ulf Moeller] 460476866Skris 460576866Skris *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 460676866Skris due to incorrect handling of multi-threading: 460776866Skris 460876866Skris 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 460976866Skris 461076866Skris 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 461176866Skris 461276866Skris 3. Count how many times MemCheck_off() has been called so that 461376866Skris nested use can be treated correctly. This also avoids 461476866Skris inband-signalling in the previous code (which relied on the 461576866Skris assumption that thread ID 0 is impossible). 461676866Skris [Bodo Moeller] 461776866Skris 461876866Skris *) Add "-rand" option also to s_client and s_server. 461976866Skris [Lutz Jaenicke] 462076866Skris 462176866Skris *) Fix CPU detection on Irix 6.x. 462276866Skris [Kurt Hockenbury <khockenb@stevens-tech.edu> and 462376866Skris "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 462476866Skris 462576866Skris *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 462676866Skris was empty. 462776866Skris [Steve Henson] 462876866Skris [This change does not apply to 0.9.7.] 462976866Skris 463076866Skris *) Use the cached encoding of an X509_NAME structure rather than 463176866Skris copying it. This is apparently the reason for the libsafe "errors" 463276866Skris but the code is actually correct. 463376866Skris [Steve Henson] 463476866Skris 463572613Skris *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 463672613Skris Bleichenbacher's DSA attack. 463776866Skris Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 463876866Skris to be set and top=0 forces the highest bit to be set; top=-1 is new 463976866Skris and leaves the highest bit random. 464076866Skris [Ulf Moeller, Bodo Moeller] 464172613Skris 464272613Skris *) In the NCONF_...-based implementations for CONF_... queries 464372613Skris (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 464472613Skris a temporary CONF structure with the data component set to NULL 464572613Skris (which gives segmentation faults in lh_retrieve). 464672613Skris Instead, use NULL for the CONF pointer in CONF_get_string and 464772613Skris CONF_get_number (which may use environment variables) and directly 464872613Skris return NULL from CONF_get_section. 464972613Skris [Bodo Moeller] 465072613Skris 465172613Skris *) Fix potential buffer overrun for EBCDIC. 465272613Skris [Ulf Moeller] 465372613Skris 465472613Skris *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 465572613Skris keyUsage if basicConstraints absent for a CA. 465672613Skris [Steve Henson] 465772613Skris 465872613Skris *) Make SMIME_write_PKCS7() write mail header values with a format that 465972613Skris is more generally accepted (no spaces before the semicolon), since 466072613Skris some programs can't parse those values properly otherwise. Also make 466172613Skris sure BIO's that break lines after each write do not create invalid 466272613Skris headers. 466372613Skris [Richard Levitte] 466472613Skris 466572613Skris *) Make the CRL encoding routines work with empty SEQUENCE OF. The 466672613Skris macros previously used would not encode an empty SEQUENCE OF 466772613Skris and break the signature. 466872613Skris [Steve Henson] 466976866Skris [This change does not apply to 0.9.7.] 467072613Skris 467172613Skris *) Zero the premaster secret after deriving the master secret in 467272613Skris DH ciphersuites. 467372613Skris [Steve Henson] 467472613Skris 467572613Skris *) Add some EVP_add_digest_alias registrations (as found in 467672613Skris OpenSSL_add_all_digests()) to SSL_library_init() 467772613Skris aka OpenSSL_add_ssl_algorithms(). This provides improved 467872613Skris compatibility with peers using X.509 certificates 467972613Skris with unconventional AlgorithmIdentifier OIDs. 468072613Skris [Bodo Moeller] 468172613Skris 468272613Skris *) Fix for Irix with NO_ASM. 468372613Skris ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 468472613Skris 468572613Skris *) ./config script fixes. 468672613Skris [Ulf Moeller, Richard Levitte] 468772613Skris 468872613Skris *) Fix 'openssl passwd -1'. 468972613Skris [Bodo Moeller] 469072613Skris 469172613Skris *) Change PKCS12_key_gen_asc() so it can cope with non null 469272613Skris terminated strings whose length is passed in the passlen 469372613Skris parameter, for example from PEM callbacks. This was done 469472613Skris by adding an extra length parameter to asc2uni(). 469572613Skris [Steve Henson, reported by <oddissey@samsung.co.kr>] 469672613Skris 469772613Skris *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 469872613Skris call failed, free the DSA structure. 469972613Skris [Bodo Moeller] 470072613Skris 470172613Skris *) Fix to uni2asc() to cope with zero length Unicode strings. 470272613Skris These are present in some PKCS#12 files. 470372613Skris [Steve Henson] 470472613Skris 470572613Skris *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 470672613Skris Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 470772613Skris when writing a 32767 byte record. 470872613Skris [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 470972613Skris 471072613Skris *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 471172613Skris obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 471272613Skris 471372613Skris (RSA objects have a reference count access to which is protected 471472613Skris by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 471572613Skris so they are meant to be shared between threads.) 471672613Skris [Bodo Moeller, Geoff Thorpe; original patch submitted by 471772613Skris "Reddie, Steven" <Steven.Reddie@ca.com>] 471872613Skris 471972613Skris *) Fix a deadlock in CRYPTO_mem_leaks(). 472072613Skris [Bodo Moeller] 472172613Skris 472276866Skris *) Use better test patterns in bntest. 472376866Skris [Ulf M�ller] 472476866Skris 472572613Skris *) rand_win.c fix for Borland C. 472672613Skris [Ulf M�ller] 472772613Skris 472872613Skris *) BN_rshift bugfix for n == 0. 472972613Skris [Bodo Moeller] 473072613Skris 473176866Skris *) Add a 'bctest' script that checks for some known 'bc' bugs 473276866Skris so that 'make test' does not abort just because 'bc' is broken. 473376866Skris [Bodo Moeller] 473476866Skris 473572613Skris *) Store verify_result within SSL_SESSION also for client side to 473672613Skris avoid potential security hole. (Re-used sessions on the client side 473772613Skris always resulted in verify_result==X509_V_OK, not using the original 473872613Skris result of the server certificate verification.) 473972613Skris [Lutz Jaenicke] 474072613Skris 474172613Skris *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 474272613Skris SSL3_RT_APPLICATION_DATA, return 0. 474372613Skris Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 474472613Skris [Bodo Moeller] 474572613Skris 474672613Skris *) Fix SSL_peek: 474772613Skris Both ssl2_peek and ssl3_peek, which were totally broken in earlier 474872613Skris releases, have been re-implemented by renaming the previous 474972613Skris implementations of ssl2_read and ssl3_read to ssl2_read_internal 475072613Skris and ssl3_read_internal, respectively, and adding 'peek' parameters 475172613Skris to them. The new ssl[23]_{read,peek} functions are calls to 475272613Skris ssl[23]_read_internal with the 'peek' flag set appropriately. 475372613Skris A 'peek' parameter has also been added to ssl3_read_bytes, which 475472613Skris does the actual work for ssl3_read_internal. 475572613Skris [Bodo Moeller] 475672613Skris 475776866Skris *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 475876866Skris the method-specific "init()" handler. Also clean up ex_data after 475976866Skris calling the method-specific "finish()" handler. Previously, this was 476076866Skris happening the other way round. 476176866Skris [Geoff Thorpe] 476276866Skris 476372613Skris *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 476472613Skris The previous value, 12, was not always sufficient for BN_mod_exp(). 476572613Skris [Bodo Moeller] 476672613Skris 476776866Skris *) Make sure that shared libraries get the internal name engine with 476876866Skris the full version number and not just 0. This should mark the 476976866Skris shared libraries as not backward compatible. Of course, this should 477076866Skris be changed again when we can guarantee backward binary compatibility. 477176866Skris [Richard Levitte] 477276866Skris 477372613Skris *) Fix typo in get_cert_by_subject() in by_dir.c 477472613Skris [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 477572613Skris 477676866Skris *) Rework the system to generate shared libraries: 477776866Skris 477876866Skris - Make note of the expected extension for the shared libraries and 477976866Skris if there is a need for symbolic links from for example libcrypto.so.0 478076866Skris to libcrypto.so.0.9.7. There is extended info in Configure for 478176866Skris that. 478276866Skris 478376866Skris - Make as few rebuilds of the shared libraries as possible. 478476866Skris 478576866Skris - Still avoid linking the OpenSSL programs with the shared libraries. 478676866Skris 478776866Skris - When installing, install the shared libraries separately from the 478876866Skris static ones. 478976866Skris [Richard Levitte] 479076866Skris 479172613Skris *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 479272613Skris 479372613Skris Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 479472613Skris and not in SSL_clear because the latter is also used by the 479572613Skris accept/connect functions; previously, the settings made by 479672613Skris SSL_set_read_ahead would be lost during the handshake. 479772613Skris [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 479872613Skris 479972613Skris *) Correct util/mkdef.pl to be selective about disabled algorithms. 480072613Skris Previously, it would create entries for disableed algorithms no 480172613Skris matter what. 480272613Skris [Richard Levitte] 480372613Skris 480472613Skris *) Added several new manual pages for SSL_* function. 480572613Skris [Lutz Jaenicke] 480672613Skris 480768651Skris Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 480868651Skris 480968651Skris *) In ssl23_get_client_hello, generate an error message when faced 481068651Skris with an initial SSL 3.0/TLS record that is too small to contain the 481168651Skris first two bytes of the ClientHello message, i.e. client_version. 481268651Skris (Note that this is a pathologic case that probably has never happened 481368651Skris in real life.) The previous approach was to use the version number 481468651Skris from the record header as a substitute; but our protocol choice 481568651Skris should not depend on that one because it is not authenticated 481668651Skris by the Finished messages. 481768651Skris [Bodo Moeller] 481868651Skris 481968651Skris *) More robust randomness gathering functions for Windows. 482068651Skris [Jeffrey Altman <jaltman@columbia.edu>] 482168651Skris 482268651Skris *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 482368651Skris not set then we don't setup the error code for issuer check errors 482468651Skris to avoid possibly overwriting other errors which the callback does 482568651Skris handle. If an application does set the flag then we assume it knows 482668651Skris what it is doing and can handle the new informational codes 482768651Skris appropriately. 482868651Skris [Steve Henson] 482968651Skris 483068651Skris *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 483168651Skris a general "ANY" type, as such it should be able to decode anything 483268651Skris including tagged types. However it didn't check the class so it would 483368651Skris wrongly interpret tagged types in the same way as their universal 483468651Skris counterpart and unknown types were just rejected. Changed so that the 483568651Skris tagged and unknown types are handled in the same way as a SEQUENCE: 483668651Skris that is the encoding is stored intact. There is also a new type 483768651Skris "V_ASN1_OTHER" which is used when the class is not universal, in this 483868651Skris case we have no idea what the actual type is so we just lump them all 483968651Skris together. 484068651Skris [Steve Henson] 484168651Skris 484268651Skris *) On VMS, stdout may very well lead to a file that is written to 484368651Skris in a record-oriented fashion. That means that every write() will 484468651Skris write a separate record, which will be read separately by the 484568651Skris programs trying to read from it. This can be very confusing. 484668651Skris 484768651Skris The solution is to put a BIO filter in the way that will buffer 484868651Skris text until a linefeed is reached, and then write everything a 484968651Skris line at a time, so every record written will be an actual line, 485068651Skris not chunks of lines and not (usually doesn't happen, but I've 485168651Skris seen it once) several lines in one record. BIO_f_linebuffer() is 485268651Skris the answer. 485368651Skris 485468651Skris Currently, it's a VMS-only method, because that's where it has 485568651Skris been tested well enough. 485668651Skris [Richard Levitte] 485768651Skris 485868651Skris *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 485968651Skris it can return incorrect results. 486068651Skris (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 486168651Skris but it was in 0.9.6-beta[12].) 486268651Skris [Bodo Moeller] 486368651Skris 486468651Skris *) Disable the check for content being present when verifying detached 486568651Skris signatures in pk7_smime.c. Some versions of Netscape (wrongly) 486668651Skris include zero length content when signing messages. 486768651Skris [Steve Henson] 486868651Skris 486968651Skris *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 487068651Skris BIO_ctrl (for BIO pairs). 487168651Skris [Bodo M�ller] 487268651Skris 487368651Skris *) Add DSO method for VMS. 487468651Skris [Richard Levitte] 487568651Skris 487668651Skris *) Bug fix: Montgomery multiplication could produce results with the 487768651Skris wrong sign. 487868651Skris [Ulf M�ller] 487968651Skris 488068651Skris *) Add RPM specification openssl.spec and modify it to build three 488168651Skris packages. The default package contains applications, application 488268651Skris documentation and run-time libraries. The devel package contains 488368651Skris include files, static libraries and function documentation. The 488468651Skris doc package contains the contents of the doc directory. The original 488568651Skris openssl.spec was provided by Damien Miller <djm@mindrot.org>. 488668651Skris [Richard Levitte] 488768651Skris 488868651Skris *) Add a large number of documentation files for many SSL routines. 488968651Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 489068651Skris 489168651Skris *) Add a configuration entry for Sony News 4. 489268651Skris [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 489368651Skris 489468651Skris *) Don't set the two most significant bits to one when generating a 489568651Skris random number < q in the DSA library. 489668651Skris [Ulf M�ller] 489768651Skris 489868651Skris *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 489968651Skris behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 490068651Skris the underlying transport is blocking) if a handshake took place. 490168651Skris (The default behaviour is needed by applications such as s_client 490268651Skris and s_server that use select() to determine when to use SSL_read; 490368651Skris but for applications that know in advance when to expect data, it 490468651Skris just makes things more complicated.) 490568651Skris [Bodo Moeller] 490668651Skris 490768651Skris *) Add RAND_egd_bytes(), which gives control over the number of bytes read 490868651Skris from EGD. 490968651Skris [Ben Laurie] 491068651Skris 491168651Skris *) Add a few more EBCDIC conditionals that make `req' and `x509' 491268651Skris work better on such systems. 491368651Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 491468651Skris 491568651Skris *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 491668651Skris Update PKCS12_parse() so it copies the friendlyName and the 491768651Skris keyid to the certificates aux info. 491868651Skris [Steve Henson] 491968651Skris 492068651Skris *) Fix bug in PKCS7_verify() which caused an infinite loop 492168651Skris if there was more than one signature. 492268651Skris [Sven Uszpelkat <su@celocom.de>] 492368651Skris 492468651Skris *) Major change in util/mkdef.pl to include extra information 492568651Skris about each symbol, as well as presentig variables as well 492668651Skris as functions. This change means that there's n more need 492768651Skris to rebuild the .num files when some algorithms are excluded. 492868651Skris [Richard Levitte] 492968651Skris 493068651Skris *) Allow the verify time to be set by an application, 493168651Skris rather than always using the current time. 493268651Skris [Steve Henson] 493368651Skris 493468651Skris *) Phase 2 verify code reorganisation. The certificate 493568651Skris verify code now looks up an issuer certificate by a 493668651Skris number of criteria: subject name, authority key id 493768651Skris and key usage. It also verifies self signed certificates 493868651Skris by the same criteria. The main comparison function is 493968651Skris X509_check_issued() which performs these checks. 494068651Skris 494168651Skris Lot of changes were necessary in order to support this 494268651Skris without completely rewriting the lookup code. 494368651Skris 494468651Skris Authority and subject key identifier are now cached. 494568651Skris 494668651Skris The LHASH 'certs' is X509_STORE has now been replaced 494768651Skris by a STACK_OF(X509_OBJECT). This is mainly because an 494868651Skris LHASH can't store or retrieve multiple objects with 494968651Skris the same hash value. 495068651Skris 495168651Skris As a result various functions (which were all internal 495268651Skris use only) have changed to handle the new X509_STORE 495368651Skris structure. This will break anything that messed round 495468651Skris with X509_STORE internally. 495568651Skris 495668651Skris The functions X509_STORE_add_cert() now checks for an 495768651Skris exact match, rather than just subject name. 495868651Skris 495968651Skris The X509_STORE API doesn't directly support the retrieval 496068651Skris of multiple certificates matching a given criteria, however 496168651Skris this can be worked round by performing a lookup first 496268651Skris (which will fill the cache with candidate certificates) 496368651Skris and then examining the cache for matches. This is probably 496468651Skris the best we can do without throwing out X509_LOOKUP 496568651Skris entirely (maybe later...). 496668651Skris 496768651Skris The X509_VERIFY_CTX structure has been enhanced considerably. 496868651Skris 496968651Skris All certificate lookup operations now go via a get_issuer() 497068651Skris callback. Although this currently uses an X509_STORE it 497168651Skris can be replaced by custom lookups. This is a simple way 497268651Skris to bypass the X509_STORE hackery necessary to make this 497368651Skris work and makes it possible to use more efficient techniques 497468651Skris in future. A very simple version which uses a simple 497568651Skris STACK for its trusted certificate store is also provided 497668651Skris using X509_STORE_CTX_trusted_stack(). 497768651Skris 497868651Skris The verify_cb() and verify() callbacks now have equivalents 497968651Skris in the X509_STORE_CTX structure. 498068651Skris 498168651Skris X509_STORE_CTX also has a 'flags' field which can be used 498268651Skris to customise the verify behaviour. 498368651Skris [Steve Henson] 498468651Skris 498568651Skris *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 498668651Skris excludes S/MIME capabilities. 498768651Skris [Steve Henson] 498868651Skris 498968651Skris *) When a certificate request is read in keep a copy of the 499068651Skris original encoding of the signed data and use it when outputing 499168651Skris again. Signatures then use the original encoding rather than 499268651Skris a decoded, encoded version which may cause problems if the 499368651Skris request is improperly encoded. 499468651Skris [Steve Henson] 499568651Skris 499668651Skris *) For consistency with other BIO_puts implementations, call 499768651Skris buffer_write(b, ...) directly in buffer_puts instead of calling 499868651Skris BIO_write(b, ...). 499968651Skris 500068651Skris In BIO_puts, increment b->num_write as in BIO_write. 500168651Skris [Peter.Sylvester@EdelWeb.fr] 500268651Skris 500368651Skris *) Fix BN_mul_word for the case where the word is 0. (We have to use 500468651Skris BN_zero, we may not return a BIGNUM with an array consisting of 500568651Skris words set to zero.) 500668651Skris [Bodo Moeller] 500768651Skris 500868651Skris *) Avoid calling abort() from within the library when problems are 500968651Skris detected, except if preprocessor symbols have been defined 501068651Skris (such as REF_CHECK, BN_DEBUG etc.). 501168651Skris [Bodo Moeller] 501268651Skris 501368651Skris *) New openssl application 'rsautl'. This utility can be 501468651Skris used for low level RSA operations. DER public key 501568651Skris BIO/fp routines also added. 501668651Skris [Steve Henson] 501768651Skris 501868651Skris *) New Configure entry and patches for compiling on QNX 4. 501968651Skris [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 502068651Skris 502168651Skris *) A demo state-machine implementation was sponsored by 502268651Skris Nuron (http://www.nuron.com/) and is now available in 502368651Skris demos/state_machine. 502468651Skris [Ben Laurie] 502568651Skris 502668651Skris *) New options added to the 'dgst' utility for signature 502768651Skris generation and verification. 502868651Skris [Steve Henson] 502968651Skris 503068651Skris *) Unrecognized PKCS#7 content types are now handled via a 503168651Skris catch all ASN1_TYPE structure. This allows unsupported 503268651Skris types to be stored as a "blob" and an application can 503368651Skris encode and decode it manually. 503468651Skris [Steve Henson] 503568651Skris 503668651Skris *) Fix various signed/unsigned issues to make a_strex.c 503768651Skris compile under VC++. 503868651Skris [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 503968651Skris 504068651Skris *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 504168651Skris length if passed a buffer. ASN1_INTEGER_to_BN failed 504268651Skris if passed a NULL BN and its argument was negative. 504368651Skris [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 504468651Skris 504568651Skris *) Modification to PKCS#7 encoding routines to output definite 504668651Skris length encoding. Since currently the whole structures are in 504768651Skris memory there's not real point in using indefinite length 504868651Skris constructed encoding. However if OpenSSL is compiled with 504968651Skris the flag PKCS7_INDEFINITE_ENCODING the old form is used. 505068651Skris [Steve Henson] 505168651Skris 505268651Skris *) Added BIO_vprintf() and BIO_vsnprintf(). 505368651Skris [Richard Levitte] 505468651Skris 505568651Skris *) Added more prefixes to parse for in the the strings written 505668651Skris through a logging bio, to cover all the levels that are available 505768651Skris through syslog. The prefixes are now: 505868651Skris 505968651Skris PANIC, EMERG, EMR => LOG_EMERG 506068651Skris ALERT, ALR => LOG_ALERT 506168651Skris CRIT, CRI => LOG_CRIT 506268651Skris ERROR, ERR => LOG_ERR 506368651Skris WARNING, WARN, WAR => LOG_WARNING 506468651Skris NOTICE, NOTE, NOT => LOG_NOTICE 506568651Skris INFO, INF => LOG_INFO 506668651Skris DEBUG, DBG => LOG_DEBUG 506768651Skris 506868651Skris and as before, if none of those prefixes are present at the 506968651Skris beginning of the string, LOG_ERR is chosen. 507068651Skris 507168651Skris On Win32, the LOG_* levels are mapped according to this: 507268651Skris 507368651Skris LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 507468651Skris LOG_WARNING => EVENTLOG_WARNING_TYPE 507568651Skris LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 507668651Skris 507768651Skris [Richard Levitte] 507868651Skris 507968651Skris *) Made it possible to reconfigure with just the configuration 508068651Skris argument "reconf" or "reconfigure". The command line arguments 508168651Skris are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 508268651Skris and are retrieved from there when reconfiguring. 508368651Skris [Richard Levitte] 508468651Skris 508568651Skris *) MD4 implemented. 508668651Skris [Assar Westerlund <assar@sics.se>, Richard Levitte] 508768651Skris 508868651Skris *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 508968651Skris [Richard Levitte] 509068651Skris 509168651Skris *) The obj_dat.pl script was messing up the sorting of object 509268651Skris names. The reason was that it compared the quoted version 509368651Skris of strings as a result "OCSP" > "OCSP Signing" because 509468651Skris " > SPACE. Changed script to store unquoted versions of 509568651Skris names and add quotes on output. It was also omitting some 509668651Skris names from the lookup table if they were given a default 509768651Skris value (that is if SN is missing it is given the same 509868651Skris value as LN and vice versa), these are now added on the 509968651Skris grounds that if an object has a name we should be able to 510068651Skris look it up. Finally added warning output when duplicate 510168651Skris short or long names are found. 510268651Skris [Steve Henson] 510368651Skris 510468651Skris *) Changes needed for Tandem NSK. 510568651Skris [Scott Uroff <scott@xypro.com>] 510668651Skris 510768651Skris *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 510868651Skris RSA_padding_check_SSLv23(), special padding was never detected 510968651Skris and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 511068651Skris version rollback attacks was not effective. 511168651Skris 511268651Skris In s23_clnt.c, don't use special rollback-attack detection padding 511368651Skris (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 511468651Skris client; similarly, in s23_srvr.c, don't do the rollback check if 511568651Skris SSL 2.0 is the only protocol enabled in the server. 511668651Skris [Bodo Moeller] 511768651Skris 511868651Skris *) Make it possible to get hexdumps of unprintable data with 'openssl 511968651Skris asn1parse'. By implication, the functions ASN1_parse_dump() and 512068651Skris BIO_dump_indent() are added. 512168651Skris [Richard Levitte] 512268651Skris 512368651Skris *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 512468651Skris these print out strings and name structures based on various 512568651Skris flags including RFC2253 support and proper handling of 512668651Skris multibyte characters. Added options to the 'x509' utility 512768651Skris to allow the various flags to be set. 512868651Skris [Steve Henson] 512968651Skris 513068651Skris *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 513168651Skris Also change the functions X509_cmp_current_time() and 513268651Skris X509_gmtime_adj() work with an ASN1_TIME structure, 513368651Skris this will enable certificates using GeneralizedTime in validity 513468651Skris dates to be checked. 513568651Skris [Steve Henson] 513668651Skris 513768651Skris *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 513868651Skris negative public key encodings) on by default, 513968651Skris NO_NEG_PUBKEY_BUG can be set to disable it. 514068651Skris [Steve Henson] 514168651Skris 514268651Skris *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 514368651Skris content octets. An i2c_ASN1_OBJECT is unnecessary because 514468651Skris the encoding can be trivially obtained from the structure. 514568651Skris [Steve Henson] 514668651Skris 514768651Skris *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 514868651Skris not read locks (CRYPTO_r_[un]lock). 514968651Skris [Bodo Moeller] 515068651Skris 515168651Skris *) A first attempt at creating official support for shared 515268651Skris libraries through configuration. I've kept it so the 515368651Skris default is static libraries only, and the OpenSSL programs 515468651Skris are always statically linked for now, but there are 515568651Skris preparations for dynamic linking in place. 515689837Skris This has been tested on Linux and Tru64. 515768651Skris [Richard Levitte] 515868651Skris 515968651Skris *) Randomness polling function for Win9x, as described in: 516068651Skris Peter Gutmann, Software Generation of Practically Strong 516168651Skris Random Numbers. 516268651Skris [Ulf M�ller] 516368651Skris 516468651Skris *) Fix so PRNG is seeded in req if using an already existing 516568651Skris DSA key. 516668651Skris [Steve Henson] 516768651Skris 516868651Skris *) New options to smime application. -inform and -outform 516968651Skris allow alternative formats for the S/MIME message including 517068651Skris PEM and DER. The -content option allows the content to be 517168651Skris specified separately. This should allow things like Netscape 517268651Skris form signing output easier to verify. 517368651Skris [Steve Henson] 517468651Skris 517568651Skris *) Fix the ASN1 encoding of tags using the 'long form'. 517668651Skris [Steve Henson] 517768651Skris 517868651Skris *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 517968651Skris STRING types. These convert content octets to and from the 518068651Skris underlying type. The actual tag and length octets are 518168651Skris already assumed to have been read in and checked. These 518268651Skris are needed because all other string types have virtually 518368651Skris identical handling apart from the tag. By having versions 518468651Skris of the ASN1 functions that just operate on content octets 518568651Skris IMPLICIT tagging can be handled properly. It also allows 518668651Skris the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 518768651Skris and ASN1_INTEGER are identical apart from the tag. 518868651Skris [Steve Henson] 518968651Skris 519068651Skris *) Change the handling of OID objects as follows: 519168651Skris 519268651Skris - New object identifiers are inserted in objects.txt, following 519368651Skris the syntax given in objects.README. 519468651Skris - objects.pl is used to process obj_mac.num and create a new 519568651Skris obj_mac.h. 519668651Skris - obj_dat.pl is used to create a new obj_dat.h, using the data in 519768651Skris obj_mac.h. 519868651Skris 519968651Skris This is currently kind of a hack, and the perl code in objects.pl 520068651Skris isn't very elegant, but it works as I intended. The simplest way 520168651Skris to check that it worked correctly is to look in obj_dat.h and 520268651Skris check the array nid_objs and make sure the objects haven't moved 520368651Skris around (this is important!). Additions are OK, as well as 520468651Skris consistent name changes. 520568651Skris [Richard Levitte] 520668651Skris 520768651Skris *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 520868651Skris [Bodo Moeller] 520968651Skris 521068651Skris *) Addition of the command line parameter '-rand file' to 'openssl req'. 521168651Skris The given file adds to whatever has already been seeded into the 521268651Skris random pool through the RANDFILE configuration file option or 521368651Skris environment variable, or the default random state file. 521468651Skris [Richard Levitte] 521568651Skris 521668651Skris *) mkstack.pl now sorts each macro group into lexical order. 521768651Skris Previously the output order depended on the order the files 521868651Skris appeared in the directory, resulting in needless rewriting 521968651Skris of safestack.h . 522068651Skris [Steve Henson] 522168651Skris 522268651Skris *) Patches to make OpenSSL compile under Win32 again. Mostly 522368651Skris work arounds for the VC++ problem that it treats func() as 522468651Skris func(void). Also stripped out the parts of mkdef.pl that 522568651Skris added extra typesafe functions: these no longer exist. 522668651Skris [Steve Henson] 522768651Skris 522868651Skris *) Reorganisation of the stack code. The macros are now all 522968651Skris collected in safestack.h . Each macro is defined in terms of 523068651Skris a "stack macro" of the form SKM_<name>(type, a, b). The 523168651Skris DEBUG_SAFESTACK is now handled in terms of function casts, 523268651Skris this has the advantage of retaining type safety without the 523368651Skris use of additional functions. If DEBUG_SAFESTACK is not defined 523468651Skris then the non typesafe macros are used instead. Also modified the 523568651Skris mkstack.pl script to handle the new form. Needs testing to see 523668651Skris if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 523768651Skris the default if no major problems. Similar behaviour for ASN1_SET_OF 523868651Skris and PKCS12_STACK_OF. 523968651Skris [Steve Henson] 524068651Skris 524168651Skris *) When some versions of IIS use the 'NET' form of private key the 524268651Skris key derivation algorithm is different. Normally MD5(password) is 524368651Skris used as a 128 bit RC4 key. In the modified case 524468651Skris MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 524568651Skris new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 524668651Skris as the old Netscape_RSA functions except they have an additional 524768651Skris 'sgckey' parameter which uses the modified algorithm. Also added 524868651Skris an -sgckey command line option to the rsa utility. Thanks to 524968651Skris Adrian Peck <bertie@ncipher.com> for posting details of the modified 525068651Skris algorithm to openssl-dev. 525168651Skris [Steve Henson] 525268651Skris 525368651Skris *) The evp_local.h macros were using 'c.##kname' which resulted in 525468651Skris invalid expansion on some systems (SCO 5.0.5 for example). 525568651Skris Corrected to 'c.kname'. 525668651Skris [Phillip Porch <root@theporch.com>] 525768651Skris 525868651Skris *) New X509_get1_email() and X509_REQ_get1_email() functions that return 525968651Skris a STACK of email addresses from a certificate or request, these look 526068651Skris in the subject name and the subject alternative name extensions and 526168651Skris omit any duplicate addresses. 526268651Skris [Steve Henson] 526368651Skris 526468651Skris *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 526568651Skris This makes DSA verification about 2 % faster. 526668651Skris [Bodo Moeller] 526768651Skris 526868651Skris *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 526968651Skris (meaning that now 2^5 values will be precomputed, which is only 4 KB 527068651Skris plus overhead for 1024 bit moduli). 527168651Skris This makes exponentiations about 0.5 % faster for 1024 bit 527268651Skris exponents (as measured by "openssl speed rsa2048"). 527368651Skris [Bodo Moeller] 527468651Skris 527568651Skris *) Rename memory handling macros to avoid conflicts with other 527668651Skris software: 527768651Skris Malloc => OPENSSL_malloc 527868651Skris Malloc_locked => OPENSSL_malloc_locked 527968651Skris Realloc => OPENSSL_realloc 528068651Skris Free => OPENSSL_free 528168651Skris [Richard Levitte] 528268651Skris 528368651Skris *) New function BN_mod_exp_mont_word for small bases (roughly 15% 528468651Skris faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 528568651Skris [Bodo Moeller] 528668651Skris 528768651Skris *) CygWin32 support. 528868651Skris [John Jarvie <jjarvie@newsguy.com>] 528968651Skris 529068651Skris *) The type-safe stack code has been rejigged. It is now only compiled 529168651Skris in when OpenSSL is configured with the DEBUG_SAFESTACK option and 529268651Skris by default all type-specific stack functions are "#define"d back to 529368651Skris standard stack functions. This results in more streamlined output 529468651Skris but retains the type-safety checking possibilities of the original 529568651Skris approach. 529668651Skris [Geoff Thorpe] 529768651Skris 529868651Skris *) The STACK code has been cleaned up, and certain type declarations 529968651Skris that didn't make a lot of sense have been brought in line. This has 530068651Skris also involved a cleanup of sorts in safestack.h to more correctly 530168651Skris map type-safe stack functions onto their plain stack counterparts. 530268651Skris This work has also resulted in a variety of "const"ifications of 530368651Skris lots of the code, especially "_cmp" operations which should normally 530468651Skris be prototyped with "const" parameters anyway. 530568651Skris [Geoff Thorpe] 530668651Skris 530768651Skris *) When generating bytes for the first time in md_rand.c, 'stir the pool' 530868651Skris by seeding with STATE_SIZE dummy bytes (with zero entropy count). 530968651Skris (The PRNG state consists of two parts, the large pool 'state' and 'md', 531068651Skris where all of 'md' is used each time the PRNG is used, but 'state' 531168651Skris is used only indexed by a cyclic counter. As entropy may not be 531268651Skris well distributed from the beginning, 'md' is important as a 531368651Skris chaining variable. However, the output function chains only half 531468651Skris of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 531568651Skris all of 'md', and seeding with STATE_SIZE dummy bytes will result 531668651Skris in all of 'state' being rewritten, with the new values depending 531768651Skris on virtually all of 'md'. This overcomes the 80 bit limitation.) 531868651Skris [Bodo Moeller] 531968651Skris 532068651Skris *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 532168651Skris the handshake is continued after ssl_verify_cert_chain(); 532268651Skris otherwise, if SSL_VERIFY_NONE is set, remaining error codes 532368651Skris can lead to 'unexplainable' connection aborts later. 532468651Skris [Bodo Moeller; problem tracked down by Lutz Jaenicke] 532568651Skris 532668651Skris *) Major EVP API cipher revision. 532768651Skris Add hooks for extra EVP features. This allows various cipher 532868651Skris parameters to be set in the EVP interface. Support added for variable 532968651Skris key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 533068651Skris setting of RC2 and RC5 parameters. 533168651Skris 533268651Skris Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 533368651Skris ciphers. 533468651Skris 533568651Skris Remove lots of duplicated code from the EVP library. For example *every* 533668651Skris cipher init() function handles the 'iv' in the same way according to the 533768651Skris cipher mode. They also all do nothing if the 'key' parameter is NULL and 533868651Skris for CFB and OFB modes they zero ctx->num. 533968651Skris 534068651Skris New functionality allows removal of S/MIME code RC2 hack. 534168651Skris 534268651Skris Most of the routines have the same form and so can be declared in terms 534368651Skris of macros. 534468651Skris 534568651Skris By shifting this to the top level EVP_CipherInit() it can be removed from 534668651Skris all individual ciphers. If the cipher wants to handle IVs or keys 534768651Skris differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 534868651Skris flags. 534968651Skris 535068651Skris Change lots of functions like EVP_EncryptUpdate() to now return a 535168651Skris value: although software versions of the algorithms cannot fail 535268651Skris any installed hardware versions can. 535368651Skris [Steve Henson] 535468651Skris 535568651Skris *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 535668651Skris this option is set, tolerate broken clients that send the negotiated 535768651Skris protocol version number instead of the requested protocol version 535868651Skris number. 535968651Skris [Bodo Moeller] 536068651Skris 536168651Skris *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 536268651Skris i.e. non-zero for export ciphersuites, zero otherwise. 536368651Skris Previous versions had this flag inverted, inconsistent with 536468651Skris rsa_tmp_cb (..._TMP_RSA_CB). 536568651Skris [Bodo Moeller; problem reported by Amit Chopra] 536668651Skris 536768651Skris *) Add missing DSA library text string. Work around for some IIS 536868651Skris key files with invalid SEQUENCE encoding. 536968651Skris [Steve Henson] 537068651Skris 537168651Skris *) Add a document (doc/standards.txt) that list all kinds of standards 537268651Skris and so on that are implemented in OpenSSL. 537368651Skris [Richard Levitte] 537468651Skris 537568651Skris *) Enhance c_rehash script. Old version would mishandle certificates 537668651Skris with the same subject name hash and wouldn't handle CRLs at all. 537768651Skris Added -fingerprint option to crl utility, to support new c_rehash 537868651Skris features. 537968651Skris [Steve Henson] 538068651Skris 538168651Skris *) Eliminate non-ANSI declarations in crypto.h and stack.h. 538268651Skris [Ulf M�ller] 538368651Skris 538468651Skris *) Fix for SSL server purpose checking. Server checking was 538568651Skris rejecting certificates which had extended key usage present 538668651Skris but no ssl client purpose. 538768651Skris [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 538868651Skris 538968651Skris *) Make PKCS#12 code work with no password. The PKCS#12 spec 539068651Skris is a little unclear about how a blank password is handled. 539168651Skris Since the password in encoded as a BMPString with terminating 539268651Skris double NULL a zero length password would end up as just the 539368651Skris double NULL. However no password at all is different and is 539468651Skris handled differently in the PKCS#12 key generation code. NS 539568651Skris treats a blank password as zero length. MSIE treats it as no 539668651Skris password on export: but it will try both on import. We now do 539768651Skris the same: PKCS12_parse() tries zero length and no password if 539868651Skris the password is set to "" or NULL (NULL is now a valid password: 539968651Skris it wasn't before) as does the pkcs12 application. 540068651Skris [Steve Henson] 540168651Skris 540268651Skris *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 540368651Skris perror when PEM_read_bio_X509_REQ fails, the error message must 540468651Skris be obtained from the error queue. 540568651Skris [Bodo Moeller] 540668651Skris 540768651Skris *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 540868651Skris it in ERR_remove_state if appropriate, and change ERR_get_state 540968651Skris accordingly to avoid race conditions (this is necessary because 541068651Skris thread_hash is no longer constant once set). 541168651Skris [Bodo Moeller] 541268651Skris 541368651Skris *) Bugfix for linux-elf makefile.one. 541468651Skris [Ulf M�ller] 541568651Skris 541668651Skris *) RSA_get_default_method() will now cause a default 541768651Skris RSA_METHOD to be chosen if one doesn't exist already. 541868651Skris Previously this was only set during a call to RSA_new() 541968651Skris or RSA_new_method(NULL) meaning it was possible for 542068651Skris RSA_get_default_method() to return NULL. 542168651Skris [Geoff Thorpe] 542268651Skris 542368651Skris *) Added native name translation to the existing DSO code 542468651Skris that will convert (if the flag to do so is set) filenames 542568651Skris that are sufficiently small and have no path information 542668651Skris into a canonical native form. Eg. "blah" converted to 542768651Skris "libblah.so" or "blah.dll" etc. 542868651Skris [Geoff Thorpe] 542968651Skris 543068651Skris *) New function ERR_error_string_n(e, buf, len) which is like 543168651Skris ERR_error_string(e, buf), but writes at most 'len' bytes 543268651Skris including the 0 terminator. For ERR_error_string_n, 'buf' 543368651Skris may not be NULL. 543468651Skris [Damien Miller <djm@mindrot.org>, Bodo Moeller] 543568651Skris 543668651Skris *) CONF library reworked to become more general. A new CONF 543768651Skris configuration file reader "class" is implemented as well as a 543868651Skris new functions (NCONF_*, for "New CONF") to handle it. The now 543968651Skris old CONF_* functions are still there, but are reimplemented to 544068651Skris work in terms of the new functions. Also, a set of functions 544168651Skris to handle the internal storage of the configuration data is 544268651Skris provided to make it easier to write new configuration file 544368651Skris reader "classes" (I can definitely see something reading a 544468651Skris configuration file in XML format, for example), called _CONF_*, 544568651Skris or "the configuration storage API"... 544668651Skris 544768651Skris The new configuration file reading functions are: 544868651Skris 544968651Skris NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 545068651Skris NCONF_get_section, NCONF_get_string, NCONF_get_numbre 545168651Skris 545268651Skris NCONF_default, NCONF_WIN32 545368651Skris 545468651Skris NCONF_dump_fp, NCONF_dump_bio 545568651Skris 545668651Skris NCONF_default and NCONF_WIN32 are method (or "class") choosers, 545768651Skris NCONF_new creates a new CONF object. This works in the same way 545868651Skris as other interfaces in OpenSSL, like the BIO interface. 545968651Skris NCONF_dump_* dump the internal storage of the configuration file, 546068651Skris which is useful for debugging. All other functions take the same 546168651Skris arguments as the old CONF_* functions wth the exception of the 546268651Skris first that must be a `CONF *' instead of a `LHASH *'. 546368651Skris 546468651Skris To make it easer to use the new classes with the old CONF_* functions, 546568651Skris the function CONF_set_default_method is provided. 546668651Skris [Richard Levitte] 546768651Skris 546868651Skris *) Add '-tls1' option to 'openssl ciphers', which was already 546968651Skris mentioned in the documentation but had not been implemented. 547068651Skris (This option is not yet really useful because even the additional 547168651Skris experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 547268651Skris [Bodo Moeller] 547368651Skris 547468651Skris *) Initial DSO code added into libcrypto for letting OpenSSL (and 547568651Skris OpenSSL-based applications) load shared libraries and bind to 547668651Skris them in a portable way. 547768651Skris [Geoff Thorpe, with contributions from Richard Levitte] 547868651Skris 547959191Skris Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 548059191Skris 548159191Skris *) Make sure _lrotl and _lrotr are only used with MSVC. 548259191Skris 548359191Skris *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 548459191Skris (the default implementation of RAND_status). 548559191Skris 548659191Skris *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 548759191Skris to '-clrext' (= clear extensions), as intended and documented. 548859191Skris [Bodo Moeller; inconsistency pointed out by Michael Attili 548959191Skris <attili@amaxo.com>] 549059191Skris 549159191Skris *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 549259191Skris was larger than the MD block size. 549359191Skris [Steve Henson, pointed out by Yost William <YostW@tce.com>] 549459191Skris 549559191Skris *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 549659191Skris fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 549759191Skris using the passed key: if the passed key was a private key the result 549859191Skris of X509_print(), for example, would be to print out all the private key 549959191Skris components. 550059191Skris [Steve Henson] 550159191Skris 550259191Skris *) des_quad_cksum() byte order bug fix. 550359191Skris [Ulf M�ller, using the problem description in krb4-0.9.7, where 550459191Skris the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 550559191Skris 550659191Skris *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 550759191Skris discouraged. 550859191Skris [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 550959191Skris 551059191Skris *) For easily testing in shell scripts whether some command 551159191Skris 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 551259191Skris returns with exit code 0 iff no command of the given name is available. 551359191Skris 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 551459191Skris the output goes to stdout and nothing is printed to stderr. 551559191Skris Additional arguments are always ignored. 551659191Skris 551759191Skris Since for each cipher there is a command of the same name, 551859191Skris the 'no-cipher' compilation switches can be tested this way. 551959191Skris 552059191Skris ('openssl no-XXX' is not able to detect pseudo-commands such 552159191Skris as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 552259191Skris [Bodo Moeller] 552359191Skris 552459191Skris *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 552559191Skris [Bodo Moeller] 552659191Skris 552759191Skris *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 552859191Skris is set; it will be thrown away anyway because each handshake creates 552959191Skris its own key. 553059191Skris ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 553159191Skris to parameters -- in previous versions (since OpenSSL 0.9.3) the 553259191Skris 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 553359191Skris you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 553459191Skris [Bodo Moeller] 553559191Skris 553659191Skris *) New s_client option -ign_eof: EOF at stdin is ignored, and 553759191Skris 'Q' and 'R' lose their special meanings (quit/renegotiate). 553859191Skris This is part of what -quiet does; unlike -quiet, -ign_eof 553959191Skris does not suppress any output. 554059191Skris [Richard Levitte] 554159191Skris 554259191Skris *) Add compatibility options to the purpose and trust code. The 554359191Skris purpose X509_PURPOSE_ANY is "any purpose" which automatically 554459191Skris accepts a certificate or CA, this was the previous behaviour, 554559191Skris with all the associated security issues. 554659191Skris 554759191Skris X509_TRUST_COMPAT is the old trust behaviour: only and 554859191Skris automatically trust self signed roots in certificate store. A 554959191Skris new trust setting X509_TRUST_DEFAULT is used to specify that 555059191Skris a purpose has no associated trust setting and it should instead 555159191Skris use the value in the default purpose. 555259191Skris [Steve Henson] 555359191Skris 555459191Skris *) Fix the PKCS#8 DSA private key code so it decodes keys again 555559191Skris and fix a memory leak. 555659191Skris [Steve Henson] 555759191Skris 555859191Skris *) In util/mkerr.pl (which implements 'make errors'), preserve 555959191Skris reason strings from the previous version of the .c file, as 556059191Skris the default to have only downcase letters (and digits) in 556159191Skris automatically generated reasons codes is not always appropriate. 556259191Skris [Bodo Moeller] 556359191Skris 556459191Skris *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 556559191Skris using strerror. Previously, ERR_reason_error_string() returned 556659191Skris library names as reason strings for SYSerr; but SYSerr is a special 556759191Skris case where small numbers are errno values, not library numbers. 556859191Skris [Bodo Moeller] 556959191Skris 557059191Skris *) Add '-dsaparam' option to 'openssl dhparam' application. This 557159191Skris converts DSA parameters into DH parameters. (When creating parameters, 557259191Skris DSA_generate_parameters is used.) 557359191Skris [Bodo Moeller] 557459191Skris 557559191Skris *) Include 'length' (recommended exponent length) in C code generated 557659191Skris by 'openssl dhparam -C'. 557759191Skris [Bodo Moeller] 557859191Skris 557959191Skris *) The second argument to set_label in perlasm was already being used 558059191Skris so couldn't be used as a "file scope" flag. Moved to third argument 558159191Skris which was free. 558259191Skris [Steve Henson] 558359191Skris 558459191Skris *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 558559191Skris instead of RAND_bytes for encryption IVs and salts. 558659191Skris [Bodo Moeller] 558759191Skris 558859191Skris *) Include RAND_status() into RAND_METHOD instead of implementing 558959191Skris it only for md_rand.c Otherwise replacing the PRNG by calling 559059191Skris RAND_set_rand_method would be impossible. 559159191Skris [Bodo Moeller] 559259191Skris 559359191Skris *) Don't let DSA_generate_key() enter an infinite loop if the random 559459191Skris number generation fails. 559559191Skris [Bodo Moeller] 559659191Skris 559759191Skris *) New 'rand' application for creating pseudo-random output. 559859191Skris [Bodo Moeller] 559959191Skris 560059191Skris *) Added configuration support for Linux/IA64 560159191Skris [Rolf Haberrecker <rolf@suse.de>] 560259191Skris 560359191Skris *) Assembler module support for Mingw32. 560459191Skris [Ulf M�ller] 560559191Skris 560659191Skris *) Shared library support for HPUX (in shlib/). 560759191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 560859191Skris 560959191Skris *) Shared library support for Solaris gcc. 561059191Skris [Lutz Behnke <behnke@trustcenter.de>] 561159191Skris 561259191Skris Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 561359191Skris 561459191Skris *) PKCS7_encrypt() was adding text MIME headers twice because they 561559191Skris were added manually and by SMIME_crlf_copy(). 561659191Skris [Steve Henson] 561759191Skris 561859191Skris *) In bntest.c don't call BN_rand with zero bits argument. 561959191Skris [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 562059191Skris 562159191Skris *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 562259191Skris case was implemented. This caused BN_div_recp() to fail occasionally. 562359191Skris [Ulf M�ller] 562459191Skris 562559191Skris *) Add an optional second argument to the set_label() in the perl 562659191Skris assembly language builder. If this argument exists and is set 562759191Skris to 1 it signals that the assembler should use a symbol whose 562859191Skris scope is the entire file, not just the current function. This 562959191Skris is needed with MASM which uses the format label:: for this scope. 563059191Skris [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 563159191Skris 563259191Skris *) Change the ASN1 types so they are typedefs by default. Before 563359191Skris almost all types were #define'd to ASN1_STRING which was causing 563459191Skris STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 563559191Skris for example. 563659191Skris [Steve Henson] 563759191Skris 563859191Skris *) Change names of new functions to the new get1/get0 naming 563959191Skris convention: After 'get1', the caller owns a reference count 564059191Skris and has to call ..._free; 'get0' returns a pointer to some 564159191Skris data structure without incrementing reference counters. 564259191Skris (Some of the existing 'get' functions increment a reference 564359191Skris counter, some don't.) 564459191Skris Similarly, 'set1' and 'add1' functions increase reference 564559191Skris counters or duplicate objects. 564659191Skris [Steve Henson] 564759191Skris 564859191Skris *) Allow for the possibility of temp RSA key generation failure: 564959191Skris the code used to assume it always worked and crashed on failure. 565059191Skris [Steve Henson] 565159191Skris 565259191Skris *) Fix potential buffer overrun problem in BIO_printf(). 565359191Skris [Ulf M�ller, using public domain code by Patrick Powell; problem 565459191Skris pointed out by David Sacerdote <das33@cornell.edu>] 565559191Skris 565659191Skris *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 565759191Skris RAND_egd() and RAND_status(). In the command line application, 565859191Skris the EGD socket can be specified like a seed file using RANDFILE 565959191Skris or -rand. 566059191Skris [Ulf M�ller] 566159191Skris 566259191Skris *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 566359191Skris Some CAs (e.g. Verisign) distribute certificates in this form. 566459191Skris [Steve Henson] 566559191Skris 566659191Skris *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 566759191Skris list to exclude them. This means that no special compilation option 566859191Skris is needed to use anonymous DH: it just needs to be included in the 566959191Skris cipher list. 567059191Skris [Steve Henson] 567159191Skris 567259191Skris *) Change the EVP_MD_CTX_type macro so its meaning consistent with 567359191Skris EVP_MD_type. The old functionality is available in a new macro called 567459191Skris EVP_MD_md(). Change code that uses it and update docs. 567559191Skris [Steve Henson] 567659191Skris 567759191Skris *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 567859191Skris where the 'void *' argument is replaced by a function pointer argument. 567959191Skris Previously 'void *' was abused to point to functions, which works on 568059191Skris many platforms, but is not correct. As these functions are usually 568159191Skris called by macros defined in OpenSSL header files, most source code 568259191Skris should work without changes. 568359191Skris [Richard Levitte] 568459191Skris 568559191Skris *) <openssl/opensslconf.h> (which is created by Configure) now contains 568659191Skris sections with information on -D... compiler switches used for 568759191Skris compiling the library so that applications can see them. To enable 568859191Skris one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 568959191Skris must be defined. E.g., 569059191Skris #define OPENSSL_ALGORITHM_DEFINES 569159191Skris #include <openssl/opensslconf.h> 569259191Skris defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 569359191Skris [Richard Levitte, Ulf and Bodo M�ller] 569459191Skris 569559191Skris *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 569659191Skris record layer. 569759191Skris [Bodo Moeller] 569859191Skris 569959191Skris *) Change the 'other' type in certificate aux info to a STACK_OF 570059191Skris X509_ALGOR. Although not an AlgorithmIdentifier as such it has 570159191Skris the required ASN1 format: arbitrary types determined by an OID. 570259191Skris [Steve Henson] 570359191Skris 570459191Skris *) Add some PEM_write_X509_REQ_NEW() functions and a command line 570559191Skris argument to 'req'. This is not because the function is newer or 570659191Skris better than others it just uses the work 'NEW' in the certificate 570759191Skris request header lines. Some software needs this. 570859191Skris [Steve Henson] 570959191Skris 571059191Skris *) Reorganise password command line arguments: now passwords can be 571159191Skris obtained from various sources. Delete the PEM_cb function and make 571259191Skris it the default behaviour: i.e. if the callback is NULL and the 571359191Skris usrdata argument is not NULL interpret it as a null terminated pass 571459191Skris phrase. If usrdata and the callback are NULL then the pass phrase 571559191Skris is prompted for as usual. 571659191Skris [Steve Henson] 571759191Skris 571859191Skris *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 571959191Skris the support is automatically enabled. The resulting binaries will 572059191Skris autodetect the card and use it if present. 572159191Skris [Ben Laurie and Compaq Inc.] 572259191Skris 572359191Skris *) Work around for Netscape hang bug. This sends certificate request 572459191Skris and server done in one record. Since this is perfectly legal in the 572559191Skris SSL/TLS protocol it isn't a "bug" option and is on by default. See 572659191Skris the bugs/SSLv3 entry for more info. 572759191Skris [Steve Henson] 572859191Skris 572959191Skris *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 573059191Skris [Andy Polyakov] 573159191Skris 573259191Skris *) Add -rand argument to smime and pkcs12 applications and read/write 573359191Skris of seed file. 573459191Skris [Steve Henson] 573559191Skris 573659191Skris *) New 'passwd' tool for crypt(3) and apr1 password hashes. 573759191Skris [Bodo Moeller] 573859191Skris 573959191Skris *) Add command line password options to the remaining applications. 574059191Skris [Steve Henson] 574159191Skris 574259191Skris *) Bug fix for BN_div_recp() for numerators with an even number of 574359191Skris bits. 574459191Skris [Ulf M�ller] 574559191Skris 574659191Skris *) More tests in bntest.c, and changed test_bn output. 574759191Skris [Ulf M�ller] 574859191Skris 574959191Skris *) ./config recognizes MacOS X now. 575059191Skris [Andy Polyakov] 575159191Skris 575259191Skris *) Bug fix for BN_div() when the first words of num and divsor are 575359191Skris equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 575459191Skris [Ulf M�ller] 575559191Skris 575659191Skris *) Add support for various broken PKCS#8 formats, and command line 575759191Skris options to produce them. 575859191Skris [Steve Henson] 575959191Skris 576059191Skris *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 576159191Skris get temporary BIGNUMs from a BN_CTX. 576259191Skris [Ulf M�ller] 576359191Skris 576459191Skris *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 576559191Skris for p == 0. 576659191Skris [Ulf M�ller] 576759191Skris 576859191Skris *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 576959191Skris include a #define from the old name to the new. The original intent 577059191Skris was that statically linked binaries could for example just call 577159191Skris SSLeay_add_all_ciphers() to just add ciphers to the table and not 577259191Skris link with digests. This never worked becayse SSLeay_add_all_digests() 577359191Skris and SSLeay_add_all_ciphers() were in the same source file so calling 577459191Skris one would link with the other. They are now in separate source files. 577559191Skris [Steve Henson] 577659191Skris 577759191Skris *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 577859191Skris [Steve Henson] 577959191Skris 578059191Skris *) Use a less unusual form of the Miller-Rabin primality test (it used 578159191Skris a binary algorithm for exponentiation integrated into the Miller-Rabin 578259191Skris loop, our standard modexp algorithms are faster). 578359191Skris [Bodo Moeller] 578459191Skris 578559191Skris *) Support for the EBCDIC character set completed. 578659191Skris [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 578759191Skris 578859191Skris *) Source code cleanups: use const where appropriate, eliminate casts, 578959191Skris use void * instead of char * in lhash. 579059191Skris [Ulf M�ller] 579159191Skris 579259191Skris *) Bugfix: ssl3_send_server_key_exchange was not restartable 579359191Skris (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 579459191Skris this the server could overwrite ephemeral keys that the client 579559191Skris has already seen). 579659191Skris [Bodo Moeller] 579759191Skris 579859191Skris *) Turn DSA_is_prime into a macro that calls BN_is_prime, 579959191Skris using 50 iterations of the Rabin-Miller test. 580059191Skris 580159191Skris DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 580259191Skris iterations of the Rabin-Miller test as required by the appendix 580359191Skris to FIPS PUB 186[-1]) instead of DSA_is_prime. 580459191Skris As BN_is_prime_fasttest includes trial division, DSA parameter 580559191Skris generation becomes much faster. 580659191Skris 580759191Skris This implies a change for the callback functions in DSA_is_prime 580859191Skris and DSA_generate_parameters: The callback function is called once 580959191Skris for each positive witness in the Rabin-Miller test, not just 581059191Skris occasionally in the inner loop; and the parameters to the 581159191Skris callback function now provide an iteration count for the outer 581259191Skris loop rather than for the current invocation of the inner loop. 581359191Skris DSA_generate_parameters additionally can call the callback 581459191Skris function with an 'iteration count' of -1, meaning that a 581559191Skris candidate has passed the trial division test (when q is generated 581659191Skris from an application-provided seed, trial division is skipped). 581759191Skris [Bodo Moeller] 581859191Skris 581959191Skris *) New function BN_is_prime_fasttest that optionally does trial 582059191Skris division before starting the Rabin-Miller test and has 582159191Skris an additional BN_CTX * argument (whereas BN_is_prime always 582259191Skris has to allocate at least one BN_CTX). 582359191Skris 'callback(1, -1, cb_arg)' is called when a number has passed the 582459191Skris trial division stage. 582559191Skris [Bodo Moeller] 582659191Skris 582759191Skris *) Fix for bug in CRL encoding. The validity dates weren't being handled 582859191Skris as ASN1_TIME. 582959191Skris [Steve Henson] 583059191Skris 583159191Skris *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 583259191Skris [Steve Henson] 583359191Skris 583459191Skris *) New function BN_pseudo_rand(). 583559191Skris [Ulf M�ller] 583659191Skris 583759191Skris *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 583859191Skris bignum version of BN_from_montgomery() with the working code from 583959191Skris SSLeay 0.9.0 (the word based version is faster anyway), and clean up 584059191Skris the comments. 584159191Skris [Ulf M�ller] 584259191Skris 584359191Skris *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 584459191Skris made it impossible to use the same SSL_SESSION data structure in 584559191Skris SSL2 clients in multiple threads. 584659191Skris [Bodo Moeller] 584759191Skris 584859191Skris *) The return value of RAND_load_file() no longer counts bytes obtained 584959191Skris by stat(). RAND_load_file(..., -1) is new and uses the complete file 585059191Skris to seed the PRNG (previously an explicit byte count was required). 585159191Skris [Ulf M�ller, Bodo M�ller] 585259191Skris 585359191Skris *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 585459191Skris used (char *) instead of (void *) and had casts all over the place. 585559191Skris [Steve Henson] 585659191Skris 585759191Skris *) Make BN_generate_prime() return NULL on error if ret!=NULL. 585859191Skris [Ulf M�ller] 585959191Skris 586059191Skris *) Retain source code compatibility for BN_prime_checks macro: 586159191Skris BN_is_prime(..., BN_prime_checks, ...) now uses 586259191Skris BN_prime_checks_for_size to determine the appropriate number of 586359191Skris Rabin-Miller iterations. 586459191Skris [Ulf M�ller] 586559191Skris 586659191Skris *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 586759191Skris DH_CHECK_P_NOT_SAFE_PRIME. 586859191Skris (Check if this is true? OpenPGP calls them "strong".) 586959191Skris [Ulf M�ller] 587059191Skris 587159191Skris *) Merge the functionality of "dh" and "gendh" programs into a new program 587259191Skris "dhparam". The old programs are retained for now but will handle DH keys 587359191Skris (instead of parameters) in future. 587459191Skris [Steve Henson] 587559191Skris 587659191Skris *) Make the ciphers, s_server and s_client programs check the return values 587759191Skris when a new cipher list is set. 587859191Skris [Steve Henson] 587959191Skris 588059191Skris *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 588159191Skris ciphers. Before when the 56bit ciphers were enabled the sorting was 588259191Skris wrong. 588359191Skris 588459191Skris The syntax for the cipher sorting has been extended to support sorting by 588559191Skris cipher-strength (using the strength_bits hard coded in the tables). 588659191Skris The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 588759191Skris 588859191Skris Fix a bug in the cipher-command parser: when supplying a cipher command 588959191Skris string with an "undefined" symbol (neither command nor alphanumeric 589059191Skris [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 589159191Skris an error is flagged. 589259191Skris 589359191Skris Due to the strength-sorting extension, the code of the 589459191Skris ssl_create_cipher_list() function was completely rearranged. I hope that 589559191Skris the readability was also increased :-) 589659191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 589759191Skris 589859191Skris *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 589959191Skris for the first serial number and places 2 in the serial number file. This 590059191Skris avoids problems when the root CA is created with serial number zero and 590159191Skris the first user certificate has the same issuer name and serial number 590259191Skris as the root CA. 590359191Skris [Steve Henson] 590459191Skris 590559191Skris *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 590659191Skris the new code. Add documentation for this stuff. 590759191Skris [Steve Henson] 590859191Skris 590959191Skris *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 591059191Skris X509_*() to X509at_*() on the grounds that they don't handle X509 591159191Skris structures and behave in an analagous way to the X509v3 functions: 591259191Skris they shouldn't be called directly but wrapper functions should be used 591359191Skris instead. 591459191Skris 591559191Skris So we also now have some wrapper functions that call the X509at functions 591659191Skris when passed certificate requests. (TO DO: similar things can be done with 591759191Skris PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 591859191Skris things. Some of these need some d2i or i2d and print functionality 591959191Skris because they handle more complex structures.) 592059191Skris [Steve Henson] 592159191Skris 592259191Skris *) Add missing #ifndefs that caused missing symbols when building libssl 592359191Skris as a shared library without RSA. Use #ifndef NO_SSL2 instead of 592459191Skris NO_RSA in ssl/s2*.c. 592559191Skris [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 592659191Skris 592759191Skris *) Precautions against using the PRNG uninitialized: RAND_bytes() now 592859191Skris has a return value which indicates the quality of the random data 592959191Skris (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 593059191Skris error queue. New function RAND_pseudo_bytes() generates output that is 593159191Skris guaranteed to be unique but not unpredictable. RAND_add is like 593259191Skris RAND_seed, but takes an extra argument for an entropy estimate 593359191Skris (RAND_seed always assumes full entropy). 593459191Skris [Ulf M�ller] 593559191Skris 593659191Skris *) Do more iterations of Rabin-Miller probable prime test (specifically, 593759191Skris 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 593859191Skris instead of only 2 for all lengths; see BN_prime_checks_for_size definition 593959191Skris in crypto/bn/bn_prime.c for the complete table). This guarantees a 594059191Skris false-positive rate of at most 2^-80 for random input. 594159191Skris [Bodo Moeller] 594259191Skris 594359191Skris *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 594459191Skris [Bodo Moeller] 594559191Skris 594659191Skris *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 594759191Skris in the 0.9.5 release), this returns the chain 594859191Skris from an X509_CTX structure with a dup of the stack and all 594959191Skris the X509 reference counts upped: so the stack will exist 595059191Skris after X509_CTX_cleanup() has been called. Modify pkcs12.c 595159191Skris to use this. 595259191Skris 595359191Skris Also make SSL_SESSION_print() print out the verify return 595459191Skris code. 595559191Skris [Steve Henson] 595659191Skris 595759191Skris *) Add manpage for the pkcs12 command. Also change the default 595859191Skris behaviour so MAC iteration counts are used unless the new 595959191Skris -nomaciter option is used. This improves file security and 596059191Skris only older versions of MSIE (4.0 for example) need it. 596159191Skris [Steve Henson] 596259191Skris 596359191Skris *) Honor the no-xxx Configure options when creating .DEF files. 596459191Skris [Ulf M�ller] 596559191Skris 596659191Skris *) Add PKCS#10 attributes to field table: challengePassword, 596759191Skris unstructuredName and unstructuredAddress. These are taken from 596859191Skris draft PKCS#9 v2.0 but are compatible with v1.2 provided no 596959191Skris international characters are used. 597059191Skris 597159191Skris More changes to X509_ATTRIBUTE code: allow the setting of types 597259191Skris based on strings. Remove the 'loc' parameter when adding 597359191Skris attributes because these will be a SET OF encoding which is sorted 597459191Skris in ASN1 order. 597559191Skris [Steve Henson] 597659191Skris 597759191Skris *) Initial changes to the 'req' utility to allow request generation 597859191Skris automation. This will allow an application to just generate a template 597959191Skris file containing all the field values and have req construct the 598059191Skris request. 598159191Skris 598259191Skris Initial support for X509_ATTRIBUTE handling. Stacks of these are 598359191Skris used all over the place including certificate requests and PKCS#7 598459191Skris structures. They are currently handled manually where necessary with 598559191Skris some primitive wrappers for PKCS#7. The new functions behave in a 598659191Skris manner analogous to the X509 extension functions: they allow 598759191Skris attributes to be looked up by NID and added. 598859191Skris 598959191Skris Later something similar to the X509V3 code would be desirable to 599059191Skris automatically handle the encoding, decoding and printing of the 599159191Skris more complex types. The string types like challengePassword can 599259191Skris be handled by the string table functions. 599359191Skris 599459191Skris Also modified the multi byte string table handling. Now there is 599559191Skris a 'global mask' which masks out certain types. The table itself 599659191Skris can use the flag STABLE_NO_MASK to ignore the mask setting: this 599759191Skris is useful when for example there is only one permissible type 599859191Skris (as in countryName) and using the mask might result in no valid 599959191Skris types at all. 600059191Skris [Steve Henson] 600159191Skris 600259191Skris *) Clean up 'Finished' handling, and add functions SSL_get_finished and 600359191Skris SSL_get_peer_finished to allow applications to obtain the latest 600459191Skris Finished messages sent to the peer or expected from the peer, 600559191Skris respectively. (SSL_get_peer_finished is usually the Finished message 600659191Skris actually received from the peer, otherwise the protocol will be aborted.) 600759191Skris 600859191Skris As the Finished message are message digests of the complete handshake 600959191Skris (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 601059191Skris be used for external authentication procedures when the authentication 601159191Skris provided by SSL/TLS is not desired or is not enough. 601259191Skris [Bodo Moeller] 601359191Skris 601459191Skris *) Enhanced support for Alpha Linux is added. Now ./config checks if 601559191Skris the host supports BWX extension and if Compaq C is present on the 601659191Skris $PATH. Just exploiting of the BWX extension results in 20-30% 601759191Skris performance kick for some algorithms, e.g. DES and RC4 to mention 601859191Skris a couple. Compaq C in turn generates ~20% faster code for MD5 and 601959191Skris SHA1. 602059191Skris [Andy Polyakov] 602159191Skris 602259191Skris *) Add support for MS "fast SGC". This is arguably a violation of the 602359191Skris SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 602459191Skris weak crypto and after checking the certificate is SGC a second one 602559191Skris with strong crypto. MS SGC stops the first handshake after receiving 602659191Skris the server certificate message and sends a second client hello. Since 602759191Skris a server will typically do all the time consuming operations before 602859191Skris expecting any further messages from the client (server key exchange 602959191Skris is the most expensive) there is little difference between the two. 603059191Skris 603159191Skris To get OpenSSL to support MS SGC we have to permit a second client 603259191Skris hello message after we have sent server done. In addition we have to 603359191Skris reset the MAC if we do get this second client hello. 603459191Skris [Steve Henson] 603559191Skris 603659191Skris *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 603759191Skris if a DER encoded private key is RSA or DSA traditional format. Changed 603859191Skris d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 603959191Skris format DER encoded private key. Newer code should use PKCS#8 format which 604059191Skris has the key type encoded in the ASN1 structure. Added DER private key 604159191Skris support to pkcs8 application. 604259191Skris [Steve Henson] 604359191Skris 604459191Skris *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 604559191Skris ciphersuites has been selected (as required by the SSL 3/TLS 1 604659191Skris specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 604759191Skris is set, we interpret this as a request to violate the specification 604859191Skris (the worst that can happen is a handshake failure, and 'correct' 604959191Skris behaviour would result in a handshake failure anyway). 605059191Skris [Bodo Moeller] 605159191Skris 605259191Skris *) In SSL_CTX_add_session, take into account that there might be multiple 605359191Skris SSL_SESSION structures with the same session ID (e.g. when two threads 605459191Skris concurrently obtain them from an external cache). 605559191Skris The internal cache can handle only one SSL_SESSION with a given ID, 605659191Skris so if there's a conflict, we now throw out the old one to achieve 605759191Skris consistency. 605859191Skris [Bodo Moeller] 605959191Skris 606059191Skris *) Add OIDs for idea and blowfish in CBC mode. This will allow both 606159191Skris to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 606259191Skris some routines that use cipher OIDs: some ciphers do not have OIDs 606359191Skris defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 606459191Skris example. 606559191Skris [Steve Henson] 606659191Skris 606759191Skris *) Simplify the trust setting structure and code. Now we just have 606859191Skris two sequences of OIDs for trusted and rejected settings. These will 606959191Skris typically have values the same as the extended key usage extension 607059191Skris and any application specific purposes. 607159191Skris 607259191Skris The trust checking code now has a default behaviour: it will just 607359191Skris check for an object with the same NID as the passed id. Functions can 607459191Skris be provided to override either the default behaviour or the behaviour 607559191Skris for a given id. SSL client, server and email already have functions 607659191Skris in place for compatibility: they check the NID and also return "trusted" 607759191Skris if the certificate is self signed. 607859191Skris [Steve Henson] 607959191Skris 608059191Skris *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 608159191Skris traditional format into an EVP_PKEY structure. 608259191Skris [Steve Henson] 608359191Skris 608459191Skris *) Add a password callback function PEM_cb() which either prompts for 608559191Skris a password if usr_data is NULL or otherwise assumes it is a null 608659191Skris terminated password. Allow passwords to be passed on command line 608759191Skris environment or config files in a few more utilities. 608859191Skris [Steve Henson] 608959191Skris 609059191Skris *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 609159191Skris keys. Add some short names for PKCS#8 PBE algorithms and allow them 609259191Skris to be specified on the command line for the pkcs8 and pkcs12 utilities. 609359191Skris Update documentation. 609459191Skris [Steve Henson] 609559191Skris 609659191Skris *) Support for ASN1 "NULL" type. This could be handled before by using 609759191Skris ASN1_TYPE but there wasn't any function that would try to read a NULL 609859191Skris and produce an error if it couldn't. For compatibility we also have 609959191Skris ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 610059191Skris don't allocate anything because they don't need to. 610159191Skris [Steve Henson] 610259191Skris 610359191Skris *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 610459191Skris for details. 610559191Skris [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 610659191Skris 610759191Skris *) Rebuild of the memory allocation routines used by OpenSSL code and 610859191Skris possibly others as well. The purpose is to make an interface that 610959191Skris provide hooks so anyone can build a separate set of allocation and 611059191Skris deallocation routines to be used by OpenSSL, for example memory 611159191Skris pool implementations, or something else, which was previously hard 611259191Skris since Malloc(), Realloc() and Free() were defined as macros having 611359191Skris the values malloc, realloc and free, respectively (except for Win32 611459191Skris compilations). The same is provided for memory debugging code. 611559191Skris OpenSSL already comes with functionality to find memory leaks, but 611659191Skris this gives people a chance to debug other memory problems. 611759191Skris 611859191Skris With these changes, a new set of functions and macros have appeared: 611959191Skris 612068651Skris CRYPTO_set_mem_debug_functions() [F] 612168651Skris CRYPTO_get_mem_debug_functions() [F] 612268651Skris CRYPTO_dbg_set_options() [F] 612368651Skris CRYPTO_dbg_get_options() [F] 612468651Skris CRYPTO_malloc_debug_init() [M] 612559191Skris 612659191Skris The memory debug functions are NULL by default, unless the library 612759191Skris is compiled with CRYPTO_MDEBUG or friends is defined. If someone 612859191Skris wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 612959191Skris gives the standard debugging functions that come with OpenSSL) or 613059191Skris CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 613159191Skris provided by the library user) must be used. When the standard 613259191Skris debugging functions are used, CRYPTO_dbg_set_options can be used to 613359191Skris request additional information: 613459191Skris CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 613559191Skris the CRYPTO_MDEBUG_xxx macro when compiling the library. 613659191Skris 613759191Skris Also, things like CRYPTO_set_mem_functions will always give the 613859191Skris expected result (the new set of functions is used for allocation 613959191Skris and deallocation) at all times, regardless of platform and compiler 614059191Skris options. 614159191Skris 614259191Skris To finish it up, some functions that were never use in any other 614359191Skris way than through macros have a new API and new semantic: 614459191Skris 614559191Skris CRYPTO_dbg_malloc() 614659191Skris CRYPTO_dbg_realloc() 614759191Skris CRYPTO_dbg_free() 614859191Skris 614959191Skris All macros of value have retained their old syntax. 615059191Skris [Richard Levitte and Bodo Moeller] 615159191Skris 615259191Skris *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 615359191Skris ordering of SMIMECapabilities wasn't in "strength order" and there 615459191Skris was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 615559191Skris algorithm. 615659191Skris [Steve Henson] 615759191Skris 615859191Skris *) Some ASN1 types with illegal zero length encoding (INTEGER, 615959191Skris ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 616059191Skris [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 616159191Skris 616259191Skris *) Merge in my S/MIME library for OpenSSL. This provides a simple 616359191Skris S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 616459191Skris functionality to handle multipart/signed properly) and a utility 616559191Skris called 'smime' to call all this stuff. This is based on code I 616659191Skris originally wrote for Celo who have kindly allowed it to be 616759191Skris included in OpenSSL. 616859191Skris [Steve Henson] 616959191Skris 617059191Skris *) Add variants des_set_key_checked and des_set_key_unchecked of 617159191Skris des_set_key (aka des_key_sched). Global variable des_check_key 617259191Skris decides which of these is called by des_set_key; this way 617359191Skris des_check_key behaves as it always did, but applications and 617459191Skris the library itself, which was buggy for des_check_key == 1, 617559191Skris have a cleaner way to pick the version they need. 617659191Skris [Bodo Moeller] 617759191Skris 617859191Skris *) New function PKCS12_newpass() which changes the password of a 617959191Skris PKCS12 structure. 618059191Skris [Steve Henson] 618159191Skris 618259191Skris *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 618359191Skris dynamic mix. In both cases the ids can be used as an index into the 618459191Skris table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 618559191Skris functions so they accept a list of the field values and the 618659191Skris application doesn't need to directly manipulate the X509_TRUST 618759191Skris structure. 618859191Skris [Steve Henson] 618959191Skris 619059191Skris *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 619159191Skris need initialising. 619259191Skris [Steve Henson] 619359191Skris 619459191Skris *) Modify the way the V3 extension code looks up extensions. This now 619559191Skris works in a similar way to the object code: we have some "standard" 619659191Skris extensions in a static table which is searched with OBJ_bsearch() 619759191Skris and the application can add dynamic ones if needed. The file 619859191Skris crypto/x509v3/ext_dat.h now has the info: this file needs to be 619959191Skris updated whenever a new extension is added to the core code and kept 620059191Skris in ext_nid order. There is a simple program 'tabtest.c' which checks 620159191Skris this. New extensions are not added too often so this file can readily 620259191Skris be maintained manually. 620359191Skris 620459191Skris There are two big advantages in doing things this way. The extensions 620559191Skris can be looked up immediately and no longer need to be "added" using 620659191Skris X509V3_add_standard_extensions(): this function now does nothing. 620759191Skris [Side note: I get *lots* of email saying the extension code doesn't 620859191Skris work because people forget to call this function] 620959191Skris Also no dynamic allocation is done unless new extensions are added: 621059191Skris so if we don't add custom extensions there is no need to call 621159191Skris X509V3_EXT_cleanup(). 621259191Skris [Steve Henson] 621359191Skris 621459191Skris *) Modify enc utility's salting as follows: make salting the default. Add a 621559191Skris magic header, so unsalted files fail gracefully instead of just decrypting 621659191Skris to garbage. This is because not salting is a big security hole, so people 621759191Skris should be discouraged from doing it. 621859191Skris [Ben Laurie] 621959191Skris 622059191Skris *) Fixes and enhancements to the 'x509' utility. It allowed a message 622159191Skris digest to be passed on the command line but it only used this 622259191Skris parameter when signing a certificate. Modified so all relevant 622359191Skris operations are affected by the digest parameter including the 622459191Skris -fingerprint and -x509toreq options. Also -x509toreq choked if a 622559191Skris DSA key was used because it didn't fix the digest. 622659191Skris [Steve Henson] 622759191Skris 622859191Skris *) Initial certificate chain verify code. Currently tests the untrusted 622959191Skris certificates for consistency with the verify purpose (which is set 623059191Skris when the X509_STORE_CTX structure is set up) and checks the pathlength. 623159191Skris 623259191Skris There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 623359191Skris this is because it will reject chains with invalid extensions whereas 623459191Skris every previous version of OpenSSL and SSLeay made no checks at all. 623559191Skris 623659191Skris Trust code: checks the root CA for the relevant trust settings. Trust 623759191Skris settings have an initial value consistent with the verify purpose: e.g. 623859191Skris if the verify purpose is for SSL client use it expects the CA to be 623959191Skris trusted for SSL client use. However the default value can be changed to 624059191Skris permit custom trust settings: one example of this would be to only trust 624159191Skris certificates from a specific "secure" set of CAs. 624259191Skris 624359191Skris Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 624459191Skris which should be used for version portability: especially since the 624559191Skris verify structure is likely to change more often now. 624659191Skris 624759191Skris SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 624859191Skris to set them. If not set then assume SSL clients will verify SSL servers 624959191Skris and vice versa. 625059191Skris 625159191Skris Two new options to the verify program: -untrusted allows a set of 625259191Skris untrusted certificates to be passed in and -purpose which sets the 625359191Skris intended purpose of the certificate. If a purpose is set then the 625459191Skris new chain verify code is used to check extension consistency. 625559191Skris [Steve Henson] 625659191Skris 625759191Skris *) Support for the authority information access extension. 625859191Skris [Steve Henson] 625959191Skris 626059191Skris *) Modify RSA and DSA PEM read routines to transparently handle 626159191Skris PKCS#8 format private keys. New *_PUBKEY_* functions that handle 626259191Skris public keys in a format compatible with certificate 626359191Skris SubjectPublicKeyInfo structures. Unfortunately there were already 626459191Skris functions called *_PublicKey_* which used various odd formats so 626559191Skris these are retained for compatibility: however the DSA variants were 626659191Skris never in a public release so they have been deleted. Changed dsa/rsa 626759191Skris utilities to handle the new format: note no releases ever handled public 626859191Skris keys so we should be OK. 626959191Skris 627059191Skris The primary motivation for this change is to avoid the same fiasco 627159191Skris that dogs private keys: there are several incompatible private key 627259191Skris formats some of which are standard and some OpenSSL specific and 627359191Skris require various evil hacks to allow partial transparent handling and 627459191Skris even then it doesn't work with DER formats. Given the option anything 627559191Skris other than PKCS#8 should be dumped: but the other formats have to 627659191Skris stay in the name of compatibility. 627759191Skris 627859191Skris With public keys and the benefit of hindsight one standard format 627959191Skris is used which works with EVP_PKEY, RSA or DSA structures: though 628059191Skris it clearly returns an error if you try to read the wrong kind of key. 628159191Skris 628259191Skris Added a -pubkey option to the 'x509' utility to output the public key. 628359191Skris Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 628459191Skris (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 628559191Skris EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 628659191Skris that do the same as the EVP_PKEY_assign_*() except they up the 628759191Skris reference count of the added key (they don't "swallow" the 628859191Skris supplied key). 628959191Skris [Steve Henson] 629059191Skris 629159191Skris *) Fixes to crypto/x509/by_file.c the code to read in certificates and 629259191Skris CRLs would fail if the file contained no certificates or no CRLs: 629359191Skris added a new function to read in both types and return the number 629459191Skris read: this means that if none are read it will be an error. The 629559191Skris DER versions of the certificate and CRL reader would always fail 629659191Skris because it isn't possible to mix certificates and CRLs in DER format 629759191Skris without choking one or the other routine. Changed this to just read 629859191Skris a certificate: this is the best we can do. Also modified the code 629959191Skris in apps/verify.c to take notice of return codes: it was previously 630059191Skris attempting to read in certificates from NULL pointers and ignoring 630159191Skris any errors: this is one reason why the cert and CRL reader seemed 630259191Skris to work. It doesn't check return codes from the default certificate 630359191Skris routines: these may well fail if the certificates aren't installed. 630459191Skris [Steve Henson] 630559191Skris 630659191Skris *) Code to support otherName option in GeneralName. 630759191Skris [Steve Henson] 630859191Skris 630959191Skris *) First update to verify code. Change the verify utility 631059191Skris so it warns if it is passed a self signed certificate: 631159191Skris for consistency with the normal behaviour. X509_verify 631259191Skris has been modified to it will now verify a self signed 631359191Skris certificate if *exactly* the same certificate appears 631459191Skris in the store: it was previously impossible to trust a 631559191Skris single self signed certificate. This means that: 631659191Skris openssl verify ss.pem 631759191Skris now gives a warning about a self signed certificate but 631859191Skris openssl verify -CAfile ss.pem ss.pem 631959191Skris is OK. 632059191Skris [Steve Henson] 632159191Skris 632259191Skris *) For servers, store verify_result in SSL_SESSION data structure 632359191Skris (and add it to external session representation). 632459191Skris This is needed when client certificate verifications fails, 632559191Skris but an application-provided verification callback (set by 632659191Skris SSL_CTX_set_cert_verify_callback) allows accepting the session 632759191Skris anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 632859191Skris but returns 1): When the session is reused, we have to set 632959191Skris ssl->verify_result to the appropriate error code to avoid 633059191Skris security holes. 633159191Skris [Bodo Moeller, problem pointed out by Lutz Jaenicke] 633259191Skris 633359191Skris *) Fix a bug in the new PKCS#7 code: it didn't consider the 633459191Skris case in PKCS7_dataInit() where the signed PKCS7 structure 633559191Skris didn't contain any existing data because it was being created. 633659191Skris [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 633759191Skris 633859191Skris *) Add a salt to the key derivation routines in enc.c. This 633959191Skris forms the first 8 bytes of the encrypted file. Also add a 634059191Skris -S option to allow a salt to be input on the command line. 634159191Skris [Steve Henson] 634259191Skris 634359191Skris *) New function X509_cmp(). Oddly enough there wasn't a function 634459191Skris to compare two certificates. We do this by working out the SHA1 634559191Skris hash and comparing that. X509_cmp() will be needed by the trust 634659191Skris code. 634759191Skris [Steve Henson] 634859191Skris 634959191Skris *) SSL_get1_session() is like SSL_get_session(), but increments 635059191Skris the reference count in the SSL_SESSION returned. 635159191Skris [Geoff Thorpe <geoff@eu.c2.net>] 635259191Skris 635359191Skris *) Fix for 'req': it was adding a null to request attributes. 635459191Skris Also change the X509_LOOKUP and X509_INFO code to handle 635559191Skris certificate auxiliary information. 635659191Skris [Steve Henson] 635759191Skris 635859191Skris *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 635959191Skris the 'enc' command. 636059191Skris [Steve Henson] 636159191Skris 636259191Skris *) Add the possibility to add extra information to the memory leak 636359191Skris detecting output, to form tracebacks, showing from where each 636459191Skris allocation was originated: CRYPTO_push_info("constant string") adds 636559191Skris the string plus current file name and line number to a per-thread 636659191Skris stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 636759191Skris is like calling CYRPTO_pop_info() until the stack is empty. 636859191Skris Also updated memory leak detection code to be multi-thread-safe. 636959191Skris [Richard Levitte] 637059191Skris 637159191Skris *) Add options -text and -noout to pkcs7 utility and delete the 637259191Skris encryption options which never did anything. Update docs. 637359191Skris [Steve Henson] 637459191Skris 637559191Skris *) Add options to some of the utilities to allow the pass phrase 637659191Skris to be included on either the command line (not recommended on 637759191Skris OSes like Unix) or read from the environment. Update the 637859191Skris manpages and fix a few bugs. 637959191Skris [Steve Henson] 638059191Skris 638159191Skris *) Add a few manpages for some of the openssl commands. 638259191Skris [Steve Henson] 638359191Skris 638459191Skris *) Fix the -revoke option in ca. It was freeing up memory twice, 638559191Skris leaking and not finding already revoked certificates. 638659191Skris [Steve Henson] 638759191Skris 638859191Skris *) Extensive changes to support certificate auxiliary information. 638959191Skris This involves the use of X509_CERT_AUX structure and X509_AUX 639059191Skris functions. An X509_AUX function such as PEM_read_X509_AUX() 639159191Skris can still read in a certificate file in the usual way but it 639259191Skris will also read in any additional "auxiliary information". By 639359191Skris doing things this way a fair degree of compatibility can be 639459191Skris retained: existing certificates can have this information added 639559191Skris using the new 'x509' options. 639659191Skris 639759191Skris Current auxiliary information includes an "alias" and some trust 639859191Skris settings. The trust settings will ultimately be used in enhanced 639959191Skris certificate chain verification routines: currently a certificate 640059191Skris can only be trusted if it is self signed and then it is trusted 640159191Skris for all purposes. 640259191Skris [Steve Henson] 640359191Skris 640459191Skris *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 640559191Skris The problem was that one of the replacement routines had not been working 640659191Skris since SSLeay releases. For now the offending routine has been replaced 640759191Skris with non-optimised assembler. Even so, this now gives around 95% 640859191Skris performance improvement for 1024 bit RSA signs. 640959191Skris [Mark Cox] 641059191Skris 641159191Skris *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 641259191Skris handling. Most clients have the effective key size in bits equal to 641359191Skris the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 641459191Skris A few however don't do this and instead use the size of the decrypted key 641559191Skris to determine the RC2 key length and the AlgorithmIdentifier to determine 641659191Skris the effective key length. In this case the effective key length can still 641759191Skris be 40 bits but the key length can be 168 bits for example. This is fixed 641859191Skris by manually forcing an RC2 key into the EVP_PKEY structure because the 641959191Skris EVP code can't currently handle unusual RC2 key sizes: it always assumes 642059191Skris the key length and effective key length are equal. 642159191Skris [Steve Henson] 642259191Skris 642359191Skris *) Add a bunch of functions that should simplify the creation of 642459191Skris X509_NAME structures. Now you should be able to do: 642559191Skris X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 642659191Skris and have it automatically work out the correct field type and fill in 642759191Skris the structures. The more adventurous can try: 642859191Skris X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 642959191Skris and it will (hopefully) work out the correct multibyte encoding. 643059191Skris [Steve Henson] 643159191Skris 643259191Skris *) Change the 'req' utility to use the new field handling and multibyte 643359191Skris copy routines. Before the DN field creation was handled in an ad hoc 643459191Skris way in req, ca, and x509 which was rather broken and didn't support 643559191Skris BMPStrings or UTF8Strings. Since some software doesn't implement 643659191Skris BMPStrings or UTF8Strings yet, they can be enabled using the config file 643759191Skris using the dirstring_type option. See the new comment in the default 643859191Skris openssl.cnf for more info. 643959191Skris [Steve Henson] 644059191Skris 644159191Skris *) Make crypto/rand/md_rand.c more robust: 644259191Skris - Assure unique random numbers after fork(). 644359191Skris - Make sure that concurrent threads access the global counter and 644459191Skris md serializably so that we never lose entropy in them 644559191Skris or use exactly the same state in multiple threads. 644659191Skris Access to the large state is not always serializable because 644759191Skris the additional locking could be a performance killer, and 644859191Skris md should be large enough anyway. 644959191Skris [Bodo Moeller] 645059191Skris 645159191Skris *) New file apps/app_rand.c with commonly needed functionality 645259191Skris for handling the random seed file. 645359191Skris 645459191Skris Use the random seed file in some applications that previously did not: 645559191Skris ca, 645659191Skris dsaparam -genkey (which also ignored its '-rand' option), 645759191Skris s_client, 645859191Skris s_server, 645959191Skris x509 (when signing). 646059191Skris Except on systems with /dev/urandom, it is crucial to have a random 646159191Skris seed file at least for key creation, DSA signing, and for DH exchanges; 646259191Skris for RSA signatures we could do without one. 646359191Skris 646459191Skris gendh and gendsa (unlike genrsa) used to read only the first byte 646559191Skris of each file listed in the '-rand' option. The function as previously 646659191Skris found in genrsa is now in app_rand.c and is used by all programs 646759191Skris that support '-rand'. 646859191Skris [Bodo Moeller] 646959191Skris 647059191Skris *) In RAND_write_file, use mode 0600 for creating files; 647159191Skris don't just chmod when it may be too late. 647259191Skris [Bodo Moeller] 647359191Skris 647459191Skris *) Report an error from X509_STORE_load_locations 647559191Skris when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 647659191Skris [Bill Perry] 647759191Skris 647859191Skris *) New function ASN1_mbstring_copy() this copies a string in either 647959191Skris ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 648059191Skris into an ASN1_STRING type. A mask of permissible types is passed 648159191Skris and it chooses the "minimal" type to use or an error if not type 648259191Skris is suitable. 648359191Skris [Steve Henson] 648459191Skris 648559191Skris *) Add function equivalents to the various macros in asn1.h. The old 648659191Skris macros are retained with an M_ prefix. Code inside the library can 648759191Skris use the M_ macros. External code (including the openssl utility) 648859191Skris should *NOT* in order to be "shared library friendly". 648959191Skris [Steve Henson] 649059191Skris 649159191Skris *) Add various functions that can check a certificate's extensions 649259191Skris to see if it usable for various purposes such as SSL client, 649359191Skris server or S/MIME and CAs of these types. This is currently 649459191Skris VERY EXPERIMENTAL but will ultimately be used for certificate chain 649559191Skris verification. Also added a -purpose flag to x509 utility to 649659191Skris print out all the purposes. 649759191Skris [Steve Henson] 649859191Skris 649959191Skris *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 650059191Skris functions. 650159191Skris [Steve Henson] 650259191Skris 650359191Skris *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 650459191Skris for, obtain and decode and extension and obtain its critical flag. 650559191Skris This allows all the necessary extension code to be handled in a 650659191Skris single function call. 650759191Skris [Steve Henson] 650859191Skris 650959191Skris *) RC4 tune-up featuring 30-40% performance improvement on most RISC 651059191Skris platforms. See crypto/rc4/rc4_enc.c for further details. 651159191Skris [Andy Polyakov] 651259191Skris 651359191Skris *) New -noout option to asn1parse. This causes no output to be produced 651459191Skris its main use is when combined with -strparse and -out to extract data 651559191Skris from a file (which may not be in ASN.1 format). 651659191Skris [Steve Henson] 651759191Skris 651859191Skris *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 651959191Skris when producing the local key id. 652059191Skris [Richard Levitte <levitte@stacken.kth.se>] 652159191Skris 652259191Skris *) New option -dhparam in s_server. This allows a DH parameter file to be 652359191Skris stated explicitly. If it is not stated then it tries the first server 652459191Skris certificate file. The previous behaviour hard coded the filename 652559191Skris "server.pem". 652659191Skris [Steve Henson] 652759191Skris 652859191Skris *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 652959191Skris a public key to be input or output. For example: 653059191Skris openssl rsa -in key.pem -pubout -out pubkey.pem 653159191Skris Also added necessary DSA public key functions to handle this. 653259191Skris [Steve Henson] 653359191Skris 653459191Skris *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 653559191Skris in the message. This was handled by allowing 653659191Skris X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 653759191Skris [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 653859191Skris 653959191Skris *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 654059191Skris to the end of the strings whereas this didn't. This would cause problems 654159191Skris if strings read with d2i_ASN1_bytes() were later modified. 654259191Skris [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 654359191Skris 654459191Skris *) Fix for base64 decode bug. When a base64 bio reads only one line of 654559191Skris data and it contains EOF it will end up returning an error. This is 654659191Skris caused by input 46 bytes long. The cause is due to the way base64 654759191Skris BIOs find the start of base64 encoded data. They do this by trying a 654859191Skris trial decode on each line until they find one that works. When they 654959191Skris do a flag is set and it starts again knowing it can pass all the 655059191Skris data directly through the decoder. Unfortunately it doesn't reset 655159191Skris the context it uses. This means that if EOF is reached an attempt 655259191Skris is made to pass two EOFs through the context and this causes the 655359191Skris resulting error. This can also cause other problems as well. As is 655459191Skris usual with these problems it takes *ages* to find and the fix is 655559191Skris trivial: move one line. 655659191Skris [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 655759191Skris 655859191Skris *) Ugly workaround to get s_client and s_server working under Windows. The 655959191Skris old code wouldn't work because it needed to select() on sockets and the 656059191Skris tty (for keypresses and to see if data could be written). Win32 only 656159191Skris supports select() on sockets so we select() with a 1s timeout on the 656259191Skris sockets and then see if any characters are waiting to be read, if none 656359191Skris are present then we retry, we also assume we can always write data to 656459191Skris the tty. This isn't nice because the code then blocks until we've 656559191Skris received a complete line of data and it is effectively polling the 656659191Skris keyboard at 1s intervals: however it's quite a bit better than not 656759191Skris working at all :-) A dedicated Windows application might handle this 656859191Skris with an event loop for example. 656959191Skris [Steve Henson] 657059191Skris 657159191Skris *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 657259191Skris and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 657359191Skris will be called when RSA_sign() and RSA_verify() are used. This is useful 657459191Skris if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 657559191Skris For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 657659191Skris should *not* be used: RSA_sign() and RSA_verify() must be used instead. 657759191Skris This necessitated the support of an extra signature type NID_md5_sha1 657859191Skris for SSL signatures and modifications to the SSL library to use it instead 657959191Skris of calling RSA_public_decrypt() and RSA_private_encrypt(). 658059191Skris [Steve Henson] 658159191Skris 658259191Skris *) Add new -verify -CAfile and -CApath options to the crl program, these 658359191Skris will lookup a CRL issuers certificate and verify the signature in a 658459191Skris similar way to the verify program. Tidy up the crl program so it 658559191Skris no longer accesses structures directly. Make the ASN1 CRL parsing a bit 658659191Skris less strict. It will now permit CRL extensions even if it is not 658759191Skris a V2 CRL: this will allow it to tolerate some broken CRLs. 658859191Skris [Steve Henson] 658959191Skris 659059191Skris *) Initialize all non-automatic variables each time one of the openssl 659159191Skris sub-programs is started (this is necessary as they may be started 659259191Skris multiple times from the "OpenSSL>" prompt). 659359191Skris [Lennart Bang, Bodo Moeller] 659459191Skris 659559191Skris *) Preliminary compilation option RSA_NULL which disables RSA crypto without 659659191Skris removing all other RSA functionality (this is what NO_RSA does). This 659759191Skris is so (for example) those in the US can disable those operations covered 659859191Skris by the RSA patent while allowing storage and parsing of RSA keys and RSA 659959191Skris key generation. 660059191Skris [Steve Henson] 660159191Skris 660259191Skris *) Non-copying interface to BIO pairs. 660359191Skris (still largely untested) 660459191Skris [Bodo Moeller] 660559191Skris 660659191Skris *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 660759191Skris ASCII string. This was handled independently in various places before. 660859191Skris [Steve Henson] 660959191Skris 661059191Skris *) New functions UTF8_getc() and UTF8_putc() that parse and generate 661159191Skris UTF8 strings a character at a time. 661259191Skris [Steve Henson] 661359191Skris 661459191Skris *) Use client_version from client hello to select the protocol 661559191Skris (s23_srvr.c) and for RSA client key exchange verification 661659191Skris (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 661759191Skris [Bodo Moeller] 661859191Skris 661959191Skris *) Add various utility functions to handle SPKACs, these were previously 662059191Skris handled by poking round in the structure internals. Added new function 662159191Skris NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 662259191Skris print, verify and generate SPKACs. Based on an original idea from 662359191Skris Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 662459191Skris [Steve Henson] 662559191Skris 662659191Skris *) RIPEMD160 is operational on all platforms and is back in 'make test'. 662759191Skris [Andy Polyakov] 662859191Skris 662959191Skris *) Allow the config file extension section to be overwritten on the 663059191Skris command line. Based on an original idea from Massimiliano Pala 663159191Skris <madwolf@comune.modena.it>. The new option is called -extensions 663259191Skris and can be applied to ca, req and x509. Also -reqexts to override 663359191Skris the request extensions in req and -crlexts to override the crl extensions 663459191Skris in ca. 663559191Skris [Steve Henson] 663659191Skris 663759191Skris *) Add new feature to the SPKAC handling in ca. Now you can include 663859191Skris the same field multiple times by preceding it by "XXXX." for example: 663959191Skris 1.OU="Unit name 1" 664059191Skris 2.OU="Unit name 2" 664159191Skris this is the same syntax as used in the req config file. 664259191Skris [Steve Henson] 664359191Skris 664459191Skris *) Allow certificate extensions to be added to certificate requests. These 664559191Skris are specified in a 'req_extensions' option of the req section of the 664659191Skris config file. They can be printed out with the -text option to req but 664759191Skris are otherwise ignored at present. 664859191Skris [Steve Henson] 664959191Skris 665059191Skris *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 665159191Skris data read consists of only the final block it would not decrypted because 665259191Skris EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 665359191Skris A misplaced 'break' also meant the decrypted final block might not be 665459191Skris copied until the next read. 665559191Skris [Steve Henson] 665659191Skris 665759191Skris *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 665859191Skris a few extra parameters to the DH structure: these will be useful if 665959191Skris for example we want the value of 'q' or implement X9.42 DH. 666059191Skris [Steve Henson] 666159191Skris 666259191Skris *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 666359191Skris provides hooks that allow the default DSA functions or functions on a 666459191Skris "per key" basis to be replaced. This allows hardware acceleration and 666559191Skris hardware key storage to be handled without major modification to the 666659191Skris library. Also added low level modexp hooks and CRYPTO_EX structure and 666759191Skris associated functions. 666859191Skris [Steve Henson] 666959191Skris 667059191Skris *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 667159191Skris as "read only": it can't be written to and the buffer it points to will 667259191Skris not be freed. Reading from a read only BIO is much more efficient than 667359191Skris a normal memory BIO. This was added because there are several times when 667459191Skris an area of memory needs to be read from a BIO. The previous method was 667559191Skris to create a memory BIO and write the data to it, this results in two 667659191Skris copies of the data and an O(n^2) reading algorithm. There is a new 667759191Skris function BIO_new_mem_buf() which creates a read only memory BIO from 667859191Skris an area of memory. Also modified the PKCS#7 routines to use read only 667959191Skris memory BIOs. 668059191Skris [Steve Henson] 668159191Skris 668259191Skris *) Bugfix: ssl23_get_client_hello did not work properly when called in 668359191Skris state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 668459191Skris a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 668559191Skris but a retry condition occured while trying to read the rest. 668659191Skris [Bodo Moeller] 668759191Skris 668859191Skris *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 668959191Skris NID_pkcs7_encrypted by default: this was wrong since this should almost 669059191Skris always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 669159191Skris the encrypted data type: this is a more sensible place to put it and it 669259191Skris allows the PKCS#12 code to be tidied up that duplicated this 669359191Skris functionality. 669459191Skris [Steve Henson] 669559191Skris 669659191Skris *) Changed obj_dat.pl script so it takes its input and output files on 669759191Skris the command line. This should avoid shell escape redirection problems 669859191Skris under Win32. 669959191Skris [Steve Henson] 670059191Skris 670159191Skris *) Initial support for certificate extension requests, these are included 670259191Skris in things like Xenroll certificate requests. Included functions to allow 670359191Skris extensions to be obtained and added. 670459191Skris [Steve Henson] 670559191Skris 670659191Skris *) -crlf option to s_client and s_server for sending newlines as 670759191Skris CRLF (as required by many protocols). 670859191Skris [Bodo Moeller] 670959191Skris 671055714Skris Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 671155714Skris 671255714Skris *) Install libRSAglue.a when OpenSSL is built with RSAref. 671355714Skris [Ralf S. Engelschall] 671455714Skris 671555714Skris *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 671655714Skris [Andrija Antonijevic <TheAntony2@bigfoot.com>] 671755714Skris 671855714Skris *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 671955714Skris program. 672055714Skris [Steve Henson] 672155714Skris 672255714Skris *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 672355714Skris DH parameters/keys (q is lost during that conversion, but the resulting 672455714Skris DH parameters contain its length). 672555714Skris 672655714Skris For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 672755714Skris much faster than DH_generate_parameters (which creates parameters 672855714Skris where p = 2*q + 1), and also the smaller q makes DH computations 672955714Skris much more efficient (160-bit exponentiation instead of 1024-bit 673055714Skris exponentiation); so this provides a convenient way to support DHE 673155714Skris ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 673255714Skris utter importance to use 673355714Skris SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 673455714Skris or 673555714Skris SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 673655714Skris when such DH parameters are used, because otherwise small subgroup 673755714Skris attacks may become possible! 673855714Skris [Bodo Moeller] 673955714Skris 674055714Skris *) Avoid memory leak in i2d_DHparams. 674155714Skris [Bodo Moeller] 674255714Skris 674355714Skris *) Allow the -k option to be used more than once in the enc program: 674455714Skris this allows the same encrypted message to be read by multiple recipients. 674555714Skris [Steve Henson] 674655714Skris 674755714Skris *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 674855714Skris an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 674955714Skris it will always use the numerical form of the OID, even if it has a short 675055714Skris or long name. 675155714Skris [Steve Henson] 675255714Skris 675355714Skris *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 675455714Skris method only got called if p,q,dmp1,dmq1,iqmp components were present, 675555714Skris otherwise bn_mod_exp was called. In the case of hardware keys for example 675655714Skris no private key components need be present and it might store extra data 675759191Skris in the RSA structure, which cannot be accessed from bn_mod_exp. 675859191Skris By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 675959191Skris private key operations. 676055714Skris [Steve Henson] 676155714Skris 676255714Skris *) Added support for SPARC Linux. 676355714Skris [Andy Polyakov] 676455714Skris 676555714Skris *) pem_password_cb function type incompatibly changed from 676655714Skris typedef int pem_password_cb(char *buf, int size, int rwflag); 676755714Skris to 676855714Skris ....(char *buf, int size, int rwflag, void *userdata); 676955714Skris so that applications can pass data to their callbacks: 677055714Skris The PEM[_ASN1]_{read,write}... functions and macros now take an 677155714Skris additional void * argument, which is just handed through whenever 677255714Skris the password callback is called. 677359191Skris [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 677455714Skris 677555714Skris New function SSL_CTX_set_default_passwd_cb_userdata. 677655714Skris 677755714Skris Compatibility note: As many C implementations push function arguments 677855714Skris onto the stack in reverse order, the new library version is likely to 677955714Skris interoperate with programs that have been compiled with the old 678055714Skris pem_password_cb definition (PEM_whatever takes some data that 678155714Skris happens to be on the stack as its last argument, and the callback 678255714Skris just ignores this garbage); but there is no guarantee whatsoever that 678355714Skris this will work. 678455714Skris 678555714Skris *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 678655714Skris (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 678755714Skris problems not only on Windows, but also on some Unix platforms. 678855714Skris To avoid problematic command lines, these definitions are now in an 678955714Skris auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 679055714Skris for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 679155714Skris [Bodo Moeller] 679255714Skris 679355714Skris *) MIPS III/IV assembler module is reimplemented. 679455714Skris [Andy Polyakov] 679555714Skris 679655714Skris *) More DES library cleanups: remove references to srand/rand and 679755714Skris delete an unused file. 679855714Skris [Ulf M�ller] 679955714Skris 680055714Skris *) Add support for the the free Netwide assembler (NASM) under Win32, 680155714Skris since not many people have MASM (ml) and it can be hard to obtain. 680255714Skris This is currently experimental but it seems to work OK and pass all 680355714Skris the tests. Check out INSTALL.W32 for info. 680455714Skris [Steve Henson] 680555714Skris 680655714Skris *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 680755714Skris without temporary keys kept an extra copy of the server key, 680855714Skris and connections with temporary keys did not free everything in case 680955714Skris of an error. 681055714Skris [Bodo Moeller] 681155714Skris 681255714Skris *) New function RSA_check_key and new openssl rsa option -check 681355714Skris for verifying the consistency of RSA keys. 681455714Skris [Ulf Moeller, Bodo Moeller] 681555714Skris 681655714Skris *) Various changes to make Win32 compile work: 681755714Skris 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 681855714Skris 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 681955714Skris comparison" warnings. 682055714Skris 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 682155714Skris [Steve Henson] 682255714Skris 682355714Skris *) Add a debugging option to PKCS#5 v2 key generation function: when 682455714Skris you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 682555714Skris derived keys are printed to stderr. 682655714Skris [Steve Henson] 682755714Skris 682855714Skris *) Copy the flags in ASN1_STRING_dup(). 682955714Skris [Roman E. Pavlov <pre@mo.msk.ru>] 683055714Skris 683155714Skris *) The x509 application mishandled signing requests containing DSA 683255714Skris keys when the signing key was also DSA and the parameters didn't match. 683355714Skris 683455714Skris It was supposed to omit the parameters when they matched the signing key: 683555714Skris the verifying software was then supposed to automatically use the CA's 683655714Skris parameters if they were absent from the end user certificate. 683755714Skris 683855714Skris Omitting parameters is no longer recommended. The test was also 683955714Skris the wrong way round! This was probably due to unusual behaviour in 684055714Skris EVP_cmp_parameters() which returns 1 if the parameters match. 684155714Skris This meant that parameters were omitted when they *didn't* match and 684255714Skris the certificate was useless. Certificates signed with 'ca' didn't have 684355714Skris this bug. 684455714Skris [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 684555714Skris 684655714Skris *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 684755714Skris The interface is as follows: 684855714Skris Applications can use 684955714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 685055714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 685155714Skris "off" is now the default. 685255714Skris The library internally uses 685355714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 685455714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 685555714Skris to disable memory-checking temporarily. 685655714Skris 685755714Skris Some inconsistent states that previously were possible (and were 685855714Skris even the default) are now avoided. 685955714Skris 686055714Skris -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 686155714Skris with each memory chunk allocated; this is occasionally more helpful 686255714Skris than just having a counter. 686355714Skris 686455714Skris -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 686555714Skris 686655714Skris -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 686755714Skris extensions. 686855714Skris [Bodo Moeller] 686955714Skris 687055714Skris *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 687155714Skris which largely parallels "options", but is for changing API behaviour, 687255714Skris whereas "options" are about protocol behaviour. 687355714Skris Initial "mode" flags are: 687455714Skris 687555714Skris SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 687655714Skris a single record has been written. 687755714Skris SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 687855714Skris retries use the same buffer location. 687955714Skris (But all of the contents must be 688055714Skris copied!) 688155714Skris [Bodo Moeller] 688255714Skris 688379998Skris *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 688455714Skris worked. 688555714Skris 688655714Skris *) Fix problems with no-hmac etc. 688755714Skris [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 688855714Skris 688955714Skris *) New functions RSA_get_default_method(), RSA_set_method() and 689055714Skris RSA_get_method(). These allows replacement of RSA_METHODs without having 689155714Skris to mess around with the internals of an RSA structure. 689255714Skris [Steve Henson] 689355714Skris 689455714Skris *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 689555714Skris Also really enable memory leak checks in openssl.c and in some 689655714Skris test programs. 689755714Skris [Chad C. Mulligan, Bodo Moeller] 689855714Skris 689955714Skris *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 690055714Skris up the length of negative integers. This has now been simplified to just 690155714Skris store the length when it is first determined and use it later, rather 690255714Skris than trying to keep track of where data is copied and updating it to 690355714Skris point to the end. 690455714Skris [Steve Henson, reported by Brien Wheeler 690555714Skris <bwheeler@authentica-security.com>] 690655714Skris 690755714Skris *) Add a new function PKCS7_signatureVerify. This allows the verification 690855714Skris of a PKCS#7 signature but with the signing certificate passed to the 690955714Skris function itself. This contrasts with PKCS7_dataVerify which assumes the 691055714Skris certificate is present in the PKCS#7 structure. This isn't always the 691155714Skris case: certificates can be omitted from a PKCS#7 structure and be 691255714Skris distributed by "out of band" means (such as a certificate database). 691355714Skris [Steve Henson] 691455714Skris 691555714Skris *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 691655714Skris function prototypes in pem.h, also change util/mkdef.pl to add the 691755714Skris necessary function names. 691855714Skris [Steve Henson] 691955714Skris 692055714Skris *) mk1mf.pl (used by Windows builds) did not properly read the 692155714Skris options set by Configure in the top level Makefile, and Configure 692255714Skris was not even able to write more than one option correctly. 692355714Skris Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 692455714Skris [Bodo Moeller] 692555714Skris 692655714Skris *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 692755714Skris file to be loaded from a BIO or FILE pointer. The BIO version will 692855714Skris for example allow memory BIOs to contain config info. 692955714Skris [Steve Henson] 693055714Skris 693155714Skris *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 693255714Skris Whoever hopes to achieve shared-library compatibility across versions 693355714Skris must use this, not the compile-time macro. 693455714Skris (Exercise 0.9.4: Which is the minimum library version required by 693555714Skris such programs?) 693655714Skris Note: All this applies only to multi-threaded programs, others don't 693755714Skris need locks. 693855714Skris [Bodo Moeller] 693955714Skris 694055714Skris *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 694155714Skris through a BIO pair triggered the default case, i.e. 694255714Skris SSLerr(...,SSL_R_UNKNOWN_STATE). 694355714Skris [Bodo Moeller] 694455714Skris 694555714Skris *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 694655714Skris can use the SSL library even if none of the specific BIOs is 694755714Skris appropriate. 694855714Skris [Bodo Moeller] 694955714Skris 695055714Skris *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 695155714Skris for the encoded length. 695255714Skris [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 695355714Skris 695455714Skris *) Add initial documentation of the X509V3 functions. 695555714Skris [Steve Henson] 695655714Skris 695755714Skris *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 695855714Skris PEM_write_bio_PKCS8PrivateKey() that are equivalent to 695955714Skris PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 696055714Skris secure PKCS#8 private key format with a high iteration count. 696155714Skris [Steve Henson] 696255714Skris 696355714Skris *) Fix determination of Perl interpreter: A perl or perl5 696455714Skris _directory_ in $PATH was also accepted as the interpreter. 696555714Skris [Ralf S. Engelschall] 696655714Skris 696755714Skris *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 696855714Skris wrong with it but it was very old and did things like calling 696955714Skris PEM_ASN1_read() directly and used MD5 for the hash not to mention some 697055714Skris unusual formatting. 697155714Skris [Steve Henson] 697255714Skris 697355714Skris *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 697455714Skris to use the new extension code. 697555714Skris [Steve Henson] 697655714Skris 697755714Skris *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 697855714Skris with macros. This should make it easier to change their form, add extra 697955714Skris arguments etc. Fix a few PEM prototypes which didn't have cipher as a 698055714Skris constant. 698155714Skris [Steve Henson] 698255714Skris 698355714Skris *) Add to configuration table a new entry that can specify an alternative 698455714Skris name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 698555714Skris according to Mark Crispin <MRC@Panda.COM>. 698655714Skris [Bodo Moeller] 698755714Skris 698855714Skris#if 0 698955714Skris *) DES CBC did not update the IV. Weird. 699055714Skris [Ben Laurie] 699155714Skris#else 699255714Skris des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 699355714Skris Changing the behaviour of the former might break existing programs -- 699455714Skris where IV updating is needed, des_ncbc_encrypt can be used. 699555714Skris#endif 699655714Skris 699755714Skris *) When bntest is run from "make test" it drives bc to check its 699855714Skris calculations, as well as internally checking them. If an internal check 699955714Skris fails, it needs to cause bc to give a non-zero result or make test carries 700055714Skris on without noticing the failure. Fixed. 700155714Skris [Ben Laurie] 700255714Skris 700355714Skris *) DES library cleanups. 700455714Skris [Ulf M�ller] 700555714Skris 700655714Skris *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 700755714Skris used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 700855714Skris ciphers. NOTE: although the key derivation function has been verified 700955714Skris against some published test vectors it has not been extensively tested 701055714Skris yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 701155714Skris of v2.0. 701255714Skris [Steve Henson] 701355714Skris 701455714Skris *) Instead of "mkdir -p", which is not fully portable, use new 701555714Skris Perl script "util/mkdir-p.pl". 701655714Skris [Bodo Moeller] 701755714Skris 701855714Skris *) Rewrite the way password based encryption (PBE) is handled. It used to 701955714Skris assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 702055714Skris structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 702155714Skris but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 702255714Skris the 'parameter' field of the AlgorithmIdentifier is passed to the 702355714Skris underlying key generation function so it must do its own ASN1 parsing. 702455714Skris This has also changed the EVP_PBE_CipherInit() function which now has a 702555714Skris 'parameter' argument instead of literal salt and iteration count values 702655714Skris and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 702755714Skris [Steve Henson] 702855714Skris 702955714Skris *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 703055714Skris and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 703155714Skris Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 703255714Skris KEY" because this clashed with PKCS#8 unencrypted string. Since this 703355714Skris value was just used as a "magic string" and not used directly its 703455714Skris value doesn't matter. 703555714Skris [Steve Henson] 703655714Skris 703755714Skris *) Introduce some semblance of const correctness to BN. Shame C doesn't 703855714Skris support mutable. 703955714Skris [Ben Laurie] 704055714Skris 704155714Skris *) "linux-sparc64" configuration (ultrapenguin). 704255714Skris [Ray Miller <ray.miller@oucs.ox.ac.uk>] 704355714Skris "linux-sparc" configuration. 704455714Skris [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 704555714Skris 704655714Skris *) config now generates no-xxx options for missing ciphers. 704755714Skris [Ulf M�ller] 704855714Skris 704955714Skris *) Support the EBCDIC character set (work in progress). 705055714Skris File ebcdic.c not yet included because it has a different license. 705155714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 705255714Skris 705355714Skris *) Support BS2000/OSD-POSIX. 705455714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 705555714Skris 705655714Skris *) Make callbacks for key generation use void * instead of char *. 705755714Skris [Ben Laurie] 705855714Skris 705955714Skris *) Make S/MIME samples compile (not yet tested). 706055714Skris [Ben Laurie] 706155714Skris 706255714Skris *) Additional typesafe stacks. 706355714Skris [Ben Laurie] 706455714Skris 706555714Skris *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 706655714Skris [Bodo Moeller] 706755714Skris 706855714Skris 706955714Skris Changes between 0.9.3 and 0.9.3a [29 May 1999] 707055714Skris 707155714Skris *) New configuration variant "sco5-gcc". 707255714Skris 707355714Skris *) Updated some demos. 707455714Skris [Sean O Riordain, Wade Scholine] 707555714Skris 707655714Skris *) Add missing BIO_free at exit of pkcs12 application. 707755714Skris [Wu Zhigang] 707855714Skris 707955714Skris *) Fix memory leak in conf.c. 708055714Skris [Steve Henson] 708155714Skris 708255714Skris *) Updates for Win32 to assembler version of MD5. 708355714Skris [Steve Henson] 708455714Skris 708555714Skris *) Set #! path to perl in apps/der_chop to where we found it 708655714Skris instead of using a fixed path. 708755714Skris [Bodo Moeller] 708855714Skris 708955714Skris *) SHA library changes for irix64-mips4-cc. 709055714Skris [Andy Polyakov] 709155714Skris 709255714Skris *) Improvements for VMS support. 709355714Skris [Richard Levitte] 709455714Skris 709555714Skris 709655714Skris Changes between 0.9.2b and 0.9.3 [24 May 1999] 709755714Skris 709855714Skris *) Bignum library bug fix. IRIX 6 passes "make test" now! 709955714Skris This also avoids the problems with SC4.2 and unpatched SC5. 710055714Skris [Andy Polyakov <appro@fy.chalmers.se>] 710155714Skris 710255714Skris *) New functions sk_num, sk_value and sk_set to replace the previous macros. 710355714Skris These are required because of the typesafe stack would otherwise break 710455714Skris existing code. If old code used a structure member which used to be STACK 710555714Skris and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 710655714Skris sk_num or sk_value it would produce an error because the num, data members 710755714Skris are not present in STACK_OF. Now it just produces a warning. sk_set 710855714Skris replaces the old method of assigning a value to sk_value 710955714Skris (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 711055714Skris that does this will no longer work (and should use sk_set instead) but 711155714Skris this could be regarded as a "questionable" behaviour anyway. 711255714Skris [Steve Henson] 711355714Skris 711455714Skris *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 711555714Skris correctly handle encrypted S/MIME data. 711655714Skris [Steve Henson] 711755714Skris 711855714Skris *) Change type of various DES function arguments from des_cblock 711955714Skris (which means, in function argument declarations, pointer to char) 712055714Skris to des_cblock * (meaning pointer to array with 8 char elements), 712155714Skris which allows the compiler to do more typechecking; it was like 712255714Skris that back in SSLeay, but with lots of ugly casts. 712355714Skris 712455714Skris Introduce new type const_des_cblock. 712555714Skris [Bodo Moeller] 712655714Skris 712755714Skris *) Reorganise the PKCS#7 library and get rid of some of the more obvious 712855714Skris problems: find RecipientInfo structure that matches recipient certificate 712955714Skris and initialise the ASN1 structures properly based on passed cipher. 713055714Skris [Steve Henson] 713155714Skris 713255714Skris *) Belatedly make the BN tests actually check the results. 713355714Skris [Ben Laurie] 713455714Skris 713555714Skris *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 713655714Skris to and from BNs: it was completely broken. New compilation option 713755714Skris NEG_PUBKEY_BUG to allow for some broken certificates that encode public 713855714Skris key elements as negative integers. 713955714Skris [Steve Henson] 714055714Skris 714155714Skris *) Reorganize and speed up MD5. 714255714Skris [Andy Polyakov <appro@fy.chalmers.se>] 714355714Skris 714455714Skris *) VMS support. 714555714Skris [Richard Levitte <richard@levitte.org>] 714655714Skris 714755714Skris *) New option -out to asn1parse to allow the parsed structure to be 714855714Skris output to a file. This is most useful when combined with the -strparse 714955714Skris option to examine the output of things like OCTET STRINGS. 715055714Skris [Steve Henson] 715155714Skris 715255714Skris *) Make SSL library a little more fool-proof by not requiring any longer 715355714Skris that SSL_set_{accept,connect}_state be called before 715455714Skris SSL_{accept,connect} may be used (SSL_set_..._state is omitted 715555714Skris in many applications because usually everything *appeared* to work as 715655714Skris intended anyway -- now it really works as intended). 715755714Skris [Bodo Moeller] 715855714Skris 715955714Skris *) Move openssl.cnf out of lib/. 716055714Skris [Ulf M�ller] 716155714Skris 716255714Skris *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 716355714Skris -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 716455714Skris -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 716555714Skris [Ralf S. Engelschall] 716655714Skris 716755714Skris *) Various fixes to the EVP and PKCS#7 code. It may now be able to 716855714Skris handle PKCS#7 enveloped data properly. 716955714Skris [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 717055714Skris 717155714Skris *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 717255714Skris copying pointers. The cert_st handling is changed by this in 717355714Skris various ways (and thus what used to be known as ctx->default_cert 717455714Skris is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 717555714Skris any longer when s->cert does not give us what we need). 717655714Skris ssl_cert_instantiate becomes obsolete by this change. 717755714Skris As soon as we've got the new code right (possibly it already is?), 717855714Skris we have solved a couple of bugs of the earlier code where s->cert 717955714Skris was used as if it could not have been shared with other SSL structures. 718055714Skris 718155714Skris Note that using the SSL API in certain dirty ways now will result 718255714Skris in different behaviour than observed with earlier library versions: 718355714Skris Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 718455714Skris does not influence s as it used to. 718555714Skris 718655714Skris In order to clean up things more thoroughly, inside SSL_SESSION 718755714Skris we don't use CERT any longer, but a new structure SESS_CERT 718855714Skris that holds per-session data (if available); currently, this is 718955714Skris the peer's certificate chain and, for clients, the server's certificate 719055714Skris and temporary key. CERT holds only those values that can have 719155714Skris meaningful defaults in an SSL_CTX. 719255714Skris [Bodo Moeller] 719355714Skris 719455714Skris *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 719555714Skris from the internal representation. Various PKCS#7 fixes: remove some 719655714Skris evil casts and set the enc_dig_alg field properly based on the signing 719755714Skris key type. 719855714Skris [Steve Henson] 719955714Skris 720055714Skris *) Allow PKCS#12 password to be set from the command line or the 720155714Skris environment. Let 'ca' get its config file name from the environment 720255714Skris variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 720355714Skris and 'x509'). 720455714Skris [Steve Henson] 720555714Skris 720655714Skris *) Allow certificate policies extension to use an IA5STRING for the 720755714Skris organization field. This is contrary to the PKIX definition but 720855714Skris VeriSign uses it and IE5 only recognises this form. Document 'x509' 720955714Skris extension option. 721055714Skris [Steve Henson] 721155714Skris 721255714Skris *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 721355714Skris without disallowing inline assembler and the like for non-pedantic builds. 721455714Skris [Ben Laurie] 721555714Skris 721655714Skris *) Support Borland C++ builder. 721755714Skris [Janez Jere <jj@void.si>, modified by Ulf M�ller] 721855714Skris 721955714Skris *) Support Mingw32. 722055714Skris [Ulf M�ller] 722155714Skris 722255714Skris *) SHA-1 cleanups and performance enhancements. 722355714Skris [Andy Polyakov <appro@fy.chalmers.se>] 722455714Skris 722555714Skris *) Sparc v8plus assembler for the bignum library. 722655714Skris [Andy Polyakov <appro@fy.chalmers.se>] 722755714Skris 722855714Skris *) Accept any -xxx and +xxx compiler options in Configure. 722955714Skris [Ulf M�ller] 723055714Skris 723155714Skris *) Update HPUX configuration. 723255714Skris [Anonymous] 723355714Skris 723455714Skris *) Add missing sk_<type>_unshift() function to safestack.h 723555714Skris [Ralf S. Engelschall] 723655714Skris 723755714Skris *) New function SSL_CTX_use_certificate_chain_file that sets the 723855714Skris "extra_cert"s in addition to the certificate. (This makes sense 723955714Skris only for "PEM" format files, as chains as a whole are not 724055714Skris DER-encoded.) 724155714Skris [Bodo Moeller] 724255714Skris 724355714Skris *) Support verify_depth from the SSL API. 724455714Skris x509_vfy.c had what can be considered an off-by-one-error: 724555714Skris Its depth (which was not part of the external interface) 724655714Skris was actually counting the number of certificates in a chain; 724755714Skris now it really counts the depth. 724855714Skris [Bodo Moeller] 724955714Skris 725055714Skris *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 725155714Skris instead of X509err, which often resulted in confusing error 725255714Skris messages since the error codes are not globally unique 725355714Skris (e.g. an alleged error in ssl3_accept when a certificate 725455714Skris didn't match the private key). 725555714Skris 725655714Skris *) New function SSL_CTX_set_session_id_context that allows to set a default 725755714Skris value (so that you don't need SSL_set_session_id_context for each 725855714Skris connection using the SSL_CTX). 725955714Skris [Bodo Moeller] 726055714Skris 726155714Skris *) OAEP decoding bug fix. 726255714Skris [Ulf M�ller] 726355714Skris 726455714Skris *) Support INSTALL_PREFIX for package builders, as proposed by 726555714Skris David Harris. 726655714Skris [Bodo Moeller] 726755714Skris 726855714Skris *) New Configure options "threads" and "no-threads". For systems 726955714Skris where the proper compiler options are known (currently Solaris 727055714Skris and Linux), "threads" is the default. 727155714Skris [Bodo Moeller] 727255714Skris 727355714Skris *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 727455714Skris [Bodo Moeller] 727555714Skris 727655714Skris *) Install various scripts to $(OPENSSLDIR)/misc, not to 727755714Skris $(INSTALLTOP)/bin -- they shouldn't clutter directories 727855714Skris such as /usr/local/bin. 727955714Skris [Bodo Moeller] 728055714Skris 728155714Skris *) "make linux-shared" to build shared libraries. 728255714Skris [Niels Poppe <niels@netbox.org>] 728355714Skris 728455714Skris *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 728555714Skris [Ulf M�ller] 728655714Skris 728755714Skris *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 728855714Skris extension adding in x509 utility. 728955714Skris [Steve Henson] 729055714Skris 729155714Skris *) Remove NOPROTO sections and error code comments. 729255714Skris [Ulf M�ller] 729355714Skris 729455714Skris *) Partial rewrite of the DEF file generator to now parse the ANSI 729555714Skris prototypes. 729655714Skris [Steve Henson] 729755714Skris 729855714Skris *) New Configure options --prefix=DIR and --openssldir=DIR. 729955714Skris [Ulf M�ller] 730055714Skris 730155714Skris *) Complete rewrite of the error code script(s). It is all now handled 730255714Skris by one script at the top level which handles error code gathering, 730355714Skris header rewriting and C source file generation. It should be much better 730455714Skris than the old method: it now uses a modified version of Ulf's parser to 730555714Skris read the ANSI prototypes in all header files (thus the old K&R definitions 730655714Skris aren't needed for error creation any more) and do a better job of 730755714Skris translating function codes into names. The old 'ASN1 error code imbedded 730855714Skris in a comment' is no longer necessary and it doesn't use .err files which 730955714Skris have now been deleted. Also the error code call doesn't have to appear all 731055714Skris on one line (which resulted in some large lines...). 731155714Skris [Steve Henson] 731255714Skris 731355714Skris *) Change #include filenames from <foo.h> to <openssl/foo.h>. 731455714Skris [Bodo Moeller] 731555714Skris 731655714Skris *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 731755714Skris 0 (which usually indicates a closed connection), but continue reading. 731855714Skris [Bodo Moeller] 731955714Skris 732055714Skris *) Fix some race conditions. 732155714Skris [Bodo Moeller] 732255714Skris 732355714Skris *) Add support for CRL distribution points extension. Add Certificate 732455714Skris Policies and CRL distribution points documentation. 732555714Skris [Steve Henson] 732655714Skris 732755714Skris *) Move the autogenerated header file parts to crypto/opensslconf.h. 732855714Skris [Ulf M�ller] 732955714Skris 733055714Skris *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 733155714Skris 8 of keying material. Merlin has also confirmed interop with this fix 733255714Skris between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 733355714Skris [Merlin Hughes <merlin@baltimore.ie>] 733455714Skris 733555714Skris *) Fix lots of warnings. 733655714Skris [Richard Levitte <levitte@stacken.kth.se>] 733755714Skris 733855714Skris *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 733955714Skris the directory spec didn't end with a LIST_SEPARATOR_CHAR. 734055714Skris [Richard Levitte <levitte@stacken.kth.se>] 734155714Skris 734255714Skris *) Fix problems with sizeof(long) == 8. 734355714Skris [Andy Polyakov <appro@fy.chalmers.se>] 734455714Skris 734555714Skris *) Change functions to ANSI C. 734655714Skris [Ulf M�ller] 734755714Skris 734855714Skris *) Fix typos in error codes. 734955714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 735055714Skris 735155714Skris *) Remove defunct assembler files from Configure. 735255714Skris [Ulf M�ller] 735355714Skris 735455714Skris *) SPARC v8 assembler BIGNUM implementation. 735555714Skris [Andy Polyakov <appro@fy.chalmers.se>] 735655714Skris 735755714Skris *) Support for Certificate Policies extension: both print and set. 735855714Skris Various additions to support the r2i method this uses. 735955714Skris [Steve Henson] 736055714Skris 736155714Skris *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 736255714Skris return a const string when you are expecting an allocated buffer. 736355714Skris [Ben Laurie] 736455714Skris 736555714Skris *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 736655714Skris types DirectoryString and DisplayText. 736755714Skris [Steve Henson] 736855714Skris 736955714Skris *) Add code to allow r2i extensions to access the configuration database, 737055714Skris add an LHASH database driver and add several ctx helper functions. 737155714Skris [Steve Henson] 737255714Skris 737355714Skris *) Fix an evil bug in bn_expand2() which caused various BN functions to 737455714Skris fail when they extended the size of a BIGNUM. 737555714Skris [Steve Henson] 737655714Skris 737755714Skris *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 737855714Skris support typesafe stack. 737955714Skris [Steve Henson] 738055714Skris 738155714Skris *) Fix typo in SSL_[gs]et_options(). 738255714Skris [Nils Frostberg <nils@medcom.se>] 738355714Skris 738455714Skris *) Delete various functions and files that belonged to the (now obsolete) 738555714Skris old X509V3 handling code. 738655714Skris [Steve Henson] 738755714Skris 738855714Skris *) New Configure option "rsaref". 738955714Skris [Ulf M�ller] 739055714Skris 739155714Skris *) Don't auto-generate pem.h. 739255714Skris [Bodo Moeller] 739355714Skris 739455714Skris *) Introduce type-safe ASN.1 SETs. 739555714Skris [Ben Laurie] 739655714Skris 739755714Skris *) Convert various additional casted stacks to type-safe STACK_OF() variants. 739855714Skris [Ben Laurie, Ralf S. Engelschall, Steve Henson] 739955714Skris 740055714Skris *) Introduce type-safe STACKs. This will almost certainly break lots of code 740155714Skris that links with OpenSSL (well at least cause lots of warnings), but fear 740255714Skris not: the conversion is trivial, and it eliminates loads of evil casts. A 740355714Skris few STACKed things have been converted already. Feel free to convert more. 740455714Skris In the fullness of time, I'll do away with the STACK type altogether. 740555714Skris [Ben Laurie] 740655714Skris 740755714Skris *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 740855714Skris specified in <certfile> by updating the entry in the index.txt file. 740955714Skris This way one no longer has to edit the index.txt file manually for 741055714Skris revoking a certificate. The -revoke option does the gory details now. 741155714Skris [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 741255714Skris 741355714Skris *) Fix `openssl crl -noout -text' combination where `-noout' killed the 741455714Skris `-text' option at all and this way the `-noout -text' combination was 741555714Skris inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 741655714Skris [Ralf S. Engelschall] 741755714Skris 741855714Skris *) Make sure a corresponding plain text error message exists for the 741955714Skris X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 742055714Skris verify callback function determined that a certificate was revoked. 742155714Skris [Ralf S. Engelschall] 742255714Skris 742355714Skris *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 742455714Skris ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 742555714Skris all available cipers including rc5, which was forgotten until now. 742655714Skris In order to let the testing shell script know which algorithms 742755714Skris are available, a new (up to now undocumented) command 742855714Skris "openssl list-cipher-commands" is used. 742955714Skris [Bodo Moeller] 743055714Skris 743155714Skris *) Bugfix: s_client occasionally would sleep in select() when 743255714Skris it should have checked SSL_pending() first. 743355714Skris [Bodo Moeller] 743455714Skris 743555714Skris *) New functions DSA_do_sign and DSA_do_verify to provide access to 743655714Skris the raw DSA values prior to ASN.1 encoding. 743755714Skris [Ulf M�ller] 743855714Skris 743955714Skris *) Tweaks to Configure 744055714Skris [Niels Poppe <niels@netbox.org>] 744155714Skris 744255714Skris *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 744355714Skris yet... 744455714Skris [Steve Henson] 744555714Skris 744655714Skris *) New variables $(RANLIB) and $(PERL) in the Makefiles. 744755714Skris [Ulf M�ller] 744855714Skris 744955714Skris *) New config option to avoid instructions that are illegal on the 80386. 745055714Skris The default code is faster, but requires at least a 486. 745155714Skris [Ulf M�ller] 745255714Skris 745355714Skris *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 745455714Skris SSL2_SERVER_VERSION (not used at all) macros, which are now the 745555714Skris same as SSL2_VERSION anyway. 745655714Skris [Bodo Moeller] 745755714Skris 745855714Skris *) New "-showcerts" option for s_client. 745955714Skris [Bodo Moeller] 746055714Skris 746155714Skris *) Still more PKCS#12 integration. Add pkcs12 application to openssl 746255714Skris application. Various cleanups and fixes. 746355714Skris [Steve Henson] 746455714Skris 746555714Skris *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 746655714Skris modify error routines to work internally. Add error codes and PBE init 746755714Skris to library startup routines. 746855714Skris [Steve Henson] 746955714Skris 747055714Skris *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 747155714Skris packing functions to asn1 and evp. Changed function names and error 747255714Skris codes along the way. 747355714Skris [Steve Henson] 747455714Skris 747555714Skris *) PKCS12 integration: and so it begins... First of several patches to 747655714Skris slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 747755714Skris objects to objects.h 747855714Skris [Steve Henson] 747955714Skris 748055714Skris *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 748155714Skris and display support for Thawte strong extranet extension. 748255714Skris [Steve Henson] 748355714Skris 748455714Skris *) Add LinuxPPC support. 748555714Skris [Jeff Dubrule <igor@pobox.org>] 748655714Skris 748755714Skris *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 748855714Skris bn_div_words in alpha.s. 748955714Skris [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 749055714Skris 749155714Skris *) Make sure the RSA OAEP test is skipped under -DRSAref because 749255714Skris OAEP isn't supported when OpenSSL is built with RSAref. 749355714Skris [Ulf Moeller <ulf@fitug.de>] 749455714Skris 749555714Skris *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 749655714Skris so they no longer are missing under -DNOPROTO. 749755714Skris [Soren S. Jorvang <soren@t.dk>] 749855714Skris 749955714Skris 750055714Skris Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 750155714Skris 750255714Skris *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 750355714Skris doesn't work when the session is reused. Coming soon! 750455714Skris [Ben Laurie] 750555714Skris 750655714Skris *) Fix a security hole, that allows sessions to be reused in the wrong 750755714Skris context thus bypassing client cert protection! All software that uses 750855714Skris client certs and session caches in multiple contexts NEEDS PATCHING to 750955714Skris allow session reuse! A fuller solution is in the works. 751055714Skris [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 751155714Skris 751255714Skris *) Some more source tree cleanups (removed obsolete files 751355714Skris crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 751455714Skris permission on "config" script to be executable) and a fix for the INSTALL 751555714Skris document. 751655714Skris [Ulf Moeller <ulf@fitug.de>] 751755714Skris 751855714Skris *) Remove some legacy and erroneous uses of malloc, free instead of 751955714Skris Malloc, Free. 752055714Skris [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 752155714Skris 752255714Skris *) Make rsa_oaep_test return non-zero on error. 752355714Skris [Ulf Moeller <ulf@fitug.de>] 752455714Skris 752555714Skris *) Add support for native Solaris shared libraries. Configure 752655714Skris solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 752755714Skris if someone would make that last step automatic. 752855714Skris [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 752955714Skris 753055714Skris *) ctx_size was not built with the right compiler during "make links". Fixed. 753155714Skris [Ben Laurie] 753255714Skris 753355714Skris *) Change the meaning of 'ALL' in the cipher list. It now means "everything 753455714Skris except NULL ciphers". This means the default cipher list will no longer 753555714Skris enable NULL ciphers. They need to be specifically enabled e.g. with 753655714Skris the string "DEFAULT:eNULL". 753755714Skris [Steve Henson] 753855714Skris 753955714Skris *) Fix to RSA private encryption routines: if p < q then it would 754055714Skris occasionally produce an invalid result. This will only happen with 754155714Skris externally generated keys because OpenSSL (and SSLeay) ensure p > q. 754255714Skris [Steve Henson] 754355714Skris 754455714Skris *) Be less restrictive and allow also `perl util/perlpath.pl 754555714Skris /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 754655714Skris because this way one can also use an interpreter named `perl5' (which is 754755714Skris usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 754855714Skris installed as `perl'). 754955714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 755055714Skris 755155714Skris *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 755255714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 755355714Skris 755455714Skris *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 755555714Skris advapi32.lib to Win32 build and change the pem test comparision 755655714Skris to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 755755714Skris suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 755855714Skris and crypto/des/ede_cbcm_enc.c. 755955714Skris [Steve Henson] 756055714Skris 756155714Skris *) DES quad checksum was broken on big-endian architectures. Fixed. 756255714Skris [Ben Laurie] 756355714Skris 756455714Skris *) Comment out two functions in bio.h that aren't implemented. Fix up the 756555714Skris Win32 test batch file so it (might) work again. The Win32 test batch file 756655714Skris is horrible: I feel ill.... 756755714Skris [Steve Henson] 756855714Skris 756955714Skris *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 757055714Skris in e_os.h. Audit of header files to check ANSI and non ANSI 757155714Skris sections: 10 functions were absent from non ANSI section and not exported 757255714Skris from Windows DLLs. Fixed up libeay.num for new functions. 757355714Skris [Steve Henson] 757455714Skris 757555714Skris *) Make `openssl version' output lines consistent. 757655714Skris [Ralf S. Engelschall] 757755714Skris 757855714Skris *) Fix Win32 symbol export lists for BIO functions: Added 757955714Skris BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 758055714Skris to ms/libeay{16,32}.def. 758155714Skris [Ralf S. Engelschall] 758255714Skris 758355714Skris *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 758455714Skris fine under Unix and passes some trivial tests I've now added. But the 758555714Skris whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 758655714Skris added to make sure no one expects that this stuff really works in the 758755714Skris OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 758855714Skris up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 758955714Skris openssl_bio.xs. 759055714Skris [Ralf S. Engelschall] 759155714Skris 759255714Skris *) Fix the generation of two part addresses in perl. 759355714Skris [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 759455714Skris 759555714Skris *) Add config entry for Linux on MIPS. 759655714Skris [John Tobey <jtobey@channel1.com>] 759755714Skris 759855714Skris *) Make links whenever Configure is run, unless we are on Windoze. 759955714Skris [Ben Laurie] 760055714Skris 760155714Skris *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 760255714Skris Currently only issuerAltName and AuthorityKeyIdentifier make any sense 760355714Skris in CRLs. 760455714Skris [Steve Henson] 760555714Skris 760655714Skris *) Add a useful kludge to allow package maintainers to specify compiler and 760755714Skris other platforms details on the command line without having to patch the 760855714Skris Configure script everytime: One now can use ``perl Configure 760955714Skris <id>:<details>'', i.e. platform ids are allowed to have details appended 761055714Skris to them (seperated by colons). This is treated as there would be a static 761155714Skris pre-configured entry in Configure's %table under key <id> with value 761255714Skris <details> and ``perl Configure <id>'' is called. So, when you want to 761355714Skris perform a quick test-compile under FreeBSD 3.1 with pgcc and without 761455714Skris assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 761555714Skris now, which overrides the FreeBSD-elf entry on-the-fly. 761655714Skris [Ralf S. Engelschall] 761755714Skris 761855714Skris *) Disable new TLS1 ciphersuites by default: they aren't official yet. 761955714Skris [Ben Laurie] 762055714Skris 762155714Skris *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 762255714Skris on the `perl Configure ...' command line. This way one can compile 762355714Skris OpenSSL libraries with Position Independent Code (PIC) which is needed 762455714Skris for linking it into DSOs. 762555714Skris [Ralf S. Engelschall] 762655714Skris 762755714Skris *) Remarkably, export ciphers were totally broken and no-one had noticed! 762855714Skris Fixed. 762955714Skris [Ben Laurie] 763055714Skris 763155714Skris *) Cleaned up the LICENSE document: The official contact for any license 763255714Skris questions now is the OpenSSL core team under openssl-core@openssl.org. 763355714Skris And add a paragraph about the dual-license situation to make sure people 763455714Skris recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 763555714Skris to the OpenSSL toolkit. 763655714Skris [Ralf S. Engelschall] 763755714Skris 763855714Skris *) General source tree makefile cleanups: Made `making xxx in yyy...' 763955714Skris display consistent in the source tree and replaced `/bin/rm' by `rm'. 764055714Skris Additonally cleaned up the `make links' target: Remove unnecessary 764155714Skris semicolons, subsequent redundant removes, inline point.sh into mklink.sh 764255714Skris to speed processing and no longer clutter the display with confusing 764355714Skris stuff. Instead only the actually done links are displayed. 764455714Skris [Ralf S. Engelschall] 764555714Skris 764655714Skris *) Permit null encryption ciphersuites, used for authentication only. It used 764755714Skris to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 764855714Skris It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 764955714Skris encryption. 765055714Skris [Ben Laurie] 765155714Skris 765255714Skris *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 765355714Skris signed attributes when verifying signatures (this would break them), 765455714Skris the detached data encoding was wrong and public keys obtained using 765555714Skris X509_get_pubkey() weren't freed. 765655714Skris [Steve Henson] 765755714Skris 765855714Skris *) Add text documentation for the BUFFER functions. Also added a work around 765955714Skris to a Win95 console bug. This was triggered by the password read stuff: the 766055714Skris last character typed gets carried over to the next fread(). If you were 766155714Skris generating a new cert request using 'req' for example then the last 766255714Skris character of the passphrase would be CR which would then enter the first 766355714Skris field as blank. 766455714Skris [Steve Henson] 766555714Skris 766655714Skris *) Added the new `Includes OpenSSL Cryptography Software' button as 766755714Skris doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 766855714Skris button and can be used by applications based on OpenSSL to show the 766955714Skris relationship to the OpenSSL project. 767055714Skris [Ralf S. Engelschall] 767155714Skris 767255714Skris *) Remove confusing variables in function signatures in files 767355714Skris ssl/ssl_lib.c and ssl/ssl.h. 767455714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 767555714Skris 767655714Skris *) Don't install bss_file.c under PREFIX/include/ 767755714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 767855714Skris 767955714Skris *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 768055714Skris functions that return function pointers and has support for NT specific 768155714Skris stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 768255714Skris #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 768355714Skris unsigned to signed types: this was killing the Win32 compile. 768455714Skris [Steve Henson] 768555714Skris 768655714Skris *) Add new certificate file to stack functions, 768755714Skris SSL_add_dir_cert_subjects_to_stack() and 768855714Skris SSL_add_file_cert_subjects_to_stack(). These largely supplant 768955714Skris SSL_load_client_CA_file(), and can be used to add multiple certs easily 769055714Skris to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 769155714Skris This means that Apache-SSL and similar packages don't have to mess around 769255714Skris to add as many CAs as they want to the preferred list. 769355714Skris [Ben Laurie] 769455714Skris 769555714Skris *) Experiment with doxygen documentation. Currently only partially applied to 769655714Skris ssl/ssl_lib.c. 769755714Skris See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 769855714Skris openssl.doxy as the configuration file. 769955714Skris [Ben Laurie] 770055714Skris 770155714Skris *) Get rid of remaining C++-style comments which strict C compilers hate. 770255714Skris [Ralf S. Engelschall, pointed out by Carlos Amengual] 770355714Skris 770455714Skris *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 770555714Skris compiled in by default: it has problems with large keys. 770655714Skris [Steve Henson] 770755714Skris 770855714Skris *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 770955714Skris DH private keys and/or callback functions which directly correspond to 771055714Skris their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 771155714Skris is needed for applications which have to configure certificates on a 771255714Skris per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 771355714Skris (e.g. s_server). 771455714Skris For the RSA certificate situation is makes no difference, but 771555714Skris for the DSA certificate situation this fixes the "no shared cipher" 771655714Skris problem where the OpenSSL cipher selection procedure failed because the 771755714Skris temporary keys were not overtaken from the context and the API provided 771855714Skris no way to reconfigure them. 771955714Skris The new functions now let applications reconfigure the stuff and they 772055714Skris are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 772155714Skris SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 772255714Skris non-public-API function ssl_cert_instantiate() is used as a helper 772355714Skris function and also to reduce code redundancy inside ssl_rsa.c. 772455714Skris [Ralf S. Engelschall] 772555714Skris 772655714Skris *) Move s_server -dcert and -dkey options out of the undocumented feature 772755714Skris area because they are useful for the DSA situation and should be 772855714Skris recognized by the users. 772955714Skris [Ralf S. Engelschall] 773055714Skris 773155714Skris *) Fix the cipher decision scheme for export ciphers: the export bits are 773255714Skris *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 773355714Skris SSL_EXP_MASK. So, the original variable has to be used instead of the 773455714Skris already masked variable. 773555714Skris [Richard Levitte <levitte@stacken.kth.se>] 773655714Skris 773755714Skris *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 773855714Skris [Richard Levitte <levitte@stacken.kth.se>] 773955714Skris 774055714Skris *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 774155714Skris from `int' to `unsigned int' because it's a length and initialized by 774255714Skris EVP_DigestFinal() which expects an `unsigned int *'. 774355714Skris [Richard Levitte <levitte@stacken.kth.se>] 774455714Skris 774555714Skris *) Don't hard-code path to Perl interpreter on shebang line of Configure 774655714Skris script. Instead use the usual Shell->Perl transition trick. 774755714Skris [Ralf S. Engelschall] 774855714Skris 774955714Skris *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 775055714Skris (in addition to RSA certificates) to match the behaviour of `openssl dsa 775155714Skris -noout -modulus' as it's already the case for `openssl rsa -noout 775255714Skris -modulus'. For RSA the -modulus is the real "modulus" while for DSA 775355714Skris currently the public key is printed (a decision which was already done by 775455714Skris `openssl dsa -modulus' in the past) which serves a similar purpose. 775555714Skris Additionally the NO_RSA no longer completely removes the whole -modulus 775655714Skris option; it now only avoids using the RSA stuff. Same applies to NO_DSA 775755714Skris now, too. 775855714Skris [Ralf S. Engelschall] 775955714Skris 776055714Skris *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 776155714Skris BIO. See the source (crypto/evp/bio_ok.c) for more info. 776255714Skris [Arne Ansper <arne@ats.cyber.ee>] 776355714Skris 776455714Skris *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 776555714Skris to be added. Now both 'req' and 'ca' can use new objects defined in the 776655714Skris config file. 776755714Skris [Steve Henson] 776855714Skris 776955714Skris *) Add cool BIO that does syslog (or event log on NT). 777055714Skris [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 777155714Skris 777255714Skris *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 777355714Skris TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 777455714Skris TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 777555714Skris Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 777655714Skris [Ben Laurie] 777755714Skris 777855714Skris *) Add preliminary config info for new extension code. 777955714Skris [Steve Henson] 778055714Skris 778155714Skris *) Make RSA_NO_PADDING really use no padding. 778255714Skris [Ulf Moeller <ulf@fitug.de>] 778355714Skris 778455714Skris *) Generate errors when private/public key check is done. 778555714Skris [Ben Laurie] 778655714Skris 778755714Skris *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 778855714Skris for some CRL extensions and new objects added. 778955714Skris [Steve Henson] 779055714Skris 779155714Skris *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 779255714Skris key usage extension and fuller support for authority key id. 779355714Skris [Steve Henson] 779455714Skris 779555714Skris *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 779655714Skris padding method for RSA, which is recommended for new applications in PKCS 779755714Skris #1 v2.0 (RFC 2437, October 1998). 779855714Skris OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 779955714Skris foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 780055714Skris against Bleichbacher's attack on RSA. 780155714Skris [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 780255714Skris Ben Laurie] 780355714Skris 780455714Skris *) Updates to the new SSL compression code 780555714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 780655714Skris 780755714Skris *) Fix so that the version number in the master secret, when passed 780855714Skris via RSA, checks that if TLS was proposed, but we roll back to SSLv3 780955714Skris (because the server will not accept higher), that the version number 781055714Skris is 0x03,0x01, not 0x03,0x00 781155714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 781255714Skris 781355714Skris *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 781455714Skris leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 781555714Skris in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 781655714Skris [Steve Henson] 781755714Skris 781855714Skris *) Support for RAW extensions where an arbitrary extension can be 781955714Skris created by including its DER encoding. See apps/openssl.cnf for 782055714Skris an example. 782155714Skris [Steve Henson] 782255714Skris 782355714Skris *) Make sure latest Perl versions don't interpret some generated C array 782455714Skris code as Perl array code in the crypto/err/err_genc.pl script. 782555714Skris [Lars Weber <3weber@informatik.uni-hamburg.de>] 782655714Skris 782755714Skris *) Modify ms/do_ms.bat to not generate assembly language makefiles since 782855714Skris not many people have the assembler. Various Win32 compilation fixes and 782955714Skris update to the INSTALL.W32 file with (hopefully) more accurate Win32 783055714Skris build instructions. 783155714Skris [Steve Henson] 783255714Skris 783355714Skris *) Modify configure script 'Configure' to automatically create crypto/date.h 783455714Skris file under Win32 and also build pem.h from pem.org. New script 783555714Skris util/mkfiles.pl to create the MINFO file on environments that can't do a 783655714Skris 'make files': perl util/mkfiles.pl >MINFO should work. 783755714Skris [Steve Henson] 783855714Skris 783955714Skris *) Major rework of DES function declarations, in the pursuit of correctness 784055714Skris and purity. As a result, many evil casts evaporated, and some weirdness, 784155714Skris too. You may find this causes warnings in your code. Zapping your evil 784255714Skris casts will probably fix them. Mostly. 784355714Skris [Ben Laurie] 784455714Skris 784555714Skris *) Fix for a typo in asn1.h. Bug fix to object creation script 784655714Skris obj_dat.pl. It considered a zero in an object definition to mean 784755714Skris "end of object": none of the objects in objects.h have any zeros 784855714Skris so it wasn't spotted. 784955714Skris [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 785055714Skris 785155714Skris *) Add support for Triple DES Cipher Block Chaining with Output Feedback 785255714Skris Masking (CBCM). In the absence of test vectors, the best I have been able 785355714Skris to do is check that the decrypt undoes the encrypt, so far. Send me test 785455714Skris vectors if you have them. 785555714Skris [Ben Laurie] 785655714Skris 785755714Skris *) Correct calculation of key length for export ciphers (too much space was 785855714Skris allocated for null ciphers). This has not been tested! 785955714Skris [Ben Laurie] 786055714Skris 786155714Skris *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 786255714Skris message is now correct (it understands "crypto" and "ssl" on its 786355714Skris command line). There is also now an "update" option. This will update 786455714Skris the util/ssleay.num and util/libeay.num files with any new functions. 786555714Skris If you do a: 786655714Skris perl util/mkdef.pl crypto ssl update 786755714Skris it will update them. 786855714Skris [Steve Henson] 786955714Skris 787055714Skris *) Overhauled the Perl interface (perl/*): 787155714Skris - ported BN stuff to OpenSSL's different BN library 787255714Skris - made the perl/ source tree CVS-aware 787355714Skris - renamed the package from SSLeay to OpenSSL (the files still contain 787455714Skris their history because I've copied them in the repository) 787555714Skris - removed obsolete files (the test scripts will be replaced 787655714Skris by better Test::Harness variants in the future) 787755714Skris [Ralf S. Engelschall] 787855714Skris 787955714Skris *) First cut for a very conservative source tree cleanup: 788055714Skris 1. merge various obsolete readme texts into doc/ssleay.txt 788155714Skris where we collect the old documents and readme texts. 788255714Skris 2. remove the first part of files where I'm already sure that we no 788355714Skris longer need them because of three reasons: either they are just temporary 788455714Skris files which were left by Eric or they are preserved original files where 788555714Skris I've verified that the diff is also available in the CVS via "cvs diff 788655714Skris -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 788755714Skris the crypto/md/ stuff). 788855714Skris [Ralf S. Engelschall] 788955714Skris 789055714Skris *) More extension code. Incomplete support for subject and issuer alt 789155714Skris name, issuer and authority key id. Change the i2v function parameters 789255714Skris and add an extra 'crl' parameter in the X509V3_CTX structure: guess 789355714Skris what that's for :-) Fix to ASN1 macro which messed up 789455714Skris IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 789555714Skris [Steve Henson] 789655714Skris 789755714Skris *) Preliminary support for ENUMERATED type. This is largely copied from the 789855714Skris INTEGER code. 789955714Skris [Steve Henson] 790055714Skris 790155714Skris *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 790255714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 790355714Skris 790455714Skris *) Make sure `make rehash' target really finds the `openssl' program. 790555714Skris [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 790655714Skris 790755714Skris *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 790855714Skris like to hear about it if this slows down other processors. 790955714Skris [Ben Laurie] 791055714Skris 791155714Skris *) Add CygWin32 platform information to Configure script. 791255714Skris [Alan Batie <batie@aahz.jf.intel.com>] 791355714Skris 791455714Skris *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 791555714Skris [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 791655714Skris 791755714Skris *) New program nseq to manipulate netscape certificate sequences 791855714Skris [Steve Henson] 791955714Skris 792055714Skris *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 792155714Skris few typos. 792255714Skris [Steve Henson] 792355714Skris 792455714Skris *) Fixes to BN code. Previously the default was to define BN_RECURSION 792555714Skris but the BN code had some problems that would cause failures when 792655714Skris doing certificate verification and some other functions. 792755714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 792855714Skris 792955714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 793055714Skris [Steve Henson] 793155714Skris 793255714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 793355714Skris [Steve Henson] 793455714Skris 793555714Skris *) Add several PKIX and private extended key usage OIDs. 793655714Skris [Steve Henson] 793755714Skris 793855714Skris *) Modify the 'ca' program to handle the new extension code. Modify 793955714Skris openssl.cnf for new extension format, add comments. 794055714Skris [Steve Henson] 794155714Skris 794255714Skris *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 794355714Skris and add a sample to openssl.cnf so req -x509 now adds appropriate 794455714Skris CA extensions. 794555714Skris [Steve Henson] 794655714Skris 794755714Skris *) Continued X509 V3 changes. Add to other makefiles, integrate with the 794855714Skris error code, add initial support to X509_print() and x509 application. 794955714Skris [Steve Henson] 795055714Skris 795155714Skris *) Takes a deep breath and start addding X509 V3 extension support code. Add 795255714Skris files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 795355714Skris stuff is currently isolated and isn't even compiled yet. 795455714Skris [Steve Henson] 795555714Skris 795655714Skris *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 795755714Skris ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 795855714Skris Removed the versions check from X509 routines when loading extensions: 795955714Skris this allows certain broken certificates that don't set the version 796055714Skris properly to be processed. 796155714Skris [Steve Henson] 796255714Skris 796355714Skris *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 796455714Skris Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 796555714Skris can still be regenerated with "make depend". 796655714Skris [Ben Laurie] 796755714Skris 796855714Skris *) Spelling mistake in C version of CAST-128. 796955714Skris [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 797055714Skris 797155714Skris *) Changes to the error generation code. The perl script err-code.pl 797255714Skris now reads in the old error codes and retains the old numbers, only 797355714Skris adding new ones if necessary. It also only changes the .err files if new 797455714Skris codes are added. The makefiles have been modified to only insert errors 797555714Skris when needed (to avoid needlessly modifying header files). This is done 797655714Skris by only inserting errors if the .err file is newer than the auto generated 797755714Skris C file. To rebuild all the error codes from scratch (the old behaviour) 797855714Skris either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 797955714Skris or delete all the .err files. 798055714Skris [Steve Henson] 798155714Skris 798255714Skris *) CAST-128 was incorrectly implemented for short keys. The C version has 798355714Skris been fixed, but is untested. The assembler versions are also fixed, but 798455714Skris new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 798555714Skris to regenerate it if needed. 798655714Skris [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 798755714Skris Hagino <itojun@kame.net>] 798855714Skris 798955714Skris *) File was opened incorrectly in randfile.c. 799055714Skris [Ulf M�ller <ulf@fitug.de>] 799155714Skris 799255714Skris *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 799355714Skris functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 799455714Skris GeneralizedTime. ASN1_TIME is the proper type used in certificates et 799555714Skris al: it's just almost always a UTCTime. Note this patch adds new error 799655714Skris codes so do a "make errors" if there are problems. 799755714Skris [Steve Henson] 799855714Skris 799955714Skris *) Correct Linux 1 recognition in config. 800055714Skris [Ulf M�ller <ulf@fitug.de>] 800155714Skris 800255714Skris *) Remove pointless MD5 hash when using DSA keys in ca. 800355714Skris [Anonymous <nobody@replay.com>] 800455714Skris 800555714Skris *) Generate an error if given an empty string as a cert directory. Also 800655714Skris generate an error if handed NULL (previously returned 0 to indicate an 800755714Skris error, but didn't set one). 800855714Skris [Ben Laurie, reported by Anonymous <nobody@replay.com>] 800955714Skris 801055714Skris *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 801155714Skris [Ben Laurie] 801255714Skris 801355714Skris *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 801455714Skris parameters. This was causing a warning which killed off the Win32 compile. 801555714Skris [Steve Henson] 801655714Skris 801755714Skris *) Remove C++ style comments from crypto/bn/bn_local.h. 801855714Skris [Neil Costigan <neil.costigan@celocom.com>] 801955714Skris 802055714Skris *) The function OBJ_txt2nid was broken. It was supposed to return a nid 802155714Skris based on a text string, looking up short and long names and finally 802255714Skris "dot" format. The "dot" format stuff didn't work. Added new function 802355714Skris OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 802455714Skris OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 802555714Skris OID is not part of the table. 802655714Skris [Steve Henson] 802755714Skris 802855714Skris *) Add prototypes to X509 lookup/verify methods, fixing a bug in 802955714Skris X509_LOOKUP_by_alias(). 803055714Skris [Ben Laurie] 803155714Skris 803255714Skris *) Sort openssl functions by name. 803355714Skris [Ben Laurie] 803455714Skris 803555714Skris *) Get the gendsa program working (hopefully) and add it to app list. Remove 803655714Skris encryption from sample DSA keys (in case anyone is interested the password 803755714Skris was "1234"). 803855714Skris [Steve Henson] 803955714Skris 804055714Skris *) Make _all_ *_free functions accept a NULL pointer. 804155714Skris [Frans Heymans <fheymans@isaserver.be>] 804255714Skris 804355714Skris *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 804455714Skris NULL pointers. 804555714Skris [Anonymous <nobody@replay.com>] 804655714Skris 804755714Skris *) s_server should send the CAfile as acceptable CAs, not its own cert. 804855714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 804955714Skris 805055714Skris *) Don't blow it for numeric -newkey arguments to apps/req. 805155714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 805255714Skris 805355714Skris *) Temp key "for export" tests were wrong in s3_srvr.c. 805455714Skris [Anonymous <nobody@replay.com>] 805555714Skris 805655714Skris *) Add prototype for temp key callback functions 805755714Skris SSL_CTX_set_tmp_{rsa,dh}_callback(). 805855714Skris [Ben Laurie] 805955714Skris 806055714Skris *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 806155714Skris DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 806255714Skris [Steve Henson] 806355714Skris 806455714Skris *) X509_name_add_entry() freed the wrong thing after an error. 806555714Skris [Arne Ansper <arne@ats.cyber.ee>] 806655714Skris 806755714Skris *) rsa_eay.c would attempt to free a NULL context. 806855714Skris [Arne Ansper <arne@ats.cyber.ee>] 806955714Skris 807055714Skris *) BIO_s_socket() had a broken should_retry() on Windoze. 807155714Skris [Arne Ansper <arne@ats.cyber.ee>] 807255714Skris 807355714Skris *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 807455714Skris [Arne Ansper <arne@ats.cyber.ee>] 807555714Skris 807655714Skris *) Make sure the already existing X509_STORE->depth variable is initialized 807755714Skris in X509_STORE_new(), but document the fact that this variable is still 807855714Skris unused in the certificate verification process. 807955714Skris [Ralf S. Engelschall] 808055714Skris 808155714Skris *) Fix the various library and apps files to free up pkeys obtained from 808255714Skris X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 808355714Skris [Steve Henson] 808455714Skris 808555714Skris *) Fix reference counting in X509_PUBKEY_get(). This makes 808655714Skris demos/maurice/example2.c work, amongst others, probably. 808755714Skris [Steve Henson and Ben Laurie] 808855714Skris 808955714Skris *) First cut of a cleanup for apps/. First the `ssleay' program is now named 809055714Skris `openssl' and second, the shortcut symlinks for the `openssl <command>' 809155714Skris are no longer created. This way we have a single and consistent command 809255714Skris line interface `openssl <command>', similar to `cvs <command>'. 809355714Skris [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 809455714Skris 809555714Skris *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 809655714Skris BIT STRING wrapper always have zero unused bits. 809755714Skris [Steve Henson] 809855714Skris 809955714Skris *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 810055714Skris [Steve Henson] 810155714Skris 810255714Skris *) Make the top-level INSTALL documentation easier to understand. 810355714Skris [Paul Sutton] 810455714Skris 810555714Skris *) Makefiles updated to exit if an error occurs in a sub-directory 810655714Skris make (including if user presses ^C) [Paul Sutton] 810755714Skris 810855714Skris *) Make Montgomery context stuff explicit in RSA data structure. 810955714Skris [Ben Laurie] 811055714Skris 811155714Skris *) Fix build order of pem and err to allow for generated pem.h. 811255714Skris [Ben Laurie] 811355714Skris 811455714Skris *) Fix renumbering bug in X509_NAME_delete_entry(). 811555714Skris [Ben Laurie] 811655714Skris 811755714Skris *) Enhanced the err-ins.pl script so it makes the error library number 811855714Skris global and can add a library name. This is needed for external ASN1 and 811955714Skris other error libraries. 812055714Skris [Steve Henson] 812155714Skris 812255714Skris *) Fixed sk_insert which never worked properly. 812355714Skris [Steve Henson] 812455714Skris 812555714Skris *) Fix ASN1 macros so they can handle indefinite length construted 812655714Skris EXPLICIT tags. Some non standard certificates use these: they can now 812755714Skris be read in. 812855714Skris [Steve Henson] 812955714Skris 813055714Skris *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 813155714Skris into a single doc/ssleay.txt bundle. This way the information is still 813255714Skris preserved but no longer messes up this directory. Now it's new room for 813355714Skris the new set of documenation files. 813455714Skris [Ralf S. Engelschall] 813555714Skris 813655714Skris *) SETs were incorrectly DER encoded. This was a major pain, because they 813755714Skris shared code with SEQUENCEs, which aren't coded the same. This means that 813855714Skris almost everything to do with SETs or SEQUENCEs has either changed name or 813955714Skris number of arguments. 814055714Skris [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 814155714Skris 814255714Skris *) Fix test data to work with the above. 814355714Skris [Ben Laurie] 814455714Skris 814555714Skris *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 814655714Skris was already fixed by Eric for 0.9.1 it seems. 814755714Skris [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 814855714Skris 814955714Skris *) Autodetect FreeBSD3. 815055714Skris [Ben Laurie] 815155714Skris 815255714Skris *) Fix various bugs in Configure. This affects the following platforms: 815355714Skris nextstep 815455714Skris ncr-scde 815555714Skris unixware-2.0 815655714Skris unixware-2.0-pentium 815755714Skris sco5-cc. 815855714Skris [Ben Laurie] 815955714Skris 816055714Skris *) Eliminate generated files from CVS. Reorder tests to regenerate files 816155714Skris before they are needed. 816255714Skris [Ben Laurie] 816355714Skris 816455714Skris *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 816555714Skris [Ben Laurie] 816655714Skris 816755714Skris 816855714Skris Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 816955714Skris 817055714Skris *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 817155714Skris changed SSLeay to OpenSSL in version strings. 817255714Skris [Ralf S. Engelschall] 817355714Skris 817455714Skris *) Some fixups to the top-level documents. 817555714Skris [Paul Sutton] 817655714Skris 817755714Skris *) Fixed the nasty bug where rsaref.h was not found under compile-time 817855714Skris because the symlink to include/ was missing. 817955714Skris [Ralf S. Engelschall] 818055714Skris 818155714Skris *) Incorporated the popular no-RSA/DSA-only patches 818255714Skris which allow to compile a RSA-free SSLeay. 818355714Skris [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 818455714Skris 818555714Skris *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 818655714Skris when "ssleay" is still not found. 818755714Skris [Ralf S. Engelschall] 818855714Skris 818955714Skris *) Added more platforms to Configure: Cray T3E, HPUX 11, 819055714Skris [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 819155714Skris 819255714Skris *) Updated the README file. 819355714Skris [Ralf S. Engelschall] 819455714Skris 819555714Skris *) Added various .cvsignore files in the CVS repository subdirs 819655714Skris to make a "cvs update" really silent. 819755714Skris [Ralf S. Engelschall] 819855714Skris 819955714Skris *) Recompiled the error-definition header files and added 820055714Skris missing symbols to the Win32 linker tables. 820155714Skris [Ralf S. Engelschall] 820255714Skris 820355714Skris *) Cleaned up the top-level documents; 820455714Skris o new files: CHANGES and LICENSE 820555714Skris o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 820655714Skris o merged COPYRIGHT into LICENSE 820755714Skris o removed obsolete TODO file 820855714Skris o renamed MICROSOFT to INSTALL.W32 820955714Skris [Ralf S. Engelschall] 821055714Skris 821155714Skris *) Removed dummy files from the 0.9.1b source tree: 821255714Skris crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 821355714Skris crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 821455714Skris crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 821555714Skris crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 821655714Skris util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 821755714Skris [Ralf S. Engelschall] 821855714Skris 821955714Skris *) Added various platform portability fixes. 822055714Skris [Mark J. Cox] 822155714Skris 822255714Skris *) The Genesis of the OpenSSL rpject: 822355714Skris We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 822455714Skris Young and Tim J. Hudson created while they were working for C2Net until 822555714Skris summer 1998. 822655714Skris [The OpenSSL Project] 822755714Skris 822855714Skris 822955714Skris Changes between 0.9.0b and 0.9.1b [not released] 823055714Skris 823155714Skris *) Updated a few CA certificates under certs/ 823255714Skris [Eric A. Young] 823355714Skris 823455714Skris *) Changed some BIGNUM api stuff. 823555714Skris [Eric A. Young] 823655714Skris 823755714Skris *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 823855714Skris DGUX x86, Linux Alpha, etc. 823955714Skris [Eric A. Young] 824055714Skris 824155714Skris *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 824255714Skris RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 824355714Skris available). 824455714Skris [Eric A. Young] 824555714Skris 824655714Skris *) Add -strparse option to asn1pars program which parses nested 824755714Skris binary structures 824855714Skris [Dr Stephen Henson <shenson@bigfoot.com>] 824955714Skris 825055714Skris *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 825155714Skris [Eric A. Young] 825255714Skris 825355714Skris *) DSA fix for "ca" program. 825455714Skris [Eric A. Young] 825555714Skris 825655714Skris *) Added "-genkey" option to "dsaparam" program. 825755714Skris [Eric A. Young] 825855714Skris 825955714Skris *) Added RIPE MD160 (rmd160) message digest. 826055714Skris [Eric A. Young] 826155714Skris 826255714Skris *) Added -a (all) option to "ssleay version" command. 826355714Skris [Eric A. Young] 826455714Skris 826555714Skris *) Added PLATFORM define which is the id given to Configure. 826655714Skris [Eric A. Young] 826755714Skris 826855714Skris *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 826955714Skris [Eric A. Young] 827055714Skris 827155714Skris *) Extended the ASN.1 parser routines. 827255714Skris [Eric A. Young] 827355714Skris 827455714Skris *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 827555714Skris [Eric A. Young] 827655714Skris 827755714Skris *) Added a BN_CTX to the BN library. 827855714Skris [Eric A. Young] 827955714Skris 828055714Skris *) Fixed the weak key values in DES library 828155714Skris [Eric A. Young] 828255714Skris 828355714Skris *) Changed API in EVP library for cipher aliases. 828455714Skris [Eric A. Young] 828555714Skris 828655714Skris *) Added support for RC2/64bit cipher. 828755714Skris [Eric A. Young] 828855714Skris 828955714Skris *) Converted the lhash library to the crypto/mem.c functions. 829055714Skris [Eric A. Young] 829155714Skris 829255714Skris *) Added more recognized ASN.1 object ids. 829355714Skris [Eric A. Young] 829455714Skris 829555714Skris *) Added more RSA padding checks for SSL/TLS. 829655714Skris [Eric A. Young] 829755714Skris 829855714Skris *) Added BIO proxy/filter functionality. 829955714Skris [Eric A. Young] 830055714Skris 830155714Skris *) Added extra_certs to SSL_CTX which can be used 830255714Skris send extra CA certificates to the client in the CA cert chain sending 830355714Skris process. It can be configured with SSL_CTX_add_extra_chain_cert(). 830455714Skris [Eric A. Young] 830555714Skris 830655714Skris *) Now Fortezza is denied in the authentication phase because 830755714Skris this is key exchange mechanism is not supported by SSLeay at all. 830855714Skris [Eric A. Young] 830955714Skris 831055714Skris *) Additional PKCS1 checks. 831155714Skris [Eric A. Young] 831255714Skris 831355714Skris *) Support the string "TLSv1" for all TLS v1 ciphers. 831455714Skris [Eric A. Young] 831555714Skris 831655714Skris *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 831755714Skris ex_data index of the SSL context in the X509_STORE_CTX ex_data. 831855714Skris [Eric A. Young] 831955714Skris 832055714Skris *) Fixed a few memory leaks. 832155714Skris [Eric A. Young] 832255714Skris 832355714Skris *) Fixed various code and comment typos. 832455714Skris [Eric A. Young] 832555714Skris 832655714Skris *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 832755714Skris bytes sent in the client random. 832855714Skris [Edward Bishop <ebishop@spyglass.com>] 832955714Skris 8330