CHANGES revision 206046
155714Skris 255714Skris OpenSSL CHANGES 355714Skris _______________ 455714Skris 5206046Ssimon Changes between 0.9.8m and 0.9.8n [24 Mar 2010] 6206046Ssimon 7206046Ssimon *) When rejecting SSL/TLS records due to an incorrect version number, never 8206046Ssimon update s->server with a new major version number. As of 9206046Ssimon - OpenSSL 0.9.8m if 'short' is a 16-bit type, 10206046Ssimon - OpenSSL 0.9.8f if 'short' is longer than 16 bits, 11206046Ssimon the previous behavior could result in a read attempt at NULL when 12206046Ssimon receiving specific incorrect SSL/TLS records once record payload 13206046Ssimon protection is active. (CVE-2010-0740) 14206046Ssimon [Bodo Moeller, Adam Langley <agl@chromium.org>] 15206046Ssimon 16206046Ssimon *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL 17206046Ssimon could be crashed if the relevant tables were not present (e.g. chrooted). 18206046Ssimon [Tomas Hoger <thoger@redhat.com>] 19206046Ssimon 20205128Ssimon Changes between 0.9.8l and 0.9.8m [25 Feb 2010] 21205128Ssimon 22205128Ssimon *) Always check bn_wexpend() return values for failure. (CVE-2009-3245) 23205128Ssimon [Martin Olsson, Neel Mehta] 24205128Ssimon 25205128Ssimon *) Fix X509_STORE locking: Every 'objs' access requires a lock (to 26205128Ssimon accommodate for stack sorting, always a write lock!). 27205128Ssimon [Bodo Moeller] 28205128Ssimon 29205128Ssimon *) On some versions of WIN32 Heap32Next is very slow. This can cause 30205128Ssimon excessive delays in the RAND_poll(): over a minute. As a workaround 31205128Ssimon include a time check in the inner Heap32Next loop too. 32205128Ssimon [Steve Henson] 33205128Ssimon 34205128Ssimon *) The code that handled flushing of data in SSL/TLS originally used the 35205128Ssimon BIO_CTRL_INFO ctrl to see if any data was pending first. This caused 36205128Ssimon the problem outlined in PR#1949. The fix suggested there however can 37205128Ssimon trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions 38205128Ssimon of Apache). So instead simplify the code to flush unconditionally. 39205128Ssimon This should be fine since flushing with no data to flush is a no op. 40205128Ssimon [Steve Henson] 41205128Ssimon 42205128Ssimon *) Handle TLS versions 2.0 and later properly and correctly use the 43205128Ssimon highest version of TLS/SSL supported. Although TLS >= 2.0 is some way 44205128Ssimon off ancient servers have a habit of sticking around for a while... 45205128Ssimon [Steve Henson] 46205128Ssimon 47205128Ssimon *) Modify compression code so it frees up structures without using the 48205128Ssimon ex_data callbacks. This works around a problem where some applications 49205128Ssimon call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when 50205128Ssimon restarting) then use compression (e.g. SSL with compression) later. 51205128Ssimon This results in significant per-connection memory leaks and 52205128Ssimon has caused some security issues including CVE-2008-1678 and 53205128Ssimon CVE-2009-4355. 54205128Ssimon [Steve Henson] 55205128Ssimon 56205128Ssimon *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't 57205128Ssimon change when encrypting or decrypting. 58205128Ssimon [Bodo Moeller] 59205128Ssimon 60205128Ssimon *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to 61205128Ssimon connect and renegotiate with servers which do not support RI. 62205128Ssimon Until RI is more widely deployed this option is enabled by default. 63205128Ssimon [Steve Henson] 64205128Ssimon 65205128Ssimon *) Add "missing" ssl ctrls to clear options and mode. 66205128Ssimon [Steve Henson] 67205128Ssimon 68205128Ssimon *) If client attempts to renegotiate and doesn't support RI respond with 69205128Ssimon a no_renegotiation alert as required by RFC5746. Some renegotiating 70205128Ssimon TLS clients will continue a connection gracefully when they receive 71205128Ssimon the alert. Unfortunately OpenSSL mishandled this alert and would hang 72205128Ssimon waiting for a server hello which it will never receive. Now we treat a 73205128Ssimon received no_renegotiation alert as a fatal error. This is because 74205128Ssimon applications requesting a renegotiation might well expect it to succeed 75205128Ssimon and would have no code in place to handle the server denying it so the 76205128Ssimon only safe thing to do is to terminate the connection. 77205128Ssimon [Steve Henson] 78205128Ssimon 79205128Ssimon *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if 80205128Ssimon peer supports secure renegotiation and 0 otherwise. Print out peer 81205128Ssimon renegotiation support in s_client/s_server. 82205128Ssimon [Steve Henson] 83205128Ssimon 84205128Ssimon *) Replace the highly broken and deprecated SPKAC certification method with 85205128Ssimon the updated NID creation version. This should correctly handle UTF8. 86205128Ssimon [Steve Henson] 87205128Ssimon 88205128Ssimon *) Implement RFC5746. Re-enable renegotiation but require the extension 89205128Ssimon as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION 90205128Ssimon turns out to be a bad idea. It has been replaced by 91205128Ssimon SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with 92205128Ssimon SSL_CTX_set_options(). This is really not recommended unless you 93205128Ssimon know what you are doing. 94205128Ssimon [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson] 95205128Ssimon 96205128Ssimon *) Fixes to stateless session resumption handling. Use initial_ctx when 97205128Ssimon issuing and attempting to decrypt tickets in case it has changed during 98205128Ssimon servername handling. Use a non-zero length session ID when attempting 99205128Ssimon stateless session resumption: this makes it possible to determine if 100205128Ssimon a resumption has occurred immediately after receiving server hello 101205128Ssimon (several places in OpenSSL subtly assume this) instead of later in 102205128Ssimon the handshake. 103205128Ssimon [Steve Henson] 104205128Ssimon 105205128Ssimon *) The functions ENGINE_ctrl(), OPENSSL_isservice(), 106205128Ssimon CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error 107205128Ssimon fixes for a few places where the return code is not checked 108205128Ssimon correctly. 109205128Ssimon [Julia Lawall <julia@diku.dk>] 110205128Ssimon 111205128Ssimon *) Add --strict-warnings option to Configure script to include devteam 112205128Ssimon warnings in other configurations. 113205128Ssimon [Steve Henson] 114205128Ssimon 115205128Ssimon *) Add support for --libdir option and LIBDIR variable in makefiles. This 116205128Ssimon makes it possible to install openssl libraries in locations which 117205128Ssimon have names other than "lib", for example "/usr/lib64" which some 118205128Ssimon systems need. 119205128Ssimon [Steve Henson, based on patch from Jeremy Utley] 120205128Ssimon 121205128Ssimon *) Don't allow the use of leading 0x80 in OIDs. This is a violation of 122205128Ssimon X690 8.9.12 and can produce some misleading textual output of OIDs. 123205128Ssimon [Steve Henson, reported by Dan Kaminsky] 124205128Ssimon 125205128Ssimon *) Delete MD2 from algorithm tables. This follows the recommendation in 126205128Ssimon several standards that it is not used in new applications due to 127205128Ssimon several cryptographic weaknesses. For binary compatibility reasons 128205128Ssimon the MD2 API is still compiled in by default. 129205128Ssimon [Steve Henson] 130205128Ssimon 131205128Ssimon *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved 132205128Ssimon and restored. 133205128Ssimon [Steve Henson] 134205128Ssimon 135205128Ssimon *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and 136205128Ssimon OPENSSL_asc2uni conditionally on Netware platforms to avoid a name 137205128Ssimon clash. 138205128Ssimon [Guenter <lists@gknw.net>] 139205128Ssimon 140205128Ssimon *) Fix the server certificate chain building code to use X509_verify_cert(), 141205128Ssimon it used to have an ad-hoc builder which was unable to cope with anything 142205128Ssimon other than a simple chain. 143205128Ssimon [David Woodhouse <dwmw2@infradead.org>, Steve Henson] 144205128Ssimon 145205128Ssimon *) Don't check self signed certificate signatures in X509_verify_cert() 146205128Ssimon by default (a flag can override this): it just wastes time without 147205128Ssimon adding any security. As a useful side effect self signed root CAs 148205128Ssimon with non-FIPS digests are now usable in FIPS mode. 149205128Ssimon [Steve Henson] 150205128Ssimon 151205128Ssimon *) In dtls1_process_out_of_seq_message() the check if the current message 152205128Ssimon is already buffered was missing. For every new message was memory 153205128Ssimon allocated, allowing an attacker to perform an denial of service attack 154205128Ssimon with sending out of seq handshake messages until there is no memory 155205128Ssimon left. Additionally every future messege was buffered, even if the 156205128Ssimon sequence number made no sense and would be part of another handshake. 157205128Ssimon So only messages with sequence numbers less than 10 in advance will be 158205128Ssimon buffered. (CVE-2009-1378) 159205128Ssimon [Robin Seggelmann, discovered by Daniel Mentz] 160205128Ssimon 161205128Ssimon *) Records are buffered if they arrive with a future epoch to be 162205128Ssimon processed after finishing the corresponding handshake. There is 163205128Ssimon currently no limitation to this buffer allowing an attacker to perform 164205128Ssimon a DOS attack with sending records with future epochs until there is no 165205128Ssimon memory left. This patch adds the pqueue_size() function to detemine 166205128Ssimon the size of a buffer and limits the record buffer to 100 entries. 167205128Ssimon (CVE-2009-1377) 168205128Ssimon [Robin Seggelmann, discovered by Daniel Mentz] 169205128Ssimon 170205128Ssimon *) Keep a copy of frag->msg_header.frag_len so it can be used after the 171205128Ssimon parent structure is freed. (CVE-2009-1379) 172205128Ssimon [Daniel Mentz] 173205128Ssimon 174205128Ssimon *) Handle non-blocking I/O properly in SSL_shutdown() call. 175205128Ssimon [Darryl Miles <darryl-mailinglists@netbauds.net>] 176205128Ssimon 177205128Ssimon *) Add 2.5.4.* OIDs 178205128Ssimon [Ilya O. <vrghost@gmail.com>] 179205128Ssimon 180205128Ssimon Changes between 0.9.8k and 0.9.8l [5 Nov 2009] 181205128Ssimon 182205128Ssimon *) Disable renegotiation completely - this fixes a severe security 183205128Ssimon problem (CVE-2009-3555) at the cost of breaking all 184205128Ssimon renegotiation. Renegotiation can be re-enabled by setting 185205128Ssimon SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at 186205128Ssimon run-time. This is really not recommended unless you know what 187205128Ssimon you're doing. 188205128Ssimon [Ben Laurie] 189205128Ssimon 190194206Ssimon Changes between 0.9.8j and 0.9.8k [25 Mar 2009] 191167612Ssimon 192194206Ssimon *) Don't set val to NULL when freeing up structures, it is freed up by 193194206Ssimon underlying code. If sizeof(void *) > sizeof(long) this can result in 194194206Ssimon zeroing past the valid field. (CVE-2009-0789) 195194206Ssimon [Paolo Ganci <Paolo.Ganci@AdNovum.CH>] 196194206Ssimon 197194206Ssimon *) Fix bug where return value of CMS_SignerInfo_verify_content() was not 198194206Ssimon checked correctly. This would allow some invalid signed attributes to 199194206Ssimon appear to verify correctly. (CVE-2009-0591) 200194206Ssimon [Ivan Nestlerode <inestlerode@us.ibm.com>] 201194206Ssimon 202194206Ssimon *) Reject UniversalString and BMPString types with invalid lengths. This 203194206Ssimon prevents a crash in ASN1_STRING_print_ex() which assumes the strings have 204194206Ssimon a legal length. (CVE-2009-0590) 205194206Ssimon [Steve Henson] 206194206Ssimon 207194206Ssimon *) Set S/MIME signing as the default purpose rather than setting it 208194206Ssimon unconditionally. This allows applications to override it at the store 209194206Ssimon level. 210194206Ssimon [Steve Henson] 211194206Ssimon 212194206Ssimon *) Permit restricted recursion of ASN1 strings. This is needed in practice 213194206Ssimon to handle some structures. 214194206Ssimon [Steve Henson] 215194206Ssimon 216194206Ssimon *) Improve efficiency of mem_gets: don't search whole buffer each time 217194206Ssimon for a '\n' 218194206Ssimon [Jeremy Shapiro <jnshapir@us.ibm.com>] 219194206Ssimon 220194206Ssimon *) New -hex option for openssl rand. 221194206Ssimon [Matthieu Herrb] 222194206Ssimon 223194206Ssimon *) Print out UTF8String and NumericString when parsing ASN1. 224194206Ssimon [Steve Henson] 225194206Ssimon 226194206Ssimon *) Support NumericString type for name components. 227194206Ssimon [Steve Henson] 228194206Ssimon 229194206Ssimon *) Allow CC in the environment to override the automatically chosen 230194206Ssimon compiler. Note that nothing is done to ensure flags work with the 231194206Ssimon chosen compiler. 232194206Ssimon [Ben Laurie] 233194206Ssimon 234194206Ssimon Changes between 0.9.8i and 0.9.8j [07 Jan 2009] 235194206Ssimon 236194206Ssimon *) Properly check EVP_VerifyFinal() and similar return values 237194206Ssimon (CVE-2008-5077). 238194206Ssimon [Ben Laurie, Bodo Moeller, Google Security Team] 239194206Ssimon 240194206Ssimon *) Enable TLS extensions by default. 241194206Ssimon [Ben Laurie] 242194206Ssimon 243194206Ssimon *) Allow the CHIL engine to be loaded, whether the application is 244194206Ssimon multithreaded or not. (This does not release the developer from the 245194206Ssimon obligation to set up the dynamic locking callbacks.) 246194206Ssimon [Sander Temme <sander@temme.net>] 247194206Ssimon 248194206Ssimon *) Use correct exit code if there is an error in dgst command. 249194206Ssimon [Steve Henson; problem pointed out by Roland Dirlewanger] 250194206Ssimon 251194206Ssimon *) Tweak Configure so that you need to say "experimental-jpake" to enable 252194206Ssimon JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications. 253194206Ssimon [Bodo Moeller] 254194206Ssimon 255194206Ssimon *) Add experimental JPAKE support, including demo authentication in 256194206Ssimon s_client and s_server. 257194206Ssimon [Ben Laurie] 258194206Ssimon 259194206Ssimon *) Set the comparison function in v3_addr_canonize(). 260194206Ssimon [Rob Austein <sra@hactrn.net>] 261194206Ssimon 262194206Ssimon *) Add support for XMPP STARTTLS in s_client. 263194206Ssimon [Philip Paeps <philip@freebsd.org>] 264194206Ssimon 265194206Ssimon *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior 266194206Ssimon to ensure that even with this option, only ciphersuites in the 267194206Ssimon server's preference list will be accepted. (Note that the option 268194206Ssimon applies only when resuming a session, so the earlier behavior was 269194206Ssimon just about the algorithm choice for symmetric cryptography.) 270194206Ssimon [Bodo Moeller] 271194206Ssimon 272194206Ssimon Changes between 0.9.8h and 0.9.8i [15 Sep 2008] 273194206Ssimon 274205128Ssimon *) Fix NULL pointer dereference if a DTLS server received 275205128Ssimon ChangeCipherSpec as first record (CVE-2009-1386). 276205128Ssimon [PR #1679] 277205128Ssimon 278194206Ssimon *) Fix a state transitition in s3_srvr.c and d1_srvr.c 279194206Ssimon (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...). 280194206Ssimon [Nagendra Modadugu] 281194206Ssimon 282194206Ssimon *) The fix in 0.9.8c that supposedly got rid of unsafe 283194206Ssimon double-checked locking was incomplete for RSA blinding, 284194206Ssimon addressing just one layer of what turns out to have been 285194206Ssimon doubly unsafe triple-checked locking. 286194206Ssimon 287194206Ssimon So now fix this for real by retiring the MONT_HELPER macro 288194206Ssimon in crypto/rsa/rsa_eay.c. 289194206Ssimon 290194206Ssimon [Bodo Moeller; problem pointed out by Marius Schilder] 291194206Ssimon 292194206Ssimon *) Various precautionary measures: 293194206Ssimon 294194206Ssimon - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h). 295194206Ssimon 296194206Ssimon - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c). 297194206Ssimon (NB: This would require knowledge of the secret session ticket key 298194206Ssimon to exploit, in which case you'd be SOL either way.) 299194206Ssimon 300194206Ssimon - Change bn_nist.c so that it will properly handle input BIGNUMs 301194206Ssimon outside the expected range. 302194206Ssimon 303194206Ssimon - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG 304194206Ssimon builds. 305194206Ssimon 306194206Ssimon [Neel Mehta, Bodo Moeller] 307194206Ssimon 308194206Ssimon *) Allow engines to be "soft loaded" - i.e. optionally don't die if 309194206Ssimon the load fails. Useful for distros. 310194206Ssimon [Ben Laurie and the FreeBSD team] 311194206Ssimon 312194206Ssimon *) Add support for Local Machine Keyset attribute in PKCS#12 files. 313194206Ssimon [Steve Henson] 314194206Ssimon 315194206Ssimon *) Fix BN_GF2m_mod_arr() top-bit cleanup code. 316194206Ssimon [Huang Ying] 317194206Ssimon 318194206Ssimon *) Expand ENGINE to support engine supplied SSL client certificate functions. 319194206Ssimon 320194206Ssimon This work was sponsored by Logica. 321194206Ssimon [Steve Henson] 322194206Ssimon 323194206Ssimon *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows 324194206Ssimon keystores. Support for SSL/TLS client authentication too. 325194206Ssimon Not compiled unless enable-capieng specified to Configure. 326194206Ssimon 327194206Ssimon This work was sponsored by Logica. 328194206Ssimon [Steve Henson] 329194206Ssimon 330194206Ssimon *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using 331194206Ssimon ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain 332194206Ssimon attribute creation routines such as certifcate requests and PKCS#12 333194206Ssimon files. 334194206Ssimon [Steve Henson] 335194206Ssimon 336194206Ssimon Changes between 0.9.8g and 0.9.8h [28 May 2008] 337194206Ssimon 338194206Ssimon *) Fix flaw if 'Server Key exchange message' is omitted from a TLS 339194206Ssimon handshake which could lead to a cilent crash as found using the 340194206Ssimon Codenomicon TLS test suite (CVE-2008-1672) 341194206Ssimon [Steve Henson, Mark Cox] 342194206Ssimon 343194206Ssimon *) Fix double free in TLS server name extensions which could lead to 344194206Ssimon a remote crash found by Codenomicon TLS test suite (CVE-2008-0891) 345194206Ssimon [Joe Orton] 346194206Ssimon 347194206Ssimon *) Clear error queue in SSL_CTX_use_certificate_chain_file() 348194206Ssimon 349194206Ssimon Clear the error queue to ensure that error entries left from 350194206Ssimon older function calls do not interfere with the correct operation. 351194206Ssimon [Lutz Jaenicke, Erik de Castro Lopo] 352194206Ssimon 353194206Ssimon *) Remove root CA certificates of commercial CAs: 354194206Ssimon 355194206Ssimon The OpenSSL project does not recommend any specific CA and does not 356194206Ssimon have any policy with respect to including or excluding any CA. 357194206Ssimon Therefore it does not make any sense to ship an arbitrary selection 358194206Ssimon of root CA certificates with the OpenSSL software. 359194206Ssimon [Lutz Jaenicke] 360194206Ssimon 361194206Ssimon *) RSA OAEP patches to fix two separate invalid memory reads. 362194206Ssimon The first one involves inputs when 'lzero' is greater than 363194206Ssimon 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes 364194206Ssimon before the beginning of from). The second one involves inputs where 365194206Ssimon the 'db' section contains nothing but zeroes (there is a one-byte 366194206Ssimon invalid read after the end of 'db'). 367194206Ssimon [Ivan Nestlerode <inestlerode@us.ibm.com>] 368194206Ssimon 369194206Ssimon *) Partial backport from 0.9.9-dev: 370194206Ssimon 371194206Ssimon Introduce bn_mul_mont (dedicated Montgomery multiplication 372194206Ssimon procedure) as a candidate for BIGNUM assembler implementation. 373194206Ssimon While 0.9.9-dev uses assembler for various architectures, only 374194206Ssimon x86_64 is available by default here in the 0.9.8 branch, and 375194206Ssimon 32-bit x86 is available through a compile-time setting. 376194206Ssimon 377194206Ssimon To try the 32-bit x86 assembler implementation, use Configure 378194206Ssimon option "enable-montasm" (which exists only for this backport). 379194206Ssimon 380194206Ssimon As "enable-montasm" for 32-bit x86 disclaims code stability 381194206Ssimon anyway, in this constellation we activate additional code 382194206Ssimon backported from 0.9.9-dev for further performance improvements, 383194206Ssimon namely BN_from_montgomery_word. (To enable this otherwise, 384194206Ssimon e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".) 385194206Ssimon 386194206Ssimon [Andy Polyakov (backport partially by Bodo Moeller)] 387194206Ssimon 388194206Ssimon *) Add TLS session ticket callback. This allows an application to set 389194206Ssimon TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed 390194206Ssimon values. This is useful for key rollover for example where several key 391194206Ssimon sets may exist with different names. 392194206Ssimon [Steve Henson] 393194206Ssimon 394194206Ssimon *) Reverse ENGINE-internal logic for caching default ENGINE handles. 395194206Ssimon This was broken until now in 0.9.8 releases, such that the only way 396194206Ssimon a registered ENGINE could be used (assuming it initialises 397194206Ssimon successfully on the host) was to explicitly set it as the default 398194206Ssimon for the relevant algorithms. This is in contradiction with 0.9.7 399194206Ssimon behaviour and the documentation. With this fix, when an ENGINE is 400194206Ssimon registered into a given algorithm's table of implementations, the 401194206Ssimon 'uptodate' flag is reset so that auto-discovery will be used next 402194206Ssimon time a new context for that algorithm attempts to select an 403194206Ssimon implementation. 404194206Ssimon [Ian Lister (tweaked by Geoff Thorpe)] 405194206Ssimon 406194206Ssimon *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9 407194206Ssimon implemention in the following ways: 408194206Ssimon 409194206Ssimon Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be 410194206Ssimon hard coded. 411194206Ssimon 412194206Ssimon Lack of BER streaming support means one pass streaming processing is 413194206Ssimon only supported if data is detached: setting the streaming flag is 414194206Ssimon ignored for embedded content. 415194206Ssimon 416194206Ssimon CMS support is disabled by default and must be explicitly enabled 417194206Ssimon with the enable-cms configuration option. 418194206Ssimon [Steve Henson] 419194206Ssimon 420194206Ssimon *) Update the GMP engine glue to do direct copies between BIGNUM and 421194206Ssimon mpz_t when openssl and GMP use the same limb size. Otherwise the 422194206Ssimon existing "conversion via a text string export" trick is still used. 423194206Ssimon [Paul Sheer <paulsheer@gmail.com>] 424194206Ssimon 425194206Ssimon *) Zlib compression BIO. This is a filter BIO which compressed and 426194206Ssimon uncompresses any data passed through it. 427194206Ssimon [Steve Henson] 428194206Ssimon 429194206Ssimon *) Add AES_wrap_key() and AES_unwrap_key() functions to implement 430194206Ssimon RFC3394 compatible AES key wrapping. 431194206Ssimon [Steve Henson] 432194206Ssimon 433194206Ssimon *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0(): 434194206Ssimon sets string data without copying. X509_ALGOR_set0() and 435194206Ssimon X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier) 436194206Ssimon data. Attribute function X509at_get0_data_by_OBJ(): retrieves data 437194206Ssimon from an X509_ATTRIBUTE structure optionally checking it occurs only 438194206Ssimon once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied 439194206Ssimon data. 440194206Ssimon [Steve Henson] 441194206Ssimon 442194206Ssimon *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set() 443194206Ssimon to get the expected BN_FLG_CONSTTIME behavior. 444194206Ssimon [Bodo Moeller (Google)] 445194206Ssimon 446194206Ssimon *) Netware support: 447194206Ssimon 448194206Ssimon - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets 449194206Ssimon - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT) 450194206Ssimon - added some more tests to do_tests.pl 451194206Ssimon - fixed RunningProcess usage so that it works with newer LIBC NDKs too 452194206Ssimon - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency 453194206Ssimon - added new Configure targets netware-clib-bsdsock, netware-clib-gcc, 454194206Ssimon netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc 455194206Ssimon - various changes to netware.pl to enable gcc-cross builds on Win32 456194206Ssimon platform 457194206Ssimon - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD) 458194206Ssimon - various changes to fix missing prototype warnings 459194206Ssimon - fixed x86nasm.pl to create correct asm files for NASM COFF output 460194206Ssimon - added AES, WHIRLPOOL and CPUID assembler code to build files 461194206Ssimon - added missing AES assembler make rules to mk1mf.pl 462194206Ssimon - fixed order of includes in apps/ocsp.c so that e_os.h settings apply 463194206Ssimon [Guenter Knauf <eflash@gmx.net>] 464194206Ssimon 465194206Ssimon *) Implement certificate status request TLS extension defined in RFC3546. 466194206Ssimon A client can set the appropriate parameters and receive the encoded 467194206Ssimon OCSP response via a callback. A server can query the supplied parameters 468194206Ssimon and set the encoded OCSP response in the callback. Add simplified examples 469194206Ssimon to s_client and s_server. 470194206Ssimon [Steve Henson] 471194206Ssimon 472194206Ssimon Changes between 0.9.8f and 0.9.8g [19 Oct 2007] 473194206Ssimon 474194206Ssimon *) Fix various bugs: 475194206Ssimon + Binary incompatibility of ssl_ctx_st structure 476194206Ssimon + DTLS interoperation with non-compliant servers 477194206Ssimon + Don't call get_session_cb() without proposed session 478194206Ssimon + Fix ia64 assembler code 479194206Ssimon [Andy Polyakov, Steve Henson] 480194206Ssimon 481194206Ssimon Changes between 0.9.8e and 0.9.8f [11 Oct 2007] 482194206Ssimon 483194206Ssimon *) DTLS Handshake overhaul. There were longstanding issues with 484194206Ssimon OpenSSL DTLS implementation, which were making it impossible for 485194206Ssimon RFC 4347 compliant client to communicate with OpenSSL server. 486194206Ssimon Unfortunately just fixing these incompatibilities would "cut off" 487194206Ssimon pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e 488194206Ssimon server keeps tolerating non RFC compliant syntax. The opposite is 489194206Ssimon not true, 0.9.8f client can not communicate with earlier server. 490194206Ssimon This update even addresses CVE-2007-4995. 491194206Ssimon [Andy Polyakov] 492194206Ssimon 493194206Ssimon *) Changes to avoid need for function casts in OpenSSL: some compilers 494194206Ssimon (gcc 4.2 and later) reject their use. 495194206Ssimon [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>, 496194206Ssimon Steve Henson] 497194206Ssimon 498194206Ssimon *) Add RFC4507 support to OpenSSL. This includes the corrections in 499194206Ssimon RFC4507bis. The encrypted ticket format is an encrypted encoded 500194206Ssimon SSL_SESSION structure, that way new session features are automatically 501194206Ssimon supported. 502194206Ssimon 503194206Ssimon If a client application caches session in an SSL_SESSION structure 504194206Ssimon support is transparent because tickets are now stored in the encoded 505194206Ssimon SSL_SESSION. 506194206Ssimon 507194206Ssimon The SSL_CTX structure automatically generates keys for ticket 508194206Ssimon protection in servers so again support should be possible 509194206Ssimon with no application modification. 510194206Ssimon 511194206Ssimon If a client or server wishes to disable RFC4507 support then the option 512194206Ssimon SSL_OP_NO_TICKET can be set. 513194206Ssimon 514194206Ssimon Add a TLS extension debugging callback to allow the contents of any client 515194206Ssimon or server extensions to be examined. 516194206Ssimon 517194206Ssimon This work was sponsored by Google. 518194206Ssimon [Steve Henson] 519194206Ssimon 520194206Ssimon *) Add initial support for TLS extensions, specifically for the server_name 521194206Ssimon extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 522194206Ssimon have new members for a host name. The SSL data structure has an 523194206Ssimon additional member SSL_CTX *initial_ctx so that new sessions can be 524194206Ssimon stored in that context to allow for session resumption, even after the 525194206Ssimon SSL has been switched to a new SSL_CTX in reaction to a client's 526194206Ssimon server_name extension. 527194206Ssimon 528194206Ssimon New functions (subject to change): 529194206Ssimon 530194206Ssimon SSL_get_servername() 531194206Ssimon SSL_get_servername_type() 532194206Ssimon SSL_set_SSL_CTX() 533194206Ssimon 534194206Ssimon New CTRL codes and macros (subject to change): 535194206Ssimon 536194206Ssimon SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 537194206Ssimon - SSL_CTX_set_tlsext_servername_callback() 538194206Ssimon SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 539194206Ssimon - SSL_CTX_set_tlsext_servername_arg() 540194206Ssimon SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 541194206Ssimon 542194206Ssimon openssl s_client has a new '-servername ...' option. 543194206Ssimon 544194206Ssimon openssl s_server has new options '-servername_host ...', '-cert2 ...', 545194206Ssimon '-key2 ...', '-servername_fatal' (subject to change). This allows 546194206Ssimon testing the HostName extension for a specific single host name ('-cert' 547194206Ssimon and '-key' remain fallbacks for handshakes without HostName 548194206Ssimon negotiation). If the unrecogninzed_name alert has to be sent, this by 549194206Ssimon default is a warning; it becomes fatal with the '-servername_fatal' 550194206Ssimon option. 551194206Ssimon 552194206Ssimon [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson] 553194206Ssimon 554194206Ssimon *) Add AES and SSE2 assembly language support to VC++ build. 555194206Ssimon [Steve Henson] 556194206Ssimon 557194206Ssimon *) Mitigate attack on final subtraction in Montgomery reduction. 558194206Ssimon [Andy Polyakov] 559194206Ssimon 560194206Ssimon *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0 561194206Ssimon (which previously caused an internal error). 562194206Ssimon [Bodo Moeller] 563194206Ssimon 564194206Ssimon *) Squeeze another 10% out of IGE mode when in != out. 565194206Ssimon [Ben Laurie] 566194206Ssimon 567194206Ssimon *) AES IGE mode speedup. 568194206Ssimon [Dean Gaudet (Google)] 569194206Ssimon 570194206Ssimon *) Add the Korean symmetric 128-bit cipher SEED (see 571194206Ssimon http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and 572194206Ssimon add SEED ciphersuites from RFC 4162: 573194206Ssimon 574194206Ssimon TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA" 575194206Ssimon TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA" 576194206Ssimon TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA" 577194206Ssimon TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA" 578194206Ssimon 579194206Ssimon To minimize changes between patchlevels in the OpenSSL 0.9.8 580194206Ssimon series, SEED remains excluded from compilation unless OpenSSL 581194206Ssimon is configured with 'enable-seed'. 582194206Ssimon [KISA, Bodo Moeller] 583194206Ssimon 584194206Ssimon *) Mitigate branch prediction attacks, which can be practical if a 585194206Ssimon single processor is shared, allowing a spy process to extract 586194206Ssimon information. For detailed background information, see 587194206Ssimon http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron, 588194206Ssimon J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL 589194206Ssimon and Necessary Software Countermeasures"). The core of the change 590194206Ssimon are new versions BN_div_no_branch() and 591194206Ssimon BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(), 592194206Ssimon respectively, which are slower, but avoid the security-relevant 593194206Ssimon conditional branches. These are automatically called by BN_div() 594194206Ssimon and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one 595194206Ssimon of the input BIGNUMs. Also, BN_is_bit_set() has been changed to 596194206Ssimon remove a conditional branch. 597194206Ssimon 598194206Ssimon BN_FLG_CONSTTIME is the new name for the previous 599194206Ssimon BN_FLG_EXP_CONSTTIME flag, since it now affects more than just 600194206Ssimon modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag 601194206Ssimon in the exponent causes BN_mod_exp_mont() to use the alternative 602194206Ssimon implementation in BN_mod_exp_mont_consttime().) The old name 603194206Ssimon remains as a deprecated alias. 604194206Ssimon 605194206Ssimon Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general 606194206Ssimon RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses 607194206Ssimon constant-time implementations for more than just exponentiation. 608194206Ssimon Here too the old name is kept as a deprecated alias. 609194206Ssimon 610194206Ssimon BN_BLINDING_new() will now use BN_dup() for the modulus so that 611194206Ssimon the BN_BLINDING structure gets an independent copy of the 612194206Ssimon modulus. This means that the previous "BIGNUM *m" argument to 613194206Ssimon BN_BLINDING_new() and to BN_BLINDING_create_param() now 614194206Ssimon essentially becomes "const BIGNUM *m", although we can't actually 615194206Ssimon change this in the header file before 0.9.9. It allows 616194206Ssimon RSA_setup_blinding() to use BN_with_flags() on the modulus to 617194206Ssimon enable BN_FLG_CONSTTIME. 618194206Ssimon 619194206Ssimon [Matthew D Wood (Intel Corp)] 620194206Ssimon 621194206Ssimon *) In the SSL/TLS server implementation, be strict about session ID 622194206Ssimon context matching (which matters if an application uses a single 623194206Ssimon external cache for different purposes). Previously, 624194206Ssimon out-of-context reuse was forbidden only if SSL_VERIFY_PEER was 625194206Ssimon set. This did ensure strict client verification, but meant that, 626194206Ssimon with applications using a single external cache for quite 627194206Ssimon different requirements, clients could circumvent ciphersuite 628194206Ssimon restrictions for a given session ID context by starting a session 629194206Ssimon in a different context. 630194206Ssimon [Bodo Moeller] 631194206Ssimon 632167612Ssimon *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 633167612Ssimon a ciphersuite string such as "DEFAULT:RSA" cannot enable 634167612Ssimon authentication-only ciphersuites. 635167612Ssimon [Bodo Moeller] 636167612Ssimon 637194206Ssimon *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was 638194206Ssimon not complete and could lead to a possible single byte overflow 639194206Ssimon (CVE-2007-5135) [Ben Laurie] 640194206Ssimon 641194206Ssimon Changes between 0.9.8d and 0.9.8e [23 Feb 2007] 642194206Ssimon 643167612Ssimon *) Since AES128 and AES256 (and similarly Camellia128 and 644167612Ssimon Camellia256) share a single mask bit in the logic of 645167612Ssimon ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 646167612Ssimon kludge to work properly if AES128 is available and AES256 isn't 647167612Ssimon (or if Camellia128 is available and Camellia256 isn't). 648167612Ssimon [Victor Duchovni] 649167612Ssimon 650167612Ssimon *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c 651167612Ssimon (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters): 652167612Ssimon When a point or a seed is encoded in a BIT STRING, we need to 653167612Ssimon prevent the removal of trailing zero bits to get the proper DER 654167612Ssimon encoding. (By default, crypto/asn1/a_bitstr.c assumes the case 655167612Ssimon of a NamedBitList, for which trailing 0 bits need to be removed.) 656167612Ssimon [Bodo Moeller] 657167612Ssimon 658167612Ssimon *) Have SSL/TLS server implementation tolerate "mismatched" record 659167612Ssimon protocol version while receiving ClientHello even if the 660167612Ssimon ClientHello is fragmented. (The server can't insist on the 661167612Ssimon particular protocol version it has chosen before the ServerHello 662167612Ssimon message has informed the client about his choice.) 663167612Ssimon [Bodo Moeller] 664167612Ssimon 665167612Ssimon *) Add RFC 3779 support. 666167612Ssimon [Rob Austein for ARIN, Ben Laurie] 667167612Ssimon 668167612Ssimon *) Load error codes if they are not already present instead of using a 669167612Ssimon static variable. This allows them to be cleanly unloaded and reloaded. 670167612Ssimon Improve header file function name parsing. 671167612Ssimon [Steve Henson] 672167612Ssimon 673167612Ssimon *) extend SMTP and IMAP protocol emulation in s_client to use EHLO 674167612Ssimon or CAPABILITY handshake as required by RFCs. 675167612Ssimon [Goetz Babin-Ebell] 676167612Ssimon 677162911Ssimon Changes between 0.9.8c and 0.9.8d [28 Sep 2006] 678162911Ssimon 679162911Ssimon *) Introduce limits to prevent malicious keys being able to 680162911Ssimon cause a denial of service. (CVE-2006-2940) 681162911Ssimon [Steve Henson, Bodo Moeller] 682162911Ssimon 683162911Ssimon *) Fix ASN.1 parsing of certain invalid structures that can result 684162911Ssimon in a denial of service. (CVE-2006-2937) [Steve Henson] 685162911Ssimon 686162911Ssimon *) Fix buffer overflow in SSL_get_shared_ciphers() function. 687162911Ssimon (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 688162911Ssimon 689162911Ssimon *) Fix SSL client code which could crash if connecting to a 690162911Ssimon malicious SSLv2 server. (CVE-2006-4343) 691162911Ssimon [Tavis Ormandy and Will Drewry, Google Security Team] 692162911Ssimon 693162911Ssimon *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites 694162911Ssimon match only those. Before that, "AES256-SHA" would be interpreted 695162911Ssimon as a pattern and match "AES128-SHA" too (since AES128-SHA got 696162911Ssimon the same strength classification in 0.9.7h) as we currently only 697162911Ssimon have a single AES bit in the ciphersuite description bitmap. 698162911Ssimon That change, however, also applied to ciphersuite strings such as 699162911Ssimon "RC4-MD5" that intentionally matched multiple ciphersuites -- 700162911Ssimon namely, SSL 2.0 ciphersuites in addition to the more common ones 701162911Ssimon from SSL 3.0/TLS 1.0. 702162911Ssimon 703162911Ssimon So we change the selection algorithm again: Naming an explicit 704162911Ssimon ciphersuite selects this one ciphersuite, and any other similar 705162911Ssimon ciphersuite (same bitmap) from *other* protocol versions. 706162911Ssimon Thus, "RC4-MD5" again will properly select both the SSL 2.0 707162911Ssimon ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite. 708162911Ssimon 709162911Ssimon Since SSL 2.0 does not have any ciphersuites for which the 710162911Ssimon 128/256 bit distinction would be relevant, this works for now. 711162911Ssimon The proper fix will be to use different bits for AES128 and 712162911Ssimon AES256, which would have avoided the problems from the beginning; 713162911Ssimon however, bits are scarce, so we can only do this in a new release 714162911Ssimon (not just a patchlevel) when we can change the SSL_CIPHER 715162911Ssimon definition to split the single 'unsigned long mask' bitmap into 716162911Ssimon multiple values to extend the available space. 717162911Ssimon 718162911Ssimon [Bodo Moeller] 719162911Ssimon 720162911Ssimon Changes between 0.9.8b and 0.9.8c [05 Sep 2006] 721162911Ssimon 722162911Ssimon *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 723162911Ssimon (CVE-2006-4339) [Ben Laurie and Google Security Team] 724162911Ssimon 725162911Ssimon *) Add AES IGE and biIGE modes. 726162911Ssimon [Ben Laurie] 727162911Ssimon 728162911Ssimon *) Change the Unix randomness entropy gathering to use poll() when 729162911Ssimon possible instead of select(), since the latter has some 730162911Ssimon undesirable limitations. 731162911Ssimon [Darryl Miles via Richard Levitte and Bodo Moeller] 732162911Ssimon 733162911Ssimon *) Disable "ECCdraft" ciphersuites more thoroughly. Now special 734162911Ssimon treatment in ssl/ssl_ciph.s makes sure that these ciphersuites 735162911Ssimon cannot be implicitly activated as part of, e.g., the "AES" alias. 736162911Ssimon However, please upgrade to OpenSSL 0.9.9[-dev] for 737162911Ssimon non-experimental use of the ECC ciphersuites to get TLS extension 738162911Ssimon support, which is required for curve and point format negotiation 739162911Ssimon to avoid potential handshake problems. 740162911Ssimon [Bodo Moeller] 741162911Ssimon 742162911Ssimon *) Disable rogue ciphersuites: 743162911Ssimon 744162911Ssimon - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 745162911Ssimon - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 746162911Ssimon - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 747162911Ssimon 748162911Ssimon The latter two were purportedly from 749162911Ssimon draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 750162911Ssimon appear there. 751162911Ssimon 752167612Ssimon Also deactivate the remaining ciphersuites from 753162911Ssimon draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 754162911Ssimon unofficial, and the ID has long expired. 755162911Ssimon [Bodo Moeller] 756162911Ssimon 757162911Ssimon *) Fix RSA blinding Heisenbug (problems sometimes occured on 758162911Ssimon dual-core machines) and other potential thread-safety issues. 759162911Ssimon [Bodo Moeller] 760162911Ssimon 761162911Ssimon *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key 762162911Ssimon versions), which is now available for royalty-free use 763162911Ssimon (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html). 764162911Ssimon Also, add Camellia TLS ciphersuites from RFC 4132. 765162911Ssimon 766162911Ssimon To minimize changes between patchlevels in the OpenSSL 0.9.8 767162911Ssimon series, Camellia remains excluded from compilation unless OpenSSL 768162911Ssimon is configured with 'enable-camellia'. 769162911Ssimon [NTT] 770162911Ssimon 771162911Ssimon *) Disable the padding bug check when compression is in use. The padding 772162911Ssimon bug check assumes the first packet is of even length, this is not 773162911Ssimon necessarily true if compresssion is enabled and can result in false 774162911Ssimon positives causing handshake failure. The actual bug test is ancient 775162911Ssimon code so it is hoped that implementations will either have fixed it by 776162911Ssimon now or any which still have the bug do not support compression. 777162911Ssimon [Steve Henson] 778162911Ssimon 779160814Ssimon Changes between 0.9.8a and 0.9.8b [04 May 2006] 780160814Ssimon 781160814Ssimon *) When applying a cipher rule check to see if string match is an explicit 782160814Ssimon cipher suite and only match that one cipher suite if it is. 783160814Ssimon [Steve Henson] 784160814Ssimon 785160814Ssimon *) Link in manifests for VC++ if needed. 786160814Ssimon [Austin Ziegler <halostatue@gmail.com>] 787160814Ssimon 788160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 789160814Ssimon draft-ietf-tls-ecc-12.txt with proposed changes (but without 790160814Ssimon TLS extensions, which are supported starting with the 0.9.9 791160814Ssimon branch, not in the OpenSSL 0.9.8 branch). 792160814Ssimon [Douglas Stebila] 793160814Ssimon 794160814Ssimon *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support 795160814Ssimon opaque EVP_CIPHER_CTX handling. 796160814Ssimon [Steve Henson] 797160814Ssimon 798160814Ssimon *) Fixes and enhancements to zlib compression code. We now only use 799160814Ssimon "zlib1.dll" and use the default __cdecl calling convention on Win32 800160814Ssimon to conform with the standards mentioned here: 801160814Ssimon http://www.zlib.net/DLL_FAQ.txt 802160814Ssimon Static zlib linking now works on Windows and the new --with-zlib-include 803160814Ssimon --with-zlib-lib options to Configure can be used to supply the location 804160814Ssimon of the headers and library. Gracefully handle case where zlib library 805160814Ssimon can't be loaded. 806160814Ssimon [Steve Henson] 807160814Ssimon 808160814Ssimon *) Several fixes and enhancements to the OID generation code. The old code 809160814Ssimon sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't 810160814Ssimon handle numbers larger than ULONG_MAX, truncated printing and had a 811160814Ssimon non standard OBJ_obj2txt() behaviour. 812160814Ssimon [Steve Henson] 813160814Ssimon 814160814Ssimon *) Add support for building of engines under engine/ as shared libraries 815160814Ssimon under VC++ build system. 816160814Ssimon [Steve Henson] 817160814Ssimon 818160814Ssimon *) Corrected the numerous bugs in the Win32 path splitter in DSO. 819160814Ssimon Hopefully, we will not see any false combination of paths any more. 820160814Ssimon [Richard Levitte] 821160814Ssimon 822160814Ssimon Changes between 0.9.8 and 0.9.8a [11 Oct 2005] 823160814Ssimon 824160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 825160814Ssimon (part of SSL_OP_ALL). This option used to disable the 826160814Ssimon countermeasure against man-in-the-middle protocol-version 827160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 828160814Ssimon idea. (CVE-2005-2969) 829160814Ssimon 830160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 831160814Ssimon for Information Security, National Institute of Advanced Industrial 832160814Ssimon Science and Technology [AIST], Japan)] 833160814Ssimon 834160814Ssimon *) Add two function to clear and return the verify parameter flags. 835160814Ssimon [Steve Henson] 836160814Ssimon 837160814Ssimon *) Keep cipherlists sorted in the source instead of sorting them at 838160814Ssimon runtime, thus removing the need for a lock. 839160814Ssimon [Nils Larsch] 840160814Ssimon 841160814Ssimon *) Avoid some small subgroup attacks in Diffie-Hellman. 842160814Ssimon [Nick Mathewson and Ben Laurie] 843160814Ssimon 844160814Ssimon *) Add functions for well-known primes. 845160814Ssimon [Nick Mathewson] 846160814Ssimon 847160814Ssimon *) Extended Windows CE support. 848160814Ssimon [Satoshi Nakamura and Andy Polyakov] 849160814Ssimon 850160814Ssimon *) Initialize SSL_METHOD structures at compile time instead of during 851160814Ssimon runtime, thus removing the need for a lock. 852160814Ssimon [Steve Henson] 853160814Ssimon 854160814Ssimon *) Make PKCS7_decrypt() work even if no certificate is supplied by 855160814Ssimon attempting to decrypt each encrypted key in turn. Add support to 856160814Ssimon smime utility. 857160814Ssimon [Steve Henson] 858160814Ssimon 859160814Ssimon Changes between 0.9.7h and 0.9.8 [05 Jul 2005] 860160814Ssimon 861162911Ssimon [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after 862162911Ssimon OpenSSL 0.9.8.] 863162911Ssimon 864160814Ssimon *) Add libcrypto.pc and libssl.pc for those who feel they need them. 865160814Ssimon [Richard Levitte] 866160814Ssimon 867160814Ssimon *) Change CA.sh and CA.pl so they don't bundle the CSR and the private 868160814Ssimon key into the same file any more. 869160814Ssimon [Richard Levitte] 870160814Ssimon 871160814Ssimon *) Add initial support for Win64, both IA64 and AMD64/x64 flavors. 872160814Ssimon [Andy Polyakov] 873160814Ssimon 874160814Ssimon *) Add -utf8 command line and config file option to 'ca'. 875160814Ssimon [Stefan <stf@udoma.org] 876160814Ssimon 877160814Ssimon *) Removed the macro des_crypt(), as it seems to conflict with some 878160814Ssimon libraries. Use DES_crypt(). 879160814Ssimon [Richard Levitte] 880160814Ssimon 881160814Ssimon *) Correct naming of the 'chil' and '4758cca' ENGINEs. This 882160814Ssimon involves renaming the source and generated shared-libs for 883160814Ssimon both. The engines will accept the corrected or legacy ids 884160814Ssimon ('ncipher' and '4758_cca' respectively) when binding. NB, 885160814Ssimon this only applies when building 'shared'. 886160814Ssimon [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe] 887160814Ssimon 888160814Ssimon *) Add attribute functions to EVP_PKEY structure. Modify 889160814Ssimon PKCS12_create() to recognize a CSP name attribute and 890160814Ssimon use it. Make -CSP option work again in pkcs12 utility. 891160814Ssimon [Steve Henson] 892160814Ssimon 893160814Ssimon *) Add new functionality to the bn blinding code: 894160814Ssimon - automatic re-creation of the BN_BLINDING parameters after 895160814Ssimon a fixed number of uses (currently 32) 896160814Ssimon - add new function for parameter creation 897160814Ssimon - introduce flags to control the update behaviour of the 898160814Ssimon BN_BLINDING parameters 899160814Ssimon - hide BN_BLINDING structure 900160814Ssimon Add a second BN_BLINDING slot to the RSA structure to improve 901160814Ssimon performance when a single RSA object is shared among several 902160814Ssimon threads. 903160814Ssimon [Nils Larsch] 904160814Ssimon 905160814Ssimon *) Add support for DTLS. 906160814Ssimon [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie] 907160814Ssimon 908160814Ssimon *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1) 909160814Ssimon to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file() 910160814Ssimon [Walter Goulet] 911160814Ssimon 912160814Ssimon *) Remove buggy and incompletet DH cert support from 913160814Ssimon ssl/ssl_rsa.c and ssl/s3_both.c 914160814Ssimon [Nils Larsch] 915160814Ssimon 916160814Ssimon *) Use SHA-1 instead of MD5 as the default digest algorithm for 917160814Ssimon the apps/openssl applications. 918160814Ssimon [Nils Larsch] 919160814Ssimon 920160814Ssimon *) Compile clean with "-Wall -Wmissing-prototypes 921160814Ssimon -Wstrict-prototypes -Wmissing-declarations -Werror". Currently 922160814Ssimon DEBUG_SAFESTACK must also be set. 923160814Ssimon [Ben Laurie] 924160814Ssimon 925160814Ssimon *) Change ./Configure so that certain algorithms can be disabled by default. 926160814Ssimon The new counterpiece to "no-xxx" is "enable-xxx". 927160814Ssimon 928160814Ssimon The patented RC5 and MDC2 algorithms will now be disabled unless 929160814Ssimon "enable-rc5" and "enable-mdc2", respectively, are specified. 930160814Ssimon 931160814Ssimon (IDEA remains enabled despite being patented. This is because IDEA 932160814Ssimon is frequently required for interoperability, and there is no license 933160814Ssimon fee for non-commercial use. As before, "no-idea" can be used to 934160814Ssimon avoid this algorithm.) 935160814Ssimon 936160814Ssimon [Bodo Moeller] 937160814Ssimon 938160814Ssimon *) Add processing of proxy certificates (see RFC 3820). This work was 939160814Ssimon sponsored by KTH (The Royal Institute of Technology in Stockholm) and 940160814Ssimon EGEE (Enabling Grids for E-science in Europe). 941160814Ssimon [Richard Levitte] 942160814Ssimon 943160814Ssimon *) RC4 performance overhaul on modern architectures/implementations, such 944160814Ssimon as Intel P4, IA-64 and AMD64. 945160814Ssimon [Andy Polyakov] 946160814Ssimon 947160814Ssimon *) New utility extract-section.pl. This can be used specify an alternative 948160814Ssimon section number in a pod file instead of having to treat each file as 949160814Ssimon a separate case in Makefile. This can be done by adding two lines to the 950160814Ssimon pod file: 951160814Ssimon 952160814Ssimon =for comment openssl_section:XXX 953160814Ssimon 954160814Ssimon The blank line is mandatory. 955160814Ssimon 956160814Ssimon [Steve Henson] 957160814Ssimon 958160814Ssimon *) New arguments -certform, -keyform and -pass for s_client and s_server 959160814Ssimon to allow alternative format key and certificate files and passphrase 960160814Ssimon sources. 961160814Ssimon [Steve Henson] 962160814Ssimon 963160814Ssimon *) New structure X509_VERIFY_PARAM which combines current verify parameters, 964160814Ssimon update associated structures and add various utility functions. 965160814Ssimon 966160814Ssimon Add new policy related verify parameters, include policy checking in 967160814Ssimon standard verify code. Enhance 'smime' application with extra parameters 968160814Ssimon to support policy checking and print out. 969160814Ssimon [Steve Henson] 970160814Ssimon 971160814Ssimon *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3 972160814Ssimon Nehemiah processors. These extensions support AES encryption in hardware 973160814Ssimon as well as RNG (though RNG support is currently disabled). 974160814Ssimon [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov] 975160814Ssimon 976160814Ssimon *) Deprecate BN_[get|set]_params() functions (they were ignored internally). 977160814Ssimon [Geoff Thorpe] 978160814Ssimon 979160814Ssimon *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented. 980160814Ssimon [Andy Polyakov and a number of other people] 981160814Ssimon 982160814Ssimon *) Improved PowerPC platform support. Most notably BIGNUM assembler 983160814Ssimon implementation contributed by IBM. 984160814Ssimon [Suresh Chari, Peter Waltenberg, Andy Polyakov] 985160814Ssimon 986160814Ssimon *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public 987160814Ssimon exponent rather than 'unsigned long'. There is a corresponding change to 988160814Ssimon the new 'rsa_keygen' element of the RSA_METHOD structure. 989160814Ssimon [Jelte Jansen, Geoff Thorpe] 990160814Ssimon 991160814Ssimon *) Functionality for creating the initial serial number file is now 992160814Ssimon moved from CA.pl to the 'ca' utility with a new option -create_serial. 993160814Ssimon 994160814Ssimon (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial 995160814Ssimon number file to 1, which is bound to cause problems. To avoid 996160814Ssimon the problems while respecting compatibility between different 0.9.7 997160814Ssimon patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in 998160814Ssimon CA.pl for serial number initialization. With the new release 0.9.8, 999160814Ssimon we can fix the problem directly in the 'ca' utility.) 1000160814Ssimon [Steve Henson] 1001160814Ssimon 1002160814Ssimon *) Reduced header interdepencies by declaring more opaque objects in 1003160814Ssimon ossl_typ.h. As a consequence, including some headers (eg. engine.h) will 1004160814Ssimon give fewer recursive includes, which could break lazy source code - so 1005160814Ssimon this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always, 1006160814Ssimon developers should define this symbol when building and using openssl to 1007160814Ssimon ensure they track the recommended behaviour, interfaces, [etc], but 1008160814Ssimon backwards-compatible behaviour prevails when this isn't defined. 1009160814Ssimon [Geoff Thorpe] 1010160814Ssimon 1011160814Ssimon *) New function X509_POLICY_NODE_print() which prints out policy nodes. 1012160814Ssimon [Steve Henson] 1013160814Ssimon 1014160814Ssimon *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality. 1015160814Ssimon This will generate a random key of the appropriate length based on the 1016160814Ssimon cipher context. The EVP_CIPHER can provide its own random key generation 1017160814Ssimon routine to support keys of a specific form. This is used in the des and 1018160814Ssimon 3des routines to generate a key of the correct parity. Update S/MIME 1019160814Ssimon code to use new functions and hence generate correct parity DES keys. 1020160814Ssimon Add EVP_CHECK_DES_KEY #define to return an error if the key is not 1021160814Ssimon valid (weak or incorrect parity). 1022160814Ssimon [Steve Henson] 1023160814Ssimon 1024160814Ssimon *) Add a local set of CRLs that can be used by X509_verify_cert() as well 1025160814Ssimon as looking them up. This is useful when the verified structure may contain 1026160814Ssimon CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs 1027160814Ssimon present unless the new PKCS7_NO_CRL flag is asserted. 1028160814Ssimon [Steve Henson] 1029160814Ssimon 1030160814Ssimon *) Extend ASN1 oid configuration module. It now additionally accepts the 1031160814Ssimon syntax: 1032160814Ssimon 1033160814Ssimon shortName = some long name, 1.2.3.4 1034160814Ssimon [Steve Henson] 1035160814Ssimon 1036160814Ssimon *) Reimplemented the BN_CTX implementation. There is now no more static 1037160814Ssimon limitation on the number of variables it can handle nor the depth of the 1038160814Ssimon "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack 1039160814Ssimon information can now expand as required, and rather than having a single 1040160814Ssimon static array of bignums, BN_CTX now uses a linked-list of such arrays 1041160814Ssimon allowing it to expand on demand whilst maintaining the usefulness of 1042160814Ssimon BN_CTX's "bundling". 1043160814Ssimon [Geoff Thorpe] 1044160814Ssimon 1045160814Ssimon *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD 1046160814Ssimon to allow all RSA operations to function using a single BN_CTX. 1047160814Ssimon [Geoff Thorpe] 1048160814Ssimon 1049160814Ssimon *) Preliminary support for certificate policy evaluation and checking. This 1050160814Ssimon is initially intended to pass the tests outlined in "Conformance Testing 1051160814Ssimon of Relying Party Client Certificate Path Processing Logic" v1.07. 1052160814Ssimon [Steve Henson] 1053160814Ssimon 1054160814Ssimon *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and 1055160814Ssimon remained unused and not that useful. A variety of other little bignum 1056160814Ssimon tweaks and fixes have also been made continuing on from the audit (see 1057160814Ssimon below). 1058160814Ssimon [Geoff Thorpe] 1059160814Ssimon 1060160814Ssimon *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with 1061160814Ssimon associated ASN1, EVP and SSL functions and old ASN1 macros. 1062160814Ssimon [Richard Levitte] 1063160814Ssimon 1064160814Ssimon *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results, 1065160814Ssimon and this should never fail. So the return value from the use of 1066160814Ssimon BN_set_word() (which can fail due to needless expansion) is now deprecated; 1067160814Ssimon if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro. 1068160814Ssimon [Geoff Thorpe] 1069160814Ssimon 1070160814Ssimon *) BN_CTX_get() should return zero-valued bignums, providing the same 1071160814Ssimon initialised value as BN_new(). 1072160814Ssimon [Geoff Thorpe, suggested by Ulf M�ller] 1073160814Ssimon 1074160814Ssimon *) Support for inhibitAnyPolicy certificate extension. 1075160814Ssimon [Steve Henson] 1076160814Ssimon 1077160814Ssimon *) An audit of the BIGNUM code is underway, for which debugging code is 1078160814Ssimon enabled when BN_DEBUG is defined. This makes stricter enforcements on what 1079160814Ssimon is considered valid when processing BIGNUMs, and causes execution to 1080160814Ssimon assert() when a problem is discovered. If BN_DEBUG_RAND is defined, 1081160814Ssimon further steps are taken to deliberately pollute unused data in BIGNUM 1082160814Ssimon structures to try and expose faulty code further on. For now, openssl will 1083160814Ssimon (in its default mode of operation) continue to tolerate the inconsistent 1084160814Ssimon forms that it has tolerated in the past, but authors and packagers should 1085160814Ssimon consider trying openssl and their own applications when compiled with 1086160814Ssimon these debugging symbols defined. It will help highlight potential bugs in 1087160814Ssimon their own code, and will improve the test coverage for OpenSSL itself. At 1088160814Ssimon some point, these tighter rules will become openssl's default to improve 1089160814Ssimon maintainability, though the assert()s and other overheads will remain only 1090160814Ssimon in debugging configurations. See bn.h for more details. 1091160814Ssimon [Geoff Thorpe, Nils Larsch, Ulf M�ller] 1092160814Ssimon 1093160814Ssimon *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure 1094160814Ssimon that can only be obtained through BN_CTX_new() (which implicitly 1095160814Ssimon initialises it). The presence of this function only made it possible 1096160814Ssimon to overwrite an existing structure (and cause memory leaks). 1097160814Ssimon [Geoff Thorpe] 1098160814Ssimon 1099160814Ssimon *) Because of the callback-based approach for implementing LHASH as a 1100160814Ssimon template type, lh_insert() adds opaque objects to hash-tables and 1101160814Ssimon lh_doall() or lh_doall_arg() are typically used with a destructor callback 1102160814Ssimon to clean up those corresponding objects before destroying the hash table 1103160814Ssimon (and losing the object pointers). So some over-zealous constifications in 1104160814Ssimon LHASH have been relaxed so that lh_insert() does not take (nor store) the 1105160814Ssimon objects as "const" and the lh_doall[_arg] callback wrappers are not 1106160814Ssimon prototyped to have "const" restrictions on the object pointers they are 1107160814Ssimon given (and so aren't required to cast them away any more). 1108160814Ssimon [Geoff Thorpe] 1109160814Ssimon 1110160814Ssimon *) The tmdiff.h API was so ugly and minimal that our own timing utility 1111160814Ssimon (speed) prefers to use its own implementation. The two implementations 1112160814Ssimon haven't been consolidated as yet (volunteers?) but the tmdiff API has had 1113160814Ssimon its object type properly exposed (MS_TM) instead of casting to/from "char 1114160814Ssimon *". This may still change yet if someone realises MS_TM and "ms_time_***" 1115160814Ssimon aren't necessarily the greatest nomenclatures - but this is what was used 1116160814Ssimon internally to the implementation so I've used that for now. 1117160814Ssimon [Geoff Thorpe] 1118160814Ssimon 1119160814Ssimon *) Ensure that deprecated functions do not get compiled when 1120160814Ssimon OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of 1121160814Ssimon the self-tests were still using deprecated key-generation functions so 1122160814Ssimon these have been updated also. 1123160814Ssimon [Geoff Thorpe] 1124160814Ssimon 1125160814Ssimon *) Reorganise PKCS#7 code to separate the digest location functionality 1126160814Ssimon into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest(). 1127160814Ssimon New function PKCS7_set_digest() to set the digest type for PKCS#7 1128160814Ssimon digestedData type. Add additional code to correctly generate the 1129160814Ssimon digestedData type and add support for this type in PKCS7 initialization 1130160814Ssimon functions. 1131160814Ssimon [Steve Henson] 1132160814Ssimon 1133160814Ssimon *) New function PKCS7_set0_type_other() this initializes a PKCS7 1134160814Ssimon structure of type "other". 1135160814Ssimon [Steve Henson] 1136160814Ssimon 1137160814Ssimon *) Fix prime generation loop in crypto/bn/bn_prime.pl by making 1138160814Ssimon sure the loop does correctly stop and breaking ("division by zero") 1139160814Ssimon modulus operations are not performed. The (pre-generated) prime 1140160814Ssimon table crypto/bn/bn_prime.h was already correct, but it could not be 1141160814Ssimon re-generated on some platforms because of the "division by zero" 1142160814Ssimon situation in the script. 1143160814Ssimon [Ralf S. Engelschall] 1144160814Ssimon 1145160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 1146160814Ssimon draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with 1147160814Ssimon SHA-1 now is only used for "small" curves (where the 1148160814Ssimon representation of a field element takes up to 24 bytes); for 1149160814Ssimon larger curves, the field element resulting from ECDH is directly 1150160814Ssimon used as premaster secret. 1151160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 1152160814Ssimon 1153160814Ssimon *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2 1154160814Ssimon curve secp160r1 to the tests. 1155160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 1156160814Ssimon 1157160814Ssimon *) Add the possibility to load symbols globally with DSO. 1158160814Ssimon [G�tz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte] 1159160814Ssimon 1160160814Ssimon *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better 1161160814Ssimon control of the error stack. 1162160814Ssimon [Richard Levitte] 1163160814Ssimon 1164160814Ssimon *) Add support for STORE in ENGINE. 1165160814Ssimon [Richard Levitte] 1166160814Ssimon 1167160814Ssimon *) Add the STORE type. The intention is to provide a common interface 1168160814Ssimon to certificate and key stores, be they simple file-based stores, or 1169160814Ssimon HSM-type store, or LDAP stores, or... 1170160814Ssimon NOTE: The code is currently UNTESTED and isn't really used anywhere. 1171160814Ssimon [Richard Levitte] 1172160814Ssimon 1173160814Ssimon *) Add a generic structure called OPENSSL_ITEM. This can be used to 1174160814Ssimon pass a list of arguments to any function as well as provide a way 1175160814Ssimon for a function to pass data back to the caller. 1176160814Ssimon [Richard Levitte] 1177160814Ssimon 1178160814Ssimon *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup() 1179160814Ssimon works like BUF_strdup() but can be used to duplicate a portion of 1180160814Ssimon a string. The copy gets NUL-terminated. BUF_memdup() duplicates 1181160814Ssimon a memory area. 1182160814Ssimon [Richard Levitte] 1183160814Ssimon 1184160814Ssimon *) Add the function sk_find_ex() which works like sk_find(), but will 1185160814Ssimon return an index to an element even if an exact match couldn't be 1186160814Ssimon found. The index is guaranteed to point at the element where the 1187160814Ssimon searched-for key would be inserted to preserve sorting order. 1188160814Ssimon [Richard Levitte] 1189160814Ssimon 1190160814Ssimon *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but 1191160814Ssimon takes an extra flags argument for optional functionality. Currently, 1192160814Ssimon the following flags are defined: 1193160814Ssimon 1194160814Ssimon OBJ_BSEARCH_VALUE_ON_NOMATCH 1195160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 1196160814Ssimon element where the comparing function returns a negative or zero 1197160814Ssimon number. 1198160814Ssimon 1199160814Ssimon OBJ_BSEARCH_FIRST_VALUE_ON_MATCH 1200160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 1201160814Ssimon element where the comparing function returns zero. This is useful 1202160814Ssimon if there are more than one element where the comparing function 1203160814Ssimon returns zero. 1204160814Ssimon [Richard Levitte] 1205160814Ssimon 1206160814Ssimon *) Make it possible to create self-signed certificates with 'openssl ca' 1207160814Ssimon in such a way that the self-signed certificate becomes part of the 1208160814Ssimon CA database and uses the same mechanisms for serial number generation 1209160814Ssimon as all other certificate signing. The new flag '-selfsign' enables 1210160814Ssimon this functionality. Adapt CA.sh and CA.pl.in. 1211160814Ssimon [Richard Levitte] 1212160814Ssimon 1213160814Ssimon *) Add functionality to check the public key of a certificate request 1214160814Ssimon against a given private. This is useful to check that a certificate 1215160814Ssimon request can be signed by that key (self-signing). 1216160814Ssimon [Richard Levitte] 1217160814Ssimon 1218160814Ssimon *) Make it possible to have multiple active certificates with the same 1219160814Ssimon subject in the CA index file. This is done only if the keyword 1220160814Ssimon 'unique_subject' is set to 'no' in the main CA section (default 1221160814Ssimon if 'CA_default') of the configuration file. The value is saved 1222160814Ssimon with the database itself in a separate index attribute file, 1223160814Ssimon named like the index file with '.attr' appended to the name. 1224160814Ssimon [Richard Levitte] 1225160814Ssimon 1226160814Ssimon *) Generate muti valued AVAs using '+' notation in config files for 1227160814Ssimon req and dirName. 1228160814Ssimon [Steve Henson] 1229160814Ssimon 1230160814Ssimon *) Support for nameConstraints certificate extension. 1231160814Ssimon [Steve Henson] 1232160814Ssimon 1233160814Ssimon *) Support for policyConstraints certificate extension. 1234160814Ssimon [Steve Henson] 1235160814Ssimon 1236160814Ssimon *) Support for policyMappings certificate extension. 1237160814Ssimon [Steve Henson] 1238160814Ssimon 1239160814Ssimon *) Make sure the default DSA_METHOD implementation only uses its 1240160814Ssimon dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL, 1241160814Ssimon and change its own handlers to be NULL so as to remove unnecessary 1242160814Ssimon indirection. This lets alternative implementations fallback to the 1243160814Ssimon default implementation more easily. 1244160814Ssimon [Geoff Thorpe] 1245160814Ssimon 1246160814Ssimon *) Support for directoryName in GeneralName related extensions 1247160814Ssimon in config files. 1248160814Ssimon [Steve Henson] 1249160814Ssimon 1250160814Ssimon *) Make it possible to link applications using Makefile.shared. 1251160814Ssimon Make that possible even when linking against static libraries! 1252160814Ssimon [Richard Levitte] 1253160814Ssimon 1254160814Ssimon *) Support for single pass processing for S/MIME signing. This now 1255160814Ssimon means that S/MIME signing can be done from a pipe, in addition 1256160814Ssimon cleartext signing (multipart/signed type) is effectively streaming 1257160814Ssimon and the signed data does not need to be all held in memory. 1258160814Ssimon 1259160814Ssimon This is done with a new flag PKCS7_STREAM. When this flag is set 1260160814Ssimon PKCS7_sign() only initializes the PKCS7 structure and the actual signing 1261160814Ssimon is done after the data is output (and digests calculated) in 1262160814Ssimon SMIME_write_PKCS7(). 1263160814Ssimon [Steve Henson] 1264160814Ssimon 1265160814Ssimon *) Add full support for -rpath/-R, both in shared libraries and 1266160814Ssimon applications, at least on the platforms where it's known how 1267160814Ssimon to do it. 1268160814Ssimon [Richard Levitte] 1269160814Ssimon 1270160814Ssimon *) In crypto/ec/ec_mult.c, implement fast point multiplication with 1271160814Ssimon precomputation, based on wNAF splitting: EC_GROUP_precompute_mult() 1272160814Ssimon will now compute a table of multiples of the generator that 1273160814Ssimon makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul() 1274160814Ssimon faster (notably in the case of a single point multiplication, 1275160814Ssimon scalar * generator). 1276160814Ssimon [Nils Larsch, Bodo Moeller] 1277160814Ssimon 1278160814Ssimon *) IPv6 support for certificate extensions. The various extensions 1279160814Ssimon which use the IP:a.b.c.d can now take IPv6 addresses using the 1280160814Ssimon formats of RFC1884 2.2 . IPv6 addresses are now also displayed 1281160814Ssimon correctly. 1282160814Ssimon [Steve Henson] 1283160814Ssimon 1284160814Ssimon *) Added an ENGINE that implements RSA by performing private key 1285160814Ssimon exponentiations with the GMP library. The conversions to and from 1286160814Ssimon GMP's mpz_t format aren't optimised nor are any montgomery forms 1287160814Ssimon cached, and on x86 it appears OpenSSL's own performance has caught up. 1288160814Ssimon However there are likely to be other architectures where GMP could 1289160814Ssimon provide a boost. This ENGINE is not built in by default, but it can be 1290160814Ssimon specified at Configure time and should be accompanied by the necessary 1291160814Ssimon linker additions, eg; 1292160814Ssimon ./config -DOPENSSL_USE_GMP -lgmp 1293160814Ssimon [Geoff Thorpe] 1294160814Ssimon 1295160814Ssimon *) "openssl engine" will not display ENGINE/DSO load failure errors when 1296160814Ssimon testing availability of engines with "-t" - the old behaviour is 1297160814Ssimon produced by increasing the feature's verbosity with "-tt". 1298160814Ssimon [Geoff Thorpe] 1299160814Ssimon 1300160814Ssimon *) ECDSA routines: under certain error conditions uninitialized BN objects 1301160814Ssimon could be freed. Solution: make sure initialization is performed early 1302160814Ssimon enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de> 1303160814Ssimon via PR#459) 1304160814Ssimon [Lutz Jaenicke] 1305160814Ssimon 1306160814Ssimon *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD 1307160814Ssimon and DH_METHOD (eg. by ENGINE implementations) to override the normal 1308160814Ssimon software implementations. For DSA and DH, parameter generation can 1309160814Ssimon also be overriden by providing the appropriate method callbacks. 1310160814Ssimon [Geoff Thorpe] 1311160814Ssimon 1312160814Ssimon *) Change the "progress" mechanism used in key-generation and 1313160814Ssimon primality testing to functions that take a new BN_GENCB pointer in 1314160814Ssimon place of callback/argument pairs. The new API functions have "_ex" 1315160814Ssimon postfixes and the older functions are reimplemented as wrappers for 1316160814Ssimon the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide 1317160814Ssimon declarations of the old functions to help (graceful) attempts to 1318160814Ssimon migrate to the new functions. Also, the new key-generation API 1319160814Ssimon functions operate on a caller-supplied key-structure and return 1320160814Ssimon success/failure rather than returning a key or NULL - this is to 1321160814Ssimon help make "keygen" another member function of RSA_METHOD etc. 1322160814Ssimon 1323160814Ssimon Example for using the new callback interface: 1324160814Ssimon 1325160814Ssimon int (*my_callback)(int a, int b, BN_GENCB *cb) = ...; 1326160814Ssimon void *my_arg = ...; 1327160814Ssimon BN_GENCB my_cb; 1328160814Ssimon 1329160814Ssimon BN_GENCB_set(&my_cb, my_callback, my_arg); 1330160814Ssimon 1331160814Ssimon return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb); 1332160814Ssimon /* For the meaning of a, b in calls to my_callback(), see the 1333160814Ssimon * documentation of the function that calls the callback. 1334160814Ssimon * cb will point to my_cb; my_arg can be retrieved as cb->arg. 1335160814Ssimon * my_callback should return 1 if it wants BN_is_prime_ex() 1336160814Ssimon * to continue, or 0 to stop. 1337160814Ssimon */ 1338160814Ssimon 1339160814Ssimon [Geoff Thorpe] 1340160814Ssimon 1341160814Ssimon *) Change the ZLIB compression method to be stateful, and make it 1342160814Ssimon available to TLS with the number defined in 1343160814Ssimon draft-ietf-tls-compression-04.txt. 1344160814Ssimon [Richard Levitte] 1345160814Ssimon 1346160814Ssimon *) Add the ASN.1 structures and functions for CertificatePair, which 1347160814Ssimon is defined as follows (according to X.509_4thEditionDraftV6.pdf): 1348160814Ssimon 1349160814Ssimon CertificatePair ::= SEQUENCE { 1350160814Ssimon forward [0] Certificate OPTIONAL, 1351160814Ssimon reverse [1] Certificate OPTIONAL, 1352160814Ssimon -- at least one of the pair shall be present -- } 1353160814Ssimon 1354160814Ssimon Also implement the PEM functions to read and write certificate 1355160814Ssimon pairs, and defined the PEM tag as "CERTIFICATE PAIR". 1356160814Ssimon 1357160814Ssimon This needed to be defined, mostly for the sake of the LDAP 1358160814Ssimon attribute crossCertificatePair, but may prove useful elsewhere as 1359160814Ssimon well. 1360160814Ssimon [Richard Levitte] 1361160814Ssimon 1362160814Ssimon *) Make it possible to inhibit symlinking of shared libraries in 1363160814Ssimon Makefile.shared, for Cygwin's sake. 1364160814Ssimon [Richard Levitte] 1365160814Ssimon 1366160814Ssimon *) Extend the BIGNUM API by creating a function 1367160814Ssimon void BN_set_negative(BIGNUM *a, int neg); 1368160814Ssimon and a macro that behave like 1369160814Ssimon int BN_is_negative(const BIGNUM *a); 1370160814Ssimon 1371160814Ssimon to avoid the need to access 'a->neg' directly in applications. 1372160814Ssimon [Nils Larsch] 1373160814Ssimon 1374160814Ssimon *) Implement fast modular reduction for pseudo-Mersenne primes 1375160814Ssimon used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c). 1376160814Ssimon EC_GROUP_new_curve_GFp() will now automatically use this 1377160814Ssimon if applicable. 1378160814Ssimon [Nils Larsch <nla@trustcenter.de>] 1379160814Ssimon 1380160814Ssimon *) Add new lock type (CRYPTO_LOCK_BN). 1381160814Ssimon [Bodo Moeller] 1382160814Ssimon 1383160814Ssimon *) Change the ENGINE framework to automatically load engines 1384160814Ssimon dynamically from specific directories unless they could be 1385160814Ssimon found to already be built in or loaded. Move all the 1386160814Ssimon current engines except for the cryptodev one to a new 1387160814Ssimon directory engines/. 1388160814Ssimon The engines in engines/ are built as shared libraries if 1389160814Ssimon the "shared" options was given to ./Configure or ./config. 1390160814Ssimon Otherwise, they are inserted in libcrypto.a. 1391160814Ssimon /usr/local/ssl/engines is the default directory for dynamic 1392160814Ssimon engines, but that can be overriden at configure time through 1393160814Ssimon the usual use of --prefix and/or --openssldir, and at run 1394160814Ssimon time with the environment variable OPENSSL_ENGINES. 1395160814Ssimon [Geoff Thorpe and Richard Levitte] 1396160814Ssimon 1397160814Ssimon *) Add Makefile.shared, a helper makefile to build shared 1398160814Ssimon libraries. Addapt Makefile.org. 1399160814Ssimon [Richard Levitte] 1400160814Ssimon 1401160814Ssimon *) Add version info to Win32 DLLs. 1402160814Ssimon [Peter 'Luna' Runestig" <peter@runestig.com>] 1403160814Ssimon 1404160814Ssimon *) Add new 'medium level' PKCS#12 API. Certificates and keys 1405160814Ssimon can be added using this API to created arbitrary PKCS#12 1406160814Ssimon files while avoiding the low level API. 1407160814Ssimon 1408160814Ssimon New options to PKCS12_create(), key or cert can be NULL and 1409160814Ssimon will then be omitted from the output file. The encryption 1410160814Ssimon algorithm NIDs can be set to -1 for no encryption, the mac 1411160814Ssimon iteration count can be set to 0 to omit the mac. 1412160814Ssimon 1413160814Ssimon Enhance pkcs12 utility by making the -nokeys and -nocerts 1414160814Ssimon options work when creating a PKCS#12 file. New option -nomac 1415160814Ssimon to omit the mac, NONE can be set for an encryption algorithm. 1416160814Ssimon New code is modified to use the enhanced PKCS12_create() 1417160814Ssimon instead of the low level API. 1418160814Ssimon [Steve Henson] 1419160814Ssimon 1420160814Ssimon *) Extend ASN1 encoder to support indefinite length constructed 1421160814Ssimon encoding. This can output sequences tags and octet strings in 1422160814Ssimon this form. Modify pk7_asn1.c to support indefinite length 1423160814Ssimon encoding. This is experimental and needs additional code to 1424160814Ssimon be useful, such as an ASN1 bio and some enhanced streaming 1425160814Ssimon PKCS#7 code. 1426160814Ssimon 1427160814Ssimon Extend template encode functionality so that tagging is passed 1428160814Ssimon down to the template encoder. 1429160814Ssimon [Steve Henson] 1430160814Ssimon 1431160814Ssimon *) Let 'openssl req' fail if an argument to '-newkey' is not 1432160814Ssimon recognized instead of using RSA as a default. 1433160814Ssimon [Bodo Moeller] 1434160814Ssimon 1435160814Ssimon *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt. 1436160814Ssimon As these are not official, they are not included in "ALL"; 1437160814Ssimon the "ECCdraft" ciphersuite group alias can be used to select them. 1438160814Ssimon [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)] 1439160814Ssimon 1440160814Ssimon *) Add ECDH engine support. 1441160814Ssimon [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)] 1442160814Ssimon 1443160814Ssimon *) Add ECDH in new directory crypto/ecdh/. 1444160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 1445160814Ssimon 1446160814Ssimon *) Let BN_rand_range() abort with an error after 100 iterations 1447160814Ssimon without success (which indicates a broken PRNG). 1448160814Ssimon [Bodo Moeller] 1449160814Ssimon 1450160814Ssimon *) Change BN_mod_sqrt() so that it verifies that the input value 1451160814Ssimon is really the square of the return value. (Previously, 1452160814Ssimon BN_mod_sqrt would show GIGO behaviour.) 1453160814Ssimon [Bodo Moeller] 1454160814Ssimon 1455160814Ssimon *) Add named elliptic curves over binary fields from X9.62, SECG, 1456160814Ssimon and WAP/WTLS; add OIDs that were still missing. 1457160814Ssimon 1458160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 1459160814Ssimon (Sun Microsystems Laboratories)] 1460160814Ssimon 1461160814Ssimon *) Extend the EC library for elliptic curves over binary fields 1462160814Ssimon (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/). 1463160814Ssimon New EC_METHOD: 1464160814Ssimon 1465160814Ssimon EC_GF2m_simple_method 1466160814Ssimon 1467160814Ssimon New API functions: 1468160814Ssimon 1469160814Ssimon EC_GROUP_new_curve_GF2m 1470160814Ssimon EC_GROUP_set_curve_GF2m 1471160814Ssimon EC_GROUP_get_curve_GF2m 1472160814Ssimon EC_POINT_set_affine_coordinates_GF2m 1473160814Ssimon EC_POINT_get_affine_coordinates_GF2m 1474160814Ssimon EC_POINT_set_compressed_coordinates_GF2m 1475160814Ssimon 1476160814Ssimon Point compression for binary fields is disabled by default for 1477160814Ssimon patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to 1478160814Ssimon enable it). 1479160814Ssimon 1480160814Ssimon As binary polynomials are represented as BIGNUMs, various members 1481160814Ssimon of the EC_GROUP and EC_POINT data structures can be shared 1482160814Ssimon between the implementations for prime fields and binary fields; 1483160814Ssimon the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m) 1484160814Ssimon are essentially identical to their ..._GFp counterparts. 1485160814Ssimon (For simplicity, the '..._GFp' prefix has been dropped from 1486160814Ssimon various internal method names.) 1487160814Ssimon 1488160814Ssimon An internal 'field_div' method (similar to 'field_mul' and 1489160814Ssimon 'field_sqr') has been added; this is used only for binary fields. 1490160814Ssimon 1491160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 1492160814Ssimon (Sun Microsystems Laboratories)] 1493160814Ssimon 1494160814Ssimon *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult() 1495160814Ssimon through methods ('mul', 'precompute_mult'). 1496160814Ssimon 1497160814Ssimon The generic implementations (now internally called 'ec_wNAF_mul' 1498160814Ssimon and 'ec_wNAF_precomputed_mult') remain the default if these 1499160814Ssimon methods are undefined. 1500160814Ssimon 1501160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 1502160814Ssimon (Sun Microsystems Laboratories)] 1503160814Ssimon 1504160814Ssimon *) New function EC_GROUP_get_degree, which is defined through 1505160814Ssimon EC_METHOD. For curves over prime fields, this returns the bit 1506160814Ssimon length of the modulus. 1507160814Ssimon 1508160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 1509160814Ssimon (Sun Microsystems Laboratories)] 1510160814Ssimon 1511160814Ssimon *) New functions EC_GROUP_dup, EC_POINT_dup. 1512160814Ssimon (These simply call ..._new and ..._copy). 1513160814Ssimon 1514160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 1515160814Ssimon (Sun Microsystems Laboratories)] 1516160814Ssimon 1517160814Ssimon *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c. 1518160814Ssimon Polynomials are represented as BIGNUMs (where the sign bit is not 1519160814Ssimon used) in the following functions [macros]: 1520160814Ssimon 1521160814Ssimon BN_GF2m_add 1522160814Ssimon BN_GF2m_sub [= BN_GF2m_add] 1523160814Ssimon BN_GF2m_mod [wrapper for BN_GF2m_mod_arr] 1524160814Ssimon BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr] 1525160814Ssimon BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr] 1526160814Ssimon BN_GF2m_mod_inv 1527160814Ssimon BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr] 1528160814Ssimon BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr] 1529160814Ssimon BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr] 1530160814Ssimon BN_GF2m_cmp [= BN_ucmp] 1531160814Ssimon 1532160814Ssimon (Note that only the 'mod' functions are actually for fields GF(2^m). 1533160814Ssimon BN_GF2m_add() is misnomer, but this is for the sake of consistency.) 1534160814Ssimon 1535160814Ssimon For some functions, an the irreducible polynomial defining a 1536160814Ssimon field can be given as an 'unsigned int[]' with strictly 1537160814Ssimon decreasing elements giving the indices of those bits that are set; 1538160814Ssimon i.e., p[] represents the polynomial 1539160814Ssimon f(t) = t^p[0] + t^p[1] + ... + t^p[k] 1540160814Ssimon where 1541160814Ssimon p[0] > p[1] > ... > p[k] = 0. 1542160814Ssimon This applies to the following functions: 1543160814Ssimon 1544160814Ssimon BN_GF2m_mod_arr 1545160814Ssimon BN_GF2m_mod_mul_arr 1546160814Ssimon BN_GF2m_mod_sqr_arr 1547160814Ssimon BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv] 1548160814Ssimon BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div] 1549160814Ssimon BN_GF2m_mod_exp_arr 1550160814Ssimon BN_GF2m_mod_sqrt_arr 1551160814Ssimon BN_GF2m_mod_solve_quad_arr 1552160814Ssimon BN_GF2m_poly2arr 1553160814Ssimon BN_GF2m_arr2poly 1554160814Ssimon 1555160814Ssimon Conversion can be performed by the following functions: 1556160814Ssimon 1557160814Ssimon BN_GF2m_poly2arr 1558160814Ssimon BN_GF2m_arr2poly 1559160814Ssimon 1560160814Ssimon bntest.c has additional tests for binary polynomial arithmetic. 1561160814Ssimon 1562160814Ssimon Two implementations for BN_GF2m_mod_div() are available. 1563160814Ssimon The default algorithm simply uses BN_GF2m_mod_inv() and 1564160814Ssimon BN_GF2m_mod_mul(). The alternative algorithm is compiled in only 1565160814Ssimon if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the 1566160814Ssimon copyright notice in crypto/bn/bn_gf2m.c before enabling it). 1567160814Ssimon 1568160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 1569160814Ssimon (Sun Microsystems Laboratories)] 1570160814Ssimon 1571160814Ssimon *) Add new error code 'ERR_R_DISABLED' that can be used when some 1572160814Ssimon functionality is disabled at compile-time. 1573160814Ssimon [Douglas Stebila <douglas.stebila@sun.com>] 1574160814Ssimon 1575160814Ssimon *) Change default behaviour of 'openssl asn1parse' so that more 1576160814Ssimon information is visible when viewing, e.g., a certificate: 1577160814Ssimon 1578160814Ssimon Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump' 1579160814Ssimon mode the content of non-printable OCTET STRINGs is output in a 1580160814Ssimon style similar to INTEGERs, but with '[HEX DUMP]' prepended to 1581160814Ssimon avoid the appearance of a printable string. 1582160814Ssimon [Nils Larsch <nla@trustcenter.de>] 1583160814Ssimon 1584160814Ssimon *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access 1585160814Ssimon functions 1586160814Ssimon EC_GROUP_set_asn1_flag() 1587160814Ssimon EC_GROUP_get_asn1_flag() 1588160814Ssimon EC_GROUP_set_point_conversion_form() 1589160814Ssimon EC_GROUP_get_point_conversion_form() 1590160814Ssimon These control ASN1 encoding details: 1591160814Ssimon - Curves (i.e., groups) are encoded explicitly unless asn1_flag 1592160814Ssimon has been set to OPENSSL_EC_NAMED_CURVE. 1593160814Ssimon - Points are encoded in uncompressed form by default; options for 1594160814Ssimon asn1_for are as for point2oct, namely 1595160814Ssimon POINT_CONVERSION_COMPRESSED 1596160814Ssimon POINT_CONVERSION_UNCOMPRESSED 1597160814Ssimon POINT_CONVERSION_HYBRID 1598160814Ssimon 1599160814Ssimon Also add 'seed' and 'seed_len' members to EC_GROUP with access 1600160814Ssimon functions 1601160814Ssimon EC_GROUP_set_seed() 1602160814Ssimon EC_GROUP_get0_seed() 1603160814Ssimon EC_GROUP_get_seed_len() 1604160814Ssimon This is used only for ASN1 purposes (so far). 1605160814Ssimon [Nils Larsch <nla@trustcenter.de>] 1606160814Ssimon 1607160814Ssimon *) Add 'field_type' member to EC_METHOD, which holds the NID 1608160814Ssimon of the appropriate field type OID. The new function 1609160814Ssimon EC_METHOD_get_field_type() returns this value. 1610160814Ssimon [Nils Larsch <nla@trustcenter.de>] 1611160814Ssimon 1612160814Ssimon *) Add functions 1613160814Ssimon EC_POINT_point2bn() 1614160814Ssimon EC_POINT_bn2point() 1615160814Ssimon EC_POINT_point2hex() 1616160814Ssimon EC_POINT_hex2point() 1617160814Ssimon providing useful interfaces to EC_POINT_point2oct() and 1618160814Ssimon EC_POINT_oct2point(). 1619160814Ssimon [Nils Larsch <nla@trustcenter.de>] 1620160814Ssimon 1621160814Ssimon *) Change internals of the EC library so that the functions 1622160814Ssimon EC_GROUP_set_generator() 1623160814Ssimon EC_GROUP_get_generator() 1624160814Ssimon EC_GROUP_get_order() 1625160814Ssimon EC_GROUP_get_cofactor() 1626160814Ssimon are implemented directly in crypto/ec/ec_lib.c and not dispatched 1627160814Ssimon to methods, which would lead to unnecessary code duplication when 1628160814Ssimon adding different types of curves. 1629160814Ssimon [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller] 1630160814Ssimon 1631160814Ssimon *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM 1632160814Ssimon arithmetic, and such that modified wNAFs are generated 1633160814Ssimon (which avoid length expansion in many cases). 1634160814Ssimon [Bodo Moeller] 1635160814Ssimon 1636160814Ssimon *) Add a function EC_GROUP_check_discriminant() (defined via 1637160814Ssimon EC_METHOD) that verifies that the curve discriminant is non-zero. 1638160814Ssimon 1639160814Ssimon Add a function EC_GROUP_check() that makes some sanity tests 1640160814Ssimon on a EC_GROUP, its generator and order. This includes 1641160814Ssimon EC_GROUP_check_discriminant(). 1642160814Ssimon [Nils Larsch <nla@trustcenter.de>] 1643160814Ssimon 1644160814Ssimon *) Add ECDSA in new directory crypto/ecdsa/. 1645160814Ssimon 1646160814Ssimon Add applications 'openssl ecparam' and 'openssl ecdsa' 1647160814Ssimon (these are based on 'openssl dsaparam' and 'openssl dsa'). 1648160814Ssimon 1649160814Ssimon ECDSA support is also included in various other files across the 1650160814Ssimon library. Most notably, 1651160814Ssimon - 'openssl req' now has a '-newkey ecdsa:file' option; 1652160814Ssimon - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA; 1653160814Ssimon - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and 1654160814Ssimon d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make 1655160814Ssimon them suitable for ECDSA where domain parameters must be 1656160814Ssimon extracted before the specific public key; 1657160814Ssimon - ECDSA engine support has been added. 1658160814Ssimon [Nils Larsch <nla@trustcenter.de>] 1659160814Ssimon 1660160814Ssimon *) Include some named elliptic curves, and add OIDs from X9.62, 1661160814Ssimon SECG, and WAP/WTLS. Each curve can be obtained from the new 1662160814Ssimon function 1663160814Ssimon EC_GROUP_new_by_curve_name(), 1664160814Ssimon and the list of available named curves can be obtained with 1665160814Ssimon EC_get_builtin_curves(). 1666160814Ssimon Also add a 'curve_name' member to EC_GROUP objects, which can be 1667160814Ssimon accessed via 1668160814Ssimon EC_GROUP_set_curve_name() 1669160814Ssimon EC_GROUP_get_curve_name() 1670160814Ssimon [Nils Larsch <larsch@trustcenter.de, Bodo Moeller] 1671160814Ssimon 1672160814Ssimon *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 1673160814Ssimon was actually never needed) and in BN_mul(). The removal in BN_mul() 1674160814Ssimon required a small change in bn_mul_part_recursive() and the addition 1675160814Ssimon of the functions bn_cmp_part_words(), bn_sub_part_words() and 1676160814Ssimon bn_add_part_words(), which do the same thing as bn_cmp_words(), 1677160814Ssimon bn_sub_words() and bn_add_words() except they take arrays with 1678160814Ssimon differing sizes. 1679160814Ssimon [Richard Levitte] 1680160814Ssimon 1681194206Ssimon Changes between 0.9.7l and 0.9.7m [23 Feb 2007] 1682194206Ssimon 1683167612Ssimon *) Cleanse PEM buffers before freeing them since they may contain 1684167612Ssimon sensitive data. 1685167612Ssimon [Benjamin Bennett <ben@psc.edu>] 1686167612Ssimon 1687167612Ssimon *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 1688167612Ssimon a ciphersuite string such as "DEFAULT:RSA" cannot enable 1689167612Ssimon authentication-only ciphersuites. 1690167612Ssimon [Bodo Moeller] 1691167612Ssimon 1692167612Ssimon *) Since AES128 and AES256 share a single mask bit in the logic of 1693167612Ssimon ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 1694167612Ssimon kludge to work properly if AES128 is available and AES256 isn't. 1695167612Ssimon [Victor Duchovni] 1696167612Ssimon 1697194206Ssimon *) Expand security boundary to match 1.1.1 module. 1698194206Ssimon [Steve Henson] 1699194206Ssimon 1700194206Ssimon *) Remove redundant features: hash file source, editing of test vectors 1701194206Ssimon modify fipsld to use external fips_premain.c signature. 1702194206Ssimon [Steve Henson] 1703194206Ssimon 1704194206Ssimon *) New perl script mkfipsscr.pl to create shell scripts or batch files to 1705194206Ssimon run algorithm test programs. 1706194206Ssimon [Steve Henson] 1707194206Ssimon 1708194206Ssimon *) Make algorithm test programs more tolerant of whitespace. 1709194206Ssimon [Steve Henson] 1710194206Ssimon 1711167612Ssimon *) Have SSL/TLS server implementation tolerate "mismatched" record 1712167612Ssimon protocol version while receiving ClientHello even if the 1713167612Ssimon ClientHello is fragmented. (The server can't insist on the 1714167612Ssimon particular protocol version it has chosen before the ServerHello 1715167612Ssimon message has informed the client about his choice.) 1716167612Ssimon [Bodo Moeller] 1717167612Ssimon 1718167612Ssimon *) Load error codes if they are not already present instead of using a 1719167612Ssimon static variable. This allows them to be cleanly unloaded and reloaded. 1720167612Ssimon [Steve Henson] 1721167612Ssimon 1722167612Ssimon Changes between 0.9.7k and 0.9.7l [28 Sep 2006] 1723167612Ssimon 1724167612Ssimon *) Introduce limits to prevent malicious keys being able to 1725167612Ssimon cause a denial of service. (CVE-2006-2940) 1726167612Ssimon [Steve Henson, Bodo Moeller] 1727167612Ssimon 1728167612Ssimon *) Fix ASN.1 parsing of certain invalid structures that can result 1729167612Ssimon in a denial of service. (CVE-2006-2937) [Steve Henson] 1730167612Ssimon 1731167612Ssimon *) Fix buffer overflow in SSL_get_shared_ciphers() function. 1732167612Ssimon (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 1733167612Ssimon 1734167612Ssimon *) Fix SSL client code which could crash if connecting to a 1735167612Ssimon malicious SSLv2 server. (CVE-2006-4343) 1736167612Ssimon [Tavis Ormandy and Will Drewry, Google Security Team] 1737167612Ssimon 1738162911Ssimon *) Change ciphersuite string processing so that an explicit 1739162911Ssimon ciphersuite selects this one ciphersuite (so that "AES256-SHA" 1740162911Ssimon will no longer include "AES128-SHA"), and any other similar 1741162911Ssimon ciphersuite (same bitmap) from *other* protocol versions (so that 1742162911Ssimon "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the 1743162911Ssimon SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining 1744162911Ssimon changes from 0.9.8b and 0.9.8d. 1745162911Ssimon [Bodo Moeller] 1746162911Ssimon 1747162911Ssimon Changes between 0.9.7j and 0.9.7k [05 Sep 2006] 1748162911Ssimon 1749162911Ssimon *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 1750162911Ssimon (CVE-2006-4339) [Ben Laurie and Google Security Team] 1751162911Ssimon 1752162911Ssimon *) Change the Unix randomness entropy gathering to use poll() when 1753162911Ssimon possible instead of select(), since the latter has some 1754162911Ssimon undesirable limitations. 1755162911Ssimon [Darryl Miles via Richard Levitte and Bodo Moeller] 1756162911Ssimon 1757162911Ssimon *) Disable rogue ciphersuites: 1758162911Ssimon 1759162911Ssimon - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 1760162911Ssimon - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 1761162911Ssimon - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 1762162911Ssimon 1763162911Ssimon The latter two were purportedly from 1764162911Ssimon draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 1765162911Ssimon appear there. 1766162911Ssimon 1767162911Ssimon Also deactive the remaining ciphersuites from 1768162911Ssimon draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 1769162911Ssimon unofficial, and the ID has long expired. 1770162911Ssimon [Bodo Moeller] 1771162911Ssimon 1772162911Ssimon *) Fix RSA blinding Heisenbug (problems sometimes occured on 1773162911Ssimon dual-core machines) and other potential thread-safety issues. 1774162911Ssimon [Bodo Moeller] 1775162911Ssimon 1776162911Ssimon Changes between 0.9.7i and 0.9.7j [04 May 2006] 1777162911Ssimon 1778162911Ssimon *) Adapt fipsld and the build system to link against the validated FIPS 1779162911Ssimon module in FIPS mode. 1780162911Ssimon [Steve Henson] 1781162911Ssimon 1782162911Ssimon *) Fixes for VC++ 2005 build under Windows. 1783162911Ssimon [Steve Henson] 1784162911Ssimon 1785162911Ssimon *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make 1786162911Ssimon from a Windows bash shell such as MSYS. It is autodetected from the 1787162911Ssimon "config" script when run from a VC++ environment. Modify standard VC++ 1788162911Ssimon build to use fipscanister.o from the GNU make build. 1789162911Ssimon [Steve Henson] 1790162911Ssimon 1791160814Ssimon Changes between 0.9.7h and 0.9.7i [14 Oct 2005] 1792160814Ssimon 1793160814Ssimon *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS. 1794160814Ssimon The value now differs depending on if you build for FIPS or not. 1795160814Ssimon BEWARE! A program linked with a shared FIPSed libcrypto can't be 1796160814Ssimon safely run with a non-FIPSed libcrypto, as it may crash because of 1797160814Ssimon the difference induced by this change. 1798160814Ssimon [Andy Polyakov] 1799160814Ssimon 1800160814Ssimon Changes between 0.9.7g and 0.9.7h [11 Oct 2005] 1801160814Ssimon 1802160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 1803160814Ssimon (part of SSL_OP_ALL). This option used to disable the 1804160814Ssimon countermeasure against man-in-the-middle protocol-version 1805160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 1806160814Ssimon idea. (CVE-2005-2969) 1807160814Ssimon 1808160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 1809160814Ssimon for Information Security, National Institute of Advanced Industrial 1810160814Ssimon Science and Technology [AIST], Japan)] 1811160814Ssimon 1812160814Ssimon *) Minimal support for X9.31 signatures and PSS padding modes. This is 1813160814Ssimon mainly for FIPS compliance and not fully integrated at this stage. 1814160814Ssimon [Steve Henson] 1815160814Ssimon 1816160814Ssimon *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform 1817160814Ssimon the exponentiation using a fixed-length exponent. (Otherwise, 1818160814Ssimon the information leaked through timing could expose the secret key 1819160814Ssimon after many signatures; cf. Bleichenbacher's attack on DSA with 1820160814Ssimon biased k.) 1821160814Ssimon [Bodo Moeller] 1822160814Ssimon 1823160814Ssimon *) Make a new fixed-window mod_exp implementation the default for 1824160814Ssimon RSA, DSA, and DH private-key operations so that the sequence of 1825160814Ssimon squares and multiplies and the memory access pattern are 1826160814Ssimon independent of the particular secret key. This will mitigate 1827160814Ssimon cache-timing and potential related attacks. 1828160814Ssimon 1829160814Ssimon BN_mod_exp_mont_consttime() is the new exponentiation implementation, 1830160814Ssimon and this is automatically used by BN_mod_exp_mont() if the new flag 1831160814Ssimon BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH 1832160814Ssimon will use this BN flag for private exponents unless the flag 1833160814Ssimon RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or 1834160814Ssimon DH_FLAG_NO_EXP_CONSTTIME, respectively, is set. 1835160814Ssimon 1836160814Ssimon [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller] 1837160814Ssimon 1838160814Ssimon *) Change the client implementation for SSLv23_method() and 1839160814Ssimon SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0 1840160814Ssimon Client Hello message format if the SSL_OP_NO_SSLv2 option is set. 1841160814Ssimon (Previously, the SSL 2.0 backwards compatible Client Hello 1842160814Ssimon message format would be used even with SSL_OP_NO_SSLv2.) 1843160814Ssimon [Bodo Moeller] 1844160814Ssimon 1845160814Ssimon *) Add support for smime-type MIME parameter in S/MIME messages which some 1846160814Ssimon clients need. 1847160814Ssimon [Steve Henson] 1848160814Ssimon 1849160814Ssimon *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in 1850160814Ssimon a threadsafe manner. Modify rsa code to use new function and add calls 1851160814Ssimon to dsa and dh code (which had race conditions before). 1852160814Ssimon [Steve Henson] 1853160814Ssimon 1854160814Ssimon *) Include the fixed error library code in the C error file definitions 1855160814Ssimon instead of fixing them up at runtime. This keeps the error code 1856160814Ssimon structures constant. 1857160814Ssimon [Steve Henson] 1858160814Ssimon 1859160814Ssimon Changes between 0.9.7f and 0.9.7g [11 Apr 2005] 1860160814Ssimon 1861160814Ssimon [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after 1862160814Ssimon OpenSSL 0.9.8.] 1863160814Ssimon 1864160814Ssimon *) Fixes for newer kerberos headers. NB: the casts are needed because 1865160814Ssimon the 'length' field is signed on one version and unsigned on another 1866160814Ssimon with no (?) obvious way to tell the difference, without these VC++ 1867160814Ssimon complains. Also the "definition" of FAR (blank) is no longer included 1868160814Ssimon nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up 1869160814Ssimon some needed definitions. 1870160814Ssimon [Steve Henson] 1871160814Ssimon 1872160814Ssimon *) Undo Cygwin change. 1873160814Ssimon [Ulf M�ller] 1874160814Ssimon 1875160814Ssimon *) Added support for proxy certificates according to RFC 3820. 1876160814Ssimon Because they may be a security thread to unaware applications, 1877160814Ssimon they must be explicitely allowed in run-time. See 1878160814Ssimon docs/HOWTO/proxy_certificates.txt for further information. 1879160814Ssimon [Richard Levitte] 1880160814Ssimon 1881160814Ssimon Changes between 0.9.7e and 0.9.7f [22 Mar 2005] 1882160814Ssimon 1883160814Ssimon *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating 1884160814Ssimon server and client random values. Previously 1885160814Ssimon (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in 1886160814Ssimon less random data when sizeof(time_t) > 4 (some 64 bit platforms). 1887160814Ssimon 1888160814Ssimon This change has negligible security impact because: 1889160814Ssimon 1890160814Ssimon 1. Server and client random values still have 24 bytes of pseudo random 1891160814Ssimon data. 1892160814Ssimon 1893160814Ssimon 2. Server and client random values are sent in the clear in the initial 1894160814Ssimon handshake. 1895160814Ssimon 1896160814Ssimon 3. The master secret is derived using the premaster secret (48 bytes in 1897160814Ssimon size for static RSA ciphersuites) as well as client server and random 1898160814Ssimon values. 1899160814Ssimon 1900160814Ssimon The OpenSSL team would like to thank the UK NISCC for bringing this issue 1901160814Ssimon to our attention. 1902160814Ssimon 1903160814Ssimon [Stephen Henson, reported by UK NISCC] 1904160814Ssimon 1905160814Ssimon *) Use Windows randomness collection on Cygwin. 1906160814Ssimon [Ulf M�ller] 1907160814Ssimon 1908160814Ssimon *) Fix hang in EGD/PRNGD query when communication socket is closed 1909160814Ssimon prematurely by EGD/PRNGD. 1910160814Ssimon [Darren Tucker <dtucker@zip.com.au> via Lutz J�nicke, resolves #1014] 1911160814Ssimon 1912160814Ssimon *) Prompt for pass phrases when appropriate for PKCS12 input format. 1913160814Ssimon [Steve Henson] 1914160814Ssimon 1915160814Ssimon *) Back-port of selected performance improvements from development 1916160814Ssimon branch, as well as improved support for PowerPC platforms. 1917160814Ssimon [Andy Polyakov] 1918160814Ssimon 1919160814Ssimon *) Add lots of checks for memory allocation failure, error codes to indicate 1920160814Ssimon failure and freeing up memory if a failure occurs. 1921160814Ssimon [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson] 1922160814Ssimon 1923160814Ssimon *) Add new -passin argument to dgst. 1924160814Ssimon [Steve Henson] 1925160814Ssimon 1926160814Ssimon *) Perform some character comparisons of different types in X509_NAME_cmp: 1927160814Ssimon this is needed for some certificates that reencode DNs into UTF8Strings 1928160814Ssimon (in violation of RFC3280) and can't or wont issue name rollover 1929160814Ssimon certificates. 1930160814Ssimon [Steve Henson] 1931160814Ssimon 1932160814Ssimon *) Make an explicit check during certificate validation to see that 1933160814Ssimon the CA setting in each certificate on the chain is correct. As a 1934160814Ssimon side effect always do the following basic checks on extensions, 1935160814Ssimon not just when there's an associated purpose to the check: 1936160814Ssimon 1937160814Ssimon - if there is an unhandled critical extension (unless the user 1938160814Ssimon has chosen to ignore this fault) 1939160814Ssimon - if the path length has been exceeded (if one is set at all) 1940160814Ssimon - that certain extensions fit the associated purpose (if one has 1941160814Ssimon been given) 1942160814Ssimon [Richard Levitte] 1943160814Ssimon 1944142425Snectar Changes between 0.9.7d and 0.9.7e [25 Oct 2004] 1945142425Snectar 1946142425Snectar *) Avoid a race condition when CRLs are checked in a multi threaded 1947142425Snectar environment. This would happen due to the reordering of the revoked 1948142425Snectar entries during signature checking and serial number lookup. Now the 1949142425Snectar encoding is cached and the serial number sort performed under a lock. 1950142425Snectar Add new STACK function sk_is_sorted(). 1951142425Snectar [Steve Henson] 1952142425Snectar 1953142425Snectar *) Add Delta CRL to the extension code. 1954142425Snectar [Steve Henson] 1955142425Snectar 1956142425Snectar *) Various fixes to s3_pkt.c so alerts are sent properly. 1957142425Snectar [David Holmes <d.holmes@f5.com>] 1958142425Snectar 1959142425Snectar *) Reduce the chances of duplicate issuer name and serial numbers (in 1960142425Snectar violation of RFC3280) using the OpenSSL certificate creation utilities. 1961142425Snectar This is done by creating a random 64 bit value for the initial serial 1962142425Snectar number when a serial number file is created or when a self signed 1963142425Snectar certificate is created using 'openssl req -x509'. The initial serial 1964142425Snectar number file is created using 'openssl x509 -next_serial' in CA.pl 1965142425Snectar rather than being initialized to 1. 1966142425Snectar [Steve Henson] 1967142425Snectar 1968127128Snectar Changes between 0.9.7c and 0.9.7d [17 Mar 2004] 1969127128Snectar 1970127128Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 1971160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 1972127128Snectar [Joe Orton, Steve Henson] 1973127128Snectar 1974127128Snectar *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites 1975160814Ssimon (CVE-2004-0112) 1976127128Snectar [Joe Orton, Steve Henson] 1977127128Snectar 1978127128Snectar *) Make it possible to have multiple active certificates with the same 1979127128Snectar subject in the CA index file. This is done only if the keyword 1980127128Snectar 'unique_subject' is set to 'no' in the main CA section (default 1981127128Snectar if 'CA_default') of the configuration file. The value is saved 1982127128Snectar with the database itself in a separate index attribute file, 1983127128Snectar named like the index file with '.attr' appended to the name. 1984127128Snectar [Richard Levitte] 1985127128Snectar 1986127128Snectar *) X509 verify fixes. Disable broken certificate workarounds when 1987127128Snectar X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if 1988127128Snectar keyUsage extension present. Don't accept CRLs with unhandled critical 1989127128Snectar extensions: since verify currently doesn't process CRL extensions this 1990127128Snectar rejects a CRL with *any* critical extensions. Add new verify error codes 1991127128Snectar for these cases. 1992127128Snectar [Steve Henson] 1993127128Snectar 1994127128Snectar *) When creating an OCSP nonce use an OCTET STRING inside the extnValue. 1995127128Snectar A clarification of RFC2560 will require the use of OCTET STRINGs and 1996127128Snectar some implementations cannot handle the current raw format. Since OpenSSL 1997127128Snectar copies and compares OCSP nonces as opaque blobs without any attempt at 1998127128Snectar parsing them this should not create any compatibility issues. 1999127128Snectar [Steve Henson] 2000127128Snectar 2001127128Snectar *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when 2002127128Snectar calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without 2003127128Snectar this HMAC (and other) operations are several times slower than OpenSSL 2004127128Snectar < 0.9.7. 2005127128Snectar [Steve Henson] 2006127128Snectar 2007127128Snectar *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex(). 2008127128Snectar [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 2009127128Snectar 2010127128Snectar *) Use the correct content when signing type "other". 2011127128Snectar [Steve Henson] 2012127128Snectar 2013120631Snectar Changes between 0.9.7b and 0.9.7c [30 Sep 2003] 2014120631Snectar 2015120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 2016120631Snectar 2017120631Snectar Stop out of bounds reads in the ASN1 code when presented with 2018160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 2019120631Snectar 2020160814Ssimon Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545). 2021120631Snectar 2022120631Snectar If verify callback ignores invalid public key errors don't try to check 2023120631Snectar certificate signature with the NULL public key. 2024120631Snectar 2025120631Snectar [Steve Henson] 2026120631Snectar 2027120631Snectar *) New -ignore_err option in ocsp application to stop the server 2028120631Snectar exiting on the first error in a request. 2029120631Snectar [Steve Henson] 2030120631Snectar 2031120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 2032120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 2033120631Snectar specifications. 2034120631Snectar [Steve Henson] 2035120631Snectar 2036120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 2037120631Snectar extra data after the compression methods not only for TLS 1.0 2038120631Snectar but also for SSL 3.0 (as required by the specification). 2039120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 2040120631Snectar 2041120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 2042120631Snectar when it's 512 *bits* long, not 512 bytes. 2043120631Snectar [Richard Levitte] 2044120631Snectar 2045120631Snectar *) Change AES_cbc_encrypt() so it outputs exact multiple of 2046120631Snectar blocks during encryption. 2047120631Snectar [Richard Levitte] 2048120631Snectar 2049120631Snectar *) Various fixes to base64 BIO and non blocking I/O. On write 2050120631Snectar flushes were not handled properly if the BIO retried. On read 2051120631Snectar data was not being buffered properly and had various logic bugs. 2052120631Snectar This also affects blocking I/O when the data being decoded is a 2053120631Snectar certain size. 2054120631Snectar [Steve Henson] 2055120631Snectar 2056120631Snectar *) Various S/MIME bugfixes and compatibility changes: 2057120631Snectar output correct application/pkcs7 MIME type if 2058120631Snectar PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures. 2059120631Snectar Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening 2060120631Snectar of files as .eml work). Correctly handle very long lines in MIME 2061120631Snectar parser. 2062120631Snectar [Steve Henson] 2063120631Snectar 2064120631Snectar Changes between 0.9.7a and 0.9.7b [10 Apr 2003] 2065120631Snectar 2066120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 2067120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 2068120631Snectar a protocol version number mismatch like a decryption error 2069120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 2070120631Snectar [Bodo Moeller] 2071120631Snectar 2072120631Snectar *) Turn on RSA blinding by default in the default implementation 2073120631Snectar to avoid a timing attack. Applications that don't want it can call 2074120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 2075120631Snectar They would be ill-advised to do so in most cases. 2076120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 2077120631Snectar 2078120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 2079120631Snectar seeded (in this case, the secret RSA exponent is abused as 2080120631Snectar an unpredictable seed -- if it is not unpredictable, there 2081120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 2082120631Snectar by remembering the creator's thread ID in rsa->blinding and 2083120631Snectar having all other threads use local one-time blinding factors 2084120631Snectar (this requires more computation than sharing rsa->blinding, but 2085120631Snectar avoids excessive locking; and if an RSA object is not shared 2086120631Snectar between threads, blinding will still be very fast). 2087120631Snectar [Bodo Moeller] 2088120631Snectar 2089120631Snectar *) Fixed a typo bug that would cause ENGINE_set_default() to set an 2090120631Snectar ENGINE as defaults for all supported algorithms irrespective of 2091120631Snectar the 'flags' parameter. 'flags' is now honoured, so applications 2092120631Snectar should make sure they are passing it correctly. 2093120631Snectar [Geoff Thorpe] 2094120631Snectar 2095120631Snectar *) Target "mingw" now allows native Windows code to be generated in 2096120631Snectar the Cygwin environment as well as with the MinGW compiler. 2097120631Snectar [Ulf Moeller] 2098120631Snectar 2099111147Snectar Changes between 0.9.7 and 0.9.7a [19 Feb 2003] 2100111147Snectar 2101111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 2102111147Snectar via timing by performing a MAC computation even if incorrrect 2103111147Snectar block cipher padding has been found. This is a countermeasure 2104111147Snectar against active attacks where the attacker has to distinguish 2105160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 2106111147Snectar 2107111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 2108111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 2109111147Snectar Martin Vuagnoux (EPFL, Ilion)] 2110111147Snectar 2111111147Snectar *) Make the no-err option work as intended. The intention with no-err 2112111147Snectar is not to have the whole error stack handling routines removed from 2113111147Snectar libcrypto, it's only intended to remove all the function name and 2114111147Snectar reason texts, thereby removing some of the footprint that may not 2115111147Snectar be interesting if those errors aren't displayed anyway. 2116111147Snectar 2117111147Snectar NOTE: it's still possible for any application or module to have it's 2118111147Snectar own set of error texts inserted. The routines are there, just not 2119111147Snectar used by default when no-err is given. 2120111147Snectar [Richard Levitte] 2121111147Snectar 2122111147Snectar *) Add support for FreeBSD on IA64. 2123111147Snectar [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454] 2124111147Snectar 2125111147Snectar *) Adjust DES_cbc_cksum() so it returns the same value as the MIT 2126111147Snectar Kerberos function mit_des_cbc_cksum(). Before this change, 2127111147Snectar the value returned by DES_cbc_cksum() was like the one from 2128111147Snectar mit_des_cbc_cksum(), except the bytes were swapped. 2129111147Snectar [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte] 2130111147Snectar 2131111147Snectar *) Allow an application to disable the automatic SSL chain building. 2132111147Snectar Before this a rather primitive chain build was always performed in 2133111147Snectar ssl3_output_cert_chain(): an application had no way to send the 2134111147Snectar correct chain if the automatic operation produced an incorrect result. 2135111147Snectar 2136111147Snectar Now the chain builder is disabled if either: 2137111147Snectar 2138111147Snectar 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert(). 2139111147Snectar 2140111147Snectar 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set. 2141111147Snectar 2142111147Snectar The reasoning behind this is that an application would not want the 2143111147Snectar auto chain building to take place if extra chain certificates are 2144111147Snectar present and it might also want a means of sending no additional 2145111147Snectar certificates (for example the chain has two certificates and the 2146111147Snectar root is omitted). 2147111147Snectar [Steve Henson] 2148111147Snectar 2149111147Snectar *) Add the possibility to build without the ENGINE framework. 2150111147Snectar [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 2151111147Snectar 2152111147Snectar *) Under Win32 gmtime() can return NULL: check return value in 2153111147Snectar OPENSSL_gmtime(). Add error code for case where gmtime() fails. 2154111147Snectar [Steve Henson] 2155111147Snectar 2156111147Snectar *) DSA routines: under certain error conditions uninitialized BN objects 2157111147Snectar could be freed. Solution: make sure initialization is performed early 2158111147Snectar enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>, 2159111147Snectar Nils Larsch <nla@trustcenter.de> via PR#459) 2160111147Snectar [Lutz Jaenicke] 2161111147Snectar 2162111147Snectar *) Another fix for SSLv2 session ID handling: the session ID was incorrectly 2163111147Snectar checked on reconnect on the client side, therefore session resumption 2164111147Snectar could still fail with a "ssl session id is different" error. This 2165111147Snectar behaviour is masked when SSL_OP_ALL is used due to 2166111147Snectar SSL_OP_MICROSOFT_SESS_ID_BUG being set. 2167111147Snectar Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 2168111147Snectar followup to PR #377. 2169111147Snectar [Lutz Jaenicke] 2170111147Snectar 2171111147Snectar *) IA-32 assembler support enhancements: unified ELF targets, support 2172111147Snectar for SCO/Caldera platforms, fix for Cygwin shared build. 2173111147Snectar [Andy Polyakov] 2174111147Snectar 2175111147Snectar *) Add support for FreeBSD on sparc64. As a consequence, support for 2176111147Snectar FreeBSD on non-x86 processors is separate from x86 processors on 2177111147Snectar the config script, much like the NetBSD support. 2178111147Snectar [Richard Levitte & Kris Kennaway <kris@obsecurity.org>] 2179111147Snectar 2180109998Smarkm Changes between 0.9.6h and 0.9.7 [31 Dec 2002] 2181109998Smarkm 2182120631Snectar [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after 2183120631Snectar OpenSSL 0.9.7.] 2184120631Snectar 2185109998Smarkm *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED 2186109998Smarkm code (06) was taken as the first octet of the session ID and the last 2187109998Smarkm octet was ignored consequently. As a result SSLv2 client side session 2188109998Smarkm caching could not have worked due to the session ID mismatch between 2189109998Smarkm client and server. 2190109998Smarkm Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 2191109998Smarkm PR #377. 2192109998Smarkm [Lutz Jaenicke] 2193109998Smarkm 2194109998Smarkm *) Change the declaration of needed Kerberos libraries to use EX_LIBS 2195109998Smarkm instead of the special (and badly supported) LIBKRB5. LIBKRB5 is 2196109998Smarkm removed entirely. 2197109998Smarkm [Richard Levitte] 2198109998Smarkm 2199109998Smarkm *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it 2200109998Smarkm seems that in spite of existing for more than a year, many application 2201109998Smarkm author have done nothing to provide the necessary callbacks, which 2202109998Smarkm means that this particular engine will not work properly anywhere. 2203109998Smarkm This is a very unfortunate situation which forces us, in the name 2204109998Smarkm of usability, to give the hw_ncipher.c a static lock, which is part 2205109998Smarkm of libcrypto. 2206109998Smarkm NOTE: This is for the 0.9.7 series ONLY. This hack will never 2207109998Smarkm appear in 0.9.8 or later. We EXPECT application authors to have 2208109998Smarkm dealt properly with this when 0.9.8 is released (unless we actually 2209109998Smarkm make such changes in the libcrypto locking code that changes will 2210109998Smarkm have to be made anyway). 2211109998Smarkm [Richard Levitte] 2212109998Smarkm 2213109998Smarkm *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content 2214109998Smarkm octets have been read, EOF or an error occurs. Without this change 2215109998Smarkm some truncated ASN1 structures will not produce an error. 2216109998Smarkm [Steve Henson] 2217109998Smarkm 2218109998Smarkm *) Disable Heimdal support, since it hasn't been fully implemented. 2219109998Smarkm Still give the possibility to force the use of Heimdal, but with 2220109998Smarkm warnings and a request that patches get sent to openssl-dev. 2221109998Smarkm [Richard Levitte] 2222109998Smarkm 2223109998Smarkm *) Add the VC-CE target, introduce the WINCE sysname, and add 2224109998Smarkm INSTALL.WCE and appropriate conditionals to make it build. 2225109998Smarkm [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 2226109998Smarkm 2227109998Smarkm *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and 2228109998Smarkm cygssl-x.y.z.dll, where x, y and z are the major, minor and 2229109998Smarkm edit numbers of the version. 2230109998Smarkm [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 2231109998Smarkm 2232109998Smarkm *) Introduce safe string copy and catenation functions 2233109998Smarkm (BUF_strlcpy() and BUF_strlcat()). 2234109998Smarkm [Ben Laurie (CHATS) and Richard Levitte] 2235109998Smarkm 2236109998Smarkm *) Avoid using fixed-size buffers for one-line DNs. 2237109998Smarkm [Ben Laurie (CHATS)] 2238109998Smarkm 2239109998Smarkm *) Add BUF_MEM_grow_clean() to avoid information leakage when 2240109998Smarkm resizing buffers containing secrets, and use where appropriate. 2241109998Smarkm [Ben Laurie (CHATS)] 2242109998Smarkm 2243109998Smarkm *) Avoid using fixed size buffers for configuration file location. 2244109998Smarkm [Ben Laurie (CHATS)] 2245109998Smarkm 2246109998Smarkm *) Avoid filename truncation for various CA files. 2247109998Smarkm [Ben Laurie (CHATS)] 2248109998Smarkm 2249109998Smarkm *) Use sizeof in preference to magic numbers. 2250109998Smarkm [Ben Laurie (CHATS)] 2251109998Smarkm 2252109998Smarkm *) Avoid filename truncation in cert requests. 2253109998Smarkm [Ben Laurie (CHATS)] 2254109998Smarkm 2255109998Smarkm *) Add assertions to check for (supposedly impossible) buffer 2256109998Smarkm overflows. 2257109998Smarkm [Ben Laurie (CHATS)] 2258109998Smarkm 2259109998Smarkm *) Don't cache truncated DNS entries in the local cache (this could 2260109998Smarkm potentially lead to a spoofing attack). 2261109998Smarkm [Ben Laurie (CHATS)] 2262109998Smarkm 2263109998Smarkm *) Fix various buffers to be large enough for hex/decimal 2264109998Smarkm representations in a platform independent manner. 2265109998Smarkm [Ben Laurie (CHATS)] 2266109998Smarkm 2267109998Smarkm *) Add CRYPTO_realloc_clean() to avoid information leakage when 2268109998Smarkm resizing buffers containing secrets, and use where appropriate. 2269109998Smarkm [Ben Laurie (CHATS)] 2270109998Smarkm 2271109998Smarkm *) Add BIO_indent() to avoid much slightly worrying code to do 2272109998Smarkm indents. 2273109998Smarkm [Ben Laurie (CHATS)] 2274109998Smarkm 2275109998Smarkm *) Convert sprintf()/BIO_puts() to BIO_printf(). 2276109998Smarkm [Ben Laurie (CHATS)] 2277109998Smarkm 2278109998Smarkm *) buffer_gets() could terminate with the buffer only half 2279109998Smarkm full. Fixed. 2280109998Smarkm [Ben Laurie (CHATS)] 2281109998Smarkm 2282109998Smarkm *) Add assertions to prevent user-supplied crypto functions from 2283109998Smarkm overflowing internal buffers by having large block sizes, etc. 2284109998Smarkm [Ben Laurie (CHATS)] 2285109998Smarkm 2286109998Smarkm *) New OPENSSL_assert() macro (similar to assert(), but enabled 2287109998Smarkm unconditionally). 2288109998Smarkm [Ben Laurie (CHATS)] 2289109998Smarkm 2290109998Smarkm *) Eliminate unused copy of key in RC4. 2291109998Smarkm [Ben Laurie (CHATS)] 2292109998Smarkm 2293109998Smarkm *) Eliminate unused and incorrectly sized buffers for IV in pem.h. 2294109998Smarkm [Ben Laurie (CHATS)] 2295109998Smarkm 2296109998Smarkm *) Fix off-by-one error in EGD path. 2297109998Smarkm [Ben Laurie (CHATS)] 2298109998Smarkm 2299109998Smarkm *) If RANDFILE path is too long, ignore instead of truncating. 2300109998Smarkm [Ben Laurie (CHATS)] 2301109998Smarkm 2302109998Smarkm *) Eliminate unused and incorrectly sized X.509 structure 2303109998Smarkm CBCParameter. 2304109998Smarkm [Ben Laurie (CHATS)] 2305109998Smarkm 2306109998Smarkm *) Eliminate unused and dangerous function knumber(). 2307109998Smarkm [Ben Laurie (CHATS)] 2308109998Smarkm 2309109998Smarkm *) Eliminate unused and dangerous structure, KSSL_ERR. 2310109998Smarkm [Ben Laurie (CHATS)] 2311109998Smarkm 2312109998Smarkm *) Protect against overlong session ID context length in an encoded 2313109998Smarkm session object. Since these are local, this does not appear to be 2314109998Smarkm exploitable. 2315109998Smarkm [Ben Laurie (CHATS)] 2316109998Smarkm 2317109998Smarkm *) Change from security patch (see 0.9.6e below) that did not affect 2318109998Smarkm the 0.9.6 release series: 2319109998Smarkm 2320109998Smarkm Remote buffer overflow in SSL3 protocol - an attacker could 2321109998Smarkm supply an oversized master key in Kerberos-enabled versions. 2322160814Ssimon (CVE-2002-0657) 2323109998Smarkm [Ben Laurie (CHATS)] 2324109998Smarkm 2325109998Smarkm *) Change the SSL kerb5 codes to match RFC 2712. 2326109998Smarkm [Richard Levitte] 2327109998Smarkm 2328109998Smarkm *) Make -nameopt work fully for req and add -reqopt switch. 2329109998Smarkm [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson] 2330109998Smarkm 2331109998Smarkm *) The "block size" for block ciphers in CFB and OFB mode should be 1. 2332109998Smarkm [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>] 2333109998Smarkm 2334109998Smarkm *) Make sure tests can be performed even if the corresponding algorithms 2335109998Smarkm have been removed entirely. This was also the last step to make 2336109998Smarkm OpenSSL compilable with DJGPP under all reasonable conditions. 2337109998Smarkm [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>] 2338109998Smarkm 2339109998Smarkm *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT 2340109998Smarkm to allow version independent disabling of normally unselected ciphers, 2341109998Smarkm which may be activated as a side-effect of selecting a single cipher. 2342109998Smarkm 2343109998Smarkm (E.g., cipher list string "RSA" enables ciphersuites that are left 2344109998Smarkm out of "ALL" because they do not provide symmetric encryption. 2345109998Smarkm "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) 2346109998Smarkm [Lutz Jaenicke, Bodo Moeller] 2347109998Smarkm 2348109998Smarkm *) Add appropriate support for separate platform-dependent build 2349109998Smarkm directories. The recommended way to make a platform-dependent 2350109998Smarkm build directory is the following (tested on Linux), maybe with 2351109998Smarkm some local tweaks: 2352109998Smarkm 2353109998Smarkm # Place yourself outside of the OpenSSL source tree. In 2354109998Smarkm # this example, the environment variable OPENSSL_SOURCE 2355109998Smarkm # is assumed to contain the absolute OpenSSL source directory. 2356109998Smarkm mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" 2357109998Smarkm cd objtree/"`uname -s`-`uname -r`-`uname -m`" 2358111147Snectar (cd $OPENSSL_SOURCE; find . -type f) | while read F; do 2359109998Smarkm mkdir -p `dirname $F` 2360109998Smarkm ln -s $OPENSSL_SOURCE/$F $F 2361109998Smarkm done 2362109998Smarkm 2363109998Smarkm To be absolutely sure not to disturb the source tree, a "make clean" 2364109998Smarkm is a good thing. If it isn't successfull, don't worry about it, 2365109998Smarkm it probably means the source directory is very clean. 2366109998Smarkm [Richard Levitte] 2367109998Smarkm 2368109998Smarkm *) Make sure any ENGINE control commands make local copies of string 2369109998Smarkm pointers passed to them whenever necessary. Otherwise it is possible 2370109998Smarkm the caller may have overwritten (or deallocated) the original string 2371109998Smarkm data when a later ENGINE operation tries to use the stored values. 2372109998Smarkm [G�tz Babin-Ebell <babinebell@trustcenter.de>] 2373109998Smarkm 2374109998Smarkm *) Improve diagnostics in file reading and command-line digests. 2375109998Smarkm [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>] 2376109998Smarkm 2377109998Smarkm *) Add AES modes CFB and OFB to the object database. Correct an 2378109998Smarkm error in AES-CFB decryption. 2379109998Smarkm [Richard Levitte] 2380109998Smarkm 2381109998Smarkm *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this 2382109998Smarkm allows existing EVP_CIPHER_CTX structures to be reused after 2383109998Smarkm calling EVP_*Final(). This behaviour is used by encryption 2384109998Smarkm BIOs and some applications. This has the side effect that 2385109998Smarkm applications must explicitly clean up cipher contexts with 2386109998Smarkm EVP_CIPHER_CTX_cleanup() or they will leak memory. 2387109998Smarkm [Steve Henson] 2388109998Smarkm 2389109998Smarkm *) Check the values of dna and dnb in bn_mul_recursive before calling 2390109998Smarkm bn_mul_comba (a non zero value means the a or b arrays do not contain 2391109998Smarkm n2 elements) and fallback to bn_mul_normal if either is not zero. 2392109998Smarkm [Steve Henson] 2393109998Smarkm 2394109998Smarkm *) Fix escaping of non-ASCII characters when using the -subj option 2395109998Smarkm of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) 2396109998Smarkm [Lutz Jaenicke] 2397109998Smarkm 2398109998Smarkm *) Make object definitions compliant to LDAP (RFC2256): SN is the short 2399109998Smarkm form for "surname", serialNumber has no short form. 2400109998Smarkm Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; 2401109998Smarkm therefore remove "mail" short name for "internet 7". 2402109998Smarkm The OID for unique identifiers in X509 certificates is 2403109998Smarkm x500UniqueIdentifier, not uniqueIdentifier. 2404109998Smarkm Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) 2405109998Smarkm [Lutz Jaenicke] 2406109998Smarkm 2407109998Smarkm *) Add an "init" command to the ENGINE config module and auto initialize 2408109998Smarkm ENGINEs. Without any "init" command the ENGINE will be initialized 2409109998Smarkm after all ctrl commands have been executed on it. If init=1 the 2410109998Smarkm ENGINE is initailized at that point (ctrls before that point are run 2411109998Smarkm on the uninitialized ENGINE and after on the initialized one). If 2412109998Smarkm init=0 then the ENGINE will not be iniatialized at all. 2413109998Smarkm [Steve Henson] 2414109998Smarkm 2415109998Smarkm *) Fix the 'app_verify_callback' interface so that the user-defined 2416109998Smarkm argument is actually passed to the callback: In the 2417109998Smarkm SSL_CTX_set_cert_verify_callback() prototype, the callback 2418109998Smarkm declaration has been changed from 2419109998Smarkm int (*cb)() 2420109998Smarkm into 2421109998Smarkm int (*cb)(X509_STORE_CTX *,void *); 2422109998Smarkm in ssl_verify_cert_chain (ssl/ssl_cert.c), the call 2423109998Smarkm i=s->ctx->app_verify_callback(&ctx) 2424109998Smarkm has been changed into 2425109998Smarkm i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). 2426109998Smarkm 2427109998Smarkm To update applications using SSL_CTX_set_cert_verify_callback(), 2428109998Smarkm a dummy argument can be added to their callback functions. 2429109998Smarkm [D. K. Smetters <smetters@parc.xerox.com>] 2430109998Smarkm 2431109998Smarkm *) Added the '4758cca' ENGINE to support IBM 4758 cards. 2432109998Smarkm [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe] 2433109998Smarkm 2434109998Smarkm *) Add and OPENSSL_LOAD_CONF define which will cause 2435109998Smarkm OpenSSL_add_all_algorithms() to load the openssl.cnf config file. 2436109998Smarkm This allows older applications to transparently support certain 2437109998Smarkm OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. 2438109998Smarkm Two new functions OPENSSL_add_all_algorithms_noconf() which will never 2439109998Smarkm load the config file and OPENSSL_add_all_algorithms_conf() which will 2440109998Smarkm always load it have also been added. 2441109998Smarkm [Steve Henson] 2442109998Smarkm 2443109998Smarkm *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. 2444109998Smarkm Adjust NIDs and EVP layer. 2445109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 2446109998Smarkm 2447109998Smarkm *) Config modules support in openssl utility. 2448109998Smarkm 2449109998Smarkm Most commands now load modules from the config file, 2450109998Smarkm though in a few (such as version) this isn't done 2451109998Smarkm because it couldn't be used for anything. 2452109998Smarkm 2453109998Smarkm In the case of ca and req the config file used is 2454109998Smarkm the same as the utility itself: that is the -config 2455109998Smarkm command line option can be used to specify an 2456109998Smarkm alternative file. 2457109998Smarkm [Steve Henson] 2458109998Smarkm 2459109998Smarkm *) Move default behaviour from OPENSSL_config(). If appname is NULL 2460109998Smarkm use "openssl_conf" if filename is NULL use default openssl config file. 2461109998Smarkm [Steve Henson] 2462109998Smarkm 2463109998Smarkm *) Add an argument to OPENSSL_config() to allow the use of an alternative 2464109998Smarkm config section name. Add a new flag to tolerate a missing config file 2465109998Smarkm and move code to CONF_modules_load_file(). 2466109998Smarkm [Steve Henson] 2467109998Smarkm 2468109998Smarkm *) Support for crypto accelerator cards from Accelerated Encryption 2469109998Smarkm Processing, www.aep.ie. (Use engine 'aep') 2470109998Smarkm The support was copied from 0.9.6c [engine] and adapted/corrected 2471109998Smarkm to work with the new engine framework. 2472109998Smarkm [AEP Inc. and Richard Levitte] 2473109998Smarkm 2474109998Smarkm *) Support for SureWare crypto accelerator cards from Baltimore 2475109998Smarkm Technologies. (Use engine 'sureware') 2476109998Smarkm The support was copied from 0.9.6c [engine] and adapted 2477109998Smarkm to work with the new engine framework. 2478109998Smarkm [Richard Levitte] 2479109998Smarkm 2480109998Smarkm *) Have the CHIL engine fork-safe (as defined by nCipher) and actually 2481109998Smarkm make the newer ENGINE framework commands for the CHIL engine work. 2482109998Smarkm [Toomas Kiisk <vix@cyber.ee> and Richard Levitte] 2483109998Smarkm 2484109998Smarkm *) Make it possible to produce shared libraries on ReliantUNIX. 2485109998Smarkm [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte] 2486109998Smarkm 2487109998Smarkm *) Add the configuration target debug-linux-ppro. 2488109998Smarkm Make 'openssl rsa' use the general key loading routines 2489109998Smarkm implemented in apps.c, and make those routines able to 2490109998Smarkm handle the key format FORMAT_NETSCAPE and the variant 2491109998Smarkm FORMAT_IISSGC. 2492109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 2493109998Smarkm 2494109998Smarkm *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 2495109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 2496109998Smarkm 2497109998Smarkm *) Add -keyform to rsautl, and document -engine. 2498109998Smarkm [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>] 2499109998Smarkm 2500109998Smarkm *) Change BIO_new_file (crypto/bio/bss_file.c) to use new 2501109998Smarkm BIO_R_NO_SUCH_FILE error code rather than the generic 2502109998Smarkm ERR_R_SYS_LIB error code if fopen() fails with ENOENT. 2503109998Smarkm [Ben Laurie] 2504109998Smarkm 2505109998Smarkm *) Add new functions 2506109998Smarkm ERR_peek_last_error 2507109998Smarkm ERR_peek_last_error_line 2508109998Smarkm ERR_peek_last_error_line_data. 2509109998Smarkm These are similar to 2510109998Smarkm ERR_peek_error 2511109998Smarkm ERR_peek_error_line 2512109998Smarkm ERR_peek_error_line_data, 2513109998Smarkm but report on the latest error recorded rather than the first one 2514109998Smarkm still in the error queue. 2515109998Smarkm [Ben Laurie, Bodo Moeller] 2516109998Smarkm 2517109998Smarkm *) default_algorithms option in ENGINE config module. This allows things 2518109998Smarkm like: 2519109998Smarkm default_algorithms = ALL 2520109998Smarkm default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS 2521109998Smarkm [Steve Henson] 2522109998Smarkm 2523109998Smarkm *) Prelminary ENGINE config module. 2524109998Smarkm [Steve Henson] 2525109998Smarkm 2526109998Smarkm *) New experimental application configuration code. 2527109998Smarkm [Steve Henson] 2528109998Smarkm 2529109998Smarkm *) Change the AES code to follow the same name structure as all other 2530109998Smarkm symmetric ciphers, and behave the same way. Move everything to 2531109998Smarkm the directory crypto/aes, thereby obsoleting crypto/rijndael. 2532109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 2533109998Smarkm 2534109998Smarkm *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. 2535109998Smarkm [Ben Laurie and Theo de Raadt] 2536109998Smarkm 2537109998Smarkm *) Add option to output public keys in req command. 2538109998Smarkm [Massimiliano Pala madwolf@openca.org] 2539109998Smarkm 2540109998Smarkm *) Use wNAFs in EC_POINTs_mul() for improved efficiency 2541109998Smarkm (up to about 10% better than before for P-192 and P-224). 2542109998Smarkm [Bodo Moeller] 2543109998Smarkm 2544109998Smarkm *) New functions/macros 2545109998Smarkm 2546109998Smarkm SSL_CTX_set_msg_callback(ctx, cb) 2547109998Smarkm SSL_CTX_set_msg_callback_arg(ctx, arg) 2548109998Smarkm SSL_set_msg_callback(ssl, cb) 2549109998Smarkm SSL_set_msg_callback_arg(ssl, arg) 2550109998Smarkm 2551109998Smarkm to request calling a callback function 2552109998Smarkm 2553109998Smarkm void cb(int write_p, int version, int content_type, 2554109998Smarkm const void *buf, size_t len, SSL *ssl, void *arg) 2555109998Smarkm 2556109998Smarkm whenever a protocol message has been completely received 2557109998Smarkm (write_p == 0) or sent (write_p == 1). Here 'version' is the 2558109998Smarkm protocol version according to which the SSL library interprets 2559109998Smarkm the current protocol message (SSL2_VERSION, SSL3_VERSION, or 2560109998Smarkm TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or 2561109998Smarkm the content type as defined in the SSL 3.0/TLS 1.0 protocol 2562109998Smarkm specification (change_cipher_spec(20), alert(21), handshake(22)). 2563109998Smarkm 'buf' and 'len' point to the actual message, 'ssl' to the 2564109998Smarkm SSL object, and 'arg' is the application-defined value set by 2565109998Smarkm SSL[_CTX]_set_msg_callback_arg(). 2566109998Smarkm 2567109998Smarkm 'openssl s_client' and 'openssl s_server' have new '-msg' options 2568109998Smarkm to enable a callback that displays all protocol messages. 2569109998Smarkm [Bodo Moeller] 2570109998Smarkm 2571109998Smarkm *) Change the shared library support so shared libraries are built as 2572109998Smarkm soon as the corresponding static library is finished, and thereby get 2573109998Smarkm openssl and the test programs linked against the shared library. 2574109998Smarkm This still only happens when the keyword "shard" has been given to 2575109998Smarkm the configuration scripts. 2576109998Smarkm 2577109998Smarkm NOTE: shared library support is still an experimental thing, and 2578109998Smarkm backward binary compatibility is still not guaranteed. 2579109998Smarkm ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte] 2580109998Smarkm 2581109998Smarkm *) Add support for Subject Information Access extension. 2582109998Smarkm [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 2583109998Smarkm 2584109998Smarkm *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero 2585109998Smarkm additional bytes when new memory had to be allocated, not just 2586109998Smarkm when reusing an existing buffer. 2587109998Smarkm [Bodo Moeller] 2588109998Smarkm 2589109998Smarkm *) New command line and configuration option 'utf8' for the req command. 2590109998Smarkm This allows field values to be specified as UTF8 strings. 2591109998Smarkm [Steve Henson] 2592109998Smarkm 2593109998Smarkm *) Add -multi and -mr options to "openssl speed" - giving multiple parallel 2594109998Smarkm runs for the former and machine-readable output for the latter. 2595109998Smarkm [Ben Laurie] 2596109998Smarkm 2597109998Smarkm *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion 2598109998Smarkm of the e-mail address in the DN (i.e., it will go into a certificate 2599109998Smarkm extension only). The new configuration file option 'email_in_dn = no' 2600109998Smarkm has the same effect. 2601109998Smarkm [Massimiliano Pala madwolf@openca.org] 2602109998Smarkm 2603109998Smarkm *) Change all functions with names starting with des_ to be starting 2604109998Smarkm with DES_ instead. Add wrappers that are compatible with libdes, 2605109998Smarkm but are named _ossl_old_des_*. Finally, add macros that map the 2606109998Smarkm des_* symbols to the corresponding _ossl_old_des_* if libdes 2607109998Smarkm compatibility is desired. If OpenSSL 0.9.6c compatibility is 2608109998Smarkm desired, the des_* symbols will be mapped to DES_*, with one 2609109998Smarkm exception. 2610109998Smarkm 2611109998Smarkm Since we provide two compatibility mappings, the user needs to 2612109998Smarkm define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes 2613109998Smarkm compatibility is desired. The default (i.e., when that macro 2614109998Smarkm isn't defined) is OpenSSL 0.9.6c compatibility. 2615109998Smarkm 2616109998Smarkm There are also macros that enable and disable the support of old 2617109998Smarkm des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT 2618109998Smarkm and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those 2619109998Smarkm are defined, the default will apply: to support the old des routines. 2620109998Smarkm 2621109998Smarkm In either case, one must include openssl/des.h to get the correct 2622109998Smarkm definitions. Do not try to just include openssl/des_old.h, that 2623109998Smarkm won't work. 2624109998Smarkm 2625109998Smarkm NOTE: This is a major break of an old API into a new one. Software 2626109998Smarkm authors are encouraged to switch to the DES_ style functions. Some 2627109998Smarkm time in the future, des_old.h and the libdes compatibility functions 2628109998Smarkm will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the 2629109998Smarkm default), and then completely removed. 2630109998Smarkm [Richard Levitte] 2631109998Smarkm 2632109998Smarkm *) Test for certificates which contain unsupported critical extensions. 2633109998Smarkm If such a certificate is found during a verify operation it is 2634109998Smarkm rejected by default: this behaviour can be overridden by either 2635109998Smarkm handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or 2636109998Smarkm by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function 2637109998Smarkm X509_supported_extension() has also been added which returns 1 if a 2638109998Smarkm particular extension is supported. 2639109998Smarkm [Steve Henson] 2640109998Smarkm 2641109998Smarkm *) Modify the behaviour of EVP cipher functions in similar way to digests 2642109998Smarkm to retain compatibility with existing code. 2643109998Smarkm [Steve Henson] 2644109998Smarkm 2645109998Smarkm *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain 2646109998Smarkm compatibility with existing code. In particular the 'ctx' parameter does 2647109998Smarkm not have to be to be initialized before the call to EVP_DigestInit() and 2648109998Smarkm it is tidied up after a call to EVP_DigestFinal(). New function 2649109998Smarkm EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function 2650109998Smarkm EVP_MD_CTX_copy() changed to not require the destination to be 2651109998Smarkm initialized valid and new function EVP_MD_CTX_copy_ex() added which 2652109998Smarkm requires the destination to be valid. 2653109998Smarkm 2654109998Smarkm Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), 2655109998Smarkm EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). 2656109998Smarkm [Steve Henson] 2657109998Smarkm 2658109998Smarkm *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it 2659109998Smarkm so that complete 'Handshake' protocol structures are kept in memory 2660109998Smarkm instead of overwriting 'msg_type' and 'length' with 'body' data. 2661109998Smarkm [Bodo Moeller] 2662109998Smarkm 2663109998Smarkm *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. 2664109998Smarkm [Massimo Santin via Richard Levitte] 2665109998Smarkm 2666109998Smarkm *) Major restructuring to the underlying ENGINE code. This includes 2667109998Smarkm reduction of linker bloat, separation of pure "ENGINE" manipulation 2668109998Smarkm (initialisation, etc) from functionality dealing with implementations 2669109998Smarkm of specific crypto iterfaces. This change also introduces integrated 2670109998Smarkm support for symmetric ciphers and digest implementations - so ENGINEs 2671109998Smarkm can now accelerate these by providing EVP_CIPHER and EVP_MD 2672109998Smarkm implementations of their own. This is detailed in crypto/engine/README 2673109998Smarkm as it couldn't be adequately described here. However, there are a few 2674109998Smarkm API changes worth noting - some RSA, DSA, DH, and RAND functions that 2675109998Smarkm were changed in the original introduction of ENGINE code have now 2676109998Smarkm reverted back - the hooking from this code to ENGINE is now a good 2677109998Smarkm deal more passive and at run-time, operations deal directly with 2678109998Smarkm RSA_METHODs, DSA_METHODs (etc) as they did before, rather than 2679109998Smarkm dereferencing through an ENGINE pointer any more. Also, the ENGINE 2680109998Smarkm functions dealing with BN_MOD_EXP[_CRT] handlers have been removed - 2681109998Smarkm they were not being used by the framework as there is no concept of a 2682109998Smarkm BIGNUM_METHOD and they could not be generalised to the new 2683109998Smarkm 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, 2684109998Smarkm ENGINE_cpy() has been removed as it cannot be consistently defined in 2685109998Smarkm the new code. 2686109998Smarkm [Geoff Thorpe] 2687109998Smarkm 2688109998Smarkm *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. 2689109998Smarkm [Steve Henson] 2690109998Smarkm 2691109998Smarkm *) Change mkdef.pl to sort symbols that get the same entry number, 2692109998Smarkm and make sure the automatically generated functions ERR_load_* 2693109998Smarkm become part of libeay.num as well. 2694109998Smarkm [Richard Levitte] 2695109998Smarkm 2696109998Smarkm *) New function SSL_renegotiate_pending(). This returns true once 2697109998Smarkm renegotiation has been requested (either SSL_renegotiate() call 2698109998Smarkm or HelloRequest/ClientHello receveived from the peer) and becomes 2699109998Smarkm false once a handshake has been completed. 2700109998Smarkm (For servers, SSL_renegotiate() followed by SSL_do_handshake() 2701109998Smarkm sends a HelloRequest, but does not ensure that a handshake takes 2702109998Smarkm place. SSL_renegotiate_pending() is useful for checking if the 2703109998Smarkm client has followed the request.) 2704109998Smarkm [Bodo Moeller] 2705109998Smarkm 2706109998Smarkm *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. 2707109998Smarkm By default, clients may request session resumption even during 2708109998Smarkm renegotiation (if session ID contexts permit); with this option, 2709109998Smarkm session resumption is possible only in the first handshake. 2710109998Smarkm 2711109998Smarkm SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes 2712109998Smarkm more bits available for options that should not be part of 2713109998Smarkm SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). 2714109998Smarkm [Bodo Moeller] 2715109998Smarkm 2716109998Smarkm *) Add some demos for certificate and certificate request creation. 2717109998Smarkm [Steve Henson] 2718109998Smarkm 2719109998Smarkm *) Make maximum certificate chain size accepted from the peer application 2720109998Smarkm settable (SSL*_get/set_max_cert_list()), as proposed by 2721109998Smarkm "Douglas E. Engert" <deengert@anl.gov>. 2722109998Smarkm [Lutz Jaenicke] 2723109998Smarkm 2724109998Smarkm *) Add support for shared libraries for Unixware-7 2725109998Smarkm (Boyd Lynn Gerber <gerberb@zenez.com>). 2726109998Smarkm [Lutz Jaenicke] 2727109998Smarkm 2728109998Smarkm *) Add a "destroy" handler to ENGINEs that allows structural cleanup to 2729109998Smarkm be done prior to destruction. Use this to unload error strings from 2730109998Smarkm ENGINEs that load their own error strings. NB: This adds two new API 2731109998Smarkm functions to "get" and "set" this destroy handler in an ENGINE. 2732109998Smarkm [Geoff Thorpe] 2733109998Smarkm 2734109998Smarkm *) Alter all existing ENGINE implementations (except "openssl" and 2735109998Smarkm "openbsd") to dynamically instantiate their own error strings. This 2736109998Smarkm makes them more flexible to be built both as statically-linked ENGINEs 2737109998Smarkm and self-contained shared-libraries loadable via the "dynamic" ENGINE. 2738109998Smarkm Also, add stub code to each that makes building them as self-contained 2739109998Smarkm shared-libraries easier (see README.ENGINE). 2740109998Smarkm [Geoff Thorpe] 2741109998Smarkm 2742109998Smarkm *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE 2743109998Smarkm implementations into applications that are completely implemented in 2744109998Smarkm self-contained shared-libraries. The "dynamic" ENGINE exposes control 2745109998Smarkm commands that can be used to configure what shared-library to load and 2746109998Smarkm to control aspects of the way it is handled. Also, made an update to 2747109998Smarkm the README.ENGINE file that brings its information up-to-date and 2748109998Smarkm provides some information and instructions on the "dynamic" ENGINE 2749109998Smarkm (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). 2750109998Smarkm [Geoff Thorpe] 2751109998Smarkm 2752109998Smarkm *) Make it possible to unload ranges of ERR strings with a new 2753109998Smarkm "ERR_unload_strings" function. 2754109998Smarkm [Geoff Thorpe] 2755109998Smarkm 2756109998Smarkm *) Add a copy() function to EVP_MD. 2757109998Smarkm [Ben Laurie] 2758109998Smarkm 2759109998Smarkm *) Make EVP_MD routines take a context pointer instead of just the 2760109998Smarkm md_data void pointer. 2761109998Smarkm [Ben Laurie] 2762109998Smarkm 2763109998Smarkm *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates 2764109998Smarkm that the digest can only process a single chunk of data 2765109998Smarkm (typically because it is provided by a piece of 2766109998Smarkm hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application 2767109998Smarkm is only going to provide a single chunk of data, and hence the 2768109998Smarkm framework needn't accumulate the data for oneshot drivers. 2769109998Smarkm [Ben Laurie] 2770109998Smarkm 2771109998Smarkm *) As with "ERR", make it possible to replace the underlying "ex_data" 2772109998Smarkm functions. This change also alters the storage and management of global 2773109998Smarkm ex_data state - it's now all inside ex_data.c and all "class" code (eg. 2774109998Smarkm RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class 2775109998Smarkm index counters. The API functions that use this state have been changed 2776109998Smarkm to take a "class_index" rather than pointers to the class's local STACK 2777109998Smarkm and counter, and there is now an API function to dynamically create new 2778109998Smarkm classes. This centralisation allows us to (a) plug a lot of the 2779109998Smarkm thread-safety problems that existed, and (b) makes it possible to clean 2780109998Smarkm up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) 2781109998Smarkm such data would previously have always leaked in application code and 2782109998Smarkm workarounds were in place to make the memory debugging turn a blind eye 2783109998Smarkm to it. Application code that doesn't use this new function will still 2784109998Smarkm leak as before, but their memory debugging output will announce it now 2785109998Smarkm rather than letting it slide. 2786109998Smarkm 2787109998Smarkm Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change 2788109998Smarkm induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now 2789109998Smarkm has a return value to indicate success or failure. 2790109998Smarkm [Geoff Thorpe] 2791109998Smarkm 2792109998Smarkm *) Make it possible to replace the underlying "ERR" functions such that the 2793109998Smarkm global state (2 LHASH tables and 2 locks) is only used by the "default" 2794109998Smarkm implementation. This change also adds two functions to "get" and "set" 2795109998Smarkm the implementation prior to it being automatically set the first time 2796109998Smarkm any other ERR function takes place. Ie. an application can call "get", 2797109998Smarkm pass the return value to a module it has just loaded, and that module 2798109998Smarkm can call its own "set" function using that value. This means the 2799109998Smarkm module's "ERR" operations will use (and modify) the error state in the 2800109998Smarkm application and not in its own statically linked copy of OpenSSL code. 2801109998Smarkm [Geoff Thorpe] 2802109998Smarkm 2803109998Smarkm *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment 2804109998Smarkm reference counts. This performs normal REF_PRINT/REF_CHECK macros on 2805109998Smarkm the operation, and provides a more encapsulated way for external code 2806109998Smarkm (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code 2807109998Smarkm to use these functions rather than manually incrementing the counts. 2808109998Smarkm 2809109998Smarkm Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". 2810109998Smarkm [Geoff Thorpe] 2811109998Smarkm 2812109998Smarkm *) Add EVP test program. 2813109998Smarkm [Ben Laurie] 2814109998Smarkm 2815109998Smarkm *) Add symmetric cipher support to ENGINE. Expect the API to change! 2816109998Smarkm [Ben Laurie] 2817109998Smarkm 2818109998Smarkm *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() 2819109998Smarkm X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), 2820109998Smarkm X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). 2821109998Smarkm These allow a CRL to be built without having to access X509_CRL fields 2822109998Smarkm directly. Modify 'ca' application to use new functions. 2823109998Smarkm [Steve Henson] 2824109998Smarkm 2825109998Smarkm *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended 2826109998Smarkm bug workarounds. Rollback attack detection is a security feature. 2827109998Smarkm The problem will only arise on OpenSSL servers when TLSv1 is not 2828109998Smarkm available (sslv3_server_method() or SSL_OP_NO_TLSv1). 2829109998Smarkm Software authors not wanting to support TLSv1 will have special reasons 2830109998Smarkm for their choice and can explicitly enable this option. 2831109998Smarkm [Bodo Moeller, Lutz Jaenicke] 2832109998Smarkm 2833109998Smarkm *) Rationalise EVP so it can be extended: don't include a union of 2834109998Smarkm cipher/digest structures, add init/cleanup functions for EVP_MD_CTX 2835109998Smarkm (similar to those existing for EVP_CIPHER_CTX). 2836109998Smarkm Usage example: 2837109998Smarkm 2838109998Smarkm EVP_MD_CTX md; 2839109998Smarkm 2840109998Smarkm EVP_MD_CTX_init(&md); /* new function call */ 2841109998Smarkm EVP_DigestInit(&md, EVP_sha1()); 2842109998Smarkm EVP_DigestUpdate(&md, in, len); 2843109998Smarkm EVP_DigestFinal(&md, out, NULL); 2844109998Smarkm EVP_MD_CTX_cleanup(&md); /* new function call */ 2845109998Smarkm 2846109998Smarkm [Ben Laurie] 2847109998Smarkm 2848109998Smarkm *) Make DES key schedule conform to the usual scheme, as well as 2849109998Smarkm correcting its structure. This means that calls to DES functions 2850109998Smarkm now have to pass a pointer to a des_key_schedule instead of a 2851109998Smarkm plain des_key_schedule (which was actually always a pointer 2852109998Smarkm anyway): E.g., 2853109998Smarkm 2854109998Smarkm des_key_schedule ks; 2855109998Smarkm 2856109998Smarkm des_set_key_checked(..., &ks); 2857109998Smarkm des_ncbc_encrypt(..., &ks, ...); 2858109998Smarkm 2859109998Smarkm (Note that a later change renames 'des_...' into 'DES_...'.) 2860109998Smarkm [Ben Laurie] 2861109998Smarkm 2862109998Smarkm *) Initial reduction of linker bloat: the use of some functions, such as 2863109998Smarkm PEM causes large amounts of unused functions to be linked in due to 2864109998Smarkm poor organisation. For example pem_all.c contains every PEM function 2865109998Smarkm which has a knock on effect of linking in large amounts of (unused) 2866109998Smarkm ASN1 code. Grouping together similar functions and splitting unrelated 2867109998Smarkm functions prevents this. 2868109998Smarkm [Steve Henson] 2869109998Smarkm 2870109998Smarkm *) Cleanup of EVP macros. 2871109998Smarkm [Ben Laurie] 2872109998Smarkm 2873109998Smarkm *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the 2874109998Smarkm correct _ecb suffix. 2875109998Smarkm [Ben Laurie] 2876109998Smarkm 2877109998Smarkm *) Add initial OCSP responder support to ocsp application. The 2878109998Smarkm revocation information is handled using the text based index 2879109998Smarkm use by the ca application. The responder can either handle 2880109998Smarkm requests generated internally, supplied in files (for example 2881109998Smarkm via a CGI script) or using an internal minimal server. 2882109998Smarkm [Steve Henson] 2883109998Smarkm 2884109998Smarkm *) Add configuration choices to get zlib compression for TLS. 2885109998Smarkm [Richard Levitte] 2886109998Smarkm 2887109998Smarkm *) Changes to Kerberos SSL for RFC 2712 compliance: 2888109998Smarkm 1. Implemented real KerberosWrapper, instead of just using 2889109998Smarkm KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] 2890109998Smarkm 2. Implemented optional authenticator field of KerberosWrapper. 2891109998Smarkm 2892109998Smarkm Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, 2893109998Smarkm and authenticator structs; see crypto/krb5/. 2894109998Smarkm 2895109998Smarkm Generalized Kerberos calls to support multiple Kerberos libraries. 2896109998Smarkm [Vern Staats <staatsvr@asc.hpc.mil>, 2897109998Smarkm Jeffrey Altman <jaltman@columbia.edu> 2898109998Smarkm via Richard Levitte] 2899109998Smarkm 2900109998Smarkm *) Cause 'openssl speed' to use fully hard-coded DSA keys as it 2901109998Smarkm already does with RSA. testdsa.h now has 'priv_key/pub_key' 2902109998Smarkm values for each of the key sizes rather than having just 2903109998Smarkm parameters (and 'speed' generating keys each time). 2904109998Smarkm [Geoff Thorpe] 2905109998Smarkm 2906109998Smarkm *) Speed up EVP routines. 2907109998Smarkm Before: 2908109998Smarkmencrypt 2909109998Smarkmtype 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 2910109998Smarkmdes-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k 2911109998Smarkmdes-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k 2912109998Smarkmdes-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k 2913109998Smarkmdecrypt 2914109998Smarkmdes-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k 2915109998Smarkmdes-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k 2916109998Smarkmdes-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k 2917109998Smarkm After: 2918109998Smarkmencrypt 2919109998Smarkmdes-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k 2920109998Smarkmdecrypt 2921109998Smarkmdes-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k 2922109998Smarkm [Ben Laurie] 2923109998Smarkm 2924109998Smarkm *) Added the OS2-EMX target. 2925109998Smarkm ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte] 2926109998Smarkm 2927109998Smarkm *) Rewrite apps to use NCONF routines instead of the old CONF. New functions 2928109998Smarkm to support NCONF routines in extension code. New function CONF_set_nconf() 2929109998Smarkm to allow functions which take an NCONF to also handle the old LHASH 2930109998Smarkm structure: this means that the old CONF compatible routines can be 2931109998Smarkm retained (in particular wrt extensions) without having to duplicate the 2932109998Smarkm code. New function X509V3_add_ext_nconf_sk to add extensions to a stack. 2933109998Smarkm [Steve Henson] 2934109998Smarkm 2935109998Smarkm *) Enhance the general user interface with mechanisms for inner control 2936109998Smarkm and with possibilities to have yes/no kind of prompts. 2937109998Smarkm [Richard Levitte] 2938109998Smarkm 2939109998Smarkm *) Change all calls to low level digest routines in the library and 2940109998Smarkm applications to use EVP. Add missing calls to HMAC_cleanup() and 2941109998Smarkm don't assume HMAC_CTX can be copied using memcpy(). 2942109998Smarkm [Verdon Walker <VWalker@novell.com>, Steve Henson] 2943109998Smarkm 2944109998Smarkm *) Add the possibility to control engines through control names but with 2945109998Smarkm arbitrary arguments instead of just a string. 2946109998Smarkm Change the key loaders to take a UI_METHOD instead of a callback 2947109998Smarkm function pointer. NOTE: this breaks binary compatibility with earlier 2948109998Smarkm versions of OpenSSL [engine]. 2949109998Smarkm Adapt the nCipher code for these new conditions and add a card insertion 2950109998Smarkm callback. 2951109998Smarkm [Richard Levitte] 2952109998Smarkm 2953109998Smarkm *) Enhance the general user interface with mechanisms to better support 2954109998Smarkm dialog box interfaces, application-defined prompts, the possibility 2955109998Smarkm to use defaults (for example default passwords from somewhere else) 2956109998Smarkm and interrupts/cancellations. 2957109998Smarkm [Richard Levitte] 2958109998Smarkm 2959109998Smarkm *) Tidy up PKCS#12 attribute handling. Add support for the CSP name 2960109998Smarkm attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. 2961109998Smarkm [Steve Henson] 2962109998Smarkm 2963109998Smarkm *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also 2964109998Smarkm tidy up some unnecessarily weird code in 'sk_new()'). 2965109998Smarkm [Geoff, reported by Diego Tartara <dtartara@novamens.com>] 2966109998Smarkm 2967109998Smarkm *) Change the key loading routines for ENGINEs to use the same kind 2968109998Smarkm callback (pem_password_cb) as all other routines that need this 2969109998Smarkm kind of callback. 2970109998Smarkm [Richard Levitte] 2971109998Smarkm 2972109998Smarkm *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with 2973109998Smarkm 256 bit (=32 byte) keys. Of course seeding with more entropy bytes 2974109998Smarkm than this minimum value is recommended. 2975109998Smarkm [Lutz Jaenicke] 2976109998Smarkm 2977109998Smarkm *) New random seeder for OpenVMS, using the system process statistics 2978109998Smarkm that are easily reachable. 2979109998Smarkm [Richard Levitte] 2980109998Smarkm 2981109998Smarkm *) Windows apparently can't transparently handle global 2982109998Smarkm variables defined in DLLs. Initialisations such as: 2983109998Smarkm 2984109998Smarkm const ASN1_ITEM *it = &ASN1_INTEGER_it; 2985109998Smarkm 2986109998Smarkm wont compile. This is used by the any applications that need to 2987109998Smarkm declare their own ASN1 modules. This was fixed by adding the option 2988109998Smarkm EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly 2989109998Smarkm needed for static libraries under Win32. 2990109998Smarkm [Steve Henson] 2991109998Smarkm 2992109998Smarkm *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle 2993109998Smarkm setting of purpose and trust fields. New X509_STORE trust and 2994109998Smarkm purpose functions and tidy up setting in other SSL functions. 2995109998Smarkm [Steve Henson] 2996109998Smarkm 2997109998Smarkm *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE 2998109998Smarkm structure. These are inherited by X509_STORE_CTX when it is 2999109998Smarkm initialised. This allows various defaults to be set in the 3000109998Smarkm X509_STORE structure (such as flags for CRL checking and custom 3001109998Smarkm purpose or trust settings) for functions which only use X509_STORE_CTX 3002109998Smarkm internally such as S/MIME. 3003109998Smarkm 3004109998Smarkm Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and 3005109998Smarkm trust settings if they are not set in X509_STORE. This allows X509_STORE 3006109998Smarkm purposes and trust (in S/MIME for example) to override any set by default. 3007109998Smarkm 3008109998Smarkm Add command line options for CRL checking to smime, s_client and s_server 3009109998Smarkm applications. 3010109998Smarkm [Steve Henson] 3011109998Smarkm 3012109998Smarkm *) Initial CRL based revocation checking. If the CRL checking flag(s) 3013109998Smarkm are set then the CRL is looked up in the X509_STORE structure and 3014109998Smarkm its validity and signature checked, then if the certificate is found 3015109998Smarkm in the CRL the verify fails with a revoked error. 3016109998Smarkm 3017109998Smarkm Various new CRL related callbacks added to X509_STORE_CTX structure. 3018109998Smarkm 3019109998Smarkm Command line options added to 'verify' application to support this. 3020109998Smarkm 3021109998Smarkm This needs some additional work, such as being able to handle multiple 3022109998Smarkm CRLs with different times, extension based lookup (rather than just 3023109998Smarkm by subject name) and ultimately more complete V2 CRL extension 3024109998Smarkm handling. 3025109998Smarkm [Steve Henson] 3026109998Smarkm 3027109998Smarkm *) Add a general user interface API (crypto/ui/). This is designed 3028109998Smarkm to replace things like des_read_password and friends (backward 3029109998Smarkm compatibility functions using this new API are provided). 3030109998Smarkm The purpose is to remove prompting functions from the DES code 3031109998Smarkm section as well as provide for prompting through dialog boxes in 3032109998Smarkm a window system and the like. 3033109998Smarkm [Richard Levitte] 3034109998Smarkm 3035109998Smarkm *) Add "ex_data" support to ENGINE so implementations can add state at a 3036109998Smarkm per-structure level rather than having to store it globally. 3037109998Smarkm [Geoff] 3038109998Smarkm 3039109998Smarkm *) Make it possible for ENGINE structures to be copied when retrieved by 3040109998Smarkm ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. 3041109998Smarkm This causes the "original" ENGINE structure to act like a template, 3042109998Smarkm analogous to the RSA vs. RSA_METHOD type of separation. Because of this 3043109998Smarkm operational state can be localised to each ENGINE structure, despite the 3044109998Smarkm fact they all share the same "methods". New ENGINE structures returned in 3045109998Smarkm this case have no functional references and the return value is the single 3046109998Smarkm structural reference. This matches the single structural reference returned 3047109998Smarkm by ENGINE_by_id() normally, when it is incremented on the pre-existing 3048109998Smarkm ENGINE structure. 3049109998Smarkm [Geoff] 3050109998Smarkm 3051109998Smarkm *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this 3052109998Smarkm needs to match any other type at all we need to manually clear the 3053109998Smarkm tag cache. 3054109998Smarkm [Steve Henson] 3055109998Smarkm 3056109998Smarkm *) Changes to the "openssl engine" utility to include; 3057109998Smarkm - verbosity levels ('-v', '-vv', and '-vvv') that provide information 3058109998Smarkm about an ENGINE's available control commands. 3059109998Smarkm - executing control commands from command line arguments using the 3060109998Smarkm '-pre' and '-post' switches. '-post' is only used if '-t' is 3061109998Smarkm specified and the ENGINE is successfully initialised. The syntax for 3062109998Smarkm the individual commands are colon-separated, for example; 3063109998Smarkm openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so 3064109998Smarkm [Geoff] 3065109998Smarkm 3066109998Smarkm *) New dynamic control command support for ENGINEs. ENGINEs can now 3067109998Smarkm declare their own commands (numbers), names (strings), descriptions, 3068109998Smarkm and input types for run-time discovery by calling applications. A 3069109998Smarkm subset of these commands are implicitly classed as "executable" 3070109998Smarkm depending on their input type, and only these can be invoked through 3071109998Smarkm the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this 3072109998Smarkm can be based on user input, config files, etc). The distinction is 3073109998Smarkm that "executable" commands cannot return anything other than a boolean 3074109998Smarkm result and can only support numeric or string input, whereas some 3075109998Smarkm discoverable commands may only be for direct use through 3076109998Smarkm ENGINE_ctrl(), eg. supporting the exchange of binary data, function 3077109998Smarkm pointers, or other custom uses. The "executable" commands are to 3078109998Smarkm support parameterisations of ENGINE behaviour that can be 3079109998Smarkm unambiguously defined by ENGINEs and used consistently across any 3080109998Smarkm OpenSSL-based application. Commands have been added to all the 3081109998Smarkm existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow 3082109998Smarkm control over shared-library paths without source code alterations. 3083109998Smarkm [Geoff] 3084109998Smarkm 3085109998Smarkm *) Changed all ENGINE implementations to dynamically allocate their 3086109998Smarkm ENGINEs rather than declaring them statically. Apart from this being 3087109998Smarkm necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, 3088109998Smarkm this also allows the implementations to compile without using the 3089109998Smarkm internal engine_int.h header. 3090109998Smarkm [Geoff] 3091109998Smarkm 3092109998Smarkm *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a 3093109998Smarkm 'const' value. Any code that should be able to modify a RAND_METHOD 3094109998Smarkm should already have non-const pointers to it (ie. they should only 3095109998Smarkm modify their own ones). 3096109998Smarkm [Geoff] 3097109998Smarkm 3098109998Smarkm *) Made a variety of little tweaks to the ENGINE code. 3099109998Smarkm - "atalla" and "ubsec" string definitions were moved from header files 3100109998Smarkm to C code. "nuron" string definitions were placed in variables 3101109998Smarkm rather than hard-coded - allowing parameterisation of these values 3102109998Smarkm later on via ctrl() commands. 3103109998Smarkm - Removed unused "#if 0"'d code. 3104109998Smarkm - Fixed engine list iteration code so it uses ENGINE_free() to release 3105109998Smarkm structural references. 3106109998Smarkm - Constified the RAND_METHOD element of ENGINE structures. 3107109998Smarkm - Constified various get/set functions as appropriate and added 3108109998Smarkm missing functions (including a catch-all ENGINE_cpy that duplicates 3109109998Smarkm all ENGINE values onto a new ENGINE except reference counts/state). 3110109998Smarkm - Removed NULL parameter checks in get/set functions. Setting a method 3111109998Smarkm or function to NULL is a way of cancelling out a previously set 3112109998Smarkm value. Passing a NULL ENGINE parameter is just plain stupid anyway 3113109998Smarkm and doesn't justify the extra error symbols and code. 3114109998Smarkm - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for 3115109998Smarkm flags from engine_int.h to engine.h. 3116109998Smarkm - Changed prototypes for ENGINE handler functions (init(), finish(), 3117109998Smarkm ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. 3118109998Smarkm [Geoff] 3119109998Smarkm 3120109998Smarkm *) Implement binary inversion algorithm for BN_mod_inverse in addition 3121109998Smarkm to the algorithm using long division. The binary algorithm can be 3122109998Smarkm used only if the modulus is odd. On 32-bit systems, it is faster 3123109998Smarkm only for relatively small moduli (roughly 20-30% for 128-bit moduli, 3124109998Smarkm roughly 5-15% for 256-bit moduli), so we use it only for moduli 3125109998Smarkm up to 450 bits. In 64-bit environments, the binary algorithm 3126109998Smarkm appears to be advantageous for much longer moduli; here we use it 3127109998Smarkm for moduli up to 2048 bits. 3128109998Smarkm [Bodo Moeller] 3129109998Smarkm 3130109998Smarkm *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code 3131109998Smarkm could not support the combine flag in choice fields. 3132109998Smarkm [Steve Henson] 3133109998Smarkm 3134109998Smarkm *) Add a 'copy_extensions' option to the 'ca' utility. This copies 3135109998Smarkm extensions from a certificate request to the certificate. 3136109998Smarkm [Steve Henson] 3137109998Smarkm 3138109998Smarkm *) Allow multiple 'certopt' and 'nameopt' options to be separated 3139109998Smarkm by commas. Add 'namopt' and 'certopt' options to the 'ca' config 3140109998Smarkm file: this allows the display of the certificate about to be 3141109998Smarkm signed to be customised, to allow certain fields to be included 3142109998Smarkm or excluded and extension details. The old system didn't display 3143109998Smarkm multicharacter strings properly, omitted fields not in the policy 3144109998Smarkm and couldn't display additional details such as extensions. 3145109998Smarkm [Steve Henson] 3146109998Smarkm 3147109998Smarkm *) Function EC_POINTs_mul for multiple scalar multiplication 3148109998Smarkm of an arbitrary number of elliptic curve points 3149109998Smarkm \sum scalars[i]*points[i], 3150109998Smarkm optionally including the generator defined for the EC_GROUP: 3151109998Smarkm scalar*generator + \sum scalars[i]*points[i]. 3152109998Smarkm 3153109998Smarkm EC_POINT_mul is a simple wrapper function for the typical case 3154109998Smarkm that the point list has just one item (besides the optional 3155109998Smarkm generator). 3156109998Smarkm [Bodo Moeller] 3157109998Smarkm 3158109998Smarkm *) First EC_METHODs for curves over GF(p): 3159109998Smarkm 3160109998Smarkm EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr 3161109998Smarkm operations and provides various method functions that can also 3162109998Smarkm operate with faster implementations of modular arithmetic. 3163109998Smarkm 3164109998Smarkm EC_GFp_mont_method() reuses most functions that are part of 3165109998Smarkm EC_GFp_simple_method, but uses Montgomery arithmetic. 3166109998Smarkm 3167109998Smarkm [Bodo Moeller; point addition and point doubling 3168109998Smarkm implementation directly derived from source code provided by 3169109998Smarkm Lenka Fibikova <fibikova@exp-math.uni-essen.de>] 3170109998Smarkm 3171109998Smarkm *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, 3172109998Smarkm crypto/ec/ec_lib.c): 3173109998Smarkm 3174109998Smarkm Curves are EC_GROUP objects (with an optional group generator) 3175109998Smarkm based on EC_METHODs that are built into the library. 3176109998Smarkm 3177109998Smarkm Points are EC_POINT objects based on EC_GROUP objects. 3178109998Smarkm 3179109998Smarkm Most of the framework would be able to handle curves over arbitrary 3180109998Smarkm finite fields, but as there are no obvious types for fields other 3181109998Smarkm than GF(p), some functions are limited to that for now. 3182109998Smarkm [Bodo Moeller] 3183109998Smarkm 3184109998Smarkm *) Add the -HTTP option to s_server. It is similar to -WWW, but requires 3185109998Smarkm that the file contains a complete HTTP response. 3186109998Smarkm [Richard Levitte] 3187109998Smarkm 3188109998Smarkm *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl 3189109998Smarkm change the def and num file printf format specifier from "%-40sXXX" 3190109998Smarkm to "%-39s XXX". The latter will always guarantee a space after the 3191109998Smarkm field while the former will cause them to run together if the field 3192109998Smarkm is 40 of more characters long. 3193109998Smarkm [Steve Henson] 3194109998Smarkm 3195109998Smarkm *) Constify the cipher and digest 'method' functions and structures 3196109998Smarkm and modify related functions to take constant EVP_MD and EVP_CIPHER 3197109998Smarkm pointers. 3198109998Smarkm [Steve Henson] 3199109998Smarkm 3200109998Smarkm *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them 3201109998Smarkm in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. 3202109998Smarkm [Bodo Moeller] 3203109998Smarkm 3204109998Smarkm *) Modify EVP_Digest*() routines so they now return values. Although the 3205109998Smarkm internal software routines can never fail additional hardware versions 3206109998Smarkm might. 3207109998Smarkm [Steve Henson] 3208109998Smarkm 3209109998Smarkm *) Clean up crypto/err/err.h and change some error codes to avoid conflicts: 3210109998Smarkm 3211109998Smarkm Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 3212109998Smarkm (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. 3213109998Smarkm 3214109998Smarkm ASN1 error codes 3215109998Smarkm ERR_R_NESTED_ASN1_ERROR 3216109998Smarkm ... 3217109998Smarkm ERR_R_MISSING_ASN1_EOS 3218109998Smarkm were 4 .. 9, conflicting with 3219109998Smarkm ERR_LIB_RSA (= ERR_R_RSA_LIB) 3220109998Smarkm ... 3221109998Smarkm ERR_LIB_PEM (= ERR_R_PEM_LIB). 3222109998Smarkm They are now 58 .. 63 (i.e., just below ERR_R_FATAL). 3223109998Smarkm 3224109998Smarkm Add new error code 'ERR_R_INTERNAL_ERROR'. 3225109998Smarkm [Bodo Moeller] 3226109998Smarkm 3227109998Smarkm *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock 3228109998Smarkm suffices. 3229109998Smarkm [Bodo Moeller] 3230109998Smarkm 3231109998Smarkm *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This 3232109998Smarkm sets the subject name for a new request or supersedes the 3233109998Smarkm subject name in a given request. Formats that can be parsed are 3234109998Smarkm 'CN=Some Name, OU=myOU, C=IT' 3235109998Smarkm and 3236109998Smarkm 'CN=Some Name/OU=myOU/C=IT'. 3237109998Smarkm 3238109998Smarkm Add options '-batch' and '-verbose' to 'openssl req'. 3239109998Smarkm [Massimiliano Pala <madwolf@hackmasters.net>] 3240109998Smarkm 3241109998Smarkm *) Introduce the possibility to access global variables through 3242109998Smarkm functions on platform were that's the best way to handle exporting 3243109998Smarkm global variables in shared libraries. To enable this functionality, 3244109998Smarkm one must configure with "EXPORT_VAR_AS_FN" or defined the C macro 3245109998Smarkm "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter 3246109998Smarkm is normally done by Configure or something similar). 3247109998Smarkm 3248109998Smarkm To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL 3249109998Smarkm in the source file (foo.c) like this: 3250109998Smarkm 3251109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; 3252109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(double,bar); 3253109998Smarkm 3254109998Smarkm To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL 3255109998Smarkm and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: 3256109998Smarkm 3257109998Smarkm OPENSSL_DECLARE_GLOBAL(int,foo); 3258109998Smarkm #define foo OPENSSL_GLOBAL_REF(foo) 3259109998Smarkm OPENSSL_DECLARE_GLOBAL(double,bar); 3260109998Smarkm #define bar OPENSSL_GLOBAL_REF(bar) 3261109998Smarkm 3262109998Smarkm The #defines are very important, and therefore so is including the 3263109998Smarkm header file everywhere where the defined globals are used. 3264109998Smarkm 3265109998Smarkm The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition 3266109998Smarkm of ASN.1 items, but that structure is a bit different. 3267109998Smarkm 3268109998Smarkm The largest change is in util/mkdef.pl which has been enhanced with 3269109998Smarkm better and easier to understand logic to choose which symbols should 3270109998Smarkm go into the Windows .def files as well as a number of fixes and code 3271109998Smarkm cleanup (among others, algorithm keywords are now sorted 3272109998Smarkm lexicographically to avoid constant rewrites). 3273109998Smarkm [Richard Levitte] 3274109998Smarkm 3275109998Smarkm *) In BN_div() keep a copy of the sign of 'num' before writing the 3276109998Smarkm result to 'rm' because if rm==num the value will be overwritten 3277109998Smarkm and produce the wrong result if 'num' is negative: this caused 3278109998Smarkm problems with BN_mod() and BN_nnmod(). 3279109998Smarkm [Steve Henson] 3280109998Smarkm 3281109998Smarkm *) Function OCSP_request_verify(). This checks the signature on an 3282109998Smarkm OCSP request and verifies the signer certificate. The signer 3283109998Smarkm certificate is just checked for a generic purpose and OCSP request 3284109998Smarkm trust settings. 3285109998Smarkm [Steve Henson] 3286109998Smarkm 3287109998Smarkm *) Add OCSP_check_validity() function to check the validity of OCSP 3288109998Smarkm responses. OCSP responses are prepared in real time and may only 3289109998Smarkm be a few seconds old. Simply checking that the current time lies 3290109998Smarkm between thisUpdate and nextUpdate max reject otherwise valid responses 3291109998Smarkm caused by either OCSP responder or client clock inaccuracy. Instead 3292109998Smarkm we allow thisUpdate and nextUpdate to fall within a certain period of 3293109998Smarkm the current time. The age of the response can also optionally be 3294109998Smarkm checked. Two new options -validity_period and -status_age added to 3295109998Smarkm ocsp utility. 3296109998Smarkm [Steve Henson] 3297109998Smarkm 3298109998Smarkm *) If signature or public key algorithm is unrecognized print out its 3299109998Smarkm OID rather that just UNKNOWN. 3300109998Smarkm [Steve Henson] 3301109998Smarkm 3302109998Smarkm *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and 3303109998Smarkm OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate 3304109998Smarkm ID to be generated from the issuer certificate alone which can then be 3305109998Smarkm passed to OCSP_id_issuer_cmp(). 3306109998Smarkm [Steve Henson] 3307109998Smarkm 3308109998Smarkm *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new 3309109998Smarkm ASN1 modules to export functions returning ASN1_ITEM pointers 3310109998Smarkm instead of the ASN1_ITEM structures themselves. This adds several 3311109998Smarkm new macros which allow the underlying ASN1 function/structure to 3312109998Smarkm be accessed transparently. As a result code should not use ASN1_ITEM 3313109998Smarkm references directly (such as &X509_it) but instead use the relevant 3314109998Smarkm macros (such as ASN1_ITEM_rptr(X509)). This option is to allow 3315109998Smarkm use of the new ASN1 code on platforms where exporting structures 3316109998Smarkm is problematical (for example in shared libraries) but exporting 3317109998Smarkm functions returning pointers to structures is not. 3318109998Smarkm [Steve Henson] 3319109998Smarkm 3320109998Smarkm *) Add support for overriding the generation of SSL/TLS session IDs. 3321109998Smarkm These callbacks can be registered either in an SSL_CTX or per SSL. 3322109998Smarkm The purpose of this is to allow applications to control, if they wish, 3323109998Smarkm the arbitrary values chosen for use as session IDs, particularly as it 3324109998Smarkm can be useful for session caching in multiple-server environments. A 3325109998Smarkm command-line switch for testing this (and any client code that wishes 3326109998Smarkm to use such a feature) has been added to "s_server". 3327109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 3328109998Smarkm 3329109998Smarkm *) Modify mkdef.pl to recognise and parse preprocessor conditionals 3330109998Smarkm of the form '#if defined(...) || defined(...) || ...' and 3331109998Smarkm '#if !defined(...) && !defined(...) && ...'. This also avoids 3332109998Smarkm the growing number of special cases it was previously handling. 3333109998Smarkm [Richard Levitte] 3334109998Smarkm 3335109998Smarkm *) Make all configuration macros available for application by making 3336109998Smarkm sure they are available in opensslconf.h, by giving them names starting 3337109998Smarkm with "OPENSSL_" to avoid conflicts with other packages and by making 3338109998Smarkm sure e_os2.h will cover all platform-specific cases together with 3339109998Smarkm opensslconf.h. 3340109998Smarkm Additionally, it is now possible to define configuration/platform- 3341109998Smarkm specific names (called "system identities"). In the C code, these 3342109998Smarkm are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another 3343109998Smarkm macro with the name beginning with "OPENSSL_SYS_", which is determined 3344109998Smarkm from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on 3345109998Smarkm what is available. 3346109998Smarkm [Richard Levitte] 3347109998Smarkm 3348109998Smarkm *) New option -set_serial to 'req' and 'x509' this allows the serial 3349109998Smarkm number to use to be specified on the command line. Previously self 3350109998Smarkm signed certificates were hard coded with serial number 0 and the 3351109998Smarkm CA options of 'x509' had to use a serial number in a file which was 3352109998Smarkm auto incremented. 3353109998Smarkm [Steve Henson] 3354109998Smarkm 3355109998Smarkm *) New options to 'ca' utility to support V2 CRL entry extensions. 3356109998Smarkm Currently CRL reason, invalidity date and hold instruction are 3357109998Smarkm supported. Add new CRL extensions to V3 code and some new objects. 3358109998Smarkm [Steve Henson] 3359109998Smarkm 3360109998Smarkm *) New function EVP_CIPHER_CTX_set_padding() this is used to 3361109998Smarkm disable standard block padding (aka PKCS#5 padding) in the EVP 3362109998Smarkm API, which was previously mandatory. This means that the data is 3363109998Smarkm not padded in any way and so the total length much be a multiple 3364109998Smarkm of the block size, otherwise an error occurs. 3365109998Smarkm [Steve Henson] 3366109998Smarkm 3367109998Smarkm *) Initial (incomplete) OCSP SSL support. 3368109998Smarkm [Steve Henson] 3369109998Smarkm 3370109998Smarkm *) New function OCSP_parse_url(). This splits up a URL into its host, 3371109998Smarkm port and path components: primarily to parse OCSP URLs. New -url 3372109998Smarkm option to ocsp utility. 3373109998Smarkm [Steve Henson] 3374109998Smarkm 3375109998Smarkm *) New nonce behavior. The return value of OCSP_check_nonce() now 3376109998Smarkm reflects the various checks performed. Applications can decide 3377109998Smarkm whether to tolerate certain situations such as an absent nonce 3378109998Smarkm in a response when one was present in a request: the ocsp application 3379109998Smarkm just prints out a warning. New function OCSP_add1_basic_nonce() 3380109998Smarkm this is to allow responders to include a nonce in a response even if 3381109998Smarkm the request is nonce-less. 3382109998Smarkm [Steve Henson] 3383109998Smarkm 3384109998Smarkm *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are 3385109998Smarkm skipped when using openssl x509 multiple times on a single input file, 3386109998Smarkm e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs". 3387109998Smarkm [Bodo Moeller] 3388109998Smarkm 3389109998Smarkm *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() 3390109998Smarkm set string type: to handle setting ASN1_TIME structures. Fix ca 3391109998Smarkm utility to correctly initialize revocation date of CRLs. 3392109998Smarkm [Steve Henson] 3393109998Smarkm 3394109998Smarkm *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override 3395109998Smarkm the clients preferred ciphersuites and rather use its own preferences. 3396109998Smarkm Should help to work around M$ SGC (Server Gated Cryptography) bug in 3397109998Smarkm Internet Explorer by ensuring unchanged hash method during stepup. 3398109998Smarkm (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) 3399109998Smarkm [Lutz Jaenicke] 3400109998Smarkm 3401109998Smarkm *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael 3402109998Smarkm to aes and add a new 'exist' option to print out symbols that don't 3403109998Smarkm appear to exist. 3404109998Smarkm [Steve Henson] 3405109998Smarkm 3406109998Smarkm *) Additional options to ocsp utility to allow flags to be set and 3407109998Smarkm additional certificates supplied. 3408109998Smarkm [Steve Henson] 3409109998Smarkm 3410109998Smarkm *) Add the option -VAfile to 'openssl ocsp', so the user can give the 3411109998Smarkm OCSP client a number of certificate to only verify the response 3412109998Smarkm signature against. 3413109998Smarkm [Richard Levitte] 3414109998Smarkm 3415109998Smarkm *) Update Rijndael code to version 3.0 and change EVP AES ciphers to 3416109998Smarkm handle the new API. Currently only ECB, CBC modes supported. Add new 3417109998Smarkm AES OIDs. 3418109998Smarkm 3419109998Smarkm Add TLS AES ciphersuites as described in RFC3268, "Advanced 3420109998Smarkm Encryption Standard (AES) Ciphersuites for Transport Layer 3421109998Smarkm Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were 3422109998Smarkm not enabled by default and were not part of the "ALL" ciphersuite 3423109998Smarkm alias because they were not yet official; they could be 3424109998Smarkm explicitly requested by specifying the "AESdraft" ciphersuite 3425109998Smarkm group alias. In the final release of OpenSSL 0.9.7, the group 3426109998Smarkm alias is called "AES" and is part of "ALL".) 3427109998Smarkm [Ben Laurie, Steve Henson, Bodo Moeller] 3428109998Smarkm 3429109998Smarkm *) New function OCSP_copy_nonce() to copy nonce value (if present) from 3430109998Smarkm request to response. 3431109998Smarkm [Steve Henson] 3432109998Smarkm 3433109998Smarkm *) Functions for OCSP responders. OCSP_request_onereq_count(), 3434109998Smarkm OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() 3435109998Smarkm extract information from a certificate request. OCSP_response_create() 3436109998Smarkm creates a response and optionally adds a basic response structure. 3437109998Smarkm OCSP_basic_add1_status() adds a complete single response to a basic 3438109998Smarkm response and returns the OCSP_SINGLERESP structure just added (to allow 3439109998Smarkm extensions to be included for example). OCSP_basic_add1_cert() adds a 3440109998Smarkm certificate to a basic response and OCSP_basic_sign() signs a basic 3441109998Smarkm response with various flags. New helper functions ASN1_TIME_check() 3442109998Smarkm (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() 3443109998Smarkm (converts ASN1_TIME to GeneralizedTime). 3444109998Smarkm [Steve Henson] 3445109998Smarkm 3446109998Smarkm *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() 3447109998Smarkm in a single operation. X509_get0_pubkey_bitstr() extracts the public_key 3448109998Smarkm structure from a certificate. X509_pubkey_digest() digests the public_key 3449109998Smarkm contents: this is used in various key identifiers. 3450109998Smarkm [Steve Henson] 3451109998Smarkm 3452109998Smarkm *) Make sk_sort() tolerate a NULL argument. 3453109998Smarkm [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>] 3454109998Smarkm 3455109998Smarkm *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates 3456109998Smarkm passed by the function are trusted implicitly. If any of them signed the 3457109998Smarkm response then it is assumed to be valid and is not verified. 3458109998Smarkm [Steve Henson] 3459109998Smarkm 3460109998Smarkm *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT 3461109998Smarkm to data. This was previously part of the PKCS7 ASN1 code. This 3462109998Smarkm was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. 3463109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 3464109998Smarkm <support@securenetterm.com>] 3465109998Smarkm 3466109998Smarkm *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 3467109998Smarkm routines: without these tracing memory leaks is very painful. 3468109998Smarkm Fix leaks in PKCS12 and PKCS7 routines. 3469109998Smarkm [Steve Henson] 3470109998Smarkm 3471109998Smarkm *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). 3472109998Smarkm Previously it initialised the 'type' argument to V_ASN1_UTCTIME which 3473109998Smarkm effectively meant GeneralizedTime would never be used. Now it 3474109998Smarkm is initialised to -1 but X509_time_adj() now has to check the value 3475109998Smarkm and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or 3476109998Smarkm V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. 3477109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 3478109998Smarkm <support@securenetterm.com>] 3479109998Smarkm 3480109998Smarkm *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously 3481109998Smarkm result in a zero length in the ASN1_INTEGER structure which was 3482109998Smarkm not consistent with the structure when d2i_ASN1_INTEGER() was used 3483109998Smarkm and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() 3484109998Smarkm to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() 3485109998Smarkm where it did not print out a minus for negative ASN1_INTEGER. 3486109998Smarkm [Steve Henson] 3487109998Smarkm 3488109998Smarkm *) Add summary printout to ocsp utility. The various functions which 3489109998Smarkm convert status values to strings have been renamed to: 3490109998Smarkm OCSP_response_status_str(), OCSP_cert_status_str() and 3491109998Smarkm OCSP_crl_reason_str() and are no longer static. New options 3492109998Smarkm to verify nonce values and to disable verification. OCSP response 3493109998Smarkm printout format cleaned up. 3494109998Smarkm [Steve Henson] 3495109998Smarkm 3496109998Smarkm *) Add additional OCSP certificate checks. These are those specified 3497109998Smarkm in RFC2560. This consists of two separate checks: the CA of the 3498109998Smarkm certificate being checked must either be the OCSP signer certificate 3499109998Smarkm or the issuer of the OCSP signer certificate. In the latter case the 3500109998Smarkm OCSP signer certificate must contain the OCSP signing extended key 3501109998Smarkm usage. This check is performed by attempting to match the OCSP 3502109998Smarkm signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash 3503109998Smarkm in the OCSP_CERTID structures of the response. 3504109998Smarkm [Steve Henson] 3505109998Smarkm 3506109998Smarkm *) Initial OCSP certificate verification added to OCSP_basic_verify() 3507109998Smarkm and related routines. This uses the standard OpenSSL certificate 3508109998Smarkm verify routines to perform initial checks (just CA validity) and 3509109998Smarkm to obtain the certificate chain. Then additional checks will be 3510109998Smarkm performed on the chain. Currently the root CA is checked to see 3511109998Smarkm if it is explicitly trusted for OCSP signing. This is used to set 3512109998Smarkm a root CA as a global signing root: that is any certificate that 3513109998Smarkm chains to that CA is an acceptable OCSP signing certificate. 3514109998Smarkm [Steve Henson] 3515109998Smarkm 3516109998Smarkm *) New '-extfile ...' option to 'openssl ca' for reading X.509v3 3517109998Smarkm extensions from a separate configuration file. 3518109998Smarkm As when reading extensions from the main configuration file, 3519109998Smarkm the '-extensions ...' option may be used for specifying the 3520109998Smarkm section to use. 3521109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 3522109998Smarkm 3523109998Smarkm *) New OCSP utility. Allows OCSP requests to be generated or 3524109998Smarkm read. The request can be sent to a responder and the output 3525109998Smarkm parsed, outputed or printed in text form. Not complete yet: 3526109998Smarkm still needs to check the OCSP response validity. 3527109998Smarkm [Steve Henson] 3528109998Smarkm 3529109998Smarkm *) New subcommands for 'openssl ca': 3530109998Smarkm 'openssl ca -status <serial>' prints the status of the cert with 3531109998Smarkm the given serial number (according to the index file). 3532109998Smarkm 'openssl ca -updatedb' updates the expiry status of certificates 3533109998Smarkm in the index file. 3534109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 3535109998Smarkm 3536109998Smarkm *) New '-newreq-nodes' command option to CA.pl. This is like 3537109998Smarkm '-newreq', but calls 'openssl req' with the '-nodes' option 3538109998Smarkm so that the resulting key is not encrypted. 3539109998Smarkm [Damien Miller <djm@mindrot.org>] 3540109998Smarkm 3541109998Smarkm *) New configuration for the GNU Hurd. 3542109998Smarkm [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte] 3543109998Smarkm 3544109998Smarkm *) Initial code to implement OCSP basic response verify. This 3545109998Smarkm is currently incomplete. Currently just finds the signer's 3546109998Smarkm certificate and verifies the signature on the response. 3547109998Smarkm [Steve Henson] 3548109998Smarkm 3549109998Smarkm *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in 3550109998Smarkm value of OPENSSLDIR. This is available via the new '-d' option 3551109998Smarkm to 'openssl version', and is also included in 'openssl version -a'. 3552109998Smarkm [Bodo Moeller] 3553109998Smarkm 3554109998Smarkm *) Allowing defining memory allocation callbacks that will be given 3555109998Smarkm file name and line number information in additional arguments 3556109998Smarkm (a const char* and an int). The basic functionality remains, as 3557109998Smarkm well as the original possibility to just replace malloc(), 3558109998Smarkm realloc() and free() by functions that do not know about these 3559109998Smarkm additional arguments. To register and find out the current 3560109998Smarkm settings for extended allocation functions, the following 3561109998Smarkm functions are provided: 3562109998Smarkm 3563109998Smarkm CRYPTO_set_mem_ex_functions 3564109998Smarkm CRYPTO_set_locked_mem_ex_functions 3565109998Smarkm CRYPTO_get_mem_ex_functions 3566109998Smarkm CRYPTO_get_locked_mem_ex_functions 3567109998Smarkm 3568109998Smarkm These work the same way as CRYPTO_set_mem_functions and friends. 3569109998Smarkm CRYPTO_get_[locked_]mem_functions now writes 0 where such an 3570109998Smarkm extended allocation function is enabled. 3571109998Smarkm Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where 3572109998Smarkm a conventional allocation function is enabled. 3573109998Smarkm [Richard Levitte, Bodo Moeller] 3574109998Smarkm 3575109998Smarkm *) Finish off removing the remaining LHASH function pointer casts. 3576109998Smarkm There should no longer be any prototype-casting required when using 3577109998Smarkm the LHASH abstraction, and any casts that remain are "bugs". See 3578109998Smarkm the callback types and macros at the head of lhash.h for details 3579109998Smarkm (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). 3580109998Smarkm [Geoff Thorpe] 3581109998Smarkm 3582109998Smarkm *) Add automatic query of EGD sockets in RAND_poll() for the unix variant. 3583109998Smarkm If /dev/[u]random devices are not available or do not return enough 3584109998Smarkm entropy, EGD style sockets (served by EGD or PRNGD) will automatically 3585109998Smarkm be queried. 3586109998Smarkm The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and 3587109998Smarkm /etc/entropy will be queried once each in this sequence, quering stops 3588109998Smarkm when enough entropy was collected without querying more sockets. 3589109998Smarkm [Lutz Jaenicke] 3590109998Smarkm 3591109998Smarkm *) Change the Unix RAND_poll() variant to be able to poll several 3592109998Smarkm random devices, as specified by DEVRANDOM, until a sufficient amount 3593109998Smarkm of data has been collected. We spend at most 10 ms on each file 3594109998Smarkm (select timeout) and read in non-blocking mode. DEVRANDOM now 3595109998Smarkm defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" 3596109998Smarkm (previously it was just the string "/dev/urandom"), so on typical 3597109998Smarkm platforms the 10 ms delay will never occur. 3598109998Smarkm Also separate out the Unix variant to its own file, rand_unix.c. 3599109998Smarkm For VMS, there's a currently-empty rand_vms.c. 3600109998Smarkm [Richard Levitte] 3601109998Smarkm 3602109998Smarkm *) Move OCSP client related routines to ocsp_cl.c. These 3603109998Smarkm provide utility functions which an application needing 3604109998Smarkm to issue a request to an OCSP responder and analyse the 3605109998Smarkm response will typically need: as opposed to those which an 3606109998Smarkm OCSP responder itself would need which will be added later. 3607109998Smarkm 3608109998Smarkm OCSP_request_sign() signs an OCSP request with an API similar 3609109998Smarkm to PKCS7_sign(). OCSP_response_status() returns status of OCSP 3610109998Smarkm response. OCSP_response_get1_basic() extracts basic response 3611109998Smarkm from response. OCSP_resp_find_status(): finds and extracts status 3612109998Smarkm information from an OCSP_CERTID structure (which will be created 3613109998Smarkm when the request structure is built). These are built from lower 3614109998Smarkm level functions which work on OCSP_SINGLERESP structures but 3615109998Smarkm wont normally be used unless the application wishes to examine 3616109998Smarkm extensions in the OCSP response for example. 3617109998Smarkm 3618109998Smarkm Replace nonce routines with a pair of functions. 3619109998Smarkm OCSP_request_add1_nonce() adds a nonce value and optionally 3620109998Smarkm generates a random value. OCSP_check_nonce() checks the 3621109998Smarkm validity of the nonce in an OCSP response. 3622109998Smarkm [Steve Henson] 3623109998Smarkm 3624109998Smarkm *) Change function OCSP_request_add() to OCSP_request_add0_id(). 3625109998Smarkm This doesn't copy the supplied OCSP_CERTID and avoids the 3626109998Smarkm need to free up the newly created id. Change return type 3627109998Smarkm to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. 3628109998Smarkm This can then be used to add extensions to the request. 3629109998Smarkm Deleted OCSP_request_new(), since most of its functionality 3630109998Smarkm is now in OCSP_REQUEST_new() (and the case insensitive name 3631109998Smarkm clash) apart from the ability to set the request name which 3632109998Smarkm will be added elsewhere. 3633109998Smarkm [Steve Henson] 3634109998Smarkm 3635109998Smarkm *) Update OCSP API. Remove obsolete extensions argument from 3636109998Smarkm various functions. Extensions are now handled using the new 3637109998Smarkm OCSP extension code. New simple OCSP HTTP function which 3638109998Smarkm can be used to send requests and parse the response. 3639109998Smarkm [Steve Henson] 3640109998Smarkm 3641109998Smarkm *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new 3642109998Smarkm ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN 3643109998Smarkm uses the special reorder version of SET OF to sort the attributes 3644109998Smarkm and reorder them to match the encoded order. This resolves a long 3645109998Smarkm standing problem: a verify on a PKCS7 structure just after signing 3646109998Smarkm it used to fail because the attribute order did not match the 3647109998Smarkm encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: 3648109998Smarkm it uses the received order. This is necessary to tolerate some broken 3649109998Smarkm software that does not order SET OF. This is handled by encoding 3650109998Smarkm as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) 3651109998Smarkm to produce the required SET OF. 3652109998Smarkm [Steve Henson] 3653109998Smarkm 3654109998Smarkm *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and 3655109998Smarkm OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header 3656109998Smarkm files to get correct declarations of the ASN.1 item variables. 3657109998Smarkm [Richard Levitte] 3658109998Smarkm 3659109998Smarkm *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many 3660109998Smarkm PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: 3661109998Smarkm asn1_check_tlen() would sometimes attempt to use 'ctx' when it was 3662109998Smarkm NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). 3663109998Smarkm New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant 3664109998Smarkm ASN1_ITEM and no wrapper functions. 3665109998Smarkm [Steve Henson] 3666109998Smarkm 3667109998Smarkm *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These 3668109998Smarkm replace the old function pointer based I/O routines. Change most of 3669109998Smarkm the *_d2i_bio() and *_d2i_fp() functions to use these. 3670109998Smarkm [Steve Henson] 3671109998Smarkm 3672109998Smarkm *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor 3673109998Smarkm lines, recognice more "algorithms" that can be deselected, and make 3674109998Smarkm it complain about algorithm deselection that isn't recognised. 3675109998Smarkm [Richard Levitte] 3676109998Smarkm 3677109998Smarkm *) New ASN1 functions to handle dup, sign, verify, digest, pack and 3678109998Smarkm unpack operations in terms of ASN1_ITEM. Modify existing wrappers 3679109998Smarkm to use new functions. Add NO_ASN1_OLD which can be set to remove 3680109998Smarkm some old style ASN1 functions: this can be used to determine if old 3681109998Smarkm code will still work when these eventually go away. 3682109998Smarkm [Steve Henson] 3683109998Smarkm 3684109998Smarkm *) New extension functions for OCSP structures, these follow the 3685109998Smarkm same conventions as certificates and CRLs. 3686109998Smarkm [Steve Henson] 3687109998Smarkm 3688109998Smarkm *) New function X509V3_add1_i2d(). This automatically encodes and 3689109998Smarkm adds an extension. Its behaviour can be customised with various 3690109998Smarkm flags to append, replace or delete. Various wrappers added for 3691109998Smarkm certifcates and CRLs. 3692109998Smarkm [Steve Henson] 3693109998Smarkm 3694109998Smarkm *) Fix to avoid calling the underlying ASN1 print routine when 3695109998Smarkm an extension cannot be parsed. Correct a typo in the 3696109998Smarkm OCSP_SERVICELOC extension. Tidy up print OCSP format. 3697109998Smarkm [Steve Henson] 3698109998Smarkm 3699109998Smarkm *) Make mkdef.pl parse some of the ASN1 macros and add apropriate 3700109998Smarkm entries for variables. 3701109998Smarkm [Steve Henson] 3702109998Smarkm 3703109998Smarkm *) Add functionality to apps/openssl.c for detecting locking 3704109998Smarkm problems: As the program is single-threaded, all we have 3705109998Smarkm to do is register a locking callback using an array for 3706109998Smarkm storing which locks are currently held by the program. 3707109998Smarkm [Bodo Moeller] 3708109998Smarkm 3709109998Smarkm *) Use a lock around the call to CRYPTO_get_ex_new_index() in 3710109998Smarkm SSL_get_ex_data_X509_STORE_idx(), which is used in 3711109998Smarkm ssl_verify_cert_chain() and thus can be called at any time 3712109998Smarkm during TLS/SSL handshakes so that thread-safety is essential. 3713109998Smarkm Unfortunately, the ex_data design is not at all suited 3714109998Smarkm for multi-threaded use, so it probably should be abolished. 3715109998Smarkm [Bodo Moeller] 3716109998Smarkm 3717109998Smarkm *) Added Broadcom "ubsec" ENGINE to OpenSSL. 3718109998Smarkm [Broadcom, tweaked and integrated by Geoff Thorpe] 3719109998Smarkm 3720109998Smarkm *) Move common extension printing code to new function 3721109998Smarkm X509V3_print_extensions(). Reorganise OCSP print routines and 3722109998Smarkm implement some needed OCSP ASN1 functions. Add OCSP extensions. 3723109998Smarkm [Steve Henson] 3724109998Smarkm 3725109998Smarkm *) New function X509_signature_print() to remove duplication in some 3726109998Smarkm print routines. 3727109998Smarkm [Steve Henson] 3728109998Smarkm 3729109998Smarkm *) Add a special meaning when SET OF and SEQUENCE OF flags are both 3730109998Smarkm set (this was treated exactly the same as SET OF previously). This 3731109998Smarkm is used to reorder the STACK representing the structure to match the 3732109998Smarkm encoding. This will be used to get round a problem where a PKCS7 3733109998Smarkm structure which was signed could not be verified because the STACK 3734109998Smarkm order did not reflect the encoded order. 3735109998Smarkm [Steve Henson] 3736109998Smarkm 3737109998Smarkm *) Reimplement the OCSP ASN1 module using the new code. 3738109998Smarkm [Steve Henson] 3739109998Smarkm 3740109998Smarkm *) Update the X509V3 code to permit the use of an ASN1_ITEM structure 3741109998Smarkm for its ASN1 operations. The old style function pointers still exist 3742109998Smarkm for now but they will eventually go away. 3743109998Smarkm [Steve Henson] 3744109998Smarkm 3745109998Smarkm *) Merge in replacement ASN1 code from the ASN1 branch. This almost 3746109998Smarkm completely replaces the old ASN1 functionality with a table driven 3747109998Smarkm encoder and decoder which interprets an ASN1_ITEM structure describing 3748109998Smarkm the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is 3749109998Smarkm largely maintained. Almost all of the old asn1_mac.h macro based ASN1 3750109998Smarkm has also been converted to the new form. 3751109998Smarkm [Steve Henson] 3752109998Smarkm 3753109998Smarkm *) Change BN_mod_exp_recp so that negative moduli are tolerated 3754109998Smarkm (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set 3755109998Smarkm so that BN_mod_exp_mont and BN_mod_exp_mont_word work 3756109998Smarkm for negative moduli. 3757109998Smarkm [Bodo Moeller] 3758109998Smarkm 3759109998Smarkm *) Fix BN_uadd and BN_usub: Always return non-negative results instead 3760109998Smarkm of not touching the result's sign bit. 3761109998Smarkm [Bodo Moeller] 3762109998Smarkm 3763109998Smarkm *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be 3764109998Smarkm set. 3765109998Smarkm [Bodo Moeller] 3766109998Smarkm 3767109998Smarkm *) Changed the LHASH code to use prototypes for callbacks, and created 3768109998Smarkm macros to declare and implement thin (optionally static) functions 3769109998Smarkm that provide type-safety and avoid function pointer casting for the 3770109998Smarkm type-specific callbacks. 3771109998Smarkm [Geoff Thorpe] 3772109998Smarkm 3773109998Smarkm *) Added Kerberos Cipher Suites to be used with TLS, as written in 3774109998Smarkm RFC 2712. 3775109998Smarkm [Veers Staats <staatsvr@asc.hpc.mil>, 3776109998Smarkm Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte] 3777109998Smarkm 3778109998Smarkm *) Reformat the FAQ so the different questions and answers can be divided 3779109998Smarkm in sections depending on the subject. 3780109998Smarkm [Richard Levitte] 3781109998Smarkm 3782109998Smarkm *) Have the zlib compression code load ZLIB.DLL dynamically under 3783109998Smarkm Windows. 3784109998Smarkm [Richard Levitte] 3785109998Smarkm 3786109998Smarkm *) New function BN_mod_sqrt for computing square roots modulo a prime 3787109998Smarkm (using the probabilistic Tonelli-Shanks algorithm unless 3788109998Smarkm p == 3 (mod 4) or p == 5 (mod 8), which are cases that can 3789109998Smarkm be handled deterministically). 3790109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 3791109998Smarkm 3792109998Smarkm *) Make BN_mod_inverse faster by explicitly handling small quotients 3793109998Smarkm in the Euclid loop. (Speed gain about 20% for small moduli [256 or 3794109998Smarkm 512 bits], about 30% for larger ones [1024 or 2048 bits].) 3795109998Smarkm [Bodo Moeller] 3796109998Smarkm 3797109998Smarkm *) New function BN_kronecker. 3798109998Smarkm [Bodo Moeller] 3799109998Smarkm 3800109998Smarkm *) Fix BN_gcd so that it works on negative inputs; the result is 3801109998Smarkm positive unless both parameters are zero. 3802109998Smarkm Previously something reasonably close to an infinite loop was 3803109998Smarkm possible because numbers could be growing instead of shrinking 3804109998Smarkm in the implementation of Euclid's algorithm. 3805109998Smarkm [Bodo Moeller] 3806109998Smarkm 3807109998Smarkm *) Fix BN_is_word() and BN_is_one() macros to take into account the 3808109998Smarkm sign of the number in question. 3809109998Smarkm 3810109998Smarkm Fix BN_is_word(a,w) to work correctly for w == 0. 3811109998Smarkm 3812109998Smarkm The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) 3813109998Smarkm because its test if the absolute value of 'a' equals 'w'. 3814109998Smarkm Note that BN_abs_is_word does *not* handle w == 0 reliably; 3815109998Smarkm it exists mostly for use in the implementations of BN_is_zero(), 3816109998Smarkm BN_is_one(), and BN_is_word(). 3817109998Smarkm [Bodo Moeller] 3818109998Smarkm 3819109998Smarkm *) New function BN_swap. 3820109998Smarkm [Bodo Moeller] 3821109998Smarkm 3822109998Smarkm *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that 3823109998Smarkm the exponentiation functions are more likely to produce reasonable 3824109998Smarkm results on negative inputs. 3825109998Smarkm [Bodo Moeller] 3826109998Smarkm 3827109998Smarkm *) Change BN_mod_mul so that the result is always non-negative. 3828109998Smarkm Previously, it could be negative if one of the factors was negative; 3829109998Smarkm I don't think anyone really wanted that behaviour. 3830109998Smarkm [Bodo Moeller] 3831109998Smarkm 3832109998Smarkm *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c 3833109998Smarkm (except for exponentiation, which stays in crypto/bn/bn_exp.c, 3834109998Smarkm and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c) 3835109998Smarkm and add new functions: 3836109998Smarkm 3837109998Smarkm BN_nnmod 3838109998Smarkm BN_mod_sqr 3839109998Smarkm BN_mod_add 3840109998Smarkm BN_mod_add_quick 3841109998Smarkm BN_mod_sub 3842109998Smarkm BN_mod_sub_quick 3843109998Smarkm BN_mod_lshift1 3844109998Smarkm BN_mod_lshift1_quick 3845109998Smarkm BN_mod_lshift 3846109998Smarkm BN_mod_lshift_quick 3847109998Smarkm 3848109998Smarkm These functions always generate non-negative results. 3849109998Smarkm 3850109998Smarkm BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r 3851109998Smarkm such that |m| < r < 0, BN_nnmod will output rem + |m| instead). 3852109998Smarkm 3853109998Smarkm BN_mod_XXX_quick(r, a, [b,] m) generates the same result as 3854109998Smarkm BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b] 3855109998Smarkm be reduced modulo m. 3856109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 3857109998Smarkm 3858111147Snectar#if 0 3859111147Snectar The following entry accidentily appeared in the CHANGES file 3860111147Snectar distributed with OpenSSL 0.9.7. The modifications described in 3861111147Snectar it do *not* apply to OpenSSL 0.9.7. 3862111147Snectar 3863109998Smarkm *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 3864109998Smarkm was actually never needed) and in BN_mul(). The removal in BN_mul() 3865109998Smarkm required a small change in bn_mul_part_recursive() and the addition 3866109998Smarkm of the functions bn_cmp_part_words(), bn_sub_part_words() and 3867109998Smarkm bn_add_part_words(), which do the same thing as bn_cmp_words(), 3868109998Smarkm bn_sub_words() and bn_add_words() except they take arrays with 3869109998Smarkm differing sizes. 3870109998Smarkm [Richard Levitte] 3871111147Snectar#endif 3872109998Smarkm 3873109998Smarkm *) In 'openssl passwd', verify passwords read from the terminal 3874109998Smarkm unless the '-salt' option is used (which usually means that 3875109998Smarkm verification would just waste user's time since the resulting 3876109998Smarkm hash is going to be compared with some given password hash) 3877109998Smarkm or the new '-noverify' option is used. 3878109998Smarkm 3879109998Smarkm This is an incompatible change, but it does not affect 3880109998Smarkm non-interactive use of 'openssl passwd' (passwords on the command 3881109998Smarkm line, '-stdin' option, '-in ...' option) and thus should not 3882109998Smarkm cause any problems. 3883109998Smarkm [Bodo Moeller] 3884109998Smarkm 3885109998Smarkm *) Remove all references to RSAref, since there's no more need for it. 3886109998Smarkm [Richard Levitte] 3887109998Smarkm 3888109998Smarkm *) Make DSO load along a path given through an environment variable 3889109998Smarkm (SHLIB_PATH) with shl_load(). 3890109998Smarkm [Richard Levitte] 3891109998Smarkm 3892109998Smarkm *) Constify the ENGINE code as a result of BIGNUM constification. 3893109998Smarkm Also constify the RSA code and most things related to it. In a 3894109998Smarkm few places, most notable in the depth of the ASN.1 code, ugly 3895109998Smarkm casts back to non-const were required (to be solved at a later 3896109998Smarkm time) 3897109998Smarkm [Richard Levitte] 3898109998Smarkm 3899109998Smarkm *) Make it so the openssl application has all engines loaded by default. 3900109998Smarkm [Richard Levitte] 3901109998Smarkm 3902109998Smarkm *) Constify the BIGNUM routines a little more. 3903109998Smarkm [Richard Levitte] 3904109998Smarkm 3905109998Smarkm *) Add the following functions: 3906109998Smarkm 3907109998Smarkm ENGINE_load_cswift() 3908109998Smarkm ENGINE_load_chil() 3909109998Smarkm ENGINE_load_atalla() 3910109998Smarkm ENGINE_load_nuron() 3911109998Smarkm ENGINE_load_builtin_engines() 3912109998Smarkm 3913109998Smarkm That way, an application can itself choose if external engines that 3914109998Smarkm are built-in in OpenSSL shall ever be used or not. The benefit is 3915109998Smarkm that applications won't have to be linked with libdl or other dso 3916109998Smarkm libraries unless it's really needed. 3917109998Smarkm 3918109998Smarkm Changed 'openssl engine' to load all engines on demand. 3919109998Smarkm Changed the engine header files to avoid the duplication of some 3920109998Smarkm declarations (they differed!). 3921109998Smarkm [Richard Levitte] 3922109998Smarkm 3923109998Smarkm *) 'openssl engine' can now list capabilities. 3924109998Smarkm [Richard Levitte] 3925109998Smarkm 3926109998Smarkm *) Better error reporting in 'openssl engine'. 3927109998Smarkm [Richard Levitte] 3928109998Smarkm 3929109998Smarkm *) Never call load_dh_param(NULL) in s_server. 3930109998Smarkm [Bodo Moeller] 3931109998Smarkm 3932109998Smarkm *) Add engine application. It can currently list engines by name and 3933109998Smarkm identity, and test if they are actually available. 3934109998Smarkm [Richard Levitte] 3935109998Smarkm 3936109998Smarkm *) Improve RPM specification file by forcing symbolic linking and making 3937109998Smarkm sure the installed documentation is also owned by root.root. 3938109998Smarkm [Damien Miller <djm@mindrot.org>] 3939109998Smarkm 3940109998Smarkm *) Give the OpenSSL applications more possibilities to make use of 3941109998Smarkm keys (public as well as private) handled by engines. 3942109998Smarkm [Richard Levitte] 3943109998Smarkm 3944109998Smarkm *) Add OCSP code that comes from CertCo. 3945109998Smarkm [Richard Levitte] 3946109998Smarkm 3947109998Smarkm *) Add VMS support for the Rijndael code. 3948109998Smarkm [Richard Levitte] 3949109998Smarkm 3950109998Smarkm *) Added untested support for Nuron crypto accelerator. 3951109998Smarkm [Ben Laurie] 3952109998Smarkm 3953109998Smarkm *) Add support for external cryptographic devices. This code was 3954109998Smarkm previously distributed separately as the "engine" branch. 3955109998Smarkm [Geoff Thorpe, Richard Levitte] 3956109998Smarkm 3957109998Smarkm *) Rework the filename-translation in the DSO code. It is now possible to 3958109998Smarkm have far greater control over how a "name" is turned into a filename 3959109998Smarkm depending on the operating environment and any oddities about the 3960109998Smarkm different shared library filenames on each system. 3961109998Smarkm [Geoff Thorpe] 3962109998Smarkm 3963109998Smarkm *) Support threads on FreeBSD-elf in Configure. 3964109998Smarkm [Richard Levitte] 3965109998Smarkm 3966109998Smarkm *) Fix for SHA1 assembly problem with MASM: it produces 3967109998Smarkm warnings about corrupt line number information when assembling 3968109998Smarkm with debugging information. This is caused by the overlapping 3969109998Smarkm of two sections. 3970109998Smarkm [Bernd Matthes <mainbug@celocom.de>, Steve Henson] 3971109998Smarkm 3972109998Smarkm *) NCONF changes. 3973109998Smarkm NCONF_get_number() has no error checking at all. As a replacement, 3974109998Smarkm NCONF_get_number_e() is defined (_e for "error checking") and is 3975109998Smarkm promoted strongly. The old NCONF_get_number is kept around for 3976109998Smarkm binary backward compatibility. 3977109998Smarkm Make it possible for methods to load from something other than a BIO, 3978109998Smarkm by providing a function pointer that is given a name instead of a BIO. 3979109998Smarkm For example, this could be used to load configuration data from an 3980109998Smarkm LDAP server. 3981109998Smarkm [Richard Levitte] 3982109998Smarkm 3983109998Smarkm *) Fix for non blocking accept BIOs. Added new I/O special reason 3984109998Smarkm BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs 3985109998Smarkm with non blocking I/O was not possible because no retry code was 3986109998Smarkm implemented. Also added new SSL code SSL_WANT_ACCEPT to cover 3987109998Smarkm this case. 3988109998Smarkm [Steve Henson] 3989109998Smarkm 3990109998Smarkm *) Added the beginnings of Rijndael support. 3991109998Smarkm [Ben Laurie] 3992109998Smarkm 3993109998Smarkm *) Fix for bug in DirectoryString mask setting. Add support for 3994109998Smarkm X509_NAME_print_ex() in 'req' and X509_print_ex() function 3995109998Smarkm to allow certificate printing to more controllable, additional 3996109998Smarkm 'certopt' option to 'x509' to allow new printing options to be 3997109998Smarkm set. 3998109998Smarkm [Steve Henson] 3999109998Smarkm 4000109998Smarkm *) Clean old EAY MD5 hack from e_os.h. 4001109998Smarkm [Richard Levitte] 4002109998Smarkm 4003142425Snectar Changes between 0.9.6l and 0.9.6m [17 Mar 2004] 4004142425Snectar 4005142425Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 4006160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 4007142425Snectar [Joe Orton, Steve Henson] 4008142425Snectar 4009142425Snectar Changes between 0.9.6k and 0.9.6l [04 Nov 2003] 4010142425Snectar 4011142425Snectar *) Fix additional bug revealed by the NISCC test suite: 4012142425Snectar 4013142425Snectar Stop bug triggering large recursion when presented with 4014160814Ssimon certain ASN.1 tags (CVE-2003-0851) 4015142425Snectar [Steve Henson] 4016142425Snectar 4017120631Snectar Changes between 0.9.6j and 0.9.6k [30 Sep 2003] 4018120631Snectar 4019120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 4020120631Snectar 4021120631Snectar Stop out of bounds reads in the ASN1 code when presented with 4022160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 4023120631Snectar 4024120631Snectar If verify callback ignores invalid public key errors don't try to check 4025120631Snectar certificate signature with the NULL public key. 4026120631Snectar 4027120631Snectar [Steve Henson] 4028120631Snectar 4029120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 4030120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 4031120631Snectar specifications. 4032120631Snectar [Steve Henson] 4033120631Snectar 4034120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 4035120631Snectar extra data after the compression methods not only for TLS 1.0 4036120631Snectar but also for SSL 3.0 (as required by the specification). 4037120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 4038120631Snectar 4039120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 4040120631Snectar when it's 512 *bits* long, not 512 bytes. 4041120631Snectar [Richard Levitte] 4042120631Snectar 4043120631Snectar Changes between 0.9.6i and 0.9.6j [10 Apr 2003] 4044120631Snectar 4045120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 4046120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 4047120631Snectar a protocol version number mismatch like a decryption error 4048120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 4049120631Snectar [Bodo Moeller] 4050120631Snectar 4051120631Snectar *) Turn on RSA blinding by default in the default implementation 4052120631Snectar to avoid a timing attack. Applications that don't want it can call 4053120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 4054120631Snectar They would be ill-advised to do so in most cases. 4055120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 4056120631Snectar 4057120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 4058120631Snectar seeded (in this case, the secret RSA exponent is abused as 4059120631Snectar an unpredictable seed -- if it is not unpredictable, there 4060120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 4061120631Snectar by remembering the creator's thread ID in rsa->blinding and 4062120631Snectar having all other threads use local one-time blinding factors 4063120631Snectar (this requires more computation than sharing rsa->blinding, but 4064120631Snectar avoids excessive locking; and if an RSA object is not shared 4065120631Snectar between threads, blinding will still be very fast). 4066120631Snectar [Bodo Moeller] 4067120631Snectar 4068111147Snectar Changes between 0.9.6h and 0.9.6i [19 Feb 2003] 4069111147Snectar 4070111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 4071111147Snectar via timing by performing a MAC computation even if incorrrect 4072111147Snectar block cipher padding has been found. This is a countermeasure 4073111147Snectar against active attacks where the attacker has to distinguish 4074160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 4075111147Snectar 4076111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 4077111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 4078111147Snectar Martin Vuagnoux (EPFL, Ilion)] 4079111147Snectar 4080109998Smarkm Changes between 0.9.6g and 0.9.6h [5 Dec 2002] 4081109998Smarkm 4082109998Smarkm *) New function OPENSSL_cleanse(), which is used to cleanse a section of 4083109998Smarkm memory from it's contents. This is done with a counter that will 4084109998Smarkm place alternating values in each byte. This can be used to solve 4085109998Smarkm two issues: 1) the removal of calls to memset() by highly optimizing 4086109998Smarkm compilers, and 2) cleansing with other values than 0, since those can 4087109998Smarkm be read through on certain media, for example a swap space on disk. 4088109998Smarkm [Geoff Thorpe] 4089109998Smarkm 4090109998Smarkm *) Bugfix: client side session caching did not work with external caching, 4091109998Smarkm because the session->cipher setting was not restored when reloading 4092109998Smarkm from the external cache. This problem was masked, when 4093109998Smarkm SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. 4094109998Smarkm (Found by Steve Haslam <steve@araqnid.ddts.net>.) 4095109998Smarkm [Lutz Jaenicke] 4096109998Smarkm 4097109998Smarkm *) Fix client_certificate (ssl/s2_clnt.c): The permissible total 4098109998Smarkm length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. 4099109998Smarkm [Zeev Lieber <zeev-l@yahoo.com>] 4100109998Smarkm 4101109998Smarkm *) Undo an undocumented change introduced in 0.9.6e which caused 4102109998Smarkm repeated calls to OpenSSL_add_all_ciphers() and 4103109998Smarkm OpenSSL_add_all_digests() to be ignored, even after calling 4104109998Smarkm EVP_cleanup(). 4105109998Smarkm [Richard Levitte] 4106109998Smarkm 4107109998Smarkm *) Change the default configuration reader to deal with last line not 4108109998Smarkm being properly terminated. 4109109998Smarkm [Richard Levitte] 4110109998Smarkm 4111109998Smarkm *) Change X509_NAME_cmp() so it applies the special rules on handling 4112109998Smarkm DN values that are of type PrintableString, as well as RDNs of type 4113109998Smarkm emailAddress where the value has the type ia5String. 4114109998Smarkm [stefank@valicert.com via Richard Levitte] 4115109998Smarkm 4116109998Smarkm *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half 4117109998Smarkm the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently 4118109998Smarkm doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be 4119109998Smarkm the bitwise-OR of the two for use by the majority of applications 4120109998Smarkm wanting this behaviour, and update the docs. The documented 4121109998Smarkm behaviour and actual behaviour were inconsistent and had been 4122109998Smarkm changing anyway, so this is more a bug-fix than a behavioural 4123109998Smarkm change. 4124109998Smarkm [Geoff Thorpe, diagnosed by Nadav Har'El] 4125109998Smarkm 4126109998Smarkm *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c 4127109998Smarkm (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). 4128109998Smarkm [Bodo Moeller] 4129109998Smarkm 4130109998Smarkm *) Fix initialization code race conditions in 4131109998Smarkm SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), 4132109998Smarkm SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), 4133109998Smarkm SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), 4134109998Smarkm TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), 4135109998Smarkm ssl2_get_cipher_by_char(), 4136109998Smarkm ssl3_get_cipher_by_char(). 4137109998Smarkm [Patrick McCormick <patrick@tellme.com>, Bodo Moeller] 4138109998Smarkm 4139109998Smarkm *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after 4140109998Smarkm the cached sessions are flushed, as the remove_cb() might use ex_data 4141109998Smarkm contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> 4142109998Smarkm (see [openssl.org #212]). 4143109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 4144109998Smarkm 4145109998Smarkm *) Fix typo in OBJ_txt2obj which incorrectly passed the content 4146109998Smarkm length, instead of the encoding length to d2i_ASN1_OBJECT. 4147109998Smarkm [Steve Henson] 4148109998Smarkm 4149101618Snectar Changes between 0.9.6f and 0.9.6g [9 Aug 2002] 4150101618Snectar 4151101618Snectar *) [In 0.9.6g-engine release:] 4152101618Snectar Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall'). 4153101618Snectar [Lynn Gazis <lgazis@rainbow.com>] 4154101618Snectar 4155101613Snectar Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 4156101613Snectar 4157101613Snectar *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 4158101613Snectar and get fix the header length calculation. 4159101613Snectar [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 4160101613Snectar Alon Kantor <alonk@checkpoint.com> (and others), 4161101613Snectar Steve Henson] 4162101613Snectar 4163101613Snectar *) Use proper error handling instead of 'assertions' in buffer 4164101613Snectar overflow checks added in 0.9.6e. This prevents DoS (the 4165101613Snectar assertions could call abort()). 4166101613Snectar [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] 4167101613Snectar 4168100936Snectar Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 4169100936Snectar 4170109998Smarkm *) Add various sanity checks to asn1_get_length() to reject 4171109998Smarkm the ASN1 length bytes if they exceed sizeof(long), will appear 4172109998Smarkm negative or the content length exceeds the length of the 4173109998Smarkm supplied buffer. 4174109998Smarkm [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 4175109998Smarkm 4176100936Snectar *) Fix cipher selection routines: ciphers without encryption had no flags 4177100936Snectar for the cipher strength set and where therefore not handled correctly 4178100936Snectar by the selection routines (PR #130). 4179100936Snectar [Lutz Jaenicke] 4180100936Snectar 4181100936Snectar *) Fix EVP_dsa_sha macro. 4182100936Snectar [Nils Larsch] 4183100936Snectar 4184100936Snectar *) New option 4185100936Snectar SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 4186100936Snectar for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 4187100936Snectar that was added in OpenSSL 0.9.6d. 4188100936Snectar 4189100936Snectar As the countermeasure turned out to be incompatible with some 4190100936Snectar broken SSL implementations, the new option is part of SSL_OP_ALL. 4191100936Snectar SSL_OP_ALL is usually employed when compatibility with weird SSL 4192100936Snectar implementations is desired (e.g. '-bugs' option to 's_client' and 4193100936Snectar 's_server'), so the new option is automatically set in many 4194100936Snectar applications. 4195100936Snectar [Bodo Moeller] 4196100936Snectar 4197100936Snectar *) Changes in security patch: 4198100936Snectar 4199100936Snectar Changes marked "(CHATS)" were sponsored by the Defense Advanced 4200100936Snectar Research Projects Agency (DARPA) and Air Force Research Laboratory, 4201100936Snectar Air Force Materiel Command, USAF, under agreement number 4202100936Snectar F30602-01-2-0537. 4203100936Snectar 4204100936Snectar *) Add various sanity checks to asn1_get_length() to reject 4205100936Snectar the ASN1 length bytes if they exceed sizeof(long), will appear 4206100936Snectar negative or the content length exceeds the length of the 4207160814Ssimon supplied buffer. (CVE-2002-0659) 4208100936Snectar [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 4209100936Snectar 4210100936Snectar *) Assertions for various potential buffer overflows, not known to 4211100936Snectar happen in practice. 4212100936Snectar [Ben Laurie (CHATS)] 4213100936Snectar 4214100936Snectar *) Various temporary buffers to hold ASCII versions of integers were 4215160814Ssimon too small for 64 bit platforms. (CVE-2002-0655) 4216100936Snectar [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 4217100936Snectar 4218100936Snectar *) Remote buffer overflow in SSL3 protocol - an attacker could 4219160814Ssimon supply an oversized session ID to a client. (CVE-2002-0656) 4220100936Snectar [Ben Laurie (CHATS)] 4221100936Snectar 4222100936Snectar *) Remote buffer overflow in SSL2 protocol - an attacker could 4223160814Ssimon supply an oversized client master key. (CVE-2002-0656) 4224100936Snectar [Ben Laurie (CHATS)] 4225100936Snectar 4226100928Snectar Changes between 0.9.6c and 0.9.6d [9 May 2002] 4227100928Snectar 4228100928Snectar *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 4229100928Snectar encoded as NULL) with id-dsa-with-sha1. 4230100928Snectar [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 4231100928Snectar 4232100928Snectar *) Check various X509_...() return values in apps/req.c. 4233100928Snectar [Nils Larsch <nla@trustcenter.de>] 4234100928Snectar 4235100928Snectar *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 4236100928Snectar an end-of-file condition would erronously be flagged, when the CRLF 4237100928Snectar was just at the end of a processed block. The bug was discovered when 4238100928Snectar processing data through a buffering memory BIO handing the data to a 4239100928Snectar BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 4240100928Snectar <ptsekov@syntrex.com> and Nedelcho Stanev. 4241100928Snectar [Lutz Jaenicke] 4242100928Snectar 4243100928Snectar *) Implement a countermeasure against a vulnerability recently found 4244100928Snectar in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 4245100928Snectar before application data chunks to avoid the use of known IVs 4246100928Snectar with data potentially chosen by the attacker. 4247100928Snectar [Bodo Moeller] 4248100928Snectar 4249100928Snectar *) Fix length checks in ssl3_get_client_hello(). 4250100928Snectar [Bodo Moeller] 4251100928Snectar 4252100928Snectar *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 4253100928Snectar to prevent ssl3_read_internal() from incorrectly assuming that 4254100928Snectar ssl3_read_bytes() found application data while handshake 4255100928Snectar processing was enabled when in fact s->s3->in_read_app_data was 4256100928Snectar merely automatically cleared during the initial handshake. 4257100928Snectar [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 4258100928Snectar 4259100928Snectar *) Fix object definitions for Private and Enterprise: they were not 4260100928Snectar recognized in their shortname (=lowercase) representation. Extend 4261100928Snectar obj_dat.pl to issue an error when using undefined keywords instead 4262100928Snectar of silently ignoring the problem (Svenning Sorensen 4263100928Snectar <sss@sss.dnsalias.net>). 4264100928Snectar [Lutz Jaenicke] 4265100928Snectar 4266100928Snectar *) Fix DH_generate_parameters() so that it works for 'non-standard' 4267100928Snectar generators, i.e. generators other than 2 and 5. (Previously, the 4268100928Snectar code did not properly initialise the 'add' and 'rem' values to 4269100928Snectar BN_generate_prime().) 4270100928Snectar 4271100928Snectar In the new general case, we do not insist that 'generator' is 4272100928Snectar actually a primitive root: This requirement is rather pointless; 4273100928Snectar a generator of the order-q subgroup is just as good, if not 4274100928Snectar better. 4275100928Snectar [Bodo Moeller] 4276100928Snectar 4277100928Snectar *) Map new X509 verification errors to alerts. Discovered and submitted by 4278100928Snectar Tom Wu <tom@arcot.com>. 4279100928Snectar [Lutz Jaenicke] 4280100928Snectar 4281100928Snectar *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 4282100928Snectar returning non-zero before the data has been completely received 4283100928Snectar when using non-blocking I/O. 4284100928Snectar [Bodo Moeller; problem pointed out by John Hughes] 4285100928Snectar 4286100928Snectar *) Some of the ciphers missed the strength entry (SSL_LOW etc). 4287100928Snectar [Ben Laurie, Lutz Jaenicke] 4288100928Snectar 4289100928Snectar *) Fix bug in SSL_clear(): bad sessions were not removed (found by 4290100928Snectar Yoram Zahavi <YoramZ@gilian.com>). 4291100928Snectar [Lutz Jaenicke] 4292100928Snectar 4293100928Snectar *) Add information about CygWin 1.3 and on, and preserve proper 4294100928Snectar configuration for the versions before that. 4295100928Snectar [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 4296100928Snectar 4297100928Snectar *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 4298100928Snectar check whether we deal with a copy of a session and do not delete from 4299100928Snectar the cache in this case. Problem reported by "Izhar Shoshani Levi" 4300100928Snectar <izhar@checkpoint.com>. 4301100928Snectar [Lutz Jaenicke] 4302100928Snectar 4303100928Snectar *) Do not store session data into the internal session cache, if it 4304100928Snectar is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 4305100928Snectar flag is set). Proposed by Aslam <aslam@funk.com>. 4306100928Snectar [Lutz Jaenicke] 4307100928Snectar 4308100928Snectar *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 4309100928Snectar value is 0. 4310100928Snectar [Richard Levitte] 4311100928Snectar 4312109998Smarkm *) [In 0.9.6d-engine release:] 4313109998Smarkm Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 4314100928Snectar [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 4315100928Snectar 4316100928Snectar *) Add the configuration target linux-s390x. 4317100928Snectar [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 4318100928Snectar 4319100928Snectar *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 4320100928Snectar ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 4321100928Snectar variable as an indication that a ClientHello message has been 4322100928Snectar received. As the flag value will be lost between multiple 4323100928Snectar invocations of ssl3_accept when using non-blocking I/O, the 4324100928Snectar function may not be aware that a handshake has actually taken 4325100928Snectar place, thus preventing a new session from being added to the 4326100928Snectar session cache. 4327100928Snectar 4328100928Snectar To avoid this problem, we now set s->new_session to 2 instead of 4329100928Snectar using a local variable. 4330100928Snectar [Lutz Jaenicke, Bodo Moeller] 4331100928Snectar 4332100928Snectar *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 4333100928Snectar if the SSL_R_LENGTH_MISMATCH error is detected. 4334100928Snectar [Geoff Thorpe, Bodo Moeller] 4335100928Snectar 4336100928Snectar *) New 'shared_ldflag' column in Configure platform table. 4337100928Snectar [Richard Levitte] 4338100928Snectar 4339100928Snectar *) Fix EVP_CIPHER_mode macro. 4340100928Snectar ["Dan S. Camper" <dan@bti.net>] 4341100928Snectar 4342100928Snectar *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 4343100928Snectar type, we must throw them away by setting rr->length to 0. 4344100928Snectar [D P Chang <dpc@qualys.com>] 4345100928Snectar 434689837Skris Changes between 0.9.6b and 0.9.6c [21 dec 2001] 434789837Skris 434889837Skris *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 434989837Skris <Dominikus.Scherkl@biodata.com>. (The previous implementation 435089837Skris worked incorrectly for those cases where range = 10..._2 and 435189837Skris 3*range is two bits longer than range.) 435289837Skris [Bodo Moeller] 435389837Skris 435489837Skris *) Only add signing time to PKCS7 structures if it is not already 435589837Skris present. 435689837Skris [Steve Henson] 435789837Skris 435889837Skris *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 435989837Skris OBJ_ld_ce should be OBJ_id_ce. 436089837Skris Also some ip-pda OIDs in crypto/objects/objects.txt were 436189837Skris incorrect (cf. RFC 3039). 436289837Skris [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 436389837Skris 436489837Skris *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 436589837Skris returns early because it has nothing to do. 436689837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 436789837Skris 436889837Skris *) [In 0.9.6c-engine release:] 436989837Skris Fix mutex callback return values in crypto/engine/hw_ncipher.c. 437089837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 437189837Skris 437289837Skris *) [In 0.9.6c-engine release:] 437389837Skris Add support for Cryptographic Appliance's keyserver technology. 437489837Skris (Use engine 'keyclient') 437589837Skris [Cryptographic Appliances and Geoff Thorpe] 437689837Skris 437789837Skris *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 437889837Skris is called via tools/c89.sh because arguments have to be 437989837Skris rearranged (all '-L' options must appear before the first object 438089837Skris modules). 438189837Skris [Richard Shapiro <rshapiro@abinitio.com>] 438289837Skris 438389837Skris *) [In 0.9.6c-engine release:] 438489837Skris Add support for Broadcom crypto accelerator cards, backported 438589837Skris from 0.9.7. 438689837Skris [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 438789837Skris 438889837Skris *) [In 0.9.6c-engine release:] 438989837Skris Add support for SureWare crypto accelerator cards from 439089837Skris Baltimore Technologies. (Use engine 'sureware') 439189837Skris [Baltimore Technologies and Mark Cox] 439289837Skris 439389837Skris *) [In 0.9.6c-engine release:] 439489837Skris Add support for crypto accelerator cards from Accelerated 439589837Skris Encryption Processing, www.aep.ie. (Use engine 'aep') 439689837Skris [AEP Inc. and Mark Cox] 439789837Skris 439889837Skris *) Add a configuration entry for gcc on UnixWare. 439989837Skris [Gary Benson <gbenson@redhat.com>] 440089837Skris 440189837Skris *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 440289837Skris messages are stored in a single piece (fixed-length part and 440389837Skris variable-length part combined) and fix various bugs found on the way. 440489837Skris [Bodo Moeller] 440589837Skris 440689837Skris *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 440789837Skris instead. BIO_gethostbyname() does not know what timeouts are 440889837Skris appropriate, so entries would stay in cache even when they have 440989837Skris become invalid. 441089837Skris [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 441189837Skris 441289837Skris *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 441389837Skris faced with a pathologically small ClientHello fragment that does 441489837Skris not contain client_version: Instead of aborting with an error, 441589837Skris simply choose the highest available protocol version (i.e., 441689837Skris TLS 1.0 unless it is disabled). In practice, ClientHello 441789837Skris messages are never sent like this, but this change gives us 441889837Skris strictly correct behaviour at least for TLS. 441989837Skris [Bodo Moeller] 442089837Skris 442189837Skris *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 442289837Skris never resets s->method to s->ctx->method when called from within 442389837Skris one of the SSL handshake functions. 442489837Skris [Bodo Moeller; problem pointed out by Niko Baric] 442589837Skris 442689837Skris *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 442789837Skris (sent using the client's version number) if client_version is 442889837Skris smaller than the protocol version in use. Also change 442989837Skris ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 443089837Skris the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 443189837Skris the client will at least see that alert. 443289837Skris [Bodo Moeller] 443389837Skris 443489837Skris *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 443589837Skris correctly. 443689837Skris [Bodo Moeller] 443789837Skris 443889837Skris *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 443989837Skris client receives HelloRequest while in a handshake. 444089837Skris [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 444189837Skris 444289837Skris *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 444389837Skris should end in 'break', not 'goto end' which circuments various 444489837Skris cleanups done in state SSL_ST_OK. But session related stuff 444589837Skris must be disabled for SSL_ST_OK in the case that we just sent a 444689837Skris HelloRequest. 444789837Skris 444889837Skris Also avoid some overhead by not calling ssl_init_wbio_buffer() 444989837Skris before just sending a HelloRequest. 445089837Skris [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 445189837Skris 445289837Skris *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 445389837Skris reveal whether illegal block cipher padding was found or a MAC 445489837Skris verification error occured. (Neither SSLerr() codes nor alerts 445589837Skris are directly visible to potential attackers, but the information 445689837Skris may leak via logfiles.) 445789837Skris 445889837Skris Similar changes are not required for the SSL 2.0 implementation 445989837Skris because the number of padding bytes is sent in clear for SSL 2.0, 446089837Skris and the extra bytes are just ignored. However ssl/s2_pkt.c 446189837Skris failed to verify that the purported number of padding bytes is in 446289837Skris the legal range. 446389837Skris [Bodo Moeller] 446489837Skris 446589837Skris *) Add OpenUNIX-8 support including shared libraries 446689837Skris (Boyd Lynn Gerber <gerberb@zenez.com>). 446789837Skris [Lutz Jaenicke] 446889837Skris 446989837Skris *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 447089837Skris 'wristwatch attack' using huge encoding parameters (cf. 447189837Skris James H. Manger's CRYPTO 2001 paper). Note that the 447289837Skris RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 447389837Skris encoding parameters and hence was not vulnerable. 447489837Skris [Bodo Moeller] 447589837Skris 447689837Skris *) BN_sqr() bug fix. 447789837Skris [Ulf M�ller, reported by Jim Ellis <jim.ellis@cavium.com>] 447889837Skris 447989837Skris *) Rabin-Miller test analyses assume uniformly distributed witnesses, 448089837Skris so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 448189837Skris followed by modular reduction. 448289837Skris [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 448389837Skris 448489837Skris *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 448589837Skris equivalent based on BN_pseudo_rand() instead of BN_rand(). 448689837Skris [Bodo Moeller] 448789837Skris 448889837Skris *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 448989837Skris This function was broken, as the check for a new client hello message 449089837Skris to handle SGC did not allow these large messages. 449189837Skris (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 449289837Skris [Lutz Jaenicke] 449389837Skris 449489837Skris *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 449589837Skris [Lutz Jaenicke] 449689837Skris 449789837Skris *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 449889837Skris for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 449989837Skris [Lutz Jaenicke] 450089837Skris 450189837Skris *) Rework the configuration and shared library support for Tru64 Unix. 450289837Skris The configuration part makes use of modern compiler features and 450389837Skris still retains old compiler behavior for those that run older versions 450489837Skris of the OS. The shared library support part includes a variant that 450589837Skris uses the RPATH feature, and is available through the special 450689837Skris configuration target "alpha-cc-rpath", which will never be selected 450789837Skris automatically. 450889837Skris [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 450989837Skris 451089837Skris *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 451189837Skris with the same message size as in ssl3_get_certificate_request(). 451289837Skris Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 451389837Skris messages might inadvertently be reject as too long. 451489837Skris [Petr Lampa <lampa@fee.vutbr.cz>] 451589837Skris 451689837Skris *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 451789837Skris [Andy Polyakov] 451889837Skris 451989837Skris *) Modified SSL library such that the verify_callback that has been set 452089837Skris specificly for an SSL object with SSL_set_verify() is actually being 452189837Skris used. Before the change, a verify_callback set with this function was 452289837Skris ignored and the verify_callback() set in the SSL_CTX at the time of 452389837Skris the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 452489837Skris to allow the necessary settings. 452589837Skris [Lutz Jaenicke] 452689837Skris 452789837Skris *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 452889837Skris explicitly to NULL, as at least on Solaris 8 this seems not always to be 452989837Skris done automatically (in contradiction to the requirements of the C 453089837Skris standard). This made problems when used from OpenSSH. 453189837Skris [Lutz Jaenicke] 453289837Skris 453389837Skris *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 453489837Skris dh->length and always used 453589837Skris 453689837Skris BN_rand_range(priv_key, dh->p). 453789837Skris 453889837Skris BN_rand_range() is not necessary for Diffie-Hellman, and this 453989837Skris specific range makes Diffie-Hellman unnecessarily inefficient if 454089837Skris dh->length (recommended exponent length) is much smaller than the 454189837Skris length of dh->p. We could use BN_rand_range() if the order of 454289837Skris the subgroup was stored in the DH structure, but we only have 454389837Skris dh->length. 454489837Skris 454589837Skris So switch back to 454689837Skris 454789837Skris BN_rand(priv_key, l, ...) 454889837Skris 454989837Skris where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 455089837Skris otherwise. 455189837Skris [Bodo Moeller] 455289837Skris 455389837Skris *) In 455489837Skris 455589837Skris RSA_eay_public_encrypt 455689837Skris RSA_eay_private_decrypt 455789837Skris RSA_eay_private_encrypt (signing) 455889837Skris RSA_eay_public_decrypt (signature verification) 455989837Skris 456089837Skris (default implementations for RSA_public_encrypt, 456189837Skris RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 456289837Skris always reject numbers >= n. 456389837Skris [Bodo Moeller] 456489837Skris 456589837Skris *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 456689837Skris to synchronize access to 'locking_thread'. This is necessary on 456789837Skris systems where access to 'locking_thread' (an 'unsigned long' 456889837Skris variable) is not atomic. 456989837Skris [Bodo Moeller] 457089837Skris 457189837Skris *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 457289837Skris *before* setting the 'crypto_lock_rand' flag. The previous code had 457389837Skris a race condition if 0 is a valid thread ID. 457489837Skris [Travis Vitek <vitek@roguewave.com>] 457589837Skris 457689837Skris *) Add support for shared libraries under Irix. 457789837Skris [Albert Chin-A-Young <china@thewrittenword.com>] 457889837Skris 457989837Skris *) Add configuration option to build on Linux on both big-endian and 458089837Skris little-endian MIPS. 458189837Skris [Ralf Baechle <ralf@uni-koblenz.de>] 458289837Skris 458389837Skris *) Add the possibility to create shared libraries on HP-UX. 458489837Skris [Richard Levitte] 458589837Skris 458679998Skris Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 458779998Skris 458879998Skris *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 458979998Skris to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 459079998Skris Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 459179998Skris PRNG state recovery was possible based on the output of 459279998Skris one PRNG request appropriately sized to gain knowledge on 459379998Skris 'md' followed by enough consecutive 1-byte PRNG requests 459479998Skris to traverse all of 'state'. 459579998Skris 459679998Skris 1. When updating 'md_local' (the current thread's copy of 'md') 459779998Skris during PRNG output generation, hash all of the previous 459879998Skris 'md_local' value, not just the half used for PRNG output. 459979998Skris 460079998Skris 2. Make the number of bytes from 'state' included into the hash 460179998Skris independent from the number of PRNG bytes requested. 460279998Skris 460379998Skris The first measure alone would be sufficient to avoid 460479998Skris Markku-Juhani's attack. (Actually it had never occurred 460579998Skris to me that the half of 'md_local' used for chaining was the 460679998Skris half from which PRNG output bytes were taken -- I had always 460779998Skris assumed that the secret half would be used.) The second 460879998Skris measure makes sure that additional data from 'state' is never 460979998Skris mixed into 'md_local' in small portions; this heuristically 461079998Skris further strengthens the PRNG. 461179998Skris [Bodo Moeller] 461279998Skris 461379998Skris *) Fix crypto/bn/asm/mips3.s. 461479998Skris [Andy Polyakov] 461579998Skris 461679998Skris *) When only the key is given to "enc", the IV is undefined. Print out 461779998Skris an error message in this case. 461879998Skris [Lutz Jaenicke] 461979998Skris 462079998Skris *) Handle special case when X509_NAME is empty in X509 printing routines. 462179998Skris [Steve Henson] 462279998Skris 462379998Skris *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 462479998Skris positive and less than q. 462579998Skris [Bodo Moeller] 462679998Skris 462779998Skris *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 462879998Skris used: it isn't thread safe and the add_lock_callback should handle 462979998Skris that itself. 463079998Skris [Paul Rose <Paul.Rose@bridge.com>] 463179998Skris 463279998Skris *) Verify that incoming data obeys the block size in 463379998Skris ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 463479998Skris [Bodo Moeller] 463579998Skris 463679998Skris *) Fix OAEP check. 463779998Skris [Ulf M�ller, Bodo M�ller] 463879998Skris 463979998Skris *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 464089837Skris RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 464179998Skris when fixing the server behaviour for backwards-compatible 'client 464279998Skris hello' messages. (Note that the attack is impractical against 464379998Skris SSL 3.0 and TLS 1.0 anyway because length and version checking 464479998Skris means that the probability of guessing a valid ciphertext is 464579998Skris around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 464679998Skris paper.) 464779998Skris 464879998Skris Before 0.9.5, the countermeasure (hide the error by generating a 464979998Skris random 'decryption result') did not work properly because 465079998Skris ERR_clear_error() was missing, meaning that SSL_get_error() would 465179998Skris detect the supposedly ignored error. 465279998Skris 465379998Skris Both problems are now fixed. 465479998Skris [Bodo Moeller] 465579998Skris 465679998Skris *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 465779998Skris (previously it was 1024). 465879998Skris [Bodo Moeller] 465979998Skris 466079998Skris *) Fix for compatibility mode trust settings: ignore trust settings 466179998Skris unless some valid trust or reject settings are present. 466279998Skris [Steve Henson] 466379998Skris 466479998Skris *) Fix for blowfish EVP: its a variable length cipher. 466579998Skris [Steve Henson] 466679998Skris 466779998Skris *) Fix various bugs related to DSA S/MIME verification. Handle missing 466879998Skris parameters in DSA public key structures and return an error in the 466979998Skris DSA routines if parameters are absent. 467079998Skris [Steve Henson] 467179998Skris 467279998Skris *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 467379998Skris in the current directory if neither $RANDFILE nor $HOME was set. 467479998Skris RAND_file_name() in 0.9.6a returned NULL in this case. This has 467579998Skris caused some confusion to Windows users who haven't defined $HOME. 467679998Skris Thus RAND_file_name() is changed again: e_os.h can define a 467779998Skris DEFAULT_HOME, which will be used if $HOME is not set. 467879998Skris For Windows, we use "C:"; on other platforms, we still require 467979998Skris environment variables. 468079998Skris 468179998Skris *) Move 'if (!initialized) RAND_poll()' into regions protected by 468279998Skris CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 468379998Skris having multiple threads call RAND_poll() concurrently. 468479998Skris [Bodo Moeller] 468579998Skris 468679998Skris *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 468779998Skris combination of a flag and a thread ID variable. 468879998Skris Otherwise while one thread is in ssleay_rand_bytes (which sets the 468979998Skris flag), *other* threads can enter ssleay_add_bytes without obeying 469089837Skris the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 469179998Skris that they do not hold after the first thread unsets add_do_not_lock). 469279998Skris [Bodo Moeller] 469379998Skris 469479998Skris *) Change bctest again: '-x' expressions are not available in all 469579998Skris versions of 'test'. 469679998Skris [Bodo Moeller] 469779998Skris 469876866Skris Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 469972613Skris 470076866Skris *) Fix a couple of memory leaks in PKCS7_dataDecode() 470176866Skris [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 470276866Skris 470376866Skris *) Change Configure and Makefiles to provide EXE_EXT, which will contain 470476866Skris the default extension for executables, if any. Also, make the perl 470576866Skris scripts that use symlink() to test if it really exists and use "cp" 470676866Skris if it doesn't. All this made OpenSSL compilable and installable in 470776866Skris CygWin. 470876866Skris [Richard Levitte] 470976866Skris 471076866Skris *) Fix for asn1_GetSequence() for indefinite length constructed data. 471176866Skris If SEQUENCE is length is indefinite just set c->slen to the total 471276866Skris amount of data available. 471376866Skris [Steve Henson, reported by shige@FreeBSD.org] 471476866Skris [This change does not apply to 0.9.7.] 471576866Skris 471676866Skris *) Change bctest to avoid here-documents inside command substitution 471776866Skris (workaround for FreeBSD /bin/sh bug). 471876866Skris For compatibility with Ultrix, avoid shell functions (introduced 471976866Skris in the bctest version that searches along $PATH). 472076866Skris [Bodo Moeller] 472176866Skris 472276866Skris *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 472376866Skris with des_encrypt() defined on some operating systems, like Solaris 472476866Skris and UnixWare. 472576866Skris [Richard Levitte] 472676866Skris 472776866Skris *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 472876866Skris On the Importance of Eliminating Errors in Cryptographic 472976866Skris Computations, J. Cryptology 14 (2001) 2, 101-119, 473076866Skris http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 473176866Skris [Ulf Moeller] 473276866Skris 473376866Skris *) MIPS assembler BIGNUM division bug fix. 473476866Skris [Andy Polyakov] 473576866Skris 473676866Skris *) Disabled incorrect Alpha assembler code. 473776866Skris [Richard Levitte] 473876866Skris 473976866Skris *) Fix PKCS#7 decode routines so they correctly update the length 474076866Skris after reading an EOC for the EXPLICIT tag. 474176866Skris [Steve Henson] 474276866Skris [This change does not apply to 0.9.7.] 474376866Skris 474476866Skris *) Fix bug in PKCS#12 key generation routines. This was triggered 474576866Skris if a 3DES key was generated with a 0 initial byte. Include 474676866Skris PKCS12_BROKEN_KEYGEN compilation option to retain the old 474776866Skris (but broken) behaviour. 474876866Skris [Steve Henson] 474976866Skris 475076866Skris *) Enhance bctest to search for a working bc along $PATH and print 475176866Skris it when found. 475276866Skris [Tim Rice <tim@multitalents.net> via Richard Levitte] 475376866Skris 475476866Skris *) Fix memory leaks in err.c: free err_data string if necessary; 475576866Skris don't write to the wrong index in ERR_set_error_data. 475676866Skris [Bodo Moeller] 475776866Skris 475876866Skris *) Implement ssl23_peek (analogous to ssl23_read), which previously 475976866Skris did not exist. 476076866Skris [Bodo Moeller] 476176866Skris 476276866Skris *) Replace rdtsc with _emit statements for VC++ version 5. 476376866Skris [Jeremy Cooper <jeremy@baymoo.org>] 476476866Skris 476576866Skris *) Make it possible to reuse SSLv2 sessions. 476676866Skris [Richard Levitte] 476776866Skris 476876866Skris *) In copy_email() check for >= 0 as a return value for 476976866Skris X509_NAME_get_index_by_NID() since 0 is a valid index. 477076866Skris [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 477176866Skris 477276866Skris *) Avoid coredump with unsupported or invalid public keys by checking if 477376866Skris X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 477476866Skris PKCS7_verify() fails with non detached data. 477576866Skris [Steve Henson] 477676866Skris 477776866Skris *) Don't use getenv in library functions when run as setuid/setgid. 477876866Skris New function OPENSSL_issetugid(). 477976866Skris [Ulf Moeller] 478076866Skris 478176866Skris *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 478276866Skris due to incorrect handling of multi-threading: 478376866Skris 478476866Skris 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 478576866Skris 478676866Skris 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 478776866Skris 478876866Skris 3. Count how many times MemCheck_off() has been called so that 478976866Skris nested use can be treated correctly. This also avoids 479076866Skris inband-signalling in the previous code (which relied on the 479176866Skris assumption that thread ID 0 is impossible). 479276866Skris [Bodo Moeller] 479376866Skris 479476866Skris *) Add "-rand" option also to s_client and s_server. 479576866Skris [Lutz Jaenicke] 479676866Skris 479776866Skris *) Fix CPU detection on Irix 6.x. 479876866Skris [Kurt Hockenbury <khockenb@stevens-tech.edu> and 479976866Skris "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 480076866Skris 480176866Skris *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 480276866Skris was empty. 480376866Skris [Steve Henson] 480476866Skris [This change does not apply to 0.9.7.] 480576866Skris 480676866Skris *) Use the cached encoding of an X509_NAME structure rather than 480776866Skris copying it. This is apparently the reason for the libsafe "errors" 480876866Skris but the code is actually correct. 480976866Skris [Steve Henson] 481076866Skris 481172613Skris *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 481272613Skris Bleichenbacher's DSA attack. 481376866Skris Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 481476866Skris to be set and top=0 forces the highest bit to be set; top=-1 is new 481576866Skris and leaves the highest bit random. 481676866Skris [Ulf Moeller, Bodo Moeller] 481772613Skris 481872613Skris *) In the NCONF_...-based implementations for CONF_... queries 481972613Skris (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 482072613Skris a temporary CONF structure with the data component set to NULL 482172613Skris (which gives segmentation faults in lh_retrieve). 482272613Skris Instead, use NULL for the CONF pointer in CONF_get_string and 482372613Skris CONF_get_number (which may use environment variables) and directly 482472613Skris return NULL from CONF_get_section. 482572613Skris [Bodo Moeller] 482672613Skris 482772613Skris *) Fix potential buffer overrun for EBCDIC. 482872613Skris [Ulf Moeller] 482972613Skris 483072613Skris *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 483172613Skris keyUsage if basicConstraints absent for a CA. 483272613Skris [Steve Henson] 483372613Skris 483472613Skris *) Make SMIME_write_PKCS7() write mail header values with a format that 483572613Skris is more generally accepted (no spaces before the semicolon), since 483672613Skris some programs can't parse those values properly otherwise. Also make 483772613Skris sure BIO's that break lines after each write do not create invalid 483872613Skris headers. 483972613Skris [Richard Levitte] 484072613Skris 484172613Skris *) Make the CRL encoding routines work with empty SEQUENCE OF. The 484272613Skris macros previously used would not encode an empty SEQUENCE OF 484372613Skris and break the signature. 484472613Skris [Steve Henson] 484576866Skris [This change does not apply to 0.9.7.] 484672613Skris 484772613Skris *) Zero the premaster secret after deriving the master secret in 484872613Skris DH ciphersuites. 484972613Skris [Steve Henson] 485072613Skris 485172613Skris *) Add some EVP_add_digest_alias registrations (as found in 485272613Skris OpenSSL_add_all_digests()) to SSL_library_init() 485372613Skris aka OpenSSL_add_ssl_algorithms(). This provides improved 485472613Skris compatibility with peers using X.509 certificates 485572613Skris with unconventional AlgorithmIdentifier OIDs. 485672613Skris [Bodo Moeller] 485772613Skris 485872613Skris *) Fix for Irix with NO_ASM. 485972613Skris ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 486072613Skris 486172613Skris *) ./config script fixes. 486272613Skris [Ulf Moeller, Richard Levitte] 486372613Skris 486472613Skris *) Fix 'openssl passwd -1'. 486572613Skris [Bodo Moeller] 486672613Skris 486772613Skris *) Change PKCS12_key_gen_asc() so it can cope with non null 486872613Skris terminated strings whose length is passed in the passlen 486972613Skris parameter, for example from PEM callbacks. This was done 487072613Skris by adding an extra length parameter to asc2uni(). 487172613Skris [Steve Henson, reported by <oddissey@samsung.co.kr>] 487272613Skris 487372613Skris *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 487472613Skris call failed, free the DSA structure. 487572613Skris [Bodo Moeller] 487672613Skris 487772613Skris *) Fix to uni2asc() to cope with zero length Unicode strings. 487872613Skris These are present in some PKCS#12 files. 487972613Skris [Steve Henson] 488072613Skris 488172613Skris *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 488272613Skris Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 488372613Skris when writing a 32767 byte record. 488472613Skris [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 488572613Skris 488672613Skris *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 488772613Skris obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 488872613Skris 488972613Skris (RSA objects have a reference count access to which is protected 489072613Skris by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 489172613Skris so they are meant to be shared between threads.) 489272613Skris [Bodo Moeller, Geoff Thorpe; original patch submitted by 489372613Skris "Reddie, Steven" <Steven.Reddie@ca.com>] 489472613Skris 489572613Skris *) Fix a deadlock in CRYPTO_mem_leaks(). 489672613Skris [Bodo Moeller] 489772613Skris 489876866Skris *) Use better test patterns in bntest. 489976866Skris [Ulf M�ller] 490076866Skris 490172613Skris *) rand_win.c fix for Borland C. 490272613Skris [Ulf M�ller] 490372613Skris 490472613Skris *) BN_rshift bugfix for n == 0. 490572613Skris [Bodo Moeller] 490672613Skris 490776866Skris *) Add a 'bctest' script that checks for some known 'bc' bugs 490876866Skris so that 'make test' does not abort just because 'bc' is broken. 490976866Skris [Bodo Moeller] 491076866Skris 491172613Skris *) Store verify_result within SSL_SESSION also for client side to 491272613Skris avoid potential security hole. (Re-used sessions on the client side 491372613Skris always resulted in verify_result==X509_V_OK, not using the original 491472613Skris result of the server certificate verification.) 491572613Skris [Lutz Jaenicke] 491672613Skris 491772613Skris *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 491872613Skris SSL3_RT_APPLICATION_DATA, return 0. 491972613Skris Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 492072613Skris [Bodo Moeller] 492172613Skris 492272613Skris *) Fix SSL_peek: 492372613Skris Both ssl2_peek and ssl3_peek, which were totally broken in earlier 492472613Skris releases, have been re-implemented by renaming the previous 492572613Skris implementations of ssl2_read and ssl3_read to ssl2_read_internal 492672613Skris and ssl3_read_internal, respectively, and adding 'peek' parameters 492772613Skris to them. The new ssl[23]_{read,peek} functions are calls to 492872613Skris ssl[23]_read_internal with the 'peek' flag set appropriately. 492972613Skris A 'peek' parameter has also been added to ssl3_read_bytes, which 493072613Skris does the actual work for ssl3_read_internal. 493172613Skris [Bodo Moeller] 493272613Skris 493376866Skris *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 493476866Skris the method-specific "init()" handler. Also clean up ex_data after 493576866Skris calling the method-specific "finish()" handler. Previously, this was 493676866Skris happening the other way round. 493776866Skris [Geoff Thorpe] 493876866Skris 493972613Skris *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 494072613Skris The previous value, 12, was not always sufficient for BN_mod_exp(). 494172613Skris [Bodo Moeller] 494272613Skris 494376866Skris *) Make sure that shared libraries get the internal name engine with 494476866Skris the full version number and not just 0. This should mark the 494576866Skris shared libraries as not backward compatible. Of course, this should 494676866Skris be changed again when we can guarantee backward binary compatibility. 494776866Skris [Richard Levitte] 494876866Skris 494972613Skris *) Fix typo in get_cert_by_subject() in by_dir.c 495072613Skris [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 495172613Skris 495276866Skris *) Rework the system to generate shared libraries: 495376866Skris 495476866Skris - Make note of the expected extension for the shared libraries and 495576866Skris if there is a need for symbolic links from for example libcrypto.so.0 495676866Skris to libcrypto.so.0.9.7. There is extended info in Configure for 495776866Skris that. 495876866Skris 495976866Skris - Make as few rebuilds of the shared libraries as possible. 496076866Skris 496176866Skris - Still avoid linking the OpenSSL programs with the shared libraries. 496276866Skris 496376866Skris - When installing, install the shared libraries separately from the 496476866Skris static ones. 496576866Skris [Richard Levitte] 496676866Skris 496772613Skris *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 496872613Skris 496972613Skris Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 497072613Skris and not in SSL_clear because the latter is also used by the 497172613Skris accept/connect functions; previously, the settings made by 497272613Skris SSL_set_read_ahead would be lost during the handshake. 497372613Skris [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 497472613Skris 497572613Skris *) Correct util/mkdef.pl to be selective about disabled algorithms. 497672613Skris Previously, it would create entries for disableed algorithms no 497772613Skris matter what. 497872613Skris [Richard Levitte] 497972613Skris 498072613Skris *) Added several new manual pages for SSL_* function. 498172613Skris [Lutz Jaenicke] 498272613Skris 498368651Skris Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 498468651Skris 498568651Skris *) In ssl23_get_client_hello, generate an error message when faced 498668651Skris with an initial SSL 3.0/TLS record that is too small to contain the 498768651Skris first two bytes of the ClientHello message, i.e. client_version. 498868651Skris (Note that this is a pathologic case that probably has never happened 498968651Skris in real life.) The previous approach was to use the version number 499068651Skris from the record header as a substitute; but our protocol choice 499168651Skris should not depend on that one because it is not authenticated 499268651Skris by the Finished messages. 499368651Skris [Bodo Moeller] 499468651Skris 499568651Skris *) More robust randomness gathering functions for Windows. 499668651Skris [Jeffrey Altman <jaltman@columbia.edu>] 499768651Skris 499868651Skris *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 499968651Skris not set then we don't setup the error code for issuer check errors 500068651Skris to avoid possibly overwriting other errors which the callback does 500168651Skris handle. If an application does set the flag then we assume it knows 500268651Skris what it is doing and can handle the new informational codes 500368651Skris appropriately. 500468651Skris [Steve Henson] 500568651Skris 500668651Skris *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 500768651Skris a general "ANY" type, as such it should be able to decode anything 500868651Skris including tagged types. However it didn't check the class so it would 500968651Skris wrongly interpret tagged types in the same way as their universal 501068651Skris counterpart and unknown types were just rejected. Changed so that the 501168651Skris tagged and unknown types are handled in the same way as a SEQUENCE: 501268651Skris that is the encoding is stored intact. There is also a new type 501368651Skris "V_ASN1_OTHER" which is used when the class is not universal, in this 501468651Skris case we have no idea what the actual type is so we just lump them all 501568651Skris together. 501668651Skris [Steve Henson] 501768651Skris 501868651Skris *) On VMS, stdout may very well lead to a file that is written to 501968651Skris in a record-oriented fashion. That means that every write() will 502068651Skris write a separate record, which will be read separately by the 502168651Skris programs trying to read from it. This can be very confusing. 502268651Skris 502368651Skris The solution is to put a BIO filter in the way that will buffer 502468651Skris text until a linefeed is reached, and then write everything a 502568651Skris line at a time, so every record written will be an actual line, 502668651Skris not chunks of lines and not (usually doesn't happen, but I've 502768651Skris seen it once) several lines in one record. BIO_f_linebuffer() is 502868651Skris the answer. 502968651Skris 503068651Skris Currently, it's a VMS-only method, because that's where it has 503168651Skris been tested well enough. 503268651Skris [Richard Levitte] 503368651Skris 503468651Skris *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 503568651Skris it can return incorrect results. 503668651Skris (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 503768651Skris but it was in 0.9.6-beta[12].) 503868651Skris [Bodo Moeller] 503968651Skris 504068651Skris *) Disable the check for content being present when verifying detached 504168651Skris signatures in pk7_smime.c. Some versions of Netscape (wrongly) 504268651Skris include zero length content when signing messages. 504368651Skris [Steve Henson] 504468651Skris 504568651Skris *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 504668651Skris BIO_ctrl (for BIO pairs). 504768651Skris [Bodo M�ller] 504868651Skris 504968651Skris *) Add DSO method for VMS. 505068651Skris [Richard Levitte] 505168651Skris 505268651Skris *) Bug fix: Montgomery multiplication could produce results with the 505368651Skris wrong sign. 505468651Skris [Ulf M�ller] 505568651Skris 505668651Skris *) Add RPM specification openssl.spec and modify it to build three 505768651Skris packages. The default package contains applications, application 505868651Skris documentation and run-time libraries. The devel package contains 505968651Skris include files, static libraries and function documentation. The 506068651Skris doc package contains the contents of the doc directory. The original 506168651Skris openssl.spec was provided by Damien Miller <djm@mindrot.org>. 506268651Skris [Richard Levitte] 506368651Skris 506468651Skris *) Add a large number of documentation files for many SSL routines. 506568651Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 506668651Skris 506768651Skris *) Add a configuration entry for Sony News 4. 506868651Skris [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 506968651Skris 507068651Skris *) Don't set the two most significant bits to one when generating a 507168651Skris random number < q in the DSA library. 507268651Skris [Ulf M�ller] 507368651Skris 507468651Skris *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 507568651Skris behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 507668651Skris the underlying transport is blocking) if a handshake took place. 507768651Skris (The default behaviour is needed by applications such as s_client 507868651Skris and s_server that use select() to determine when to use SSL_read; 507968651Skris but for applications that know in advance when to expect data, it 508068651Skris just makes things more complicated.) 508168651Skris [Bodo Moeller] 508268651Skris 508368651Skris *) Add RAND_egd_bytes(), which gives control over the number of bytes read 508468651Skris from EGD. 508568651Skris [Ben Laurie] 508668651Skris 508768651Skris *) Add a few more EBCDIC conditionals that make `req' and `x509' 508868651Skris work better on such systems. 508968651Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 509068651Skris 509168651Skris *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 509268651Skris Update PKCS12_parse() so it copies the friendlyName and the 509368651Skris keyid to the certificates aux info. 509468651Skris [Steve Henson] 509568651Skris 509668651Skris *) Fix bug in PKCS7_verify() which caused an infinite loop 509768651Skris if there was more than one signature. 509868651Skris [Sven Uszpelkat <su@celocom.de>] 509968651Skris 510068651Skris *) Major change in util/mkdef.pl to include extra information 510168651Skris about each symbol, as well as presentig variables as well 510268651Skris as functions. This change means that there's n more need 510368651Skris to rebuild the .num files when some algorithms are excluded. 510468651Skris [Richard Levitte] 510568651Skris 510668651Skris *) Allow the verify time to be set by an application, 510768651Skris rather than always using the current time. 510868651Skris [Steve Henson] 510968651Skris 511068651Skris *) Phase 2 verify code reorganisation. The certificate 511168651Skris verify code now looks up an issuer certificate by a 511268651Skris number of criteria: subject name, authority key id 511368651Skris and key usage. It also verifies self signed certificates 511468651Skris by the same criteria. The main comparison function is 511568651Skris X509_check_issued() which performs these checks. 511668651Skris 511768651Skris Lot of changes were necessary in order to support this 511868651Skris without completely rewriting the lookup code. 511968651Skris 512068651Skris Authority and subject key identifier are now cached. 512168651Skris 512268651Skris The LHASH 'certs' is X509_STORE has now been replaced 512368651Skris by a STACK_OF(X509_OBJECT). This is mainly because an 512468651Skris LHASH can't store or retrieve multiple objects with 512568651Skris the same hash value. 512668651Skris 512768651Skris As a result various functions (which were all internal 512868651Skris use only) have changed to handle the new X509_STORE 512968651Skris structure. This will break anything that messed round 513068651Skris with X509_STORE internally. 513168651Skris 513268651Skris The functions X509_STORE_add_cert() now checks for an 513368651Skris exact match, rather than just subject name. 513468651Skris 513568651Skris The X509_STORE API doesn't directly support the retrieval 513668651Skris of multiple certificates matching a given criteria, however 513768651Skris this can be worked round by performing a lookup first 513868651Skris (which will fill the cache with candidate certificates) 513968651Skris and then examining the cache for matches. This is probably 514068651Skris the best we can do without throwing out X509_LOOKUP 514168651Skris entirely (maybe later...). 514268651Skris 514368651Skris The X509_VERIFY_CTX structure has been enhanced considerably. 514468651Skris 514568651Skris All certificate lookup operations now go via a get_issuer() 514668651Skris callback. Although this currently uses an X509_STORE it 514768651Skris can be replaced by custom lookups. This is a simple way 514868651Skris to bypass the X509_STORE hackery necessary to make this 514968651Skris work and makes it possible to use more efficient techniques 515068651Skris in future. A very simple version which uses a simple 515168651Skris STACK for its trusted certificate store is also provided 515268651Skris using X509_STORE_CTX_trusted_stack(). 515368651Skris 515468651Skris The verify_cb() and verify() callbacks now have equivalents 515568651Skris in the X509_STORE_CTX structure. 515668651Skris 515768651Skris X509_STORE_CTX also has a 'flags' field which can be used 515868651Skris to customise the verify behaviour. 515968651Skris [Steve Henson] 516068651Skris 516168651Skris *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 516268651Skris excludes S/MIME capabilities. 516368651Skris [Steve Henson] 516468651Skris 516568651Skris *) When a certificate request is read in keep a copy of the 516668651Skris original encoding of the signed data and use it when outputing 516768651Skris again. Signatures then use the original encoding rather than 516868651Skris a decoded, encoded version which may cause problems if the 516968651Skris request is improperly encoded. 517068651Skris [Steve Henson] 517168651Skris 517268651Skris *) For consistency with other BIO_puts implementations, call 517368651Skris buffer_write(b, ...) directly in buffer_puts instead of calling 517468651Skris BIO_write(b, ...). 517568651Skris 517668651Skris In BIO_puts, increment b->num_write as in BIO_write. 517768651Skris [Peter.Sylvester@EdelWeb.fr] 517868651Skris 517968651Skris *) Fix BN_mul_word for the case where the word is 0. (We have to use 518068651Skris BN_zero, we may not return a BIGNUM with an array consisting of 518168651Skris words set to zero.) 518268651Skris [Bodo Moeller] 518368651Skris 518468651Skris *) Avoid calling abort() from within the library when problems are 518568651Skris detected, except if preprocessor symbols have been defined 518668651Skris (such as REF_CHECK, BN_DEBUG etc.). 518768651Skris [Bodo Moeller] 518868651Skris 518968651Skris *) New openssl application 'rsautl'. This utility can be 519068651Skris used for low level RSA operations. DER public key 519168651Skris BIO/fp routines also added. 519268651Skris [Steve Henson] 519368651Skris 519468651Skris *) New Configure entry and patches for compiling on QNX 4. 519568651Skris [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 519668651Skris 519768651Skris *) A demo state-machine implementation was sponsored by 519868651Skris Nuron (http://www.nuron.com/) and is now available in 519968651Skris demos/state_machine. 520068651Skris [Ben Laurie] 520168651Skris 520268651Skris *) New options added to the 'dgst' utility for signature 520368651Skris generation and verification. 520468651Skris [Steve Henson] 520568651Skris 520668651Skris *) Unrecognized PKCS#7 content types are now handled via a 520768651Skris catch all ASN1_TYPE structure. This allows unsupported 520868651Skris types to be stored as a "blob" and an application can 520968651Skris encode and decode it manually. 521068651Skris [Steve Henson] 521168651Skris 521268651Skris *) Fix various signed/unsigned issues to make a_strex.c 521368651Skris compile under VC++. 521468651Skris [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 521568651Skris 521668651Skris *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 521768651Skris length if passed a buffer. ASN1_INTEGER_to_BN failed 521868651Skris if passed a NULL BN and its argument was negative. 521968651Skris [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 522068651Skris 522168651Skris *) Modification to PKCS#7 encoding routines to output definite 522268651Skris length encoding. Since currently the whole structures are in 522368651Skris memory there's not real point in using indefinite length 522468651Skris constructed encoding. However if OpenSSL is compiled with 522568651Skris the flag PKCS7_INDEFINITE_ENCODING the old form is used. 522668651Skris [Steve Henson] 522768651Skris 522868651Skris *) Added BIO_vprintf() and BIO_vsnprintf(). 522968651Skris [Richard Levitte] 523068651Skris 523168651Skris *) Added more prefixes to parse for in the the strings written 523268651Skris through a logging bio, to cover all the levels that are available 523368651Skris through syslog. The prefixes are now: 523468651Skris 523568651Skris PANIC, EMERG, EMR => LOG_EMERG 523668651Skris ALERT, ALR => LOG_ALERT 523768651Skris CRIT, CRI => LOG_CRIT 523868651Skris ERROR, ERR => LOG_ERR 523968651Skris WARNING, WARN, WAR => LOG_WARNING 524068651Skris NOTICE, NOTE, NOT => LOG_NOTICE 524168651Skris INFO, INF => LOG_INFO 524268651Skris DEBUG, DBG => LOG_DEBUG 524368651Skris 524468651Skris and as before, if none of those prefixes are present at the 524568651Skris beginning of the string, LOG_ERR is chosen. 524668651Skris 524768651Skris On Win32, the LOG_* levels are mapped according to this: 524868651Skris 524968651Skris LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 525068651Skris LOG_WARNING => EVENTLOG_WARNING_TYPE 525168651Skris LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 525268651Skris 525368651Skris [Richard Levitte] 525468651Skris 525568651Skris *) Made it possible to reconfigure with just the configuration 525668651Skris argument "reconf" or "reconfigure". The command line arguments 525768651Skris are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 525868651Skris and are retrieved from there when reconfiguring. 525968651Skris [Richard Levitte] 526068651Skris 526168651Skris *) MD4 implemented. 526268651Skris [Assar Westerlund <assar@sics.se>, Richard Levitte] 526368651Skris 526468651Skris *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 526568651Skris [Richard Levitte] 526668651Skris 526768651Skris *) The obj_dat.pl script was messing up the sorting of object 526868651Skris names. The reason was that it compared the quoted version 526968651Skris of strings as a result "OCSP" > "OCSP Signing" because 527068651Skris " > SPACE. Changed script to store unquoted versions of 527168651Skris names and add quotes on output. It was also omitting some 527268651Skris names from the lookup table if they were given a default 527368651Skris value (that is if SN is missing it is given the same 527468651Skris value as LN and vice versa), these are now added on the 527568651Skris grounds that if an object has a name we should be able to 527668651Skris look it up. Finally added warning output when duplicate 527768651Skris short or long names are found. 527868651Skris [Steve Henson] 527968651Skris 528068651Skris *) Changes needed for Tandem NSK. 528168651Skris [Scott Uroff <scott@xypro.com>] 528268651Skris 528368651Skris *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 528468651Skris RSA_padding_check_SSLv23(), special padding was never detected 528568651Skris and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 528668651Skris version rollback attacks was not effective. 528768651Skris 528868651Skris In s23_clnt.c, don't use special rollback-attack detection padding 528968651Skris (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 529068651Skris client; similarly, in s23_srvr.c, don't do the rollback check if 529168651Skris SSL 2.0 is the only protocol enabled in the server. 529268651Skris [Bodo Moeller] 529368651Skris 529468651Skris *) Make it possible to get hexdumps of unprintable data with 'openssl 529568651Skris asn1parse'. By implication, the functions ASN1_parse_dump() and 529668651Skris BIO_dump_indent() are added. 529768651Skris [Richard Levitte] 529868651Skris 529968651Skris *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 530068651Skris these print out strings and name structures based on various 530168651Skris flags including RFC2253 support and proper handling of 530268651Skris multibyte characters. Added options to the 'x509' utility 530368651Skris to allow the various flags to be set. 530468651Skris [Steve Henson] 530568651Skris 530668651Skris *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 530768651Skris Also change the functions X509_cmp_current_time() and 530868651Skris X509_gmtime_adj() work with an ASN1_TIME structure, 530968651Skris this will enable certificates using GeneralizedTime in validity 531068651Skris dates to be checked. 531168651Skris [Steve Henson] 531268651Skris 531368651Skris *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 531468651Skris negative public key encodings) on by default, 531568651Skris NO_NEG_PUBKEY_BUG can be set to disable it. 531668651Skris [Steve Henson] 531768651Skris 531868651Skris *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 531968651Skris content octets. An i2c_ASN1_OBJECT is unnecessary because 532068651Skris the encoding can be trivially obtained from the structure. 532168651Skris [Steve Henson] 532268651Skris 532368651Skris *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 532468651Skris not read locks (CRYPTO_r_[un]lock). 532568651Skris [Bodo Moeller] 532668651Skris 532768651Skris *) A first attempt at creating official support for shared 532868651Skris libraries through configuration. I've kept it so the 532968651Skris default is static libraries only, and the OpenSSL programs 533068651Skris are always statically linked for now, but there are 533168651Skris preparations for dynamic linking in place. 533289837Skris This has been tested on Linux and Tru64. 533368651Skris [Richard Levitte] 533468651Skris 533568651Skris *) Randomness polling function for Win9x, as described in: 533668651Skris Peter Gutmann, Software Generation of Practically Strong 533768651Skris Random Numbers. 533868651Skris [Ulf M�ller] 533968651Skris 534068651Skris *) Fix so PRNG is seeded in req if using an already existing 534168651Skris DSA key. 534268651Skris [Steve Henson] 534368651Skris 534468651Skris *) New options to smime application. -inform and -outform 534568651Skris allow alternative formats for the S/MIME message including 534668651Skris PEM and DER. The -content option allows the content to be 534768651Skris specified separately. This should allow things like Netscape 534868651Skris form signing output easier to verify. 534968651Skris [Steve Henson] 535068651Skris 535168651Skris *) Fix the ASN1 encoding of tags using the 'long form'. 535268651Skris [Steve Henson] 535368651Skris 535468651Skris *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 535568651Skris STRING types. These convert content octets to and from the 535668651Skris underlying type. The actual tag and length octets are 535768651Skris already assumed to have been read in and checked. These 535868651Skris are needed because all other string types have virtually 535968651Skris identical handling apart from the tag. By having versions 536068651Skris of the ASN1 functions that just operate on content octets 536168651Skris IMPLICIT tagging can be handled properly. It also allows 536268651Skris the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 536368651Skris and ASN1_INTEGER are identical apart from the tag. 536468651Skris [Steve Henson] 536568651Skris 536668651Skris *) Change the handling of OID objects as follows: 536768651Skris 536868651Skris - New object identifiers are inserted in objects.txt, following 536968651Skris the syntax given in objects.README. 537068651Skris - objects.pl is used to process obj_mac.num and create a new 537168651Skris obj_mac.h. 537268651Skris - obj_dat.pl is used to create a new obj_dat.h, using the data in 537368651Skris obj_mac.h. 537468651Skris 537568651Skris This is currently kind of a hack, and the perl code in objects.pl 537668651Skris isn't very elegant, but it works as I intended. The simplest way 537768651Skris to check that it worked correctly is to look in obj_dat.h and 537868651Skris check the array nid_objs and make sure the objects haven't moved 537968651Skris around (this is important!). Additions are OK, as well as 538068651Skris consistent name changes. 538168651Skris [Richard Levitte] 538268651Skris 538368651Skris *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 538468651Skris [Bodo Moeller] 538568651Skris 538668651Skris *) Addition of the command line parameter '-rand file' to 'openssl req'. 538768651Skris The given file adds to whatever has already been seeded into the 538868651Skris random pool through the RANDFILE configuration file option or 538968651Skris environment variable, or the default random state file. 539068651Skris [Richard Levitte] 539168651Skris 539268651Skris *) mkstack.pl now sorts each macro group into lexical order. 539368651Skris Previously the output order depended on the order the files 539468651Skris appeared in the directory, resulting in needless rewriting 539568651Skris of safestack.h . 539668651Skris [Steve Henson] 539768651Skris 539868651Skris *) Patches to make OpenSSL compile under Win32 again. Mostly 539968651Skris work arounds for the VC++ problem that it treats func() as 540068651Skris func(void). Also stripped out the parts of mkdef.pl that 540168651Skris added extra typesafe functions: these no longer exist. 540268651Skris [Steve Henson] 540368651Skris 540468651Skris *) Reorganisation of the stack code. The macros are now all 540568651Skris collected in safestack.h . Each macro is defined in terms of 540668651Skris a "stack macro" of the form SKM_<name>(type, a, b). The 540768651Skris DEBUG_SAFESTACK is now handled in terms of function casts, 540868651Skris this has the advantage of retaining type safety without the 540968651Skris use of additional functions. If DEBUG_SAFESTACK is not defined 541068651Skris then the non typesafe macros are used instead. Also modified the 541168651Skris mkstack.pl script to handle the new form. Needs testing to see 541268651Skris if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 541368651Skris the default if no major problems. Similar behaviour for ASN1_SET_OF 541468651Skris and PKCS12_STACK_OF. 541568651Skris [Steve Henson] 541668651Skris 541768651Skris *) When some versions of IIS use the 'NET' form of private key the 541868651Skris key derivation algorithm is different. Normally MD5(password) is 541968651Skris used as a 128 bit RC4 key. In the modified case 542068651Skris MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 542168651Skris new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 542268651Skris as the old Netscape_RSA functions except they have an additional 542368651Skris 'sgckey' parameter which uses the modified algorithm. Also added 542468651Skris an -sgckey command line option to the rsa utility. Thanks to 542568651Skris Adrian Peck <bertie@ncipher.com> for posting details of the modified 542668651Skris algorithm to openssl-dev. 542768651Skris [Steve Henson] 542868651Skris 542968651Skris *) The evp_local.h macros were using 'c.##kname' which resulted in 543068651Skris invalid expansion on some systems (SCO 5.0.5 for example). 543168651Skris Corrected to 'c.kname'. 543268651Skris [Phillip Porch <root@theporch.com>] 543368651Skris 543468651Skris *) New X509_get1_email() and X509_REQ_get1_email() functions that return 543568651Skris a STACK of email addresses from a certificate or request, these look 543668651Skris in the subject name and the subject alternative name extensions and 543768651Skris omit any duplicate addresses. 543868651Skris [Steve Henson] 543968651Skris 544068651Skris *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 544168651Skris This makes DSA verification about 2 % faster. 544268651Skris [Bodo Moeller] 544368651Skris 544468651Skris *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 544568651Skris (meaning that now 2^5 values will be precomputed, which is only 4 KB 544668651Skris plus overhead for 1024 bit moduli). 544768651Skris This makes exponentiations about 0.5 % faster for 1024 bit 544868651Skris exponents (as measured by "openssl speed rsa2048"). 544968651Skris [Bodo Moeller] 545068651Skris 545168651Skris *) Rename memory handling macros to avoid conflicts with other 545268651Skris software: 545368651Skris Malloc => OPENSSL_malloc 545468651Skris Malloc_locked => OPENSSL_malloc_locked 545568651Skris Realloc => OPENSSL_realloc 545668651Skris Free => OPENSSL_free 545768651Skris [Richard Levitte] 545868651Skris 545968651Skris *) New function BN_mod_exp_mont_word for small bases (roughly 15% 546068651Skris faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 546168651Skris [Bodo Moeller] 546268651Skris 546368651Skris *) CygWin32 support. 546468651Skris [John Jarvie <jjarvie@newsguy.com>] 546568651Skris 546668651Skris *) The type-safe stack code has been rejigged. It is now only compiled 546768651Skris in when OpenSSL is configured with the DEBUG_SAFESTACK option and 546868651Skris by default all type-specific stack functions are "#define"d back to 546968651Skris standard stack functions. This results in more streamlined output 547068651Skris but retains the type-safety checking possibilities of the original 547168651Skris approach. 547268651Skris [Geoff Thorpe] 547368651Skris 547468651Skris *) The STACK code has been cleaned up, and certain type declarations 547568651Skris that didn't make a lot of sense have been brought in line. This has 547668651Skris also involved a cleanup of sorts in safestack.h to more correctly 547768651Skris map type-safe stack functions onto their plain stack counterparts. 547868651Skris This work has also resulted in a variety of "const"ifications of 547968651Skris lots of the code, especially "_cmp" operations which should normally 548068651Skris be prototyped with "const" parameters anyway. 548168651Skris [Geoff Thorpe] 548268651Skris 548368651Skris *) When generating bytes for the first time in md_rand.c, 'stir the pool' 548468651Skris by seeding with STATE_SIZE dummy bytes (with zero entropy count). 548568651Skris (The PRNG state consists of two parts, the large pool 'state' and 'md', 548668651Skris where all of 'md' is used each time the PRNG is used, but 'state' 548768651Skris is used only indexed by a cyclic counter. As entropy may not be 548868651Skris well distributed from the beginning, 'md' is important as a 548968651Skris chaining variable. However, the output function chains only half 549068651Skris of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 549168651Skris all of 'md', and seeding with STATE_SIZE dummy bytes will result 549268651Skris in all of 'state' being rewritten, with the new values depending 549368651Skris on virtually all of 'md'. This overcomes the 80 bit limitation.) 549468651Skris [Bodo Moeller] 549568651Skris 549668651Skris *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 549768651Skris the handshake is continued after ssl_verify_cert_chain(); 549868651Skris otherwise, if SSL_VERIFY_NONE is set, remaining error codes 549968651Skris can lead to 'unexplainable' connection aborts later. 550068651Skris [Bodo Moeller; problem tracked down by Lutz Jaenicke] 550168651Skris 550268651Skris *) Major EVP API cipher revision. 550368651Skris Add hooks for extra EVP features. This allows various cipher 550468651Skris parameters to be set in the EVP interface. Support added for variable 550568651Skris key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 550668651Skris setting of RC2 and RC5 parameters. 550768651Skris 550868651Skris Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 550968651Skris ciphers. 551068651Skris 551168651Skris Remove lots of duplicated code from the EVP library. For example *every* 551268651Skris cipher init() function handles the 'iv' in the same way according to the 551368651Skris cipher mode. They also all do nothing if the 'key' parameter is NULL and 551468651Skris for CFB and OFB modes they zero ctx->num. 551568651Skris 551668651Skris New functionality allows removal of S/MIME code RC2 hack. 551768651Skris 551868651Skris Most of the routines have the same form and so can be declared in terms 551968651Skris of macros. 552068651Skris 552168651Skris By shifting this to the top level EVP_CipherInit() it can be removed from 552268651Skris all individual ciphers. If the cipher wants to handle IVs or keys 552368651Skris differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 552468651Skris flags. 552568651Skris 552668651Skris Change lots of functions like EVP_EncryptUpdate() to now return a 552768651Skris value: although software versions of the algorithms cannot fail 552868651Skris any installed hardware versions can. 552968651Skris [Steve Henson] 553068651Skris 553168651Skris *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 553268651Skris this option is set, tolerate broken clients that send the negotiated 553368651Skris protocol version number instead of the requested protocol version 553468651Skris number. 553568651Skris [Bodo Moeller] 553668651Skris 553768651Skris *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 553868651Skris i.e. non-zero for export ciphersuites, zero otherwise. 553968651Skris Previous versions had this flag inverted, inconsistent with 554068651Skris rsa_tmp_cb (..._TMP_RSA_CB). 554168651Skris [Bodo Moeller; problem reported by Amit Chopra] 554268651Skris 554368651Skris *) Add missing DSA library text string. Work around for some IIS 554468651Skris key files with invalid SEQUENCE encoding. 554568651Skris [Steve Henson] 554668651Skris 554768651Skris *) Add a document (doc/standards.txt) that list all kinds of standards 554868651Skris and so on that are implemented in OpenSSL. 554968651Skris [Richard Levitte] 555068651Skris 555168651Skris *) Enhance c_rehash script. Old version would mishandle certificates 555268651Skris with the same subject name hash and wouldn't handle CRLs at all. 555368651Skris Added -fingerprint option to crl utility, to support new c_rehash 555468651Skris features. 555568651Skris [Steve Henson] 555668651Skris 555768651Skris *) Eliminate non-ANSI declarations in crypto.h and stack.h. 555868651Skris [Ulf M�ller] 555968651Skris 556068651Skris *) Fix for SSL server purpose checking. Server checking was 556168651Skris rejecting certificates which had extended key usage present 556268651Skris but no ssl client purpose. 556368651Skris [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 556468651Skris 556568651Skris *) Make PKCS#12 code work with no password. The PKCS#12 spec 556668651Skris is a little unclear about how a blank password is handled. 556768651Skris Since the password in encoded as a BMPString with terminating 556868651Skris double NULL a zero length password would end up as just the 556968651Skris double NULL. However no password at all is different and is 557068651Skris handled differently in the PKCS#12 key generation code. NS 557168651Skris treats a blank password as zero length. MSIE treats it as no 557268651Skris password on export: but it will try both on import. We now do 557368651Skris the same: PKCS12_parse() tries zero length and no password if 557468651Skris the password is set to "" or NULL (NULL is now a valid password: 557568651Skris it wasn't before) as does the pkcs12 application. 557668651Skris [Steve Henson] 557768651Skris 557868651Skris *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 557968651Skris perror when PEM_read_bio_X509_REQ fails, the error message must 558068651Skris be obtained from the error queue. 558168651Skris [Bodo Moeller] 558268651Skris 558368651Skris *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 558468651Skris it in ERR_remove_state if appropriate, and change ERR_get_state 558568651Skris accordingly to avoid race conditions (this is necessary because 558668651Skris thread_hash is no longer constant once set). 558768651Skris [Bodo Moeller] 558868651Skris 558968651Skris *) Bugfix for linux-elf makefile.one. 559068651Skris [Ulf M�ller] 559168651Skris 559268651Skris *) RSA_get_default_method() will now cause a default 559368651Skris RSA_METHOD to be chosen if one doesn't exist already. 559468651Skris Previously this was only set during a call to RSA_new() 559568651Skris or RSA_new_method(NULL) meaning it was possible for 559668651Skris RSA_get_default_method() to return NULL. 559768651Skris [Geoff Thorpe] 559868651Skris 559968651Skris *) Added native name translation to the existing DSO code 560068651Skris that will convert (if the flag to do so is set) filenames 560168651Skris that are sufficiently small and have no path information 560268651Skris into a canonical native form. Eg. "blah" converted to 560368651Skris "libblah.so" or "blah.dll" etc. 560468651Skris [Geoff Thorpe] 560568651Skris 560668651Skris *) New function ERR_error_string_n(e, buf, len) which is like 560768651Skris ERR_error_string(e, buf), but writes at most 'len' bytes 560868651Skris including the 0 terminator. For ERR_error_string_n, 'buf' 560968651Skris may not be NULL. 561068651Skris [Damien Miller <djm@mindrot.org>, Bodo Moeller] 561168651Skris 561268651Skris *) CONF library reworked to become more general. A new CONF 561368651Skris configuration file reader "class" is implemented as well as a 561468651Skris new functions (NCONF_*, for "New CONF") to handle it. The now 561568651Skris old CONF_* functions are still there, but are reimplemented to 561668651Skris work in terms of the new functions. Also, a set of functions 561768651Skris to handle the internal storage of the configuration data is 561868651Skris provided to make it easier to write new configuration file 561968651Skris reader "classes" (I can definitely see something reading a 562068651Skris configuration file in XML format, for example), called _CONF_*, 562168651Skris or "the configuration storage API"... 562268651Skris 562368651Skris The new configuration file reading functions are: 562468651Skris 562568651Skris NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 562668651Skris NCONF_get_section, NCONF_get_string, NCONF_get_numbre 562768651Skris 562868651Skris NCONF_default, NCONF_WIN32 562968651Skris 563068651Skris NCONF_dump_fp, NCONF_dump_bio 563168651Skris 563268651Skris NCONF_default and NCONF_WIN32 are method (or "class") choosers, 563368651Skris NCONF_new creates a new CONF object. This works in the same way 563468651Skris as other interfaces in OpenSSL, like the BIO interface. 563568651Skris NCONF_dump_* dump the internal storage of the configuration file, 563668651Skris which is useful for debugging. All other functions take the same 563768651Skris arguments as the old CONF_* functions wth the exception of the 563868651Skris first that must be a `CONF *' instead of a `LHASH *'. 563968651Skris 564068651Skris To make it easer to use the new classes with the old CONF_* functions, 564168651Skris the function CONF_set_default_method is provided. 564268651Skris [Richard Levitte] 564368651Skris 564468651Skris *) Add '-tls1' option to 'openssl ciphers', which was already 564568651Skris mentioned in the documentation but had not been implemented. 564668651Skris (This option is not yet really useful because even the additional 564768651Skris experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 564868651Skris [Bodo Moeller] 564968651Skris 565068651Skris *) Initial DSO code added into libcrypto for letting OpenSSL (and 565168651Skris OpenSSL-based applications) load shared libraries and bind to 565268651Skris them in a portable way. 565368651Skris [Geoff Thorpe, with contributions from Richard Levitte] 565468651Skris 565559191Skris Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 565659191Skris 565759191Skris *) Make sure _lrotl and _lrotr are only used with MSVC. 565859191Skris 565959191Skris *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 566059191Skris (the default implementation of RAND_status). 566159191Skris 566259191Skris *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 566359191Skris to '-clrext' (= clear extensions), as intended and documented. 566459191Skris [Bodo Moeller; inconsistency pointed out by Michael Attili 566559191Skris <attili@amaxo.com>] 566659191Skris 566759191Skris *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 566859191Skris was larger than the MD block size. 566959191Skris [Steve Henson, pointed out by Yost William <YostW@tce.com>] 567059191Skris 567159191Skris *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 567259191Skris fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 567359191Skris using the passed key: if the passed key was a private key the result 567459191Skris of X509_print(), for example, would be to print out all the private key 567559191Skris components. 567659191Skris [Steve Henson] 567759191Skris 567859191Skris *) des_quad_cksum() byte order bug fix. 567959191Skris [Ulf M�ller, using the problem description in krb4-0.9.7, where 568059191Skris the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 568159191Skris 568259191Skris *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 568359191Skris discouraged. 568459191Skris [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 568559191Skris 568659191Skris *) For easily testing in shell scripts whether some command 568759191Skris 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 568859191Skris returns with exit code 0 iff no command of the given name is available. 568959191Skris 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 569059191Skris the output goes to stdout and nothing is printed to stderr. 569159191Skris Additional arguments are always ignored. 569259191Skris 569359191Skris Since for each cipher there is a command of the same name, 569459191Skris the 'no-cipher' compilation switches can be tested this way. 569559191Skris 569659191Skris ('openssl no-XXX' is not able to detect pseudo-commands such 569759191Skris as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 569859191Skris [Bodo Moeller] 569959191Skris 570059191Skris *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 570159191Skris [Bodo Moeller] 570259191Skris 570359191Skris *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 570459191Skris is set; it will be thrown away anyway because each handshake creates 570559191Skris its own key. 570659191Skris ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 570759191Skris to parameters -- in previous versions (since OpenSSL 0.9.3) the 570859191Skris 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 570959191Skris you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 571059191Skris [Bodo Moeller] 571159191Skris 571259191Skris *) New s_client option -ign_eof: EOF at stdin is ignored, and 571359191Skris 'Q' and 'R' lose their special meanings (quit/renegotiate). 571459191Skris This is part of what -quiet does; unlike -quiet, -ign_eof 571559191Skris does not suppress any output. 571659191Skris [Richard Levitte] 571759191Skris 571859191Skris *) Add compatibility options to the purpose and trust code. The 571959191Skris purpose X509_PURPOSE_ANY is "any purpose" which automatically 572059191Skris accepts a certificate or CA, this was the previous behaviour, 572159191Skris with all the associated security issues. 572259191Skris 572359191Skris X509_TRUST_COMPAT is the old trust behaviour: only and 572459191Skris automatically trust self signed roots in certificate store. A 572559191Skris new trust setting X509_TRUST_DEFAULT is used to specify that 572659191Skris a purpose has no associated trust setting and it should instead 572759191Skris use the value in the default purpose. 572859191Skris [Steve Henson] 572959191Skris 573059191Skris *) Fix the PKCS#8 DSA private key code so it decodes keys again 573159191Skris and fix a memory leak. 573259191Skris [Steve Henson] 573359191Skris 573459191Skris *) In util/mkerr.pl (which implements 'make errors'), preserve 573559191Skris reason strings from the previous version of the .c file, as 573659191Skris the default to have only downcase letters (and digits) in 573759191Skris automatically generated reasons codes is not always appropriate. 573859191Skris [Bodo Moeller] 573959191Skris 574059191Skris *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 574159191Skris using strerror. Previously, ERR_reason_error_string() returned 574259191Skris library names as reason strings for SYSerr; but SYSerr is a special 574359191Skris case where small numbers are errno values, not library numbers. 574459191Skris [Bodo Moeller] 574559191Skris 574659191Skris *) Add '-dsaparam' option to 'openssl dhparam' application. This 574759191Skris converts DSA parameters into DH parameters. (When creating parameters, 574859191Skris DSA_generate_parameters is used.) 574959191Skris [Bodo Moeller] 575059191Skris 575159191Skris *) Include 'length' (recommended exponent length) in C code generated 575259191Skris by 'openssl dhparam -C'. 575359191Skris [Bodo Moeller] 575459191Skris 575559191Skris *) The second argument to set_label in perlasm was already being used 575659191Skris so couldn't be used as a "file scope" flag. Moved to third argument 575759191Skris which was free. 575859191Skris [Steve Henson] 575959191Skris 576059191Skris *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 576159191Skris instead of RAND_bytes for encryption IVs and salts. 576259191Skris [Bodo Moeller] 576359191Skris 576459191Skris *) Include RAND_status() into RAND_METHOD instead of implementing 576559191Skris it only for md_rand.c Otherwise replacing the PRNG by calling 576659191Skris RAND_set_rand_method would be impossible. 576759191Skris [Bodo Moeller] 576859191Skris 576959191Skris *) Don't let DSA_generate_key() enter an infinite loop if the random 577059191Skris number generation fails. 577159191Skris [Bodo Moeller] 577259191Skris 577359191Skris *) New 'rand' application for creating pseudo-random output. 577459191Skris [Bodo Moeller] 577559191Skris 577659191Skris *) Added configuration support for Linux/IA64 577759191Skris [Rolf Haberrecker <rolf@suse.de>] 577859191Skris 577959191Skris *) Assembler module support for Mingw32. 578059191Skris [Ulf M�ller] 578159191Skris 578259191Skris *) Shared library support for HPUX (in shlib/). 578359191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 578459191Skris 578559191Skris *) Shared library support for Solaris gcc. 578659191Skris [Lutz Behnke <behnke@trustcenter.de>] 578759191Skris 578859191Skris Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 578959191Skris 579059191Skris *) PKCS7_encrypt() was adding text MIME headers twice because they 579159191Skris were added manually and by SMIME_crlf_copy(). 579259191Skris [Steve Henson] 579359191Skris 579459191Skris *) In bntest.c don't call BN_rand with zero bits argument. 579559191Skris [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 579659191Skris 579759191Skris *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 579859191Skris case was implemented. This caused BN_div_recp() to fail occasionally. 579959191Skris [Ulf M�ller] 580059191Skris 580159191Skris *) Add an optional second argument to the set_label() in the perl 580259191Skris assembly language builder. If this argument exists and is set 580359191Skris to 1 it signals that the assembler should use a symbol whose 580459191Skris scope is the entire file, not just the current function. This 580559191Skris is needed with MASM which uses the format label:: for this scope. 580659191Skris [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 580759191Skris 580859191Skris *) Change the ASN1 types so they are typedefs by default. Before 580959191Skris almost all types were #define'd to ASN1_STRING which was causing 581059191Skris STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 581159191Skris for example. 581259191Skris [Steve Henson] 581359191Skris 581459191Skris *) Change names of new functions to the new get1/get0 naming 581559191Skris convention: After 'get1', the caller owns a reference count 581659191Skris and has to call ..._free; 'get0' returns a pointer to some 581759191Skris data structure without incrementing reference counters. 581859191Skris (Some of the existing 'get' functions increment a reference 581959191Skris counter, some don't.) 582059191Skris Similarly, 'set1' and 'add1' functions increase reference 582159191Skris counters or duplicate objects. 582259191Skris [Steve Henson] 582359191Skris 582459191Skris *) Allow for the possibility of temp RSA key generation failure: 582559191Skris the code used to assume it always worked and crashed on failure. 582659191Skris [Steve Henson] 582759191Skris 582859191Skris *) Fix potential buffer overrun problem in BIO_printf(). 582959191Skris [Ulf M�ller, using public domain code by Patrick Powell; problem 583059191Skris pointed out by David Sacerdote <das33@cornell.edu>] 583159191Skris 583259191Skris *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 583359191Skris RAND_egd() and RAND_status(). In the command line application, 583459191Skris the EGD socket can be specified like a seed file using RANDFILE 583559191Skris or -rand. 583659191Skris [Ulf M�ller] 583759191Skris 583859191Skris *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 583959191Skris Some CAs (e.g. Verisign) distribute certificates in this form. 584059191Skris [Steve Henson] 584159191Skris 584259191Skris *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 584359191Skris list to exclude them. This means that no special compilation option 584459191Skris is needed to use anonymous DH: it just needs to be included in the 584559191Skris cipher list. 584659191Skris [Steve Henson] 584759191Skris 584859191Skris *) Change the EVP_MD_CTX_type macro so its meaning consistent with 584959191Skris EVP_MD_type. The old functionality is available in a new macro called 585059191Skris EVP_MD_md(). Change code that uses it and update docs. 585159191Skris [Steve Henson] 585259191Skris 585359191Skris *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 585459191Skris where the 'void *' argument is replaced by a function pointer argument. 585559191Skris Previously 'void *' was abused to point to functions, which works on 585659191Skris many platforms, but is not correct. As these functions are usually 585759191Skris called by macros defined in OpenSSL header files, most source code 585859191Skris should work without changes. 585959191Skris [Richard Levitte] 586059191Skris 586159191Skris *) <openssl/opensslconf.h> (which is created by Configure) now contains 586259191Skris sections with information on -D... compiler switches used for 586359191Skris compiling the library so that applications can see them. To enable 586459191Skris one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 586559191Skris must be defined. E.g., 586659191Skris #define OPENSSL_ALGORITHM_DEFINES 586759191Skris #include <openssl/opensslconf.h> 586859191Skris defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 586959191Skris [Richard Levitte, Ulf and Bodo M�ller] 587059191Skris 587159191Skris *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 587259191Skris record layer. 587359191Skris [Bodo Moeller] 587459191Skris 587559191Skris *) Change the 'other' type in certificate aux info to a STACK_OF 587659191Skris X509_ALGOR. Although not an AlgorithmIdentifier as such it has 587759191Skris the required ASN1 format: arbitrary types determined by an OID. 587859191Skris [Steve Henson] 587959191Skris 588059191Skris *) Add some PEM_write_X509_REQ_NEW() functions and a command line 588159191Skris argument to 'req'. This is not because the function is newer or 588259191Skris better than others it just uses the work 'NEW' in the certificate 588359191Skris request header lines. Some software needs this. 588459191Skris [Steve Henson] 588559191Skris 588659191Skris *) Reorganise password command line arguments: now passwords can be 588759191Skris obtained from various sources. Delete the PEM_cb function and make 588859191Skris it the default behaviour: i.e. if the callback is NULL and the 588959191Skris usrdata argument is not NULL interpret it as a null terminated pass 589059191Skris phrase. If usrdata and the callback are NULL then the pass phrase 589159191Skris is prompted for as usual. 589259191Skris [Steve Henson] 589359191Skris 589459191Skris *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 589559191Skris the support is automatically enabled. The resulting binaries will 589659191Skris autodetect the card and use it if present. 589759191Skris [Ben Laurie and Compaq Inc.] 589859191Skris 589959191Skris *) Work around for Netscape hang bug. This sends certificate request 590059191Skris and server done in one record. Since this is perfectly legal in the 590159191Skris SSL/TLS protocol it isn't a "bug" option and is on by default. See 590259191Skris the bugs/SSLv3 entry for more info. 590359191Skris [Steve Henson] 590459191Skris 590559191Skris *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 590659191Skris [Andy Polyakov] 590759191Skris 590859191Skris *) Add -rand argument to smime and pkcs12 applications and read/write 590959191Skris of seed file. 591059191Skris [Steve Henson] 591159191Skris 591259191Skris *) New 'passwd' tool for crypt(3) and apr1 password hashes. 591359191Skris [Bodo Moeller] 591459191Skris 591559191Skris *) Add command line password options to the remaining applications. 591659191Skris [Steve Henson] 591759191Skris 591859191Skris *) Bug fix for BN_div_recp() for numerators with an even number of 591959191Skris bits. 592059191Skris [Ulf M�ller] 592159191Skris 592259191Skris *) More tests in bntest.c, and changed test_bn output. 592359191Skris [Ulf M�ller] 592459191Skris 592559191Skris *) ./config recognizes MacOS X now. 592659191Skris [Andy Polyakov] 592759191Skris 592859191Skris *) Bug fix for BN_div() when the first words of num and divsor are 592959191Skris equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 593059191Skris [Ulf M�ller] 593159191Skris 593259191Skris *) Add support for various broken PKCS#8 formats, and command line 593359191Skris options to produce them. 593459191Skris [Steve Henson] 593559191Skris 593659191Skris *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 593759191Skris get temporary BIGNUMs from a BN_CTX. 593859191Skris [Ulf M�ller] 593959191Skris 594059191Skris *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 594159191Skris for p == 0. 594259191Skris [Ulf M�ller] 594359191Skris 594459191Skris *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 594559191Skris include a #define from the old name to the new. The original intent 594659191Skris was that statically linked binaries could for example just call 594759191Skris SSLeay_add_all_ciphers() to just add ciphers to the table and not 594859191Skris link with digests. This never worked becayse SSLeay_add_all_digests() 594959191Skris and SSLeay_add_all_ciphers() were in the same source file so calling 595059191Skris one would link with the other. They are now in separate source files. 595159191Skris [Steve Henson] 595259191Skris 595359191Skris *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 595459191Skris [Steve Henson] 595559191Skris 595659191Skris *) Use a less unusual form of the Miller-Rabin primality test (it used 595759191Skris a binary algorithm for exponentiation integrated into the Miller-Rabin 595859191Skris loop, our standard modexp algorithms are faster). 595959191Skris [Bodo Moeller] 596059191Skris 596159191Skris *) Support for the EBCDIC character set completed. 596259191Skris [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 596359191Skris 596459191Skris *) Source code cleanups: use const where appropriate, eliminate casts, 596559191Skris use void * instead of char * in lhash. 596659191Skris [Ulf M�ller] 596759191Skris 596859191Skris *) Bugfix: ssl3_send_server_key_exchange was not restartable 596959191Skris (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 597059191Skris this the server could overwrite ephemeral keys that the client 597159191Skris has already seen). 597259191Skris [Bodo Moeller] 597359191Skris 597459191Skris *) Turn DSA_is_prime into a macro that calls BN_is_prime, 597559191Skris using 50 iterations of the Rabin-Miller test. 597659191Skris 597759191Skris DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 597859191Skris iterations of the Rabin-Miller test as required by the appendix 597959191Skris to FIPS PUB 186[-1]) instead of DSA_is_prime. 598059191Skris As BN_is_prime_fasttest includes trial division, DSA parameter 598159191Skris generation becomes much faster. 598259191Skris 598359191Skris This implies a change for the callback functions in DSA_is_prime 598459191Skris and DSA_generate_parameters: The callback function is called once 598559191Skris for each positive witness in the Rabin-Miller test, not just 598659191Skris occasionally in the inner loop; and the parameters to the 598759191Skris callback function now provide an iteration count for the outer 598859191Skris loop rather than for the current invocation of the inner loop. 598959191Skris DSA_generate_parameters additionally can call the callback 599059191Skris function with an 'iteration count' of -1, meaning that a 599159191Skris candidate has passed the trial division test (when q is generated 599259191Skris from an application-provided seed, trial division is skipped). 599359191Skris [Bodo Moeller] 599459191Skris 599559191Skris *) New function BN_is_prime_fasttest that optionally does trial 599659191Skris division before starting the Rabin-Miller test and has 599759191Skris an additional BN_CTX * argument (whereas BN_is_prime always 599859191Skris has to allocate at least one BN_CTX). 599959191Skris 'callback(1, -1, cb_arg)' is called when a number has passed the 600059191Skris trial division stage. 600159191Skris [Bodo Moeller] 600259191Skris 600359191Skris *) Fix for bug in CRL encoding. The validity dates weren't being handled 600459191Skris as ASN1_TIME. 600559191Skris [Steve Henson] 600659191Skris 600759191Skris *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 600859191Skris [Steve Henson] 600959191Skris 601059191Skris *) New function BN_pseudo_rand(). 601159191Skris [Ulf M�ller] 601259191Skris 601359191Skris *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 601459191Skris bignum version of BN_from_montgomery() with the working code from 601559191Skris SSLeay 0.9.0 (the word based version is faster anyway), and clean up 601659191Skris the comments. 601759191Skris [Ulf M�ller] 601859191Skris 601959191Skris *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 602059191Skris made it impossible to use the same SSL_SESSION data structure in 602159191Skris SSL2 clients in multiple threads. 602259191Skris [Bodo Moeller] 602359191Skris 602459191Skris *) The return value of RAND_load_file() no longer counts bytes obtained 602559191Skris by stat(). RAND_load_file(..., -1) is new and uses the complete file 602659191Skris to seed the PRNG (previously an explicit byte count was required). 602759191Skris [Ulf M�ller, Bodo M�ller] 602859191Skris 602959191Skris *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 603059191Skris used (char *) instead of (void *) and had casts all over the place. 603159191Skris [Steve Henson] 603259191Skris 603359191Skris *) Make BN_generate_prime() return NULL on error if ret!=NULL. 603459191Skris [Ulf M�ller] 603559191Skris 603659191Skris *) Retain source code compatibility for BN_prime_checks macro: 603759191Skris BN_is_prime(..., BN_prime_checks, ...) now uses 603859191Skris BN_prime_checks_for_size to determine the appropriate number of 603959191Skris Rabin-Miller iterations. 604059191Skris [Ulf M�ller] 604159191Skris 604259191Skris *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 604359191Skris DH_CHECK_P_NOT_SAFE_PRIME. 604459191Skris (Check if this is true? OpenPGP calls them "strong".) 604559191Skris [Ulf M�ller] 604659191Skris 604759191Skris *) Merge the functionality of "dh" and "gendh" programs into a new program 604859191Skris "dhparam". The old programs are retained for now but will handle DH keys 604959191Skris (instead of parameters) in future. 605059191Skris [Steve Henson] 605159191Skris 605259191Skris *) Make the ciphers, s_server and s_client programs check the return values 605359191Skris when a new cipher list is set. 605459191Skris [Steve Henson] 605559191Skris 605659191Skris *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 605759191Skris ciphers. Before when the 56bit ciphers were enabled the sorting was 605859191Skris wrong. 605959191Skris 606059191Skris The syntax for the cipher sorting has been extended to support sorting by 606159191Skris cipher-strength (using the strength_bits hard coded in the tables). 606259191Skris The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 606359191Skris 606459191Skris Fix a bug in the cipher-command parser: when supplying a cipher command 606559191Skris string with an "undefined" symbol (neither command nor alphanumeric 606659191Skris [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 606759191Skris an error is flagged. 606859191Skris 606959191Skris Due to the strength-sorting extension, the code of the 607059191Skris ssl_create_cipher_list() function was completely rearranged. I hope that 607159191Skris the readability was also increased :-) 607259191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 607359191Skris 607459191Skris *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 607559191Skris for the first serial number and places 2 in the serial number file. This 607659191Skris avoids problems when the root CA is created with serial number zero and 607759191Skris the first user certificate has the same issuer name and serial number 607859191Skris as the root CA. 607959191Skris [Steve Henson] 608059191Skris 608159191Skris *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 608259191Skris the new code. Add documentation for this stuff. 608359191Skris [Steve Henson] 608459191Skris 608559191Skris *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 608659191Skris X509_*() to X509at_*() on the grounds that they don't handle X509 608759191Skris structures and behave in an analagous way to the X509v3 functions: 608859191Skris they shouldn't be called directly but wrapper functions should be used 608959191Skris instead. 609059191Skris 609159191Skris So we also now have some wrapper functions that call the X509at functions 609259191Skris when passed certificate requests. (TO DO: similar things can be done with 609359191Skris PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 609459191Skris things. Some of these need some d2i or i2d and print functionality 609559191Skris because they handle more complex structures.) 609659191Skris [Steve Henson] 609759191Skris 609859191Skris *) Add missing #ifndefs that caused missing symbols when building libssl 609959191Skris as a shared library without RSA. Use #ifndef NO_SSL2 instead of 610059191Skris NO_RSA in ssl/s2*.c. 610159191Skris [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 610259191Skris 610359191Skris *) Precautions against using the PRNG uninitialized: RAND_bytes() now 610459191Skris has a return value which indicates the quality of the random data 610559191Skris (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 610659191Skris error queue. New function RAND_pseudo_bytes() generates output that is 610759191Skris guaranteed to be unique but not unpredictable. RAND_add is like 610859191Skris RAND_seed, but takes an extra argument for an entropy estimate 610959191Skris (RAND_seed always assumes full entropy). 611059191Skris [Ulf M�ller] 611159191Skris 611259191Skris *) Do more iterations of Rabin-Miller probable prime test (specifically, 611359191Skris 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 611459191Skris instead of only 2 for all lengths; see BN_prime_checks_for_size definition 611559191Skris in crypto/bn/bn_prime.c for the complete table). This guarantees a 611659191Skris false-positive rate of at most 2^-80 for random input. 611759191Skris [Bodo Moeller] 611859191Skris 611959191Skris *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 612059191Skris [Bodo Moeller] 612159191Skris 612259191Skris *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 612359191Skris in the 0.9.5 release), this returns the chain 612459191Skris from an X509_CTX structure with a dup of the stack and all 612559191Skris the X509 reference counts upped: so the stack will exist 612659191Skris after X509_CTX_cleanup() has been called. Modify pkcs12.c 612759191Skris to use this. 612859191Skris 612959191Skris Also make SSL_SESSION_print() print out the verify return 613059191Skris code. 613159191Skris [Steve Henson] 613259191Skris 613359191Skris *) Add manpage for the pkcs12 command. Also change the default 613459191Skris behaviour so MAC iteration counts are used unless the new 613559191Skris -nomaciter option is used. This improves file security and 613659191Skris only older versions of MSIE (4.0 for example) need it. 613759191Skris [Steve Henson] 613859191Skris 613959191Skris *) Honor the no-xxx Configure options when creating .DEF files. 614059191Skris [Ulf M�ller] 614159191Skris 614259191Skris *) Add PKCS#10 attributes to field table: challengePassword, 614359191Skris unstructuredName and unstructuredAddress. These are taken from 614459191Skris draft PKCS#9 v2.0 but are compatible with v1.2 provided no 614559191Skris international characters are used. 614659191Skris 614759191Skris More changes to X509_ATTRIBUTE code: allow the setting of types 614859191Skris based on strings. Remove the 'loc' parameter when adding 614959191Skris attributes because these will be a SET OF encoding which is sorted 615059191Skris in ASN1 order. 615159191Skris [Steve Henson] 615259191Skris 615359191Skris *) Initial changes to the 'req' utility to allow request generation 615459191Skris automation. This will allow an application to just generate a template 615559191Skris file containing all the field values and have req construct the 615659191Skris request. 615759191Skris 615859191Skris Initial support for X509_ATTRIBUTE handling. Stacks of these are 615959191Skris used all over the place including certificate requests and PKCS#7 616059191Skris structures. They are currently handled manually where necessary with 616159191Skris some primitive wrappers for PKCS#7. The new functions behave in a 616259191Skris manner analogous to the X509 extension functions: they allow 616359191Skris attributes to be looked up by NID and added. 616459191Skris 616559191Skris Later something similar to the X509V3 code would be desirable to 616659191Skris automatically handle the encoding, decoding and printing of the 616759191Skris more complex types. The string types like challengePassword can 616859191Skris be handled by the string table functions. 616959191Skris 617059191Skris Also modified the multi byte string table handling. Now there is 617159191Skris a 'global mask' which masks out certain types. The table itself 617259191Skris can use the flag STABLE_NO_MASK to ignore the mask setting: this 617359191Skris is useful when for example there is only one permissible type 617459191Skris (as in countryName) and using the mask might result in no valid 617559191Skris types at all. 617659191Skris [Steve Henson] 617759191Skris 617859191Skris *) Clean up 'Finished' handling, and add functions SSL_get_finished and 617959191Skris SSL_get_peer_finished to allow applications to obtain the latest 618059191Skris Finished messages sent to the peer or expected from the peer, 618159191Skris respectively. (SSL_get_peer_finished is usually the Finished message 618259191Skris actually received from the peer, otherwise the protocol will be aborted.) 618359191Skris 618459191Skris As the Finished message are message digests of the complete handshake 618559191Skris (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 618659191Skris be used for external authentication procedures when the authentication 618759191Skris provided by SSL/TLS is not desired or is not enough. 618859191Skris [Bodo Moeller] 618959191Skris 619059191Skris *) Enhanced support for Alpha Linux is added. Now ./config checks if 619159191Skris the host supports BWX extension and if Compaq C is present on the 619259191Skris $PATH. Just exploiting of the BWX extension results in 20-30% 619359191Skris performance kick for some algorithms, e.g. DES and RC4 to mention 619459191Skris a couple. Compaq C in turn generates ~20% faster code for MD5 and 619559191Skris SHA1. 619659191Skris [Andy Polyakov] 619759191Skris 619859191Skris *) Add support for MS "fast SGC". This is arguably a violation of the 619959191Skris SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 620059191Skris weak crypto and after checking the certificate is SGC a second one 620159191Skris with strong crypto. MS SGC stops the first handshake after receiving 620259191Skris the server certificate message and sends a second client hello. Since 620359191Skris a server will typically do all the time consuming operations before 620459191Skris expecting any further messages from the client (server key exchange 620559191Skris is the most expensive) there is little difference between the two. 620659191Skris 620759191Skris To get OpenSSL to support MS SGC we have to permit a second client 620859191Skris hello message after we have sent server done. In addition we have to 620959191Skris reset the MAC if we do get this second client hello. 621059191Skris [Steve Henson] 621159191Skris 621259191Skris *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 621359191Skris if a DER encoded private key is RSA or DSA traditional format. Changed 621459191Skris d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 621559191Skris format DER encoded private key. Newer code should use PKCS#8 format which 621659191Skris has the key type encoded in the ASN1 structure. Added DER private key 621759191Skris support to pkcs8 application. 621859191Skris [Steve Henson] 621959191Skris 622059191Skris *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 622159191Skris ciphersuites has been selected (as required by the SSL 3/TLS 1 622259191Skris specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 622359191Skris is set, we interpret this as a request to violate the specification 622459191Skris (the worst that can happen is a handshake failure, and 'correct' 622559191Skris behaviour would result in a handshake failure anyway). 622659191Skris [Bodo Moeller] 622759191Skris 622859191Skris *) In SSL_CTX_add_session, take into account that there might be multiple 622959191Skris SSL_SESSION structures with the same session ID (e.g. when two threads 623059191Skris concurrently obtain them from an external cache). 623159191Skris The internal cache can handle only one SSL_SESSION with a given ID, 623259191Skris so if there's a conflict, we now throw out the old one to achieve 623359191Skris consistency. 623459191Skris [Bodo Moeller] 623559191Skris 623659191Skris *) Add OIDs for idea and blowfish in CBC mode. This will allow both 623759191Skris to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 623859191Skris some routines that use cipher OIDs: some ciphers do not have OIDs 623959191Skris defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 624059191Skris example. 624159191Skris [Steve Henson] 624259191Skris 624359191Skris *) Simplify the trust setting structure and code. Now we just have 624459191Skris two sequences of OIDs for trusted and rejected settings. These will 624559191Skris typically have values the same as the extended key usage extension 624659191Skris and any application specific purposes. 624759191Skris 624859191Skris The trust checking code now has a default behaviour: it will just 624959191Skris check for an object with the same NID as the passed id. Functions can 625059191Skris be provided to override either the default behaviour or the behaviour 625159191Skris for a given id. SSL client, server and email already have functions 625259191Skris in place for compatibility: they check the NID and also return "trusted" 625359191Skris if the certificate is self signed. 625459191Skris [Steve Henson] 625559191Skris 625659191Skris *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 625759191Skris traditional format into an EVP_PKEY structure. 625859191Skris [Steve Henson] 625959191Skris 626059191Skris *) Add a password callback function PEM_cb() which either prompts for 626159191Skris a password if usr_data is NULL or otherwise assumes it is a null 626259191Skris terminated password. Allow passwords to be passed on command line 626359191Skris environment or config files in a few more utilities. 626459191Skris [Steve Henson] 626559191Skris 626659191Skris *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 626759191Skris keys. Add some short names for PKCS#8 PBE algorithms and allow them 626859191Skris to be specified on the command line for the pkcs8 and pkcs12 utilities. 626959191Skris Update documentation. 627059191Skris [Steve Henson] 627159191Skris 627259191Skris *) Support for ASN1 "NULL" type. This could be handled before by using 627359191Skris ASN1_TYPE but there wasn't any function that would try to read a NULL 627459191Skris and produce an error if it couldn't. For compatibility we also have 627559191Skris ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 627659191Skris don't allocate anything because they don't need to. 627759191Skris [Steve Henson] 627859191Skris 627959191Skris *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 628059191Skris for details. 628159191Skris [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 628259191Skris 628359191Skris *) Rebuild of the memory allocation routines used by OpenSSL code and 628459191Skris possibly others as well. The purpose is to make an interface that 628559191Skris provide hooks so anyone can build a separate set of allocation and 628659191Skris deallocation routines to be used by OpenSSL, for example memory 628759191Skris pool implementations, or something else, which was previously hard 628859191Skris since Malloc(), Realloc() and Free() were defined as macros having 628959191Skris the values malloc, realloc and free, respectively (except for Win32 629059191Skris compilations). The same is provided for memory debugging code. 629159191Skris OpenSSL already comes with functionality to find memory leaks, but 629259191Skris this gives people a chance to debug other memory problems. 629359191Skris 629459191Skris With these changes, a new set of functions and macros have appeared: 629559191Skris 629668651Skris CRYPTO_set_mem_debug_functions() [F] 629768651Skris CRYPTO_get_mem_debug_functions() [F] 629868651Skris CRYPTO_dbg_set_options() [F] 629968651Skris CRYPTO_dbg_get_options() [F] 630068651Skris CRYPTO_malloc_debug_init() [M] 630159191Skris 630259191Skris The memory debug functions are NULL by default, unless the library 630359191Skris is compiled with CRYPTO_MDEBUG or friends is defined. If someone 630459191Skris wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 630559191Skris gives the standard debugging functions that come with OpenSSL) or 630659191Skris CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 630759191Skris provided by the library user) must be used. When the standard 630859191Skris debugging functions are used, CRYPTO_dbg_set_options can be used to 630959191Skris request additional information: 631059191Skris CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 631159191Skris the CRYPTO_MDEBUG_xxx macro when compiling the library. 631259191Skris 631359191Skris Also, things like CRYPTO_set_mem_functions will always give the 631459191Skris expected result (the new set of functions is used for allocation 631559191Skris and deallocation) at all times, regardless of platform and compiler 631659191Skris options. 631759191Skris 631859191Skris To finish it up, some functions that were never use in any other 631959191Skris way than through macros have a new API and new semantic: 632059191Skris 632159191Skris CRYPTO_dbg_malloc() 632259191Skris CRYPTO_dbg_realloc() 632359191Skris CRYPTO_dbg_free() 632459191Skris 632559191Skris All macros of value have retained their old syntax. 632659191Skris [Richard Levitte and Bodo Moeller] 632759191Skris 632859191Skris *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 632959191Skris ordering of SMIMECapabilities wasn't in "strength order" and there 633059191Skris was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 633159191Skris algorithm. 633259191Skris [Steve Henson] 633359191Skris 633459191Skris *) Some ASN1 types with illegal zero length encoding (INTEGER, 633559191Skris ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 633659191Skris [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 633759191Skris 633859191Skris *) Merge in my S/MIME library for OpenSSL. This provides a simple 633959191Skris S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 634059191Skris functionality to handle multipart/signed properly) and a utility 634159191Skris called 'smime' to call all this stuff. This is based on code I 634259191Skris originally wrote for Celo who have kindly allowed it to be 634359191Skris included in OpenSSL. 634459191Skris [Steve Henson] 634559191Skris 634659191Skris *) Add variants des_set_key_checked and des_set_key_unchecked of 634759191Skris des_set_key (aka des_key_sched). Global variable des_check_key 634859191Skris decides which of these is called by des_set_key; this way 634959191Skris des_check_key behaves as it always did, but applications and 635059191Skris the library itself, which was buggy for des_check_key == 1, 635159191Skris have a cleaner way to pick the version they need. 635259191Skris [Bodo Moeller] 635359191Skris 635459191Skris *) New function PKCS12_newpass() which changes the password of a 635559191Skris PKCS12 structure. 635659191Skris [Steve Henson] 635759191Skris 635859191Skris *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 635959191Skris dynamic mix. In both cases the ids can be used as an index into the 636059191Skris table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 636159191Skris functions so they accept a list of the field values and the 636259191Skris application doesn't need to directly manipulate the X509_TRUST 636359191Skris structure. 636459191Skris [Steve Henson] 636559191Skris 636659191Skris *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 636759191Skris need initialising. 636859191Skris [Steve Henson] 636959191Skris 637059191Skris *) Modify the way the V3 extension code looks up extensions. This now 637159191Skris works in a similar way to the object code: we have some "standard" 637259191Skris extensions in a static table which is searched with OBJ_bsearch() 637359191Skris and the application can add dynamic ones if needed. The file 637459191Skris crypto/x509v3/ext_dat.h now has the info: this file needs to be 637559191Skris updated whenever a new extension is added to the core code and kept 637659191Skris in ext_nid order. There is a simple program 'tabtest.c' which checks 637759191Skris this. New extensions are not added too often so this file can readily 637859191Skris be maintained manually. 637959191Skris 638059191Skris There are two big advantages in doing things this way. The extensions 638159191Skris can be looked up immediately and no longer need to be "added" using 638259191Skris X509V3_add_standard_extensions(): this function now does nothing. 638359191Skris [Side note: I get *lots* of email saying the extension code doesn't 638459191Skris work because people forget to call this function] 638559191Skris Also no dynamic allocation is done unless new extensions are added: 638659191Skris so if we don't add custom extensions there is no need to call 638759191Skris X509V3_EXT_cleanup(). 638859191Skris [Steve Henson] 638959191Skris 639059191Skris *) Modify enc utility's salting as follows: make salting the default. Add a 639159191Skris magic header, so unsalted files fail gracefully instead of just decrypting 639259191Skris to garbage. This is because not salting is a big security hole, so people 639359191Skris should be discouraged from doing it. 639459191Skris [Ben Laurie] 639559191Skris 639659191Skris *) Fixes and enhancements to the 'x509' utility. It allowed a message 639759191Skris digest to be passed on the command line but it only used this 639859191Skris parameter when signing a certificate. Modified so all relevant 639959191Skris operations are affected by the digest parameter including the 640059191Skris -fingerprint and -x509toreq options. Also -x509toreq choked if a 640159191Skris DSA key was used because it didn't fix the digest. 640259191Skris [Steve Henson] 640359191Skris 640459191Skris *) Initial certificate chain verify code. Currently tests the untrusted 640559191Skris certificates for consistency with the verify purpose (which is set 640659191Skris when the X509_STORE_CTX structure is set up) and checks the pathlength. 640759191Skris 640859191Skris There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 640959191Skris this is because it will reject chains with invalid extensions whereas 641059191Skris every previous version of OpenSSL and SSLeay made no checks at all. 641159191Skris 641259191Skris Trust code: checks the root CA for the relevant trust settings. Trust 641359191Skris settings have an initial value consistent with the verify purpose: e.g. 641459191Skris if the verify purpose is for SSL client use it expects the CA to be 641559191Skris trusted for SSL client use. However the default value can be changed to 641659191Skris permit custom trust settings: one example of this would be to only trust 641759191Skris certificates from a specific "secure" set of CAs. 641859191Skris 641959191Skris Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 642059191Skris which should be used for version portability: especially since the 642159191Skris verify structure is likely to change more often now. 642259191Skris 642359191Skris SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 642459191Skris to set them. If not set then assume SSL clients will verify SSL servers 642559191Skris and vice versa. 642659191Skris 642759191Skris Two new options to the verify program: -untrusted allows a set of 642859191Skris untrusted certificates to be passed in and -purpose which sets the 642959191Skris intended purpose of the certificate. If a purpose is set then the 643059191Skris new chain verify code is used to check extension consistency. 643159191Skris [Steve Henson] 643259191Skris 643359191Skris *) Support for the authority information access extension. 643459191Skris [Steve Henson] 643559191Skris 643659191Skris *) Modify RSA and DSA PEM read routines to transparently handle 643759191Skris PKCS#8 format private keys. New *_PUBKEY_* functions that handle 643859191Skris public keys in a format compatible with certificate 643959191Skris SubjectPublicKeyInfo structures. Unfortunately there were already 644059191Skris functions called *_PublicKey_* which used various odd formats so 644159191Skris these are retained for compatibility: however the DSA variants were 644259191Skris never in a public release so they have been deleted. Changed dsa/rsa 644359191Skris utilities to handle the new format: note no releases ever handled public 644459191Skris keys so we should be OK. 644559191Skris 644659191Skris The primary motivation for this change is to avoid the same fiasco 644759191Skris that dogs private keys: there are several incompatible private key 644859191Skris formats some of which are standard and some OpenSSL specific and 644959191Skris require various evil hacks to allow partial transparent handling and 645059191Skris even then it doesn't work with DER formats. Given the option anything 645159191Skris other than PKCS#8 should be dumped: but the other formats have to 645259191Skris stay in the name of compatibility. 645359191Skris 645459191Skris With public keys and the benefit of hindsight one standard format 645559191Skris is used which works with EVP_PKEY, RSA or DSA structures: though 645659191Skris it clearly returns an error if you try to read the wrong kind of key. 645759191Skris 645859191Skris Added a -pubkey option to the 'x509' utility to output the public key. 645959191Skris Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 646059191Skris (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 646159191Skris EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 646259191Skris that do the same as the EVP_PKEY_assign_*() except they up the 646359191Skris reference count of the added key (they don't "swallow" the 646459191Skris supplied key). 646559191Skris [Steve Henson] 646659191Skris 646759191Skris *) Fixes to crypto/x509/by_file.c the code to read in certificates and 646859191Skris CRLs would fail if the file contained no certificates or no CRLs: 646959191Skris added a new function to read in both types and return the number 647059191Skris read: this means that if none are read it will be an error. The 647159191Skris DER versions of the certificate and CRL reader would always fail 647259191Skris because it isn't possible to mix certificates and CRLs in DER format 647359191Skris without choking one or the other routine. Changed this to just read 647459191Skris a certificate: this is the best we can do. Also modified the code 647559191Skris in apps/verify.c to take notice of return codes: it was previously 647659191Skris attempting to read in certificates from NULL pointers and ignoring 647759191Skris any errors: this is one reason why the cert and CRL reader seemed 647859191Skris to work. It doesn't check return codes from the default certificate 647959191Skris routines: these may well fail if the certificates aren't installed. 648059191Skris [Steve Henson] 648159191Skris 648259191Skris *) Code to support otherName option in GeneralName. 648359191Skris [Steve Henson] 648459191Skris 648559191Skris *) First update to verify code. Change the verify utility 648659191Skris so it warns if it is passed a self signed certificate: 648759191Skris for consistency with the normal behaviour. X509_verify 648859191Skris has been modified to it will now verify a self signed 648959191Skris certificate if *exactly* the same certificate appears 649059191Skris in the store: it was previously impossible to trust a 649159191Skris single self signed certificate. This means that: 649259191Skris openssl verify ss.pem 649359191Skris now gives a warning about a self signed certificate but 649459191Skris openssl verify -CAfile ss.pem ss.pem 649559191Skris is OK. 649659191Skris [Steve Henson] 649759191Skris 649859191Skris *) For servers, store verify_result in SSL_SESSION data structure 649959191Skris (and add it to external session representation). 650059191Skris This is needed when client certificate verifications fails, 650159191Skris but an application-provided verification callback (set by 650259191Skris SSL_CTX_set_cert_verify_callback) allows accepting the session 650359191Skris anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 650459191Skris but returns 1): When the session is reused, we have to set 650559191Skris ssl->verify_result to the appropriate error code to avoid 650659191Skris security holes. 650759191Skris [Bodo Moeller, problem pointed out by Lutz Jaenicke] 650859191Skris 650959191Skris *) Fix a bug in the new PKCS#7 code: it didn't consider the 651059191Skris case in PKCS7_dataInit() where the signed PKCS7 structure 651159191Skris didn't contain any existing data because it was being created. 651259191Skris [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 651359191Skris 651459191Skris *) Add a salt to the key derivation routines in enc.c. This 651559191Skris forms the first 8 bytes of the encrypted file. Also add a 651659191Skris -S option to allow a salt to be input on the command line. 651759191Skris [Steve Henson] 651859191Skris 651959191Skris *) New function X509_cmp(). Oddly enough there wasn't a function 652059191Skris to compare two certificates. We do this by working out the SHA1 652159191Skris hash and comparing that. X509_cmp() will be needed by the trust 652259191Skris code. 652359191Skris [Steve Henson] 652459191Skris 652559191Skris *) SSL_get1_session() is like SSL_get_session(), but increments 652659191Skris the reference count in the SSL_SESSION returned. 652759191Skris [Geoff Thorpe <geoff@eu.c2.net>] 652859191Skris 652959191Skris *) Fix for 'req': it was adding a null to request attributes. 653059191Skris Also change the X509_LOOKUP and X509_INFO code to handle 653159191Skris certificate auxiliary information. 653259191Skris [Steve Henson] 653359191Skris 653459191Skris *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 653559191Skris the 'enc' command. 653659191Skris [Steve Henson] 653759191Skris 653859191Skris *) Add the possibility to add extra information to the memory leak 653959191Skris detecting output, to form tracebacks, showing from where each 654059191Skris allocation was originated: CRYPTO_push_info("constant string") adds 654159191Skris the string plus current file name and line number to a per-thread 654259191Skris stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 654359191Skris is like calling CYRPTO_pop_info() until the stack is empty. 654459191Skris Also updated memory leak detection code to be multi-thread-safe. 654559191Skris [Richard Levitte] 654659191Skris 654759191Skris *) Add options -text and -noout to pkcs7 utility and delete the 654859191Skris encryption options which never did anything. Update docs. 654959191Skris [Steve Henson] 655059191Skris 655159191Skris *) Add options to some of the utilities to allow the pass phrase 655259191Skris to be included on either the command line (not recommended on 655359191Skris OSes like Unix) or read from the environment. Update the 655459191Skris manpages and fix a few bugs. 655559191Skris [Steve Henson] 655659191Skris 655759191Skris *) Add a few manpages for some of the openssl commands. 655859191Skris [Steve Henson] 655959191Skris 656059191Skris *) Fix the -revoke option in ca. It was freeing up memory twice, 656159191Skris leaking and not finding already revoked certificates. 656259191Skris [Steve Henson] 656359191Skris 656459191Skris *) Extensive changes to support certificate auxiliary information. 656559191Skris This involves the use of X509_CERT_AUX structure and X509_AUX 656659191Skris functions. An X509_AUX function such as PEM_read_X509_AUX() 656759191Skris can still read in a certificate file in the usual way but it 656859191Skris will also read in any additional "auxiliary information". By 656959191Skris doing things this way a fair degree of compatibility can be 657059191Skris retained: existing certificates can have this information added 657159191Skris using the new 'x509' options. 657259191Skris 657359191Skris Current auxiliary information includes an "alias" and some trust 657459191Skris settings. The trust settings will ultimately be used in enhanced 657559191Skris certificate chain verification routines: currently a certificate 657659191Skris can only be trusted if it is self signed and then it is trusted 657759191Skris for all purposes. 657859191Skris [Steve Henson] 657959191Skris 658059191Skris *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 658159191Skris The problem was that one of the replacement routines had not been working 658259191Skris since SSLeay releases. For now the offending routine has been replaced 658359191Skris with non-optimised assembler. Even so, this now gives around 95% 658459191Skris performance improvement for 1024 bit RSA signs. 658559191Skris [Mark Cox] 658659191Skris 658759191Skris *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 658859191Skris handling. Most clients have the effective key size in bits equal to 658959191Skris the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 659059191Skris A few however don't do this and instead use the size of the decrypted key 659159191Skris to determine the RC2 key length and the AlgorithmIdentifier to determine 659259191Skris the effective key length. In this case the effective key length can still 659359191Skris be 40 bits but the key length can be 168 bits for example. This is fixed 659459191Skris by manually forcing an RC2 key into the EVP_PKEY structure because the 659559191Skris EVP code can't currently handle unusual RC2 key sizes: it always assumes 659659191Skris the key length and effective key length are equal. 659759191Skris [Steve Henson] 659859191Skris 659959191Skris *) Add a bunch of functions that should simplify the creation of 660059191Skris X509_NAME structures. Now you should be able to do: 660159191Skris X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 660259191Skris and have it automatically work out the correct field type and fill in 660359191Skris the structures. The more adventurous can try: 660459191Skris X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 660559191Skris and it will (hopefully) work out the correct multibyte encoding. 660659191Skris [Steve Henson] 660759191Skris 660859191Skris *) Change the 'req' utility to use the new field handling and multibyte 660959191Skris copy routines. Before the DN field creation was handled in an ad hoc 661059191Skris way in req, ca, and x509 which was rather broken and didn't support 661159191Skris BMPStrings or UTF8Strings. Since some software doesn't implement 661259191Skris BMPStrings or UTF8Strings yet, they can be enabled using the config file 661359191Skris using the dirstring_type option. See the new comment in the default 661459191Skris openssl.cnf for more info. 661559191Skris [Steve Henson] 661659191Skris 661759191Skris *) Make crypto/rand/md_rand.c more robust: 661859191Skris - Assure unique random numbers after fork(). 661959191Skris - Make sure that concurrent threads access the global counter and 662059191Skris md serializably so that we never lose entropy in them 662159191Skris or use exactly the same state in multiple threads. 662259191Skris Access to the large state is not always serializable because 662359191Skris the additional locking could be a performance killer, and 662459191Skris md should be large enough anyway. 662559191Skris [Bodo Moeller] 662659191Skris 662759191Skris *) New file apps/app_rand.c with commonly needed functionality 662859191Skris for handling the random seed file. 662959191Skris 663059191Skris Use the random seed file in some applications that previously did not: 663159191Skris ca, 663259191Skris dsaparam -genkey (which also ignored its '-rand' option), 663359191Skris s_client, 663459191Skris s_server, 663559191Skris x509 (when signing). 663659191Skris Except on systems with /dev/urandom, it is crucial to have a random 663759191Skris seed file at least for key creation, DSA signing, and for DH exchanges; 663859191Skris for RSA signatures we could do without one. 663959191Skris 664059191Skris gendh and gendsa (unlike genrsa) used to read only the first byte 664159191Skris of each file listed in the '-rand' option. The function as previously 664259191Skris found in genrsa is now in app_rand.c and is used by all programs 664359191Skris that support '-rand'. 664459191Skris [Bodo Moeller] 664559191Skris 664659191Skris *) In RAND_write_file, use mode 0600 for creating files; 664759191Skris don't just chmod when it may be too late. 664859191Skris [Bodo Moeller] 664959191Skris 665059191Skris *) Report an error from X509_STORE_load_locations 665159191Skris when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 665259191Skris [Bill Perry] 665359191Skris 665459191Skris *) New function ASN1_mbstring_copy() this copies a string in either 665559191Skris ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 665659191Skris into an ASN1_STRING type. A mask of permissible types is passed 665759191Skris and it chooses the "minimal" type to use or an error if not type 665859191Skris is suitable. 665959191Skris [Steve Henson] 666059191Skris 666159191Skris *) Add function equivalents to the various macros in asn1.h. The old 666259191Skris macros are retained with an M_ prefix. Code inside the library can 666359191Skris use the M_ macros. External code (including the openssl utility) 666459191Skris should *NOT* in order to be "shared library friendly". 666559191Skris [Steve Henson] 666659191Skris 666759191Skris *) Add various functions that can check a certificate's extensions 666859191Skris to see if it usable for various purposes such as SSL client, 666959191Skris server or S/MIME and CAs of these types. This is currently 667059191Skris VERY EXPERIMENTAL but will ultimately be used for certificate chain 667159191Skris verification. Also added a -purpose flag to x509 utility to 667259191Skris print out all the purposes. 667359191Skris [Steve Henson] 667459191Skris 667559191Skris *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 667659191Skris functions. 667759191Skris [Steve Henson] 667859191Skris 667959191Skris *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 668059191Skris for, obtain and decode and extension and obtain its critical flag. 668159191Skris This allows all the necessary extension code to be handled in a 668259191Skris single function call. 668359191Skris [Steve Henson] 668459191Skris 668559191Skris *) RC4 tune-up featuring 30-40% performance improvement on most RISC 668659191Skris platforms. See crypto/rc4/rc4_enc.c for further details. 668759191Skris [Andy Polyakov] 668859191Skris 668959191Skris *) New -noout option to asn1parse. This causes no output to be produced 669059191Skris its main use is when combined with -strparse and -out to extract data 669159191Skris from a file (which may not be in ASN.1 format). 669259191Skris [Steve Henson] 669359191Skris 669459191Skris *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 669559191Skris when producing the local key id. 669659191Skris [Richard Levitte <levitte@stacken.kth.se>] 669759191Skris 669859191Skris *) New option -dhparam in s_server. This allows a DH parameter file to be 669959191Skris stated explicitly. If it is not stated then it tries the first server 670059191Skris certificate file. The previous behaviour hard coded the filename 670159191Skris "server.pem". 670259191Skris [Steve Henson] 670359191Skris 670459191Skris *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 670559191Skris a public key to be input or output. For example: 670659191Skris openssl rsa -in key.pem -pubout -out pubkey.pem 670759191Skris Also added necessary DSA public key functions to handle this. 670859191Skris [Steve Henson] 670959191Skris 671059191Skris *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 671159191Skris in the message. This was handled by allowing 671259191Skris X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 671359191Skris [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 671459191Skris 671559191Skris *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 671659191Skris to the end of the strings whereas this didn't. This would cause problems 671759191Skris if strings read with d2i_ASN1_bytes() were later modified. 671859191Skris [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 671959191Skris 672059191Skris *) Fix for base64 decode bug. When a base64 bio reads only one line of 672159191Skris data and it contains EOF it will end up returning an error. This is 672259191Skris caused by input 46 bytes long. The cause is due to the way base64 672359191Skris BIOs find the start of base64 encoded data. They do this by trying a 672459191Skris trial decode on each line until they find one that works. When they 672559191Skris do a flag is set and it starts again knowing it can pass all the 672659191Skris data directly through the decoder. Unfortunately it doesn't reset 672759191Skris the context it uses. This means that if EOF is reached an attempt 672859191Skris is made to pass two EOFs through the context and this causes the 672959191Skris resulting error. This can also cause other problems as well. As is 673059191Skris usual with these problems it takes *ages* to find and the fix is 673159191Skris trivial: move one line. 673259191Skris [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 673359191Skris 673459191Skris *) Ugly workaround to get s_client and s_server working under Windows. The 673559191Skris old code wouldn't work because it needed to select() on sockets and the 673659191Skris tty (for keypresses and to see if data could be written). Win32 only 673759191Skris supports select() on sockets so we select() with a 1s timeout on the 673859191Skris sockets and then see if any characters are waiting to be read, if none 673959191Skris are present then we retry, we also assume we can always write data to 674059191Skris the tty. This isn't nice because the code then blocks until we've 674159191Skris received a complete line of data and it is effectively polling the 674259191Skris keyboard at 1s intervals: however it's quite a bit better than not 674359191Skris working at all :-) A dedicated Windows application might handle this 674459191Skris with an event loop for example. 674559191Skris [Steve Henson] 674659191Skris 674759191Skris *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 674859191Skris and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 674959191Skris will be called when RSA_sign() and RSA_verify() are used. This is useful 675059191Skris if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 675159191Skris For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 675259191Skris should *not* be used: RSA_sign() and RSA_verify() must be used instead. 675359191Skris This necessitated the support of an extra signature type NID_md5_sha1 675459191Skris for SSL signatures and modifications to the SSL library to use it instead 675559191Skris of calling RSA_public_decrypt() and RSA_private_encrypt(). 675659191Skris [Steve Henson] 675759191Skris 675859191Skris *) Add new -verify -CAfile and -CApath options to the crl program, these 675959191Skris will lookup a CRL issuers certificate and verify the signature in a 676059191Skris similar way to the verify program. Tidy up the crl program so it 676159191Skris no longer accesses structures directly. Make the ASN1 CRL parsing a bit 676259191Skris less strict. It will now permit CRL extensions even if it is not 676359191Skris a V2 CRL: this will allow it to tolerate some broken CRLs. 676459191Skris [Steve Henson] 676559191Skris 676659191Skris *) Initialize all non-automatic variables each time one of the openssl 676759191Skris sub-programs is started (this is necessary as they may be started 676859191Skris multiple times from the "OpenSSL>" prompt). 676959191Skris [Lennart Bang, Bodo Moeller] 677059191Skris 677159191Skris *) Preliminary compilation option RSA_NULL which disables RSA crypto without 677259191Skris removing all other RSA functionality (this is what NO_RSA does). This 677359191Skris is so (for example) those in the US can disable those operations covered 677459191Skris by the RSA patent while allowing storage and parsing of RSA keys and RSA 677559191Skris key generation. 677659191Skris [Steve Henson] 677759191Skris 677859191Skris *) Non-copying interface to BIO pairs. 677959191Skris (still largely untested) 678059191Skris [Bodo Moeller] 678159191Skris 678259191Skris *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 678359191Skris ASCII string. This was handled independently in various places before. 678459191Skris [Steve Henson] 678559191Skris 678659191Skris *) New functions UTF8_getc() and UTF8_putc() that parse and generate 678759191Skris UTF8 strings a character at a time. 678859191Skris [Steve Henson] 678959191Skris 679059191Skris *) Use client_version from client hello to select the protocol 679159191Skris (s23_srvr.c) and for RSA client key exchange verification 679259191Skris (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 679359191Skris [Bodo Moeller] 679459191Skris 679559191Skris *) Add various utility functions to handle SPKACs, these were previously 679659191Skris handled by poking round in the structure internals. Added new function 679759191Skris NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 679859191Skris print, verify and generate SPKACs. Based on an original idea from 679959191Skris Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 680059191Skris [Steve Henson] 680159191Skris 680259191Skris *) RIPEMD160 is operational on all platforms and is back in 'make test'. 680359191Skris [Andy Polyakov] 680459191Skris 680559191Skris *) Allow the config file extension section to be overwritten on the 680659191Skris command line. Based on an original idea from Massimiliano Pala 680759191Skris <madwolf@comune.modena.it>. The new option is called -extensions 680859191Skris and can be applied to ca, req and x509. Also -reqexts to override 680959191Skris the request extensions in req and -crlexts to override the crl extensions 681059191Skris in ca. 681159191Skris [Steve Henson] 681259191Skris 681359191Skris *) Add new feature to the SPKAC handling in ca. Now you can include 681459191Skris the same field multiple times by preceding it by "XXXX." for example: 681559191Skris 1.OU="Unit name 1" 681659191Skris 2.OU="Unit name 2" 681759191Skris this is the same syntax as used in the req config file. 681859191Skris [Steve Henson] 681959191Skris 682059191Skris *) Allow certificate extensions to be added to certificate requests. These 682159191Skris are specified in a 'req_extensions' option of the req section of the 682259191Skris config file. They can be printed out with the -text option to req but 682359191Skris are otherwise ignored at present. 682459191Skris [Steve Henson] 682559191Skris 682659191Skris *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 682759191Skris data read consists of only the final block it would not decrypted because 682859191Skris EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 682959191Skris A misplaced 'break' also meant the decrypted final block might not be 683059191Skris copied until the next read. 683159191Skris [Steve Henson] 683259191Skris 683359191Skris *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 683459191Skris a few extra parameters to the DH structure: these will be useful if 683559191Skris for example we want the value of 'q' or implement X9.42 DH. 683659191Skris [Steve Henson] 683759191Skris 683859191Skris *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 683959191Skris provides hooks that allow the default DSA functions or functions on a 684059191Skris "per key" basis to be replaced. This allows hardware acceleration and 684159191Skris hardware key storage to be handled without major modification to the 684259191Skris library. Also added low level modexp hooks and CRYPTO_EX structure and 684359191Skris associated functions. 684459191Skris [Steve Henson] 684559191Skris 684659191Skris *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 684759191Skris as "read only": it can't be written to and the buffer it points to will 684859191Skris not be freed. Reading from a read only BIO is much more efficient than 684959191Skris a normal memory BIO. This was added because there are several times when 685059191Skris an area of memory needs to be read from a BIO. The previous method was 685159191Skris to create a memory BIO and write the data to it, this results in two 685259191Skris copies of the data and an O(n^2) reading algorithm. There is a new 685359191Skris function BIO_new_mem_buf() which creates a read only memory BIO from 685459191Skris an area of memory. Also modified the PKCS#7 routines to use read only 685559191Skris memory BIOs. 685659191Skris [Steve Henson] 685759191Skris 685859191Skris *) Bugfix: ssl23_get_client_hello did not work properly when called in 685959191Skris state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 686059191Skris a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 686159191Skris but a retry condition occured while trying to read the rest. 686259191Skris [Bodo Moeller] 686359191Skris 686459191Skris *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 686559191Skris NID_pkcs7_encrypted by default: this was wrong since this should almost 686659191Skris always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 686759191Skris the encrypted data type: this is a more sensible place to put it and it 686859191Skris allows the PKCS#12 code to be tidied up that duplicated this 686959191Skris functionality. 687059191Skris [Steve Henson] 687159191Skris 687259191Skris *) Changed obj_dat.pl script so it takes its input and output files on 687359191Skris the command line. This should avoid shell escape redirection problems 687459191Skris under Win32. 687559191Skris [Steve Henson] 687659191Skris 687759191Skris *) Initial support for certificate extension requests, these are included 687859191Skris in things like Xenroll certificate requests. Included functions to allow 687959191Skris extensions to be obtained and added. 688059191Skris [Steve Henson] 688159191Skris 688259191Skris *) -crlf option to s_client and s_server for sending newlines as 688359191Skris CRLF (as required by many protocols). 688459191Skris [Bodo Moeller] 688559191Skris 688655714Skris Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 688755714Skris 688855714Skris *) Install libRSAglue.a when OpenSSL is built with RSAref. 688955714Skris [Ralf S. Engelschall] 689055714Skris 689155714Skris *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 689255714Skris [Andrija Antonijevic <TheAntony2@bigfoot.com>] 689355714Skris 689455714Skris *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 689555714Skris program. 689655714Skris [Steve Henson] 689755714Skris 689855714Skris *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 689955714Skris DH parameters/keys (q is lost during that conversion, but the resulting 690055714Skris DH parameters contain its length). 690155714Skris 690255714Skris For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 690355714Skris much faster than DH_generate_parameters (which creates parameters 690455714Skris where p = 2*q + 1), and also the smaller q makes DH computations 690555714Skris much more efficient (160-bit exponentiation instead of 1024-bit 690655714Skris exponentiation); so this provides a convenient way to support DHE 690755714Skris ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 690855714Skris utter importance to use 690955714Skris SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 691055714Skris or 691155714Skris SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 691255714Skris when such DH parameters are used, because otherwise small subgroup 691355714Skris attacks may become possible! 691455714Skris [Bodo Moeller] 691555714Skris 691655714Skris *) Avoid memory leak in i2d_DHparams. 691755714Skris [Bodo Moeller] 691855714Skris 691955714Skris *) Allow the -k option to be used more than once in the enc program: 692055714Skris this allows the same encrypted message to be read by multiple recipients. 692155714Skris [Steve Henson] 692255714Skris 692355714Skris *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 692455714Skris an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 692555714Skris it will always use the numerical form of the OID, even if it has a short 692655714Skris or long name. 692755714Skris [Steve Henson] 692855714Skris 692955714Skris *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 693055714Skris method only got called if p,q,dmp1,dmq1,iqmp components were present, 693155714Skris otherwise bn_mod_exp was called. In the case of hardware keys for example 693255714Skris no private key components need be present and it might store extra data 693359191Skris in the RSA structure, which cannot be accessed from bn_mod_exp. 693459191Skris By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 693559191Skris private key operations. 693655714Skris [Steve Henson] 693755714Skris 693855714Skris *) Added support for SPARC Linux. 693955714Skris [Andy Polyakov] 694055714Skris 694155714Skris *) pem_password_cb function type incompatibly changed from 694255714Skris typedef int pem_password_cb(char *buf, int size, int rwflag); 694355714Skris to 694455714Skris ....(char *buf, int size, int rwflag, void *userdata); 694555714Skris so that applications can pass data to their callbacks: 694655714Skris The PEM[_ASN1]_{read,write}... functions and macros now take an 694755714Skris additional void * argument, which is just handed through whenever 694855714Skris the password callback is called. 694959191Skris [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 695055714Skris 695155714Skris New function SSL_CTX_set_default_passwd_cb_userdata. 695255714Skris 695355714Skris Compatibility note: As many C implementations push function arguments 695455714Skris onto the stack in reverse order, the new library version is likely to 695555714Skris interoperate with programs that have been compiled with the old 695655714Skris pem_password_cb definition (PEM_whatever takes some data that 695755714Skris happens to be on the stack as its last argument, and the callback 695855714Skris just ignores this garbage); but there is no guarantee whatsoever that 695955714Skris this will work. 696055714Skris 696155714Skris *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 696255714Skris (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 696355714Skris problems not only on Windows, but also on some Unix platforms. 696455714Skris To avoid problematic command lines, these definitions are now in an 696555714Skris auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 696655714Skris for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 696755714Skris [Bodo Moeller] 696855714Skris 696955714Skris *) MIPS III/IV assembler module is reimplemented. 697055714Skris [Andy Polyakov] 697155714Skris 697255714Skris *) More DES library cleanups: remove references to srand/rand and 697355714Skris delete an unused file. 697455714Skris [Ulf M�ller] 697555714Skris 697655714Skris *) Add support for the the free Netwide assembler (NASM) under Win32, 697755714Skris since not many people have MASM (ml) and it can be hard to obtain. 697855714Skris This is currently experimental but it seems to work OK and pass all 697955714Skris the tests. Check out INSTALL.W32 for info. 698055714Skris [Steve Henson] 698155714Skris 698255714Skris *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 698355714Skris without temporary keys kept an extra copy of the server key, 698455714Skris and connections with temporary keys did not free everything in case 698555714Skris of an error. 698655714Skris [Bodo Moeller] 698755714Skris 698855714Skris *) New function RSA_check_key and new openssl rsa option -check 698955714Skris for verifying the consistency of RSA keys. 699055714Skris [Ulf Moeller, Bodo Moeller] 699155714Skris 699255714Skris *) Various changes to make Win32 compile work: 699355714Skris 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 699455714Skris 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 699555714Skris comparison" warnings. 699655714Skris 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 699755714Skris [Steve Henson] 699855714Skris 699955714Skris *) Add a debugging option to PKCS#5 v2 key generation function: when 700055714Skris you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 700155714Skris derived keys are printed to stderr. 700255714Skris [Steve Henson] 700355714Skris 700455714Skris *) Copy the flags in ASN1_STRING_dup(). 700555714Skris [Roman E. Pavlov <pre@mo.msk.ru>] 700655714Skris 700755714Skris *) The x509 application mishandled signing requests containing DSA 700855714Skris keys when the signing key was also DSA and the parameters didn't match. 700955714Skris 701055714Skris It was supposed to omit the parameters when they matched the signing key: 701155714Skris the verifying software was then supposed to automatically use the CA's 701255714Skris parameters if they were absent from the end user certificate. 701355714Skris 701455714Skris Omitting parameters is no longer recommended. The test was also 701555714Skris the wrong way round! This was probably due to unusual behaviour in 701655714Skris EVP_cmp_parameters() which returns 1 if the parameters match. 701755714Skris This meant that parameters were omitted when they *didn't* match and 701855714Skris the certificate was useless. Certificates signed with 'ca' didn't have 701955714Skris this bug. 702055714Skris [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 702155714Skris 702255714Skris *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 702355714Skris The interface is as follows: 702455714Skris Applications can use 702555714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 702655714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 702755714Skris "off" is now the default. 702855714Skris The library internally uses 702955714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 703055714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 703155714Skris to disable memory-checking temporarily. 703255714Skris 703355714Skris Some inconsistent states that previously were possible (and were 703455714Skris even the default) are now avoided. 703555714Skris 703655714Skris -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 703755714Skris with each memory chunk allocated; this is occasionally more helpful 703855714Skris than just having a counter. 703955714Skris 704055714Skris -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 704155714Skris 704255714Skris -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 704355714Skris extensions. 704455714Skris [Bodo Moeller] 704555714Skris 704655714Skris *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 704755714Skris which largely parallels "options", but is for changing API behaviour, 704855714Skris whereas "options" are about protocol behaviour. 704955714Skris Initial "mode" flags are: 705055714Skris 705155714Skris SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 705255714Skris a single record has been written. 705355714Skris SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 705455714Skris retries use the same buffer location. 705555714Skris (But all of the contents must be 705655714Skris copied!) 705755714Skris [Bodo Moeller] 705855714Skris 705979998Skris *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 706055714Skris worked. 706155714Skris 706255714Skris *) Fix problems with no-hmac etc. 706355714Skris [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 706455714Skris 706555714Skris *) New functions RSA_get_default_method(), RSA_set_method() and 706655714Skris RSA_get_method(). These allows replacement of RSA_METHODs without having 706755714Skris to mess around with the internals of an RSA structure. 706855714Skris [Steve Henson] 706955714Skris 707055714Skris *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 707155714Skris Also really enable memory leak checks in openssl.c and in some 707255714Skris test programs. 707355714Skris [Chad C. Mulligan, Bodo Moeller] 707455714Skris 707555714Skris *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 707655714Skris up the length of negative integers. This has now been simplified to just 707755714Skris store the length when it is first determined and use it later, rather 707855714Skris than trying to keep track of where data is copied and updating it to 707955714Skris point to the end. 708055714Skris [Steve Henson, reported by Brien Wheeler 708155714Skris <bwheeler@authentica-security.com>] 708255714Skris 708355714Skris *) Add a new function PKCS7_signatureVerify. This allows the verification 708455714Skris of a PKCS#7 signature but with the signing certificate passed to the 708555714Skris function itself. This contrasts with PKCS7_dataVerify which assumes the 708655714Skris certificate is present in the PKCS#7 structure. This isn't always the 708755714Skris case: certificates can be omitted from a PKCS#7 structure and be 708855714Skris distributed by "out of band" means (such as a certificate database). 708955714Skris [Steve Henson] 709055714Skris 709155714Skris *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 709255714Skris function prototypes in pem.h, also change util/mkdef.pl to add the 709355714Skris necessary function names. 709455714Skris [Steve Henson] 709555714Skris 709655714Skris *) mk1mf.pl (used by Windows builds) did not properly read the 709755714Skris options set by Configure in the top level Makefile, and Configure 709855714Skris was not even able to write more than one option correctly. 709955714Skris Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 710055714Skris [Bodo Moeller] 710155714Skris 710255714Skris *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 710355714Skris file to be loaded from a BIO or FILE pointer. The BIO version will 710455714Skris for example allow memory BIOs to contain config info. 710555714Skris [Steve Henson] 710655714Skris 710755714Skris *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 710855714Skris Whoever hopes to achieve shared-library compatibility across versions 710955714Skris must use this, not the compile-time macro. 711055714Skris (Exercise 0.9.4: Which is the minimum library version required by 711155714Skris such programs?) 711255714Skris Note: All this applies only to multi-threaded programs, others don't 711355714Skris need locks. 711455714Skris [Bodo Moeller] 711555714Skris 711655714Skris *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 711755714Skris through a BIO pair triggered the default case, i.e. 711855714Skris SSLerr(...,SSL_R_UNKNOWN_STATE). 711955714Skris [Bodo Moeller] 712055714Skris 712155714Skris *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 712255714Skris can use the SSL library even if none of the specific BIOs is 712355714Skris appropriate. 712455714Skris [Bodo Moeller] 712555714Skris 712655714Skris *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 712755714Skris for the encoded length. 712855714Skris [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 712955714Skris 713055714Skris *) Add initial documentation of the X509V3 functions. 713155714Skris [Steve Henson] 713255714Skris 713355714Skris *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 713455714Skris PEM_write_bio_PKCS8PrivateKey() that are equivalent to 713555714Skris PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 713655714Skris secure PKCS#8 private key format with a high iteration count. 713755714Skris [Steve Henson] 713855714Skris 713955714Skris *) Fix determination of Perl interpreter: A perl or perl5 714055714Skris _directory_ in $PATH was also accepted as the interpreter. 714155714Skris [Ralf S. Engelschall] 714255714Skris 714355714Skris *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 714455714Skris wrong with it but it was very old and did things like calling 714555714Skris PEM_ASN1_read() directly and used MD5 for the hash not to mention some 714655714Skris unusual formatting. 714755714Skris [Steve Henson] 714855714Skris 714955714Skris *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 715055714Skris to use the new extension code. 715155714Skris [Steve Henson] 715255714Skris 715355714Skris *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 715455714Skris with macros. This should make it easier to change their form, add extra 715555714Skris arguments etc. Fix a few PEM prototypes which didn't have cipher as a 715655714Skris constant. 715755714Skris [Steve Henson] 715855714Skris 715955714Skris *) Add to configuration table a new entry that can specify an alternative 716055714Skris name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 716155714Skris according to Mark Crispin <MRC@Panda.COM>. 716255714Skris [Bodo Moeller] 716355714Skris 716455714Skris#if 0 716555714Skris *) DES CBC did not update the IV. Weird. 716655714Skris [Ben Laurie] 716755714Skris#else 716855714Skris des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 716955714Skris Changing the behaviour of the former might break existing programs -- 717055714Skris where IV updating is needed, des_ncbc_encrypt can be used. 717155714Skris#endif 717255714Skris 717355714Skris *) When bntest is run from "make test" it drives bc to check its 717455714Skris calculations, as well as internally checking them. If an internal check 717555714Skris fails, it needs to cause bc to give a non-zero result or make test carries 717655714Skris on without noticing the failure. Fixed. 717755714Skris [Ben Laurie] 717855714Skris 717955714Skris *) DES library cleanups. 718055714Skris [Ulf M�ller] 718155714Skris 718255714Skris *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 718355714Skris used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 718455714Skris ciphers. NOTE: although the key derivation function has been verified 718555714Skris against some published test vectors it has not been extensively tested 718655714Skris yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 718755714Skris of v2.0. 718855714Skris [Steve Henson] 718955714Skris 719055714Skris *) Instead of "mkdir -p", which is not fully portable, use new 719155714Skris Perl script "util/mkdir-p.pl". 719255714Skris [Bodo Moeller] 719355714Skris 719455714Skris *) Rewrite the way password based encryption (PBE) is handled. It used to 719555714Skris assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 719655714Skris structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 719755714Skris but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 719855714Skris the 'parameter' field of the AlgorithmIdentifier is passed to the 719955714Skris underlying key generation function so it must do its own ASN1 parsing. 720055714Skris This has also changed the EVP_PBE_CipherInit() function which now has a 720155714Skris 'parameter' argument instead of literal salt and iteration count values 720255714Skris and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 720355714Skris [Steve Henson] 720455714Skris 720555714Skris *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 720655714Skris and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 720755714Skris Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 720855714Skris KEY" because this clashed with PKCS#8 unencrypted string. Since this 720955714Skris value was just used as a "magic string" and not used directly its 721055714Skris value doesn't matter. 721155714Skris [Steve Henson] 721255714Skris 721355714Skris *) Introduce some semblance of const correctness to BN. Shame C doesn't 721455714Skris support mutable. 721555714Skris [Ben Laurie] 721655714Skris 721755714Skris *) "linux-sparc64" configuration (ultrapenguin). 721855714Skris [Ray Miller <ray.miller@oucs.ox.ac.uk>] 721955714Skris "linux-sparc" configuration. 722055714Skris [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 722155714Skris 722255714Skris *) config now generates no-xxx options for missing ciphers. 722355714Skris [Ulf M�ller] 722455714Skris 722555714Skris *) Support the EBCDIC character set (work in progress). 722655714Skris File ebcdic.c not yet included because it has a different license. 722755714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 722855714Skris 722955714Skris *) Support BS2000/OSD-POSIX. 723055714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 723155714Skris 723255714Skris *) Make callbacks for key generation use void * instead of char *. 723355714Skris [Ben Laurie] 723455714Skris 723555714Skris *) Make S/MIME samples compile (not yet tested). 723655714Skris [Ben Laurie] 723755714Skris 723855714Skris *) Additional typesafe stacks. 723955714Skris [Ben Laurie] 724055714Skris 724155714Skris *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 724255714Skris [Bodo Moeller] 724355714Skris 724455714Skris 724555714Skris Changes between 0.9.3 and 0.9.3a [29 May 1999] 724655714Skris 724755714Skris *) New configuration variant "sco5-gcc". 724855714Skris 724955714Skris *) Updated some demos. 725055714Skris [Sean O Riordain, Wade Scholine] 725155714Skris 725255714Skris *) Add missing BIO_free at exit of pkcs12 application. 725355714Skris [Wu Zhigang] 725455714Skris 725555714Skris *) Fix memory leak in conf.c. 725655714Skris [Steve Henson] 725755714Skris 725855714Skris *) Updates for Win32 to assembler version of MD5. 725955714Skris [Steve Henson] 726055714Skris 726155714Skris *) Set #! path to perl in apps/der_chop to where we found it 726255714Skris instead of using a fixed path. 726355714Skris [Bodo Moeller] 726455714Skris 726555714Skris *) SHA library changes for irix64-mips4-cc. 726655714Skris [Andy Polyakov] 726755714Skris 726855714Skris *) Improvements for VMS support. 726955714Skris [Richard Levitte] 727055714Skris 727155714Skris 727255714Skris Changes between 0.9.2b and 0.9.3 [24 May 1999] 727355714Skris 727455714Skris *) Bignum library bug fix. IRIX 6 passes "make test" now! 727555714Skris This also avoids the problems with SC4.2 and unpatched SC5. 727655714Skris [Andy Polyakov <appro@fy.chalmers.se>] 727755714Skris 727855714Skris *) New functions sk_num, sk_value and sk_set to replace the previous macros. 727955714Skris These are required because of the typesafe stack would otherwise break 728055714Skris existing code. If old code used a structure member which used to be STACK 728155714Skris and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 728255714Skris sk_num or sk_value it would produce an error because the num, data members 728355714Skris are not present in STACK_OF. Now it just produces a warning. sk_set 728455714Skris replaces the old method of assigning a value to sk_value 728555714Skris (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 728655714Skris that does this will no longer work (and should use sk_set instead) but 728755714Skris this could be regarded as a "questionable" behaviour anyway. 728855714Skris [Steve Henson] 728955714Skris 729055714Skris *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 729155714Skris correctly handle encrypted S/MIME data. 729255714Skris [Steve Henson] 729355714Skris 729455714Skris *) Change type of various DES function arguments from des_cblock 729555714Skris (which means, in function argument declarations, pointer to char) 729655714Skris to des_cblock * (meaning pointer to array with 8 char elements), 729755714Skris which allows the compiler to do more typechecking; it was like 729855714Skris that back in SSLeay, but with lots of ugly casts. 729955714Skris 730055714Skris Introduce new type const_des_cblock. 730155714Skris [Bodo Moeller] 730255714Skris 730355714Skris *) Reorganise the PKCS#7 library and get rid of some of the more obvious 730455714Skris problems: find RecipientInfo structure that matches recipient certificate 730555714Skris and initialise the ASN1 structures properly based on passed cipher. 730655714Skris [Steve Henson] 730755714Skris 730855714Skris *) Belatedly make the BN tests actually check the results. 730955714Skris [Ben Laurie] 731055714Skris 731155714Skris *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 731255714Skris to and from BNs: it was completely broken. New compilation option 731355714Skris NEG_PUBKEY_BUG to allow for some broken certificates that encode public 731455714Skris key elements as negative integers. 731555714Skris [Steve Henson] 731655714Skris 731755714Skris *) Reorganize and speed up MD5. 731855714Skris [Andy Polyakov <appro@fy.chalmers.se>] 731955714Skris 732055714Skris *) VMS support. 732155714Skris [Richard Levitte <richard@levitte.org>] 732255714Skris 732355714Skris *) New option -out to asn1parse to allow the parsed structure to be 732455714Skris output to a file. This is most useful when combined with the -strparse 732555714Skris option to examine the output of things like OCTET STRINGS. 732655714Skris [Steve Henson] 732755714Skris 732855714Skris *) Make SSL library a little more fool-proof by not requiring any longer 732955714Skris that SSL_set_{accept,connect}_state be called before 733055714Skris SSL_{accept,connect} may be used (SSL_set_..._state is omitted 733155714Skris in many applications because usually everything *appeared* to work as 733255714Skris intended anyway -- now it really works as intended). 733355714Skris [Bodo Moeller] 733455714Skris 733555714Skris *) Move openssl.cnf out of lib/. 733655714Skris [Ulf M�ller] 733755714Skris 733855714Skris *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 733955714Skris -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 734055714Skris -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 734155714Skris [Ralf S. Engelschall] 734255714Skris 734355714Skris *) Various fixes to the EVP and PKCS#7 code. It may now be able to 734455714Skris handle PKCS#7 enveloped data properly. 734555714Skris [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 734655714Skris 734755714Skris *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 734855714Skris copying pointers. The cert_st handling is changed by this in 734955714Skris various ways (and thus what used to be known as ctx->default_cert 735055714Skris is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 735155714Skris any longer when s->cert does not give us what we need). 735255714Skris ssl_cert_instantiate becomes obsolete by this change. 735355714Skris As soon as we've got the new code right (possibly it already is?), 735455714Skris we have solved a couple of bugs of the earlier code where s->cert 735555714Skris was used as if it could not have been shared with other SSL structures. 735655714Skris 735755714Skris Note that using the SSL API in certain dirty ways now will result 735855714Skris in different behaviour than observed with earlier library versions: 735955714Skris Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 736055714Skris does not influence s as it used to. 736155714Skris 736255714Skris In order to clean up things more thoroughly, inside SSL_SESSION 736355714Skris we don't use CERT any longer, but a new structure SESS_CERT 736455714Skris that holds per-session data (if available); currently, this is 736555714Skris the peer's certificate chain and, for clients, the server's certificate 736655714Skris and temporary key. CERT holds only those values that can have 736755714Skris meaningful defaults in an SSL_CTX. 736855714Skris [Bodo Moeller] 736955714Skris 737055714Skris *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 737155714Skris from the internal representation. Various PKCS#7 fixes: remove some 737255714Skris evil casts and set the enc_dig_alg field properly based on the signing 737355714Skris key type. 737455714Skris [Steve Henson] 737555714Skris 737655714Skris *) Allow PKCS#12 password to be set from the command line or the 737755714Skris environment. Let 'ca' get its config file name from the environment 737855714Skris variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 737955714Skris and 'x509'). 738055714Skris [Steve Henson] 738155714Skris 738255714Skris *) Allow certificate policies extension to use an IA5STRING for the 738355714Skris organization field. This is contrary to the PKIX definition but 738455714Skris VeriSign uses it and IE5 only recognises this form. Document 'x509' 738555714Skris extension option. 738655714Skris [Steve Henson] 738755714Skris 738855714Skris *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 738955714Skris without disallowing inline assembler and the like for non-pedantic builds. 739055714Skris [Ben Laurie] 739155714Skris 739255714Skris *) Support Borland C++ builder. 739355714Skris [Janez Jere <jj@void.si>, modified by Ulf M�ller] 739455714Skris 739555714Skris *) Support Mingw32. 739655714Skris [Ulf M�ller] 739755714Skris 739855714Skris *) SHA-1 cleanups and performance enhancements. 739955714Skris [Andy Polyakov <appro@fy.chalmers.se>] 740055714Skris 740155714Skris *) Sparc v8plus assembler for the bignum library. 740255714Skris [Andy Polyakov <appro@fy.chalmers.se>] 740355714Skris 740455714Skris *) Accept any -xxx and +xxx compiler options in Configure. 740555714Skris [Ulf M�ller] 740655714Skris 740755714Skris *) Update HPUX configuration. 740855714Skris [Anonymous] 740955714Skris 741055714Skris *) Add missing sk_<type>_unshift() function to safestack.h 741155714Skris [Ralf S. Engelschall] 741255714Skris 741355714Skris *) New function SSL_CTX_use_certificate_chain_file that sets the 741455714Skris "extra_cert"s in addition to the certificate. (This makes sense 741555714Skris only for "PEM" format files, as chains as a whole are not 741655714Skris DER-encoded.) 741755714Skris [Bodo Moeller] 741855714Skris 741955714Skris *) Support verify_depth from the SSL API. 742055714Skris x509_vfy.c had what can be considered an off-by-one-error: 742155714Skris Its depth (which was not part of the external interface) 742255714Skris was actually counting the number of certificates in a chain; 742355714Skris now it really counts the depth. 742455714Skris [Bodo Moeller] 742555714Skris 742655714Skris *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 742755714Skris instead of X509err, which often resulted in confusing error 742855714Skris messages since the error codes are not globally unique 742955714Skris (e.g. an alleged error in ssl3_accept when a certificate 743055714Skris didn't match the private key). 743155714Skris 743255714Skris *) New function SSL_CTX_set_session_id_context that allows to set a default 743355714Skris value (so that you don't need SSL_set_session_id_context for each 743455714Skris connection using the SSL_CTX). 743555714Skris [Bodo Moeller] 743655714Skris 743755714Skris *) OAEP decoding bug fix. 743855714Skris [Ulf M�ller] 743955714Skris 744055714Skris *) Support INSTALL_PREFIX for package builders, as proposed by 744155714Skris David Harris. 744255714Skris [Bodo Moeller] 744355714Skris 744455714Skris *) New Configure options "threads" and "no-threads". For systems 744555714Skris where the proper compiler options are known (currently Solaris 744655714Skris and Linux), "threads" is the default. 744755714Skris [Bodo Moeller] 744855714Skris 744955714Skris *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 745055714Skris [Bodo Moeller] 745155714Skris 745255714Skris *) Install various scripts to $(OPENSSLDIR)/misc, not to 745355714Skris $(INSTALLTOP)/bin -- they shouldn't clutter directories 745455714Skris such as /usr/local/bin. 745555714Skris [Bodo Moeller] 745655714Skris 745755714Skris *) "make linux-shared" to build shared libraries. 745855714Skris [Niels Poppe <niels@netbox.org>] 745955714Skris 746055714Skris *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 746155714Skris [Ulf M�ller] 746255714Skris 746355714Skris *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 746455714Skris extension adding in x509 utility. 746555714Skris [Steve Henson] 746655714Skris 746755714Skris *) Remove NOPROTO sections and error code comments. 746855714Skris [Ulf M�ller] 746955714Skris 747055714Skris *) Partial rewrite of the DEF file generator to now parse the ANSI 747155714Skris prototypes. 747255714Skris [Steve Henson] 747355714Skris 747455714Skris *) New Configure options --prefix=DIR and --openssldir=DIR. 747555714Skris [Ulf M�ller] 747655714Skris 747755714Skris *) Complete rewrite of the error code script(s). It is all now handled 747855714Skris by one script at the top level which handles error code gathering, 747955714Skris header rewriting and C source file generation. It should be much better 748055714Skris than the old method: it now uses a modified version of Ulf's parser to 748155714Skris read the ANSI prototypes in all header files (thus the old K&R definitions 748255714Skris aren't needed for error creation any more) and do a better job of 748355714Skris translating function codes into names. The old 'ASN1 error code imbedded 748455714Skris in a comment' is no longer necessary and it doesn't use .err files which 748555714Skris have now been deleted. Also the error code call doesn't have to appear all 748655714Skris on one line (which resulted in some large lines...). 748755714Skris [Steve Henson] 748855714Skris 748955714Skris *) Change #include filenames from <foo.h> to <openssl/foo.h>. 749055714Skris [Bodo Moeller] 749155714Skris 749255714Skris *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 749355714Skris 0 (which usually indicates a closed connection), but continue reading. 749455714Skris [Bodo Moeller] 749555714Skris 749655714Skris *) Fix some race conditions. 749755714Skris [Bodo Moeller] 749855714Skris 749955714Skris *) Add support for CRL distribution points extension. Add Certificate 750055714Skris Policies and CRL distribution points documentation. 750155714Skris [Steve Henson] 750255714Skris 750355714Skris *) Move the autogenerated header file parts to crypto/opensslconf.h. 750455714Skris [Ulf M�ller] 750555714Skris 750655714Skris *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 750755714Skris 8 of keying material. Merlin has also confirmed interop with this fix 750855714Skris between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 750955714Skris [Merlin Hughes <merlin@baltimore.ie>] 751055714Skris 751155714Skris *) Fix lots of warnings. 751255714Skris [Richard Levitte <levitte@stacken.kth.se>] 751355714Skris 751455714Skris *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 751555714Skris the directory spec didn't end with a LIST_SEPARATOR_CHAR. 751655714Skris [Richard Levitte <levitte@stacken.kth.se>] 751755714Skris 751855714Skris *) Fix problems with sizeof(long) == 8. 751955714Skris [Andy Polyakov <appro@fy.chalmers.se>] 752055714Skris 752155714Skris *) Change functions to ANSI C. 752255714Skris [Ulf M�ller] 752355714Skris 752455714Skris *) Fix typos in error codes. 752555714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 752655714Skris 752755714Skris *) Remove defunct assembler files from Configure. 752855714Skris [Ulf M�ller] 752955714Skris 753055714Skris *) SPARC v8 assembler BIGNUM implementation. 753155714Skris [Andy Polyakov <appro@fy.chalmers.se>] 753255714Skris 753355714Skris *) Support for Certificate Policies extension: both print and set. 753455714Skris Various additions to support the r2i method this uses. 753555714Skris [Steve Henson] 753655714Skris 753755714Skris *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 753855714Skris return a const string when you are expecting an allocated buffer. 753955714Skris [Ben Laurie] 754055714Skris 754155714Skris *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 754255714Skris types DirectoryString and DisplayText. 754355714Skris [Steve Henson] 754455714Skris 754555714Skris *) Add code to allow r2i extensions to access the configuration database, 754655714Skris add an LHASH database driver and add several ctx helper functions. 754755714Skris [Steve Henson] 754855714Skris 754955714Skris *) Fix an evil bug in bn_expand2() which caused various BN functions to 755055714Skris fail when they extended the size of a BIGNUM. 755155714Skris [Steve Henson] 755255714Skris 755355714Skris *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 755455714Skris support typesafe stack. 755555714Skris [Steve Henson] 755655714Skris 755755714Skris *) Fix typo in SSL_[gs]et_options(). 755855714Skris [Nils Frostberg <nils@medcom.se>] 755955714Skris 756055714Skris *) Delete various functions and files that belonged to the (now obsolete) 756155714Skris old X509V3 handling code. 756255714Skris [Steve Henson] 756355714Skris 756455714Skris *) New Configure option "rsaref". 756555714Skris [Ulf M�ller] 756655714Skris 756755714Skris *) Don't auto-generate pem.h. 756855714Skris [Bodo Moeller] 756955714Skris 757055714Skris *) Introduce type-safe ASN.1 SETs. 757155714Skris [Ben Laurie] 757255714Skris 757355714Skris *) Convert various additional casted stacks to type-safe STACK_OF() variants. 757455714Skris [Ben Laurie, Ralf S. Engelschall, Steve Henson] 757555714Skris 757655714Skris *) Introduce type-safe STACKs. This will almost certainly break lots of code 757755714Skris that links with OpenSSL (well at least cause lots of warnings), but fear 757855714Skris not: the conversion is trivial, and it eliminates loads of evil casts. A 757955714Skris few STACKed things have been converted already. Feel free to convert more. 758055714Skris In the fullness of time, I'll do away with the STACK type altogether. 758155714Skris [Ben Laurie] 758255714Skris 758355714Skris *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 758455714Skris specified in <certfile> by updating the entry in the index.txt file. 758555714Skris This way one no longer has to edit the index.txt file manually for 758655714Skris revoking a certificate. The -revoke option does the gory details now. 758755714Skris [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 758855714Skris 758955714Skris *) Fix `openssl crl -noout -text' combination where `-noout' killed the 759055714Skris `-text' option at all and this way the `-noout -text' combination was 759155714Skris inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 759255714Skris [Ralf S. Engelschall] 759355714Skris 759455714Skris *) Make sure a corresponding plain text error message exists for the 759555714Skris X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 759655714Skris verify callback function determined that a certificate was revoked. 759755714Skris [Ralf S. Engelschall] 759855714Skris 759955714Skris *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 760055714Skris ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 760155714Skris all available cipers including rc5, which was forgotten until now. 760255714Skris In order to let the testing shell script know which algorithms 760355714Skris are available, a new (up to now undocumented) command 760455714Skris "openssl list-cipher-commands" is used. 760555714Skris [Bodo Moeller] 760655714Skris 760755714Skris *) Bugfix: s_client occasionally would sleep in select() when 760855714Skris it should have checked SSL_pending() first. 760955714Skris [Bodo Moeller] 761055714Skris 761155714Skris *) New functions DSA_do_sign and DSA_do_verify to provide access to 761255714Skris the raw DSA values prior to ASN.1 encoding. 761355714Skris [Ulf M�ller] 761455714Skris 761555714Skris *) Tweaks to Configure 761655714Skris [Niels Poppe <niels@netbox.org>] 761755714Skris 761855714Skris *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 761955714Skris yet... 762055714Skris [Steve Henson] 762155714Skris 762255714Skris *) New variables $(RANLIB) and $(PERL) in the Makefiles. 762355714Skris [Ulf M�ller] 762455714Skris 762555714Skris *) New config option to avoid instructions that are illegal on the 80386. 762655714Skris The default code is faster, but requires at least a 486. 762755714Skris [Ulf M�ller] 762855714Skris 762955714Skris *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 763055714Skris SSL2_SERVER_VERSION (not used at all) macros, which are now the 763155714Skris same as SSL2_VERSION anyway. 763255714Skris [Bodo Moeller] 763355714Skris 763455714Skris *) New "-showcerts" option for s_client. 763555714Skris [Bodo Moeller] 763655714Skris 763755714Skris *) Still more PKCS#12 integration. Add pkcs12 application to openssl 763855714Skris application. Various cleanups and fixes. 763955714Skris [Steve Henson] 764055714Skris 764155714Skris *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 764255714Skris modify error routines to work internally. Add error codes and PBE init 764355714Skris to library startup routines. 764455714Skris [Steve Henson] 764555714Skris 764655714Skris *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 764755714Skris packing functions to asn1 and evp. Changed function names and error 764855714Skris codes along the way. 764955714Skris [Steve Henson] 765055714Skris 765155714Skris *) PKCS12 integration: and so it begins... First of several patches to 765255714Skris slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 765355714Skris objects to objects.h 765455714Skris [Steve Henson] 765555714Skris 765655714Skris *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 765755714Skris and display support for Thawte strong extranet extension. 765855714Skris [Steve Henson] 765955714Skris 766055714Skris *) Add LinuxPPC support. 766155714Skris [Jeff Dubrule <igor@pobox.org>] 766255714Skris 766355714Skris *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 766455714Skris bn_div_words in alpha.s. 766555714Skris [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 766655714Skris 766755714Skris *) Make sure the RSA OAEP test is skipped under -DRSAref because 766855714Skris OAEP isn't supported when OpenSSL is built with RSAref. 766955714Skris [Ulf Moeller <ulf@fitug.de>] 767055714Skris 767155714Skris *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 767255714Skris so they no longer are missing under -DNOPROTO. 767355714Skris [Soren S. Jorvang <soren@t.dk>] 767455714Skris 767555714Skris 767655714Skris Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 767755714Skris 767855714Skris *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 767955714Skris doesn't work when the session is reused. Coming soon! 768055714Skris [Ben Laurie] 768155714Skris 768255714Skris *) Fix a security hole, that allows sessions to be reused in the wrong 768355714Skris context thus bypassing client cert protection! All software that uses 768455714Skris client certs and session caches in multiple contexts NEEDS PATCHING to 768555714Skris allow session reuse! A fuller solution is in the works. 768655714Skris [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 768755714Skris 768855714Skris *) Some more source tree cleanups (removed obsolete files 768955714Skris crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 769055714Skris permission on "config" script to be executable) and a fix for the INSTALL 769155714Skris document. 769255714Skris [Ulf Moeller <ulf@fitug.de>] 769355714Skris 769455714Skris *) Remove some legacy and erroneous uses of malloc, free instead of 769555714Skris Malloc, Free. 769655714Skris [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 769755714Skris 769855714Skris *) Make rsa_oaep_test return non-zero on error. 769955714Skris [Ulf Moeller <ulf@fitug.de>] 770055714Skris 770155714Skris *) Add support for native Solaris shared libraries. Configure 770255714Skris solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 770355714Skris if someone would make that last step automatic. 770455714Skris [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 770555714Skris 770655714Skris *) ctx_size was not built with the right compiler during "make links". Fixed. 770755714Skris [Ben Laurie] 770855714Skris 770955714Skris *) Change the meaning of 'ALL' in the cipher list. It now means "everything 771055714Skris except NULL ciphers". This means the default cipher list will no longer 771155714Skris enable NULL ciphers. They need to be specifically enabled e.g. with 771255714Skris the string "DEFAULT:eNULL". 771355714Skris [Steve Henson] 771455714Skris 771555714Skris *) Fix to RSA private encryption routines: if p < q then it would 771655714Skris occasionally produce an invalid result. This will only happen with 771755714Skris externally generated keys because OpenSSL (and SSLeay) ensure p > q. 771855714Skris [Steve Henson] 771955714Skris 772055714Skris *) Be less restrictive and allow also `perl util/perlpath.pl 772155714Skris /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 772255714Skris because this way one can also use an interpreter named `perl5' (which is 772355714Skris usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 772455714Skris installed as `perl'). 772555714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 772655714Skris 772755714Skris *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 772855714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 772955714Skris 773055714Skris *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 773155714Skris advapi32.lib to Win32 build and change the pem test comparision 773255714Skris to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 773355714Skris suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 773455714Skris and crypto/des/ede_cbcm_enc.c. 773555714Skris [Steve Henson] 773655714Skris 773755714Skris *) DES quad checksum was broken on big-endian architectures. Fixed. 773855714Skris [Ben Laurie] 773955714Skris 774055714Skris *) Comment out two functions in bio.h that aren't implemented. Fix up the 774155714Skris Win32 test batch file so it (might) work again. The Win32 test batch file 774255714Skris is horrible: I feel ill.... 774355714Skris [Steve Henson] 774455714Skris 774555714Skris *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 774655714Skris in e_os.h. Audit of header files to check ANSI and non ANSI 774755714Skris sections: 10 functions were absent from non ANSI section and not exported 774855714Skris from Windows DLLs. Fixed up libeay.num for new functions. 774955714Skris [Steve Henson] 775055714Skris 775155714Skris *) Make `openssl version' output lines consistent. 775255714Skris [Ralf S. Engelschall] 775355714Skris 775455714Skris *) Fix Win32 symbol export lists for BIO functions: Added 775555714Skris BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 775655714Skris to ms/libeay{16,32}.def. 775755714Skris [Ralf S. Engelschall] 775855714Skris 775955714Skris *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 776055714Skris fine under Unix and passes some trivial tests I've now added. But the 776155714Skris whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 776255714Skris added to make sure no one expects that this stuff really works in the 776355714Skris OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 776455714Skris up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 776555714Skris openssl_bio.xs. 776655714Skris [Ralf S. Engelschall] 776755714Skris 776855714Skris *) Fix the generation of two part addresses in perl. 776955714Skris [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 777055714Skris 777155714Skris *) Add config entry for Linux on MIPS. 777255714Skris [John Tobey <jtobey@channel1.com>] 777355714Skris 777455714Skris *) Make links whenever Configure is run, unless we are on Windoze. 777555714Skris [Ben Laurie] 777655714Skris 777755714Skris *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 777855714Skris Currently only issuerAltName and AuthorityKeyIdentifier make any sense 777955714Skris in CRLs. 778055714Skris [Steve Henson] 778155714Skris 778255714Skris *) Add a useful kludge to allow package maintainers to specify compiler and 778355714Skris other platforms details on the command line without having to patch the 778455714Skris Configure script everytime: One now can use ``perl Configure 778555714Skris <id>:<details>'', i.e. platform ids are allowed to have details appended 778655714Skris to them (seperated by colons). This is treated as there would be a static 778755714Skris pre-configured entry in Configure's %table under key <id> with value 778855714Skris <details> and ``perl Configure <id>'' is called. So, when you want to 778955714Skris perform a quick test-compile under FreeBSD 3.1 with pgcc and without 779055714Skris assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 779155714Skris now, which overrides the FreeBSD-elf entry on-the-fly. 779255714Skris [Ralf S. Engelschall] 779355714Skris 779455714Skris *) Disable new TLS1 ciphersuites by default: they aren't official yet. 779555714Skris [Ben Laurie] 779655714Skris 779755714Skris *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 779855714Skris on the `perl Configure ...' command line. This way one can compile 779955714Skris OpenSSL libraries with Position Independent Code (PIC) which is needed 780055714Skris for linking it into DSOs. 780155714Skris [Ralf S. Engelschall] 780255714Skris 780355714Skris *) Remarkably, export ciphers were totally broken and no-one had noticed! 780455714Skris Fixed. 780555714Skris [Ben Laurie] 780655714Skris 780755714Skris *) Cleaned up the LICENSE document: The official contact for any license 780855714Skris questions now is the OpenSSL core team under openssl-core@openssl.org. 780955714Skris And add a paragraph about the dual-license situation to make sure people 781055714Skris recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 781155714Skris to the OpenSSL toolkit. 781255714Skris [Ralf S. Engelschall] 781355714Skris 781455714Skris *) General source tree makefile cleanups: Made `making xxx in yyy...' 781555714Skris display consistent in the source tree and replaced `/bin/rm' by `rm'. 781655714Skris Additonally cleaned up the `make links' target: Remove unnecessary 781755714Skris semicolons, subsequent redundant removes, inline point.sh into mklink.sh 781855714Skris to speed processing and no longer clutter the display with confusing 781955714Skris stuff. Instead only the actually done links are displayed. 782055714Skris [Ralf S. Engelschall] 782155714Skris 782255714Skris *) Permit null encryption ciphersuites, used for authentication only. It used 782355714Skris to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 782455714Skris It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 782555714Skris encryption. 782655714Skris [Ben Laurie] 782755714Skris 782855714Skris *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 782955714Skris signed attributes when verifying signatures (this would break them), 783055714Skris the detached data encoding was wrong and public keys obtained using 783155714Skris X509_get_pubkey() weren't freed. 783255714Skris [Steve Henson] 783355714Skris 783455714Skris *) Add text documentation for the BUFFER functions. Also added a work around 783555714Skris to a Win95 console bug. This was triggered by the password read stuff: the 783655714Skris last character typed gets carried over to the next fread(). If you were 783755714Skris generating a new cert request using 'req' for example then the last 783855714Skris character of the passphrase would be CR which would then enter the first 783955714Skris field as blank. 784055714Skris [Steve Henson] 784155714Skris 784255714Skris *) Added the new `Includes OpenSSL Cryptography Software' button as 784355714Skris doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 784455714Skris button and can be used by applications based on OpenSSL to show the 784555714Skris relationship to the OpenSSL project. 784655714Skris [Ralf S. Engelschall] 784755714Skris 784855714Skris *) Remove confusing variables in function signatures in files 784955714Skris ssl/ssl_lib.c and ssl/ssl.h. 785055714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 785155714Skris 785255714Skris *) Don't install bss_file.c under PREFIX/include/ 785355714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 785455714Skris 785555714Skris *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 785655714Skris functions that return function pointers and has support for NT specific 785755714Skris stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 785855714Skris #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 785955714Skris unsigned to signed types: this was killing the Win32 compile. 786055714Skris [Steve Henson] 786155714Skris 786255714Skris *) Add new certificate file to stack functions, 786355714Skris SSL_add_dir_cert_subjects_to_stack() and 786455714Skris SSL_add_file_cert_subjects_to_stack(). These largely supplant 786555714Skris SSL_load_client_CA_file(), and can be used to add multiple certs easily 786655714Skris to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 786755714Skris This means that Apache-SSL and similar packages don't have to mess around 786855714Skris to add as many CAs as they want to the preferred list. 786955714Skris [Ben Laurie] 787055714Skris 787155714Skris *) Experiment with doxygen documentation. Currently only partially applied to 787255714Skris ssl/ssl_lib.c. 787355714Skris See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 787455714Skris openssl.doxy as the configuration file. 787555714Skris [Ben Laurie] 787655714Skris 787755714Skris *) Get rid of remaining C++-style comments which strict C compilers hate. 787855714Skris [Ralf S. Engelschall, pointed out by Carlos Amengual] 787955714Skris 788055714Skris *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 788155714Skris compiled in by default: it has problems with large keys. 788255714Skris [Steve Henson] 788355714Skris 788455714Skris *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 788555714Skris DH private keys and/or callback functions which directly correspond to 788655714Skris their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 788755714Skris is needed for applications which have to configure certificates on a 788855714Skris per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 788955714Skris (e.g. s_server). 789055714Skris For the RSA certificate situation is makes no difference, but 789155714Skris for the DSA certificate situation this fixes the "no shared cipher" 789255714Skris problem where the OpenSSL cipher selection procedure failed because the 789355714Skris temporary keys were not overtaken from the context and the API provided 789455714Skris no way to reconfigure them. 789555714Skris The new functions now let applications reconfigure the stuff and they 789655714Skris are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 789755714Skris SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 789855714Skris non-public-API function ssl_cert_instantiate() is used as a helper 789955714Skris function and also to reduce code redundancy inside ssl_rsa.c. 790055714Skris [Ralf S. Engelschall] 790155714Skris 790255714Skris *) Move s_server -dcert and -dkey options out of the undocumented feature 790355714Skris area because they are useful for the DSA situation and should be 790455714Skris recognized by the users. 790555714Skris [Ralf S. Engelschall] 790655714Skris 790755714Skris *) Fix the cipher decision scheme for export ciphers: the export bits are 790855714Skris *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 790955714Skris SSL_EXP_MASK. So, the original variable has to be used instead of the 791055714Skris already masked variable. 791155714Skris [Richard Levitte <levitte@stacken.kth.se>] 791255714Skris 791355714Skris *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 791455714Skris [Richard Levitte <levitte@stacken.kth.se>] 791555714Skris 791655714Skris *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 791755714Skris from `int' to `unsigned int' because it's a length and initialized by 791855714Skris EVP_DigestFinal() which expects an `unsigned int *'. 791955714Skris [Richard Levitte <levitte@stacken.kth.se>] 792055714Skris 792155714Skris *) Don't hard-code path to Perl interpreter on shebang line of Configure 792255714Skris script. Instead use the usual Shell->Perl transition trick. 792355714Skris [Ralf S. Engelschall] 792455714Skris 792555714Skris *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 792655714Skris (in addition to RSA certificates) to match the behaviour of `openssl dsa 792755714Skris -noout -modulus' as it's already the case for `openssl rsa -noout 792855714Skris -modulus'. For RSA the -modulus is the real "modulus" while for DSA 792955714Skris currently the public key is printed (a decision which was already done by 793055714Skris `openssl dsa -modulus' in the past) which serves a similar purpose. 793155714Skris Additionally the NO_RSA no longer completely removes the whole -modulus 793255714Skris option; it now only avoids using the RSA stuff. Same applies to NO_DSA 793355714Skris now, too. 793455714Skris [Ralf S. Engelschall] 793555714Skris 793655714Skris *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 793755714Skris BIO. See the source (crypto/evp/bio_ok.c) for more info. 793855714Skris [Arne Ansper <arne@ats.cyber.ee>] 793955714Skris 794055714Skris *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 794155714Skris to be added. Now both 'req' and 'ca' can use new objects defined in the 794255714Skris config file. 794355714Skris [Steve Henson] 794455714Skris 794555714Skris *) Add cool BIO that does syslog (or event log on NT). 794655714Skris [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 794755714Skris 794855714Skris *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 794955714Skris TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 795055714Skris TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 795155714Skris Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 795255714Skris [Ben Laurie] 795355714Skris 795455714Skris *) Add preliminary config info for new extension code. 795555714Skris [Steve Henson] 795655714Skris 795755714Skris *) Make RSA_NO_PADDING really use no padding. 795855714Skris [Ulf Moeller <ulf@fitug.de>] 795955714Skris 796055714Skris *) Generate errors when private/public key check is done. 796155714Skris [Ben Laurie] 796255714Skris 796355714Skris *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 796455714Skris for some CRL extensions and new objects added. 796555714Skris [Steve Henson] 796655714Skris 796755714Skris *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 796855714Skris key usage extension and fuller support for authority key id. 796955714Skris [Steve Henson] 797055714Skris 797155714Skris *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 797255714Skris padding method for RSA, which is recommended for new applications in PKCS 797355714Skris #1 v2.0 (RFC 2437, October 1998). 797455714Skris OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 797555714Skris foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 797655714Skris against Bleichbacher's attack on RSA. 797755714Skris [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 797855714Skris Ben Laurie] 797955714Skris 798055714Skris *) Updates to the new SSL compression code 798155714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 798255714Skris 798355714Skris *) Fix so that the version number in the master secret, when passed 798455714Skris via RSA, checks that if TLS was proposed, but we roll back to SSLv3 798555714Skris (because the server will not accept higher), that the version number 798655714Skris is 0x03,0x01, not 0x03,0x00 798755714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 798855714Skris 798955714Skris *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 799055714Skris leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 799155714Skris in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 799255714Skris [Steve Henson] 799355714Skris 799455714Skris *) Support for RAW extensions where an arbitrary extension can be 799555714Skris created by including its DER encoding. See apps/openssl.cnf for 799655714Skris an example. 799755714Skris [Steve Henson] 799855714Skris 799955714Skris *) Make sure latest Perl versions don't interpret some generated C array 800055714Skris code as Perl array code in the crypto/err/err_genc.pl script. 800155714Skris [Lars Weber <3weber@informatik.uni-hamburg.de>] 800255714Skris 800355714Skris *) Modify ms/do_ms.bat to not generate assembly language makefiles since 800455714Skris not many people have the assembler. Various Win32 compilation fixes and 800555714Skris update to the INSTALL.W32 file with (hopefully) more accurate Win32 800655714Skris build instructions. 800755714Skris [Steve Henson] 800855714Skris 800955714Skris *) Modify configure script 'Configure' to automatically create crypto/date.h 801055714Skris file under Win32 and also build pem.h from pem.org. New script 801155714Skris util/mkfiles.pl to create the MINFO file on environments that can't do a 801255714Skris 'make files': perl util/mkfiles.pl >MINFO should work. 801355714Skris [Steve Henson] 801455714Skris 801555714Skris *) Major rework of DES function declarations, in the pursuit of correctness 801655714Skris and purity. As a result, many evil casts evaporated, and some weirdness, 801755714Skris too. You may find this causes warnings in your code. Zapping your evil 801855714Skris casts will probably fix them. Mostly. 801955714Skris [Ben Laurie] 802055714Skris 802155714Skris *) Fix for a typo in asn1.h. Bug fix to object creation script 802255714Skris obj_dat.pl. It considered a zero in an object definition to mean 802355714Skris "end of object": none of the objects in objects.h have any zeros 802455714Skris so it wasn't spotted. 802555714Skris [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 802655714Skris 802755714Skris *) Add support for Triple DES Cipher Block Chaining with Output Feedback 802855714Skris Masking (CBCM). In the absence of test vectors, the best I have been able 802955714Skris to do is check that the decrypt undoes the encrypt, so far. Send me test 803055714Skris vectors if you have them. 803155714Skris [Ben Laurie] 803255714Skris 803355714Skris *) Correct calculation of key length for export ciphers (too much space was 803455714Skris allocated for null ciphers). This has not been tested! 803555714Skris [Ben Laurie] 803655714Skris 803755714Skris *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 803855714Skris message is now correct (it understands "crypto" and "ssl" on its 803955714Skris command line). There is also now an "update" option. This will update 804055714Skris the util/ssleay.num and util/libeay.num files with any new functions. 804155714Skris If you do a: 804255714Skris perl util/mkdef.pl crypto ssl update 804355714Skris it will update them. 804455714Skris [Steve Henson] 804555714Skris 804655714Skris *) Overhauled the Perl interface (perl/*): 804755714Skris - ported BN stuff to OpenSSL's different BN library 804855714Skris - made the perl/ source tree CVS-aware 804955714Skris - renamed the package from SSLeay to OpenSSL (the files still contain 805055714Skris their history because I've copied them in the repository) 805155714Skris - removed obsolete files (the test scripts will be replaced 805255714Skris by better Test::Harness variants in the future) 805355714Skris [Ralf S. Engelschall] 805455714Skris 805555714Skris *) First cut for a very conservative source tree cleanup: 805655714Skris 1. merge various obsolete readme texts into doc/ssleay.txt 805755714Skris where we collect the old documents and readme texts. 805855714Skris 2. remove the first part of files where I'm already sure that we no 805955714Skris longer need them because of three reasons: either they are just temporary 806055714Skris files which were left by Eric or they are preserved original files where 806155714Skris I've verified that the diff is also available in the CVS via "cvs diff 806255714Skris -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 806355714Skris the crypto/md/ stuff). 806455714Skris [Ralf S. Engelschall] 806555714Skris 806655714Skris *) More extension code. Incomplete support for subject and issuer alt 806755714Skris name, issuer and authority key id. Change the i2v function parameters 806855714Skris and add an extra 'crl' parameter in the X509V3_CTX structure: guess 806955714Skris what that's for :-) Fix to ASN1 macro which messed up 807055714Skris IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 807155714Skris [Steve Henson] 807255714Skris 807355714Skris *) Preliminary support for ENUMERATED type. This is largely copied from the 807455714Skris INTEGER code. 807555714Skris [Steve Henson] 807655714Skris 807755714Skris *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 807855714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 807955714Skris 808055714Skris *) Make sure `make rehash' target really finds the `openssl' program. 808155714Skris [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 808255714Skris 808355714Skris *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 808455714Skris like to hear about it if this slows down other processors. 808555714Skris [Ben Laurie] 808655714Skris 808755714Skris *) Add CygWin32 platform information to Configure script. 808855714Skris [Alan Batie <batie@aahz.jf.intel.com>] 808955714Skris 809055714Skris *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 809155714Skris [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 809255714Skris 809355714Skris *) New program nseq to manipulate netscape certificate sequences 809455714Skris [Steve Henson] 809555714Skris 809655714Skris *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 809755714Skris few typos. 809855714Skris [Steve Henson] 809955714Skris 810055714Skris *) Fixes to BN code. Previously the default was to define BN_RECURSION 810155714Skris but the BN code had some problems that would cause failures when 810255714Skris doing certificate verification and some other functions. 810355714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 810455714Skris 810555714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 810655714Skris [Steve Henson] 810755714Skris 810855714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 810955714Skris [Steve Henson] 811055714Skris 811155714Skris *) Add several PKIX and private extended key usage OIDs. 811255714Skris [Steve Henson] 811355714Skris 811455714Skris *) Modify the 'ca' program to handle the new extension code. Modify 811555714Skris openssl.cnf for new extension format, add comments. 811655714Skris [Steve Henson] 811755714Skris 811855714Skris *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 811955714Skris and add a sample to openssl.cnf so req -x509 now adds appropriate 812055714Skris CA extensions. 812155714Skris [Steve Henson] 812255714Skris 812355714Skris *) Continued X509 V3 changes. Add to other makefiles, integrate with the 812455714Skris error code, add initial support to X509_print() and x509 application. 812555714Skris [Steve Henson] 812655714Skris 812755714Skris *) Takes a deep breath and start addding X509 V3 extension support code. Add 812855714Skris files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 812955714Skris stuff is currently isolated and isn't even compiled yet. 813055714Skris [Steve Henson] 813155714Skris 813255714Skris *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 813355714Skris ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 813455714Skris Removed the versions check from X509 routines when loading extensions: 813555714Skris this allows certain broken certificates that don't set the version 813655714Skris properly to be processed. 813755714Skris [Steve Henson] 813855714Skris 813955714Skris *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 814055714Skris Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 814155714Skris can still be regenerated with "make depend". 814255714Skris [Ben Laurie] 814355714Skris 814455714Skris *) Spelling mistake in C version of CAST-128. 814555714Skris [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 814655714Skris 814755714Skris *) Changes to the error generation code. The perl script err-code.pl 814855714Skris now reads in the old error codes and retains the old numbers, only 814955714Skris adding new ones if necessary. It also only changes the .err files if new 815055714Skris codes are added. The makefiles have been modified to only insert errors 815155714Skris when needed (to avoid needlessly modifying header files). This is done 815255714Skris by only inserting errors if the .err file is newer than the auto generated 815355714Skris C file. To rebuild all the error codes from scratch (the old behaviour) 815455714Skris either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 815555714Skris or delete all the .err files. 815655714Skris [Steve Henson] 815755714Skris 815855714Skris *) CAST-128 was incorrectly implemented for short keys. The C version has 815955714Skris been fixed, but is untested. The assembler versions are also fixed, but 816055714Skris new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 816155714Skris to regenerate it if needed. 816255714Skris [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 816355714Skris Hagino <itojun@kame.net>] 816455714Skris 816555714Skris *) File was opened incorrectly in randfile.c. 816655714Skris [Ulf M�ller <ulf@fitug.de>] 816755714Skris 816855714Skris *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 816955714Skris functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 817055714Skris GeneralizedTime. ASN1_TIME is the proper type used in certificates et 817155714Skris al: it's just almost always a UTCTime. Note this patch adds new error 817255714Skris codes so do a "make errors" if there are problems. 817355714Skris [Steve Henson] 817455714Skris 817555714Skris *) Correct Linux 1 recognition in config. 817655714Skris [Ulf M�ller <ulf@fitug.de>] 817755714Skris 817855714Skris *) Remove pointless MD5 hash when using DSA keys in ca. 817955714Skris [Anonymous <nobody@replay.com>] 818055714Skris 818155714Skris *) Generate an error if given an empty string as a cert directory. Also 818255714Skris generate an error if handed NULL (previously returned 0 to indicate an 818355714Skris error, but didn't set one). 818455714Skris [Ben Laurie, reported by Anonymous <nobody@replay.com>] 818555714Skris 818655714Skris *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 818755714Skris [Ben Laurie] 818855714Skris 818955714Skris *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 819055714Skris parameters. This was causing a warning which killed off the Win32 compile. 819155714Skris [Steve Henson] 819255714Skris 819355714Skris *) Remove C++ style comments from crypto/bn/bn_local.h. 819455714Skris [Neil Costigan <neil.costigan@celocom.com>] 819555714Skris 819655714Skris *) The function OBJ_txt2nid was broken. It was supposed to return a nid 819755714Skris based on a text string, looking up short and long names and finally 819855714Skris "dot" format. The "dot" format stuff didn't work. Added new function 819955714Skris OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 820055714Skris OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 820155714Skris OID is not part of the table. 820255714Skris [Steve Henson] 820355714Skris 820455714Skris *) Add prototypes to X509 lookup/verify methods, fixing a bug in 820555714Skris X509_LOOKUP_by_alias(). 820655714Skris [Ben Laurie] 820755714Skris 820855714Skris *) Sort openssl functions by name. 820955714Skris [Ben Laurie] 821055714Skris 821155714Skris *) Get the gendsa program working (hopefully) and add it to app list. Remove 821255714Skris encryption from sample DSA keys (in case anyone is interested the password 821355714Skris was "1234"). 821455714Skris [Steve Henson] 821555714Skris 821655714Skris *) Make _all_ *_free functions accept a NULL pointer. 821755714Skris [Frans Heymans <fheymans@isaserver.be>] 821855714Skris 821955714Skris *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 822055714Skris NULL pointers. 822155714Skris [Anonymous <nobody@replay.com>] 822255714Skris 822355714Skris *) s_server should send the CAfile as acceptable CAs, not its own cert. 822455714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 822555714Skris 822655714Skris *) Don't blow it for numeric -newkey arguments to apps/req. 822755714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 822855714Skris 822955714Skris *) Temp key "for export" tests were wrong in s3_srvr.c. 823055714Skris [Anonymous <nobody@replay.com>] 823155714Skris 823255714Skris *) Add prototype for temp key callback functions 823355714Skris SSL_CTX_set_tmp_{rsa,dh}_callback(). 823455714Skris [Ben Laurie] 823555714Skris 823655714Skris *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 823755714Skris DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 823855714Skris [Steve Henson] 823955714Skris 824055714Skris *) X509_name_add_entry() freed the wrong thing after an error. 824155714Skris [Arne Ansper <arne@ats.cyber.ee>] 824255714Skris 824355714Skris *) rsa_eay.c would attempt to free a NULL context. 824455714Skris [Arne Ansper <arne@ats.cyber.ee>] 824555714Skris 824655714Skris *) BIO_s_socket() had a broken should_retry() on Windoze. 824755714Skris [Arne Ansper <arne@ats.cyber.ee>] 824855714Skris 824955714Skris *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 825055714Skris [Arne Ansper <arne@ats.cyber.ee>] 825155714Skris 825255714Skris *) Make sure the already existing X509_STORE->depth variable is initialized 825355714Skris in X509_STORE_new(), but document the fact that this variable is still 825455714Skris unused in the certificate verification process. 825555714Skris [Ralf S. Engelschall] 825655714Skris 825755714Skris *) Fix the various library and apps files to free up pkeys obtained from 825855714Skris X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 825955714Skris [Steve Henson] 826055714Skris 826155714Skris *) Fix reference counting in X509_PUBKEY_get(). This makes 826255714Skris demos/maurice/example2.c work, amongst others, probably. 826355714Skris [Steve Henson and Ben Laurie] 826455714Skris 826555714Skris *) First cut of a cleanup for apps/. First the `ssleay' program is now named 826655714Skris `openssl' and second, the shortcut symlinks for the `openssl <command>' 826755714Skris are no longer created. This way we have a single and consistent command 826855714Skris line interface `openssl <command>', similar to `cvs <command>'. 826955714Skris [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 827055714Skris 827155714Skris *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 827255714Skris BIT STRING wrapper always have zero unused bits. 827355714Skris [Steve Henson] 827455714Skris 827555714Skris *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 827655714Skris [Steve Henson] 827755714Skris 827855714Skris *) Make the top-level INSTALL documentation easier to understand. 827955714Skris [Paul Sutton] 828055714Skris 828155714Skris *) Makefiles updated to exit if an error occurs in a sub-directory 828255714Skris make (including if user presses ^C) [Paul Sutton] 828355714Skris 828455714Skris *) Make Montgomery context stuff explicit in RSA data structure. 828555714Skris [Ben Laurie] 828655714Skris 828755714Skris *) Fix build order of pem and err to allow for generated pem.h. 828855714Skris [Ben Laurie] 828955714Skris 829055714Skris *) Fix renumbering bug in X509_NAME_delete_entry(). 829155714Skris [Ben Laurie] 829255714Skris 829355714Skris *) Enhanced the err-ins.pl script so it makes the error library number 829455714Skris global and can add a library name. This is needed for external ASN1 and 829555714Skris other error libraries. 829655714Skris [Steve Henson] 829755714Skris 829855714Skris *) Fixed sk_insert which never worked properly. 829955714Skris [Steve Henson] 830055714Skris 830155714Skris *) Fix ASN1 macros so they can handle indefinite length construted 830255714Skris EXPLICIT tags. Some non standard certificates use these: they can now 830355714Skris be read in. 830455714Skris [Steve Henson] 830555714Skris 830655714Skris *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 830755714Skris into a single doc/ssleay.txt bundle. This way the information is still 830855714Skris preserved but no longer messes up this directory. Now it's new room for 830955714Skris the new set of documenation files. 831055714Skris [Ralf S. Engelschall] 831155714Skris 831255714Skris *) SETs were incorrectly DER encoded. This was a major pain, because they 831355714Skris shared code with SEQUENCEs, which aren't coded the same. This means that 831455714Skris almost everything to do with SETs or SEQUENCEs has either changed name or 831555714Skris number of arguments. 831655714Skris [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 831755714Skris 831855714Skris *) Fix test data to work with the above. 831955714Skris [Ben Laurie] 832055714Skris 832155714Skris *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 832255714Skris was already fixed by Eric for 0.9.1 it seems. 832355714Skris [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 832455714Skris 832555714Skris *) Autodetect FreeBSD3. 832655714Skris [Ben Laurie] 832755714Skris 832855714Skris *) Fix various bugs in Configure. This affects the following platforms: 832955714Skris nextstep 833055714Skris ncr-scde 833155714Skris unixware-2.0 833255714Skris unixware-2.0-pentium 833355714Skris sco5-cc. 833455714Skris [Ben Laurie] 833555714Skris 833655714Skris *) Eliminate generated files from CVS. Reorder tests to regenerate files 833755714Skris before they are needed. 833855714Skris [Ben Laurie] 833955714Skris 834055714Skris *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 834155714Skris [Ben Laurie] 834255714Skris 834355714Skris 834455714Skris Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 834555714Skris 834655714Skris *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 834755714Skris changed SSLeay to OpenSSL in version strings. 834855714Skris [Ralf S. Engelschall] 834955714Skris 835055714Skris *) Some fixups to the top-level documents. 835155714Skris [Paul Sutton] 835255714Skris 835355714Skris *) Fixed the nasty bug where rsaref.h was not found under compile-time 835455714Skris because the symlink to include/ was missing. 835555714Skris [Ralf S. Engelschall] 835655714Skris 835755714Skris *) Incorporated the popular no-RSA/DSA-only patches 835855714Skris which allow to compile a RSA-free SSLeay. 835955714Skris [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 836055714Skris 836155714Skris *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 836255714Skris when "ssleay" is still not found. 836355714Skris [Ralf S. Engelschall] 836455714Skris 836555714Skris *) Added more platforms to Configure: Cray T3E, HPUX 11, 836655714Skris [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 836755714Skris 836855714Skris *) Updated the README file. 836955714Skris [Ralf S. Engelschall] 837055714Skris 837155714Skris *) Added various .cvsignore files in the CVS repository subdirs 837255714Skris to make a "cvs update" really silent. 837355714Skris [Ralf S. Engelschall] 837455714Skris 837555714Skris *) Recompiled the error-definition header files and added 837655714Skris missing symbols to the Win32 linker tables. 837755714Skris [Ralf S. Engelschall] 837855714Skris 837955714Skris *) Cleaned up the top-level documents; 838055714Skris o new files: CHANGES and LICENSE 838155714Skris o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 838255714Skris o merged COPYRIGHT into LICENSE 838355714Skris o removed obsolete TODO file 838455714Skris o renamed MICROSOFT to INSTALL.W32 838555714Skris [Ralf S. Engelschall] 838655714Skris 838755714Skris *) Removed dummy files from the 0.9.1b source tree: 838855714Skris crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 838955714Skris crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 839055714Skris crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 839155714Skris crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 839255714Skris util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 839355714Skris [Ralf S. Engelschall] 839455714Skris 839555714Skris *) Added various platform portability fixes. 839655714Skris [Mark J. Cox] 839755714Skris 839855714Skris *) The Genesis of the OpenSSL rpject: 839955714Skris We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 840055714Skris Young and Tim J. Hudson created while they were working for C2Net until 840155714Skris summer 1998. 840255714Skris [The OpenSSL Project] 840355714Skris 840455714Skris 840555714Skris Changes between 0.9.0b and 0.9.1b [not released] 840655714Skris 840755714Skris *) Updated a few CA certificates under certs/ 840855714Skris [Eric A. Young] 840955714Skris 841055714Skris *) Changed some BIGNUM api stuff. 841155714Skris [Eric A. Young] 841255714Skris 841355714Skris *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 841455714Skris DGUX x86, Linux Alpha, etc. 841555714Skris [Eric A. Young] 841655714Skris 841755714Skris *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 841855714Skris RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 841955714Skris available). 842055714Skris [Eric A. Young] 842155714Skris 842255714Skris *) Add -strparse option to asn1pars program which parses nested 842355714Skris binary structures 842455714Skris [Dr Stephen Henson <shenson@bigfoot.com>] 842555714Skris 842655714Skris *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 842755714Skris [Eric A. Young] 842855714Skris 842955714Skris *) DSA fix for "ca" program. 843055714Skris [Eric A. Young] 843155714Skris 843255714Skris *) Added "-genkey" option to "dsaparam" program. 843355714Skris [Eric A. Young] 843455714Skris 843555714Skris *) Added RIPE MD160 (rmd160) message digest. 843655714Skris [Eric A. Young] 843755714Skris 843855714Skris *) Added -a (all) option to "ssleay version" command. 843955714Skris [Eric A. Young] 844055714Skris 844155714Skris *) Added PLATFORM define which is the id given to Configure. 844255714Skris [Eric A. Young] 844355714Skris 844455714Skris *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 844555714Skris [Eric A. Young] 844655714Skris 844755714Skris *) Extended the ASN.1 parser routines. 844855714Skris [Eric A. Young] 844955714Skris 845055714Skris *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 845155714Skris [Eric A. Young] 845255714Skris 845355714Skris *) Added a BN_CTX to the BN library. 845455714Skris [Eric A. Young] 845555714Skris 845655714Skris *) Fixed the weak key values in DES library 845755714Skris [Eric A. Young] 845855714Skris 845955714Skris *) Changed API in EVP library for cipher aliases. 846055714Skris [Eric A. Young] 846155714Skris 846255714Skris *) Added support for RC2/64bit cipher. 846355714Skris [Eric A. Young] 846455714Skris 846555714Skris *) Converted the lhash library to the crypto/mem.c functions. 846655714Skris [Eric A. Young] 846755714Skris 846855714Skris *) Added more recognized ASN.1 object ids. 846955714Skris [Eric A. Young] 847055714Skris 847155714Skris *) Added more RSA padding checks for SSL/TLS. 847255714Skris [Eric A. Young] 847355714Skris 847455714Skris *) Added BIO proxy/filter functionality. 847555714Skris [Eric A. Young] 847655714Skris 847755714Skris *) Added extra_certs to SSL_CTX which can be used 847855714Skris send extra CA certificates to the client in the CA cert chain sending 847955714Skris process. It can be configured with SSL_CTX_add_extra_chain_cert(). 848055714Skris [Eric A. Young] 848155714Skris 848255714Skris *) Now Fortezza is denied in the authentication phase because 848355714Skris this is key exchange mechanism is not supported by SSLeay at all. 848455714Skris [Eric A. Young] 848555714Skris 848655714Skris *) Additional PKCS1 checks. 848755714Skris [Eric A. Young] 848855714Skris 848955714Skris *) Support the string "TLSv1" for all TLS v1 ciphers. 849055714Skris [Eric A. Young] 849155714Skris 849255714Skris *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 849355714Skris ex_data index of the SSL context in the X509_STORE_CTX ex_data. 849455714Skris [Eric A. Young] 849555714Skris 849655714Skris *) Fixed a few memory leaks. 849755714Skris [Eric A. Young] 849855714Skris 849955714Skris *) Fixed various code and comment typos. 850055714Skris [Eric A. Young] 850155714Skris 850255714Skris *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 850355714Skris bytes sent in the client random. 850455714Skris [Edward Bishop <ebishop@spyglass.com>] 850555714Skris 8506