CHANGES revision 215697
155714Skris 255714Skris OpenSSL CHANGES 355714Skris _______________ 455714Skris 5215697Ssimon Changes between 0.9.8o and 0.9.8p [16 Nov 2010] 6215697Ssimon 7215697Ssimon *) Fix extension code to avoid race conditions which can result in a buffer 8215697Ssimon overrun vulnerability: resumed sessions must not be modified as they can 9215697Ssimon be shared by multiple threads. CVE-2010-3864 10215697Ssimon [Steve Henson] 11215697Ssimon 12215697Ssimon *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939 13215697Ssimon [Steve Henson] 14215697Ssimon 15215697Ssimon *) Don't reencode certificate when calculating signature: cache and use 16215697Ssimon the original encoding instead. This makes signature verification of 17215697Ssimon some broken encodings work correctly. 18215697Ssimon [Steve Henson] 19215697Ssimon 20215697Ssimon *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT 21215697Ssimon is also one of the inputs. 22215697Ssimon [Emilia K�sper <emilia.kasper@esat.kuleuven.be> (Google)] 23215697Ssimon 24215697Ssimon *) Don't repeatedly append PBE algorithms to table if they already exist. 25215697Ssimon Sort table on each new add. This effectively makes the table read only 26215697Ssimon after all algorithms are added and subsequent calls to PKCS12_pbe_add 27215697Ssimon etc are non-op. 28215697Ssimon [Steve Henson] 29215697Ssimon 30215697Ssimon Changes between 0.9.8n and 0.9.8o [01 Jun 2010] 31215697Ssimon 32215697Ssimon [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after 33215697Ssimon OpenSSL 1.0.0.] 34215697Ssimon 35215697Ssimon *) Correct a typo in the CMS ASN1 module which can result in invalid memory 36215697Ssimon access or freeing data twice (CVE-2010-0742) 37215697Ssimon [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>] 38215697Ssimon 39215697Ssimon *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more 40215697Ssimon common in certificates and some applications which only call 41215697Ssimon SSL_library_init and not OpenSSL_add_all_algorithms() will fail. 42215697Ssimon [Steve Henson] 43215697Ssimon 44215697Ssimon *) VMS fixes: 45215697Ssimon Reduce copying into .apps and .test in makevms.com 46215697Ssimon Don't try to use blank CA certificate in CA.com 47215697Ssimon Allow use of C files from original directories in maketests.com 48215697Ssimon [Steven M. Schweda" <sms@antinode.info>] 49215697Ssimon 50206046Ssimon Changes between 0.9.8m and 0.9.8n [24 Mar 2010] 51206046Ssimon 52206046Ssimon *) When rejecting SSL/TLS records due to an incorrect version number, never 53206046Ssimon update s->server with a new major version number. As of 54206046Ssimon - OpenSSL 0.9.8m if 'short' is a 16-bit type, 55206046Ssimon - OpenSSL 0.9.8f if 'short' is longer than 16 bits, 56206046Ssimon the previous behavior could result in a read attempt at NULL when 57206046Ssimon receiving specific incorrect SSL/TLS records once record payload 58206046Ssimon protection is active. (CVE-2010-0740) 59206046Ssimon [Bodo Moeller, Adam Langley <agl@chromium.org>] 60206046Ssimon 61206046Ssimon *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL 62206046Ssimon could be crashed if the relevant tables were not present (e.g. chrooted). 63206046Ssimon [Tomas Hoger <thoger@redhat.com>] 64206046Ssimon 65205128Ssimon Changes between 0.9.8l and 0.9.8m [25 Feb 2010] 66205128Ssimon 67205128Ssimon *) Always check bn_wexpend() return values for failure. (CVE-2009-3245) 68205128Ssimon [Martin Olsson, Neel Mehta] 69205128Ssimon 70205128Ssimon *) Fix X509_STORE locking: Every 'objs' access requires a lock (to 71205128Ssimon accommodate for stack sorting, always a write lock!). 72205128Ssimon [Bodo Moeller] 73205128Ssimon 74205128Ssimon *) On some versions of WIN32 Heap32Next is very slow. This can cause 75205128Ssimon excessive delays in the RAND_poll(): over a minute. As a workaround 76205128Ssimon include a time check in the inner Heap32Next loop too. 77205128Ssimon [Steve Henson] 78205128Ssimon 79205128Ssimon *) The code that handled flushing of data in SSL/TLS originally used the 80205128Ssimon BIO_CTRL_INFO ctrl to see if any data was pending first. This caused 81205128Ssimon the problem outlined in PR#1949. The fix suggested there however can 82205128Ssimon trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions 83205128Ssimon of Apache). So instead simplify the code to flush unconditionally. 84205128Ssimon This should be fine since flushing with no data to flush is a no op. 85205128Ssimon [Steve Henson] 86205128Ssimon 87205128Ssimon *) Handle TLS versions 2.0 and later properly and correctly use the 88205128Ssimon highest version of TLS/SSL supported. Although TLS >= 2.0 is some way 89205128Ssimon off ancient servers have a habit of sticking around for a while... 90205128Ssimon [Steve Henson] 91205128Ssimon 92205128Ssimon *) Modify compression code so it frees up structures without using the 93205128Ssimon ex_data callbacks. This works around a problem where some applications 94205128Ssimon call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when 95205128Ssimon restarting) then use compression (e.g. SSL with compression) later. 96205128Ssimon This results in significant per-connection memory leaks and 97205128Ssimon has caused some security issues including CVE-2008-1678 and 98205128Ssimon CVE-2009-4355. 99205128Ssimon [Steve Henson] 100205128Ssimon 101205128Ssimon *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't 102205128Ssimon change when encrypting or decrypting. 103205128Ssimon [Bodo Moeller] 104205128Ssimon 105205128Ssimon *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to 106205128Ssimon connect and renegotiate with servers which do not support RI. 107205128Ssimon Until RI is more widely deployed this option is enabled by default. 108205128Ssimon [Steve Henson] 109205128Ssimon 110205128Ssimon *) Add "missing" ssl ctrls to clear options and mode. 111205128Ssimon [Steve Henson] 112205128Ssimon 113205128Ssimon *) If client attempts to renegotiate and doesn't support RI respond with 114205128Ssimon a no_renegotiation alert as required by RFC5746. Some renegotiating 115205128Ssimon TLS clients will continue a connection gracefully when they receive 116205128Ssimon the alert. Unfortunately OpenSSL mishandled this alert and would hang 117205128Ssimon waiting for a server hello which it will never receive. Now we treat a 118205128Ssimon received no_renegotiation alert as a fatal error. This is because 119205128Ssimon applications requesting a renegotiation might well expect it to succeed 120205128Ssimon and would have no code in place to handle the server denying it so the 121205128Ssimon only safe thing to do is to terminate the connection. 122205128Ssimon [Steve Henson] 123205128Ssimon 124205128Ssimon *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if 125205128Ssimon peer supports secure renegotiation and 0 otherwise. Print out peer 126205128Ssimon renegotiation support in s_client/s_server. 127205128Ssimon [Steve Henson] 128205128Ssimon 129205128Ssimon *) Replace the highly broken and deprecated SPKAC certification method with 130205128Ssimon the updated NID creation version. This should correctly handle UTF8. 131205128Ssimon [Steve Henson] 132205128Ssimon 133205128Ssimon *) Implement RFC5746. Re-enable renegotiation but require the extension 134205128Ssimon as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION 135205128Ssimon turns out to be a bad idea. It has been replaced by 136205128Ssimon SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with 137205128Ssimon SSL_CTX_set_options(). This is really not recommended unless you 138205128Ssimon know what you are doing. 139205128Ssimon [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson] 140205128Ssimon 141205128Ssimon *) Fixes to stateless session resumption handling. Use initial_ctx when 142205128Ssimon issuing and attempting to decrypt tickets in case it has changed during 143205128Ssimon servername handling. Use a non-zero length session ID when attempting 144205128Ssimon stateless session resumption: this makes it possible to determine if 145205128Ssimon a resumption has occurred immediately after receiving server hello 146205128Ssimon (several places in OpenSSL subtly assume this) instead of later in 147205128Ssimon the handshake. 148205128Ssimon [Steve Henson] 149205128Ssimon 150205128Ssimon *) The functions ENGINE_ctrl(), OPENSSL_isservice(), 151205128Ssimon CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error 152205128Ssimon fixes for a few places where the return code is not checked 153205128Ssimon correctly. 154205128Ssimon [Julia Lawall <julia@diku.dk>] 155205128Ssimon 156205128Ssimon *) Add --strict-warnings option to Configure script to include devteam 157205128Ssimon warnings in other configurations. 158205128Ssimon [Steve Henson] 159205128Ssimon 160205128Ssimon *) Add support for --libdir option and LIBDIR variable in makefiles. This 161205128Ssimon makes it possible to install openssl libraries in locations which 162205128Ssimon have names other than "lib", for example "/usr/lib64" which some 163205128Ssimon systems need. 164205128Ssimon [Steve Henson, based on patch from Jeremy Utley] 165205128Ssimon 166205128Ssimon *) Don't allow the use of leading 0x80 in OIDs. This is a violation of 167205128Ssimon X690 8.9.12 and can produce some misleading textual output of OIDs. 168205128Ssimon [Steve Henson, reported by Dan Kaminsky] 169205128Ssimon 170205128Ssimon *) Delete MD2 from algorithm tables. This follows the recommendation in 171205128Ssimon several standards that it is not used in new applications due to 172205128Ssimon several cryptographic weaknesses. For binary compatibility reasons 173205128Ssimon the MD2 API is still compiled in by default. 174205128Ssimon [Steve Henson] 175205128Ssimon 176205128Ssimon *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved 177205128Ssimon and restored. 178205128Ssimon [Steve Henson] 179205128Ssimon 180205128Ssimon *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and 181205128Ssimon OPENSSL_asc2uni conditionally on Netware platforms to avoid a name 182205128Ssimon clash. 183205128Ssimon [Guenter <lists@gknw.net>] 184205128Ssimon 185205128Ssimon *) Fix the server certificate chain building code to use X509_verify_cert(), 186205128Ssimon it used to have an ad-hoc builder which was unable to cope with anything 187205128Ssimon other than a simple chain. 188205128Ssimon [David Woodhouse <dwmw2@infradead.org>, Steve Henson] 189205128Ssimon 190205128Ssimon *) Don't check self signed certificate signatures in X509_verify_cert() 191205128Ssimon by default (a flag can override this): it just wastes time without 192205128Ssimon adding any security. As a useful side effect self signed root CAs 193205128Ssimon with non-FIPS digests are now usable in FIPS mode. 194205128Ssimon [Steve Henson] 195205128Ssimon 196205128Ssimon *) In dtls1_process_out_of_seq_message() the check if the current message 197205128Ssimon is already buffered was missing. For every new message was memory 198205128Ssimon allocated, allowing an attacker to perform an denial of service attack 199205128Ssimon with sending out of seq handshake messages until there is no memory 200205128Ssimon left. Additionally every future messege was buffered, even if the 201205128Ssimon sequence number made no sense and would be part of another handshake. 202205128Ssimon So only messages with sequence numbers less than 10 in advance will be 203205128Ssimon buffered. (CVE-2009-1378) 204205128Ssimon [Robin Seggelmann, discovered by Daniel Mentz] 205205128Ssimon 206205128Ssimon *) Records are buffered if they arrive with a future epoch to be 207205128Ssimon processed after finishing the corresponding handshake. There is 208205128Ssimon currently no limitation to this buffer allowing an attacker to perform 209205128Ssimon a DOS attack with sending records with future epochs until there is no 210205128Ssimon memory left. This patch adds the pqueue_size() function to detemine 211205128Ssimon the size of a buffer and limits the record buffer to 100 entries. 212205128Ssimon (CVE-2009-1377) 213205128Ssimon [Robin Seggelmann, discovered by Daniel Mentz] 214205128Ssimon 215205128Ssimon *) Keep a copy of frag->msg_header.frag_len so it can be used after the 216205128Ssimon parent structure is freed. (CVE-2009-1379) 217205128Ssimon [Daniel Mentz] 218205128Ssimon 219205128Ssimon *) Handle non-blocking I/O properly in SSL_shutdown() call. 220205128Ssimon [Darryl Miles <darryl-mailinglists@netbauds.net>] 221205128Ssimon 222205128Ssimon *) Add 2.5.4.* OIDs 223205128Ssimon [Ilya O. <vrghost@gmail.com>] 224205128Ssimon 225205128Ssimon Changes between 0.9.8k and 0.9.8l [5 Nov 2009] 226205128Ssimon 227205128Ssimon *) Disable renegotiation completely - this fixes a severe security 228205128Ssimon problem (CVE-2009-3555) at the cost of breaking all 229205128Ssimon renegotiation. Renegotiation can be re-enabled by setting 230205128Ssimon SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at 231205128Ssimon run-time. This is really not recommended unless you know what 232205128Ssimon you're doing. 233205128Ssimon [Ben Laurie] 234205128Ssimon 235194206Ssimon Changes between 0.9.8j and 0.9.8k [25 Mar 2009] 236167612Ssimon 237194206Ssimon *) Don't set val to NULL when freeing up structures, it is freed up by 238194206Ssimon underlying code. If sizeof(void *) > sizeof(long) this can result in 239194206Ssimon zeroing past the valid field. (CVE-2009-0789) 240194206Ssimon [Paolo Ganci <Paolo.Ganci@AdNovum.CH>] 241194206Ssimon 242194206Ssimon *) Fix bug where return value of CMS_SignerInfo_verify_content() was not 243194206Ssimon checked correctly. This would allow some invalid signed attributes to 244194206Ssimon appear to verify correctly. (CVE-2009-0591) 245194206Ssimon [Ivan Nestlerode <inestlerode@us.ibm.com>] 246194206Ssimon 247194206Ssimon *) Reject UniversalString and BMPString types with invalid lengths. This 248194206Ssimon prevents a crash in ASN1_STRING_print_ex() which assumes the strings have 249194206Ssimon a legal length. (CVE-2009-0590) 250194206Ssimon [Steve Henson] 251194206Ssimon 252194206Ssimon *) Set S/MIME signing as the default purpose rather than setting it 253194206Ssimon unconditionally. This allows applications to override it at the store 254194206Ssimon level. 255194206Ssimon [Steve Henson] 256194206Ssimon 257194206Ssimon *) Permit restricted recursion of ASN1 strings. This is needed in practice 258194206Ssimon to handle some structures. 259194206Ssimon [Steve Henson] 260194206Ssimon 261194206Ssimon *) Improve efficiency of mem_gets: don't search whole buffer each time 262194206Ssimon for a '\n' 263194206Ssimon [Jeremy Shapiro <jnshapir@us.ibm.com>] 264194206Ssimon 265194206Ssimon *) New -hex option for openssl rand. 266194206Ssimon [Matthieu Herrb] 267194206Ssimon 268194206Ssimon *) Print out UTF8String and NumericString when parsing ASN1. 269194206Ssimon [Steve Henson] 270194206Ssimon 271194206Ssimon *) Support NumericString type for name components. 272194206Ssimon [Steve Henson] 273194206Ssimon 274194206Ssimon *) Allow CC in the environment to override the automatically chosen 275194206Ssimon compiler. Note that nothing is done to ensure flags work with the 276194206Ssimon chosen compiler. 277194206Ssimon [Ben Laurie] 278194206Ssimon 279194206Ssimon Changes between 0.9.8i and 0.9.8j [07 Jan 2009] 280194206Ssimon 281194206Ssimon *) Properly check EVP_VerifyFinal() and similar return values 282194206Ssimon (CVE-2008-5077). 283194206Ssimon [Ben Laurie, Bodo Moeller, Google Security Team] 284194206Ssimon 285194206Ssimon *) Enable TLS extensions by default. 286194206Ssimon [Ben Laurie] 287194206Ssimon 288194206Ssimon *) Allow the CHIL engine to be loaded, whether the application is 289194206Ssimon multithreaded or not. (This does not release the developer from the 290194206Ssimon obligation to set up the dynamic locking callbacks.) 291194206Ssimon [Sander Temme <sander@temme.net>] 292194206Ssimon 293194206Ssimon *) Use correct exit code if there is an error in dgst command. 294194206Ssimon [Steve Henson; problem pointed out by Roland Dirlewanger] 295194206Ssimon 296194206Ssimon *) Tweak Configure so that you need to say "experimental-jpake" to enable 297194206Ssimon JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications. 298194206Ssimon [Bodo Moeller] 299194206Ssimon 300194206Ssimon *) Add experimental JPAKE support, including demo authentication in 301194206Ssimon s_client and s_server. 302194206Ssimon [Ben Laurie] 303194206Ssimon 304194206Ssimon *) Set the comparison function in v3_addr_canonize(). 305194206Ssimon [Rob Austein <sra@hactrn.net>] 306194206Ssimon 307194206Ssimon *) Add support for XMPP STARTTLS in s_client. 308194206Ssimon [Philip Paeps <philip@freebsd.org>] 309194206Ssimon 310194206Ssimon *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior 311194206Ssimon to ensure that even with this option, only ciphersuites in the 312194206Ssimon server's preference list will be accepted. (Note that the option 313194206Ssimon applies only when resuming a session, so the earlier behavior was 314194206Ssimon just about the algorithm choice for symmetric cryptography.) 315194206Ssimon [Bodo Moeller] 316194206Ssimon 317194206Ssimon Changes between 0.9.8h and 0.9.8i [15 Sep 2008] 318194206Ssimon 319205128Ssimon *) Fix NULL pointer dereference if a DTLS server received 320205128Ssimon ChangeCipherSpec as first record (CVE-2009-1386). 321205128Ssimon [PR #1679] 322205128Ssimon 323194206Ssimon *) Fix a state transitition in s3_srvr.c and d1_srvr.c 324194206Ssimon (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...). 325194206Ssimon [Nagendra Modadugu] 326194206Ssimon 327194206Ssimon *) The fix in 0.9.8c that supposedly got rid of unsafe 328194206Ssimon double-checked locking was incomplete for RSA blinding, 329194206Ssimon addressing just one layer of what turns out to have been 330194206Ssimon doubly unsafe triple-checked locking. 331194206Ssimon 332194206Ssimon So now fix this for real by retiring the MONT_HELPER macro 333194206Ssimon in crypto/rsa/rsa_eay.c. 334194206Ssimon 335194206Ssimon [Bodo Moeller; problem pointed out by Marius Schilder] 336194206Ssimon 337194206Ssimon *) Various precautionary measures: 338194206Ssimon 339194206Ssimon - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h). 340194206Ssimon 341194206Ssimon - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c). 342194206Ssimon (NB: This would require knowledge of the secret session ticket key 343194206Ssimon to exploit, in which case you'd be SOL either way.) 344194206Ssimon 345194206Ssimon - Change bn_nist.c so that it will properly handle input BIGNUMs 346194206Ssimon outside the expected range. 347194206Ssimon 348194206Ssimon - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG 349194206Ssimon builds. 350194206Ssimon 351194206Ssimon [Neel Mehta, Bodo Moeller] 352194206Ssimon 353194206Ssimon *) Allow engines to be "soft loaded" - i.e. optionally don't die if 354194206Ssimon the load fails. Useful for distros. 355194206Ssimon [Ben Laurie and the FreeBSD team] 356194206Ssimon 357194206Ssimon *) Add support for Local Machine Keyset attribute in PKCS#12 files. 358194206Ssimon [Steve Henson] 359194206Ssimon 360194206Ssimon *) Fix BN_GF2m_mod_arr() top-bit cleanup code. 361194206Ssimon [Huang Ying] 362194206Ssimon 363194206Ssimon *) Expand ENGINE to support engine supplied SSL client certificate functions. 364194206Ssimon 365194206Ssimon This work was sponsored by Logica. 366194206Ssimon [Steve Henson] 367194206Ssimon 368194206Ssimon *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows 369194206Ssimon keystores. Support for SSL/TLS client authentication too. 370194206Ssimon Not compiled unless enable-capieng specified to Configure. 371194206Ssimon 372194206Ssimon This work was sponsored by Logica. 373194206Ssimon [Steve Henson] 374194206Ssimon 375194206Ssimon *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using 376194206Ssimon ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain 377194206Ssimon attribute creation routines such as certifcate requests and PKCS#12 378194206Ssimon files. 379194206Ssimon [Steve Henson] 380194206Ssimon 381194206Ssimon Changes between 0.9.8g and 0.9.8h [28 May 2008] 382194206Ssimon 383194206Ssimon *) Fix flaw if 'Server Key exchange message' is omitted from a TLS 384194206Ssimon handshake which could lead to a cilent crash as found using the 385194206Ssimon Codenomicon TLS test suite (CVE-2008-1672) 386194206Ssimon [Steve Henson, Mark Cox] 387194206Ssimon 388194206Ssimon *) Fix double free in TLS server name extensions which could lead to 389194206Ssimon a remote crash found by Codenomicon TLS test suite (CVE-2008-0891) 390194206Ssimon [Joe Orton] 391194206Ssimon 392194206Ssimon *) Clear error queue in SSL_CTX_use_certificate_chain_file() 393194206Ssimon 394194206Ssimon Clear the error queue to ensure that error entries left from 395194206Ssimon older function calls do not interfere with the correct operation. 396194206Ssimon [Lutz Jaenicke, Erik de Castro Lopo] 397194206Ssimon 398194206Ssimon *) Remove root CA certificates of commercial CAs: 399194206Ssimon 400194206Ssimon The OpenSSL project does not recommend any specific CA and does not 401194206Ssimon have any policy with respect to including or excluding any CA. 402194206Ssimon Therefore it does not make any sense to ship an arbitrary selection 403194206Ssimon of root CA certificates with the OpenSSL software. 404194206Ssimon [Lutz Jaenicke] 405194206Ssimon 406194206Ssimon *) RSA OAEP patches to fix two separate invalid memory reads. 407194206Ssimon The first one involves inputs when 'lzero' is greater than 408194206Ssimon 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes 409194206Ssimon before the beginning of from). The second one involves inputs where 410194206Ssimon the 'db' section contains nothing but zeroes (there is a one-byte 411194206Ssimon invalid read after the end of 'db'). 412194206Ssimon [Ivan Nestlerode <inestlerode@us.ibm.com>] 413194206Ssimon 414194206Ssimon *) Partial backport from 0.9.9-dev: 415194206Ssimon 416194206Ssimon Introduce bn_mul_mont (dedicated Montgomery multiplication 417194206Ssimon procedure) as a candidate for BIGNUM assembler implementation. 418194206Ssimon While 0.9.9-dev uses assembler for various architectures, only 419194206Ssimon x86_64 is available by default here in the 0.9.8 branch, and 420194206Ssimon 32-bit x86 is available through a compile-time setting. 421194206Ssimon 422194206Ssimon To try the 32-bit x86 assembler implementation, use Configure 423194206Ssimon option "enable-montasm" (which exists only for this backport). 424194206Ssimon 425194206Ssimon As "enable-montasm" for 32-bit x86 disclaims code stability 426194206Ssimon anyway, in this constellation we activate additional code 427194206Ssimon backported from 0.9.9-dev for further performance improvements, 428194206Ssimon namely BN_from_montgomery_word. (To enable this otherwise, 429194206Ssimon e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".) 430194206Ssimon 431194206Ssimon [Andy Polyakov (backport partially by Bodo Moeller)] 432194206Ssimon 433194206Ssimon *) Add TLS session ticket callback. This allows an application to set 434194206Ssimon TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed 435194206Ssimon values. This is useful for key rollover for example where several key 436194206Ssimon sets may exist with different names. 437194206Ssimon [Steve Henson] 438194206Ssimon 439194206Ssimon *) Reverse ENGINE-internal logic for caching default ENGINE handles. 440194206Ssimon This was broken until now in 0.9.8 releases, such that the only way 441194206Ssimon a registered ENGINE could be used (assuming it initialises 442194206Ssimon successfully on the host) was to explicitly set it as the default 443194206Ssimon for the relevant algorithms. This is in contradiction with 0.9.7 444194206Ssimon behaviour and the documentation. With this fix, when an ENGINE is 445194206Ssimon registered into a given algorithm's table of implementations, the 446194206Ssimon 'uptodate' flag is reset so that auto-discovery will be used next 447194206Ssimon time a new context for that algorithm attempts to select an 448194206Ssimon implementation. 449194206Ssimon [Ian Lister (tweaked by Geoff Thorpe)] 450194206Ssimon 451194206Ssimon *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9 452194206Ssimon implemention in the following ways: 453194206Ssimon 454194206Ssimon Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be 455194206Ssimon hard coded. 456194206Ssimon 457194206Ssimon Lack of BER streaming support means one pass streaming processing is 458194206Ssimon only supported if data is detached: setting the streaming flag is 459194206Ssimon ignored for embedded content. 460194206Ssimon 461194206Ssimon CMS support is disabled by default and must be explicitly enabled 462194206Ssimon with the enable-cms configuration option. 463194206Ssimon [Steve Henson] 464194206Ssimon 465194206Ssimon *) Update the GMP engine glue to do direct copies between BIGNUM and 466194206Ssimon mpz_t when openssl and GMP use the same limb size. Otherwise the 467194206Ssimon existing "conversion via a text string export" trick is still used. 468194206Ssimon [Paul Sheer <paulsheer@gmail.com>] 469194206Ssimon 470194206Ssimon *) Zlib compression BIO. This is a filter BIO which compressed and 471194206Ssimon uncompresses any data passed through it. 472194206Ssimon [Steve Henson] 473194206Ssimon 474194206Ssimon *) Add AES_wrap_key() and AES_unwrap_key() functions to implement 475194206Ssimon RFC3394 compatible AES key wrapping. 476194206Ssimon [Steve Henson] 477194206Ssimon 478194206Ssimon *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0(): 479194206Ssimon sets string data without copying. X509_ALGOR_set0() and 480194206Ssimon X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier) 481194206Ssimon data. Attribute function X509at_get0_data_by_OBJ(): retrieves data 482194206Ssimon from an X509_ATTRIBUTE structure optionally checking it occurs only 483194206Ssimon once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied 484194206Ssimon data. 485194206Ssimon [Steve Henson] 486194206Ssimon 487194206Ssimon *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set() 488194206Ssimon to get the expected BN_FLG_CONSTTIME behavior. 489194206Ssimon [Bodo Moeller (Google)] 490194206Ssimon 491194206Ssimon *) Netware support: 492194206Ssimon 493194206Ssimon - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets 494194206Ssimon - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT) 495194206Ssimon - added some more tests to do_tests.pl 496194206Ssimon - fixed RunningProcess usage so that it works with newer LIBC NDKs too 497194206Ssimon - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency 498194206Ssimon - added new Configure targets netware-clib-bsdsock, netware-clib-gcc, 499194206Ssimon netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc 500194206Ssimon - various changes to netware.pl to enable gcc-cross builds on Win32 501194206Ssimon platform 502194206Ssimon - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD) 503194206Ssimon - various changes to fix missing prototype warnings 504194206Ssimon - fixed x86nasm.pl to create correct asm files for NASM COFF output 505194206Ssimon - added AES, WHIRLPOOL and CPUID assembler code to build files 506194206Ssimon - added missing AES assembler make rules to mk1mf.pl 507194206Ssimon - fixed order of includes in apps/ocsp.c so that e_os.h settings apply 508194206Ssimon [Guenter Knauf <eflash@gmx.net>] 509194206Ssimon 510194206Ssimon *) Implement certificate status request TLS extension defined in RFC3546. 511194206Ssimon A client can set the appropriate parameters and receive the encoded 512194206Ssimon OCSP response via a callback. A server can query the supplied parameters 513194206Ssimon and set the encoded OCSP response in the callback. Add simplified examples 514194206Ssimon to s_client and s_server. 515194206Ssimon [Steve Henson] 516194206Ssimon 517194206Ssimon Changes between 0.9.8f and 0.9.8g [19 Oct 2007] 518194206Ssimon 519194206Ssimon *) Fix various bugs: 520194206Ssimon + Binary incompatibility of ssl_ctx_st structure 521194206Ssimon + DTLS interoperation with non-compliant servers 522194206Ssimon + Don't call get_session_cb() without proposed session 523194206Ssimon + Fix ia64 assembler code 524194206Ssimon [Andy Polyakov, Steve Henson] 525194206Ssimon 526194206Ssimon Changes between 0.9.8e and 0.9.8f [11 Oct 2007] 527194206Ssimon 528194206Ssimon *) DTLS Handshake overhaul. There were longstanding issues with 529194206Ssimon OpenSSL DTLS implementation, which were making it impossible for 530194206Ssimon RFC 4347 compliant client to communicate with OpenSSL server. 531194206Ssimon Unfortunately just fixing these incompatibilities would "cut off" 532194206Ssimon pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e 533194206Ssimon server keeps tolerating non RFC compliant syntax. The opposite is 534194206Ssimon not true, 0.9.8f client can not communicate with earlier server. 535194206Ssimon This update even addresses CVE-2007-4995. 536194206Ssimon [Andy Polyakov] 537194206Ssimon 538194206Ssimon *) Changes to avoid need for function casts in OpenSSL: some compilers 539194206Ssimon (gcc 4.2 and later) reject their use. 540194206Ssimon [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>, 541194206Ssimon Steve Henson] 542194206Ssimon 543194206Ssimon *) Add RFC4507 support to OpenSSL. This includes the corrections in 544194206Ssimon RFC4507bis. The encrypted ticket format is an encrypted encoded 545194206Ssimon SSL_SESSION structure, that way new session features are automatically 546194206Ssimon supported. 547194206Ssimon 548194206Ssimon If a client application caches session in an SSL_SESSION structure 549194206Ssimon support is transparent because tickets are now stored in the encoded 550194206Ssimon SSL_SESSION. 551194206Ssimon 552194206Ssimon The SSL_CTX structure automatically generates keys for ticket 553194206Ssimon protection in servers so again support should be possible 554194206Ssimon with no application modification. 555194206Ssimon 556194206Ssimon If a client or server wishes to disable RFC4507 support then the option 557194206Ssimon SSL_OP_NO_TICKET can be set. 558194206Ssimon 559194206Ssimon Add a TLS extension debugging callback to allow the contents of any client 560194206Ssimon or server extensions to be examined. 561194206Ssimon 562194206Ssimon This work was sponsored by Google. 563194206Ssimon [Steve Henson] 564194206Ssimon 565194206Ssimon *) Add initial support for TLS extensions, specifically for the server_name 566194206Ssimon extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 567194206Ssimon have new members for a host name. The SSL data structure has an 568194206Ssimon additional member SSL_CTX *initial_ctx so that new sessions can be 569194206Ssimon stored in that context to allow for session resumption, even after the 570194206Ssimon SSL has been switched to a new SSL_CTX in reaction to a client's 571194206Ssimon server_name extension. 572194206Ssimon 573194206Ssimon New functions (subject to change): 574194206Ssimon 575194206Ssimon SSL_get_servername() 576194206Ssimon SSL_get_servername_type() 577194206Ssimon SSL_set_SSL_CTX() 578194206Ssimon 579194206Ssimon New CTRL codes and macros (subject to change): 580194206Ssimon 581194206Ssimon SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 582194206Ssimon - SSL_CTX_set_tlsext_servername_callback() 583194206Ssimon SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 584194206Ssimon - SSL_CTX_set_tlsext_servername_arg() 585194206Ssimon SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 586194206Ssimon 587194206Ssimon openssl s_client has a new '-servername ...' option. 588194206Ssimon 589194206Ssimon openssl s_server has new options '-servername_host ...', '-cert2 ...', 590194206Ssimon '-key2 ...', '-servername_fatal' (subject to change). This allows 591194206Ssimon testing the HostName extension for a specific single host name ('-cert' 592194206Ssimon and '-key' remain fallbacks for handshakes without HostName 593194206Ssimon negotiation). If the unrecogninzed_name alert has to be sent, this by 594194206Ssimon default is a warning; it becomes fatal with the '-servername_fatal' 595194206Ssimon option. 596194206Ssimon 597194206Ssimon [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson] 598194206Ssimon 599194206Ssimon *) Add AES and SSE2 assembly language support to VC++ build. 600194206Ssimon [Steve Henson] 601194206Ssimon 602194206Ssimon *) Mitigate attack on final subtraction in Montgomery reduction. 603194206Ssimon [Andy Polyakov] 604194206Ssimon 605194206Ssimon *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0 606194206Ssimon (which previously caused an internal error). 607194206Ssimon [Bodo Moeller] 608194206Ssimon 609194206Ssimon *) Squeeze another 10% out of IGE mode when in != out. 610194206Ssimon [Ben Laurie] 611194206Ssimon 612194206Ssimon *) AES IGE mode speedup. 613194206Ssimon [Dean Gaudet (Google)] 614194206Ssimon 615194206Ssimon *) Add the Korean symmetric 128-bit cipher SEED (see 616194206Ssimon http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and 617194206Ssimon add SEED ciphersuites from RFC 4162: 618194206Ssimon 619194206Ssimon TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA" 620194206Ssimon TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA" 621194206Ssimon TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA" 622194206Ssimon TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA" 623194206Ssimon 624194206Ssimon To minimize changes between patchlevels in the OpenSSL 0.9.8 625194206Ssimon series, SEED remains excluded from compilation unless OpenSSL 626194206Ssimon is configured with 'enable-seed'. 627194206Ssimon [KISA, Bodo Moeller] 628194206Ssimon 629194206Ssimon *) Mitigate branch prediction attacks, which can be practical if a 630194206Ssimon single processor is shared, allowing a spy process to extract 631194206Ssimon information. For detailed background information, see 632194206Ssimon http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron, 633194206Ssimon J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL 634194206Ssimon and Necessary Software Countermeasures"). The core of the change 635194206Ssimon are new versions BN_div_no_branch() and 636194206Ssimon BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(), 637194206Ssimon respectively, which are slower, but avoid the security-relevant 638194206Ssimon conditional branches. These are automatically called by BN_div() 639194206Ssimon and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one 640194206Ssimon of the input BIGNUMs. Also, BN_is_bit_set() has been changed to 641194206Ssimon remove a conditional branch. 642194206Ssimon 643194206Ssimon BN_FLG_CONSTTIME is the new name for the previous 644194206Ssimon BN_FLG_EXP_CONSTTIME flag, since it now affects more than just 645194206Ssimon modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag 646194206Ssimon in the exponent causes BN_mod_exp_mont() to use the alternative 647194206Ssimon implementation in BN_mod_exp_mont_consttime().) The old name 648194206Ssimon remains as a deprecated alias. 649194206Ssimon 650194206Ssimon Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general 651194206Ssimon RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses 652194206Ssimon constant-time implementations for more than just exponentiation. 653194206Ssimon Here too the old name is kept as a deprecated alias. 654194206Ssimon 655194206Ssimon BN_BLINDING_new() will now use BN_dup() for the modulus so that 656194206Ssimon the BN_BLINDING structure gets an independent copy of the 657194206Ssimon modulus. This means that the previous "BIGNUM *m" argument to 658194206Ssimon BN_BLINDING_new() and to BN_BLINDING_create_param() now 659194206Ssimon essentially becomes "const BIGNUM *m", although we can't actually 660194206Ssimon change this in the header file before 0.9.9. It allows 661194206Ssimon RSA_setup_blinding() to use BN_with_flags() on the modulus to 662194206Ssimon enable BN_FLG_CONSTTIME. 663194206Ssimon 664194206Ssimon [Matthew D Wood (Intel Corp)] 665194206Ssimon 666194206Ssimon *) In the SSL/TLS server implementation, be strict about session ID 667194206Ssimon context matching (which matters if an application uses a single 668194206Ssimon external cache for different purposes). Previously, 669194206Ssimon out-of-context reuse was forbidden only if SSL_VERIFY_PEER was 670194206Ssimon set. This did ensure strict client verification, but meant that, 671194206Ssimon with applications using a single external cache for quite 672194206Ssimon different requirements, clients could circumvent ciphersuite 673194206Ssimon restrictions for a given session ID context by starting a session 674194206Ssimon in a different context. 675194206Ssimon [Bodo Moeller] 676194206Ssimon 677167612Ssimon *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 678167612Ssimon a ciphersuite string such as "DEFAULT:RSA" cannot enable 679167612Ssimon authentication-only ciphersuites. 680167612Ssimon [Bodo Moeller] 681167612Ssimon 682194206Ssimon *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was 683194206Ssimon not complete and could lead to a possible single byte overflow 684194206Ssimon (CVE-2007-5135) [Ben Laurie] 685194206Ssimon 686194206Ssimon Changes between 0.9.8d and 0.9.8e [23 Feb 2007] 687194206Ssimon 688167612Ssimon *) Since AES128 and AES256 (and similarly Camellia128 and 689167612Ssimon Camellia256) share a single mask bit in the logic of 690167612Ssimon ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 691167612Ssimon kludge to work properly if AES128 is available and AES256 isn't 692167612Ssimon (or if Camellia128 is available and Camellia256 isn't). 693167612Ssimon [Victor Duchovni] 694167612Ssimon 695167612Ssimon *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c 696167612Ssimon (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters): 697167612Ssimon When a point or a seed is encoded in a BIT STRING, we need to 698167612Ssimon prevent the removal of trailing zero bits to get the proper DER 699167612Ssimon encoding. (By default, crypto/asn1/a_bitstr.c assumes the case 700167612Ssimon of a NamedBitList, for which trailing 0 bits need to be removed.) 701167612Ssimon [Bodo Moeller] 702167612Ssimon 703167612Ssimon *) Have SSL/TLS server implementation tolerate "mismatched" record 704167612Ssimon protocol version while receiving ClientHello even if the 705167612Ssimon ClientHello is fragmented. (The server can't insist on the 706167612Ssimon particular protocol version it has chosen before the ServerHello 707167612Ssimon message has informed the client about his choice.) 708167612Ssimon [Bodo Moeller] 709167612Ssimon 710167612Ssimon *) Add RFC 3779 support. 711167612Ssimon [Rob Austein for ARIN, Ben Laurie] 712167612Ssimon 713167612Ssimon *) Load error codes if they are not already present instead of using a 714167612Ssimon static variable. This allows them to be cleanly unloaded and reloaded. 715167612Ssimon Improve header file function name parsing. 716167612Ssimon [Steve Henson] 717167612Ssimon 718167612Ssimon *) extend SMTP and IMAP protocol emulation in s_client to use EHLO 719167612Ssimon or CAPABILITY handshake as required by RFCs. 720167612Ssimon [Goetz Babin-Ebell] 721167612Ssimon 722162911Ssimon Changes between 0.9.8c and 0.9.8d [28 Sep 2006] 723162911Ssimon 724162911Ssimon *) Introduce limits to prevent malicious keys being able to 725162911Ssimon cause a denial of service. (CVE-2006-2940) 726162911Ssimon [Steve Henson, Bodo Moeller] 727162911Ssimon 728162911Ssimon *) Fix ASN.1 parsing of certain invalid structures that can result 729162911Ssimon in a denial of service. (CVE-2006-2937) [Steve Henson] 730162911Ssimon 731162911Ssimon *) Fix buffer overflow in SSL_get_shared_ciphers() function. 732162911Ssimon (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 733162911Ssimon 734162911Ssimon *) Fix SSL client code which could crash if connecting to a 735162911Ssimon malicious SSLv2 server. (CVE-2006-4343) 736162911Ssimon [Tavis Ormandy and Will Drewry, Google Security Team] 737162911Ssimon 738162911Ssimon *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites 739162911Ssimon match only those. Before that, "AES256-SHA" would be interpreted 740162911Ssimon as a pattern and match "AES128-SHA" too (since AES128-SHA got 741162911Ssimon the same strength classification in 0.9.7h) as we currently only 742162911Ssimon have a single AES bit in the ciphersuite description bitmap. 743162911Ssimon That change, however, also applied to ciphersuite strings such as 744162911Ssimon "RC4-MD5" that intentionally matched multiple ciphersuites -- 745162911Ssimon namely, SSL 2.0 ciphersuites in addition to the more common ones 746162911Ssimon from SSL 3.0/TLS 1.0. 747162911Ssimon 748162911Ssimon So we change the selection algorithm again: Naming an explicit 749162911Ssimon ciphersuite selects this one ciphersuite, and any other similar 750162911Ssimon ciphersuite (same bitmap) from *other* protocol versions. 751162911Ssimon Thus, "RC4-MD5" again will properly select both the SSL 2.0 752162911Ssimon ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite. 753162911Ssimon 754162911Ssimon Since SSL 2.0 does not have any ciphersuites for which the 755162911Ssimon 128/256 bit distinction would be relevant, this works for now. 756162911Ssimon The proper fix will be to use different bits for AES128 and 757162911Ssimon AES256, which would have avoided the problems from the beginning; 758162911Ssimon however, bits are scarce, so we can only do this in a new release 759162911Ssimon (not just a patchlevel) when we can change the SSL_CIPHER 760162911Ssimon definition to split the single 'unsigned long mask' bitmap into 761162911Ssimon multiple values to extend the available space. 762162911Ssimon 763162911Ssimon [Bodo Moeller] 764162911Ssimon 765162911Ssimon Changes between 0.9.8b and 0.9.8c [05 Sep 2006] 766162911Ssimon 767162911Ssimon *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 768162911Ssimon (CVE-2006-4339) [Ben Laurie and Google Security Team] 769162911Ssimon 770162911Ssimon *) Add AES IGE and biIGE modes. 771162911Ssimon [Ben Laurie] 772162911Ssimon 773162911Ssimon *) Change the Unix randomness entropy gathering to use poll() when 774162911Ssimon possible instead of select(), since the latter has some 775162911Ssimon undesirable limitations. 776162911Ssimon [Darryl Miles via Richard Levitte and Bodo Moeller] 777162911Ssimon 778162911Ssimon *) Disable "ECCdraft" ciphersuites more thoroughly. Now special 779162911Ssimon treatment in ssl/ssl_ciph.s makes sure that these ciphersuites 780162911Ssimon cannot be implicitly activated as part of, e.g., the "AES" alias. 781162911Ssimon However, please upgrade to OpenSSL 0.9.9[-dev] for 782162911Ssimon non-experimental use of the ECC ciphersuites to get TLS extension 783162911Ssimon support, which is required for curve and point format negotiation 784162911Ssimon to avoid potential handshake problems. 785162911Ssimon [Bodo Moeller] 786162911Ssimon 787162911Ssimon *) Disable rogue ciphersuites: 788162911Ssimon 789162911Ssimon - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 790162911Ssimon - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 791162911Ssimon - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 792162911Ssimon 793162911Ssimon The latter two were purportedly from 794162911Ssimon draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 795162911Ssimon appear there. 796162911Ssimon 797167612Ssimon Also deactivate the remaining ciphersuites from 798162911Ssimon draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 799162911Ssimon unofficial, and the ID has long expired. 800162911Ssimon [Bodo Moeller] 801162911Ssimon 802162911Ssimon *) Fix RSA blinding Heisenbug (problems sometimes occured on 803162911Ssimon dual-core machines) and other potential thread-safety issues. 804162911Ssimon [Bodo Moeller] 805162911Ssimon 806162911Ssimon *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key 807162911Ssimon versions), which is now available for royalty-free use 808162911Ssimon (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html). 809162911Ssimon Also, add Camellia TLS ciphersuites from RFC 4132. 810162911Ssimon 811162911Ssimon To minimize changes between patchlevels in the OpenSSL 0.9.8 812162911Ssimon series, Camellia remains excluded from compilation unless OpenSSL 813162911Ssimon is configured with 'enable-camellia'. 814162911Ssimon [NTT] 815162911Ssimon 816162911Ssimon *) Disable the padding bug check when compression is in use. The padding 817162911Ssimon bug check assumes the first packet is of even length, this is not 818162911Ssimon necessarily true if compresssion is enabled and can result in false 819162911Ssimon positives causing handshake failure. The actual bug test is ancient 820162911Ssimon code so it is hoped that implementations will either have fixed it by 821162911Ssimon now or any which still have the bug do not support compression. 822162911Ssimon [Steve Henson] 823162911Ssimon 824160814Ssimon Changes between 0.9.8a and 0.9.8b [04 May 2006] 825160814Ssimon 826160814Ssimon *) When applying a cipher rule check to see if string match is an explicit 827160814Ssimon cipher suite and only match that one cipher suite if it is. 828160814Ssimon [Steve Henson] 829160814Ssimon 830160814Ssimon *) Link in manifests for VC++ if needed. 831160814Ssimon [Austin Ziegler <halostatue@gmail.com>] 832160814Ssimon 833160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 834160814Ssimon draft-ietf-tls-ecc-12.txt with proposed changes (but without 835160814Ssimon TLS extensions, which are supported starting with the 0.9.9 836160814Ssimon branch, not in the OpenSSL 0.9.8 branch). 837160814Ssimon [Douglas Stebila] 838160814Ssimon 839160814Ssimon *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support 840160814Ssimon opaque EVP_CIPHER_CTX handling. 841160814Ssimon [Steve Henson] 842160814Ssimon 843160814Ssimon *) Fixes and enhancements to zlib compression code. We now only use 844160814Ssimon "zlib1.dll" and use the default __cdecl calling convention on Win32 845160814Ssimon to conform with the standards mentioned here: 846160814Ssimon http://www.zlib.net/DLL_FAQ.txt 847160814Ssimon Static zlib linking now works on Windows and the new --with-zlib-include 848160814Ssimon --with-zlib-lib options to Configure can be used to supply the location 849160814Ssimon of the headers and library. Gracefully handle case where zlib library 850160814Ssimon can't be loaded. 851160814Ssimon [Steve Henson] 852160814Ssimon 853160814Ssimon *) Several fixes and enhancements to the OID generation code. The old code 854160814Ssimon sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't 855160814Ssimon handle numbers larger than ULONG_MAX, truncated printing and had a 856160814Ssimon non standard OBJ_obj2txt() behaviour. 857160814Ssimon [Steve Henson] 858160814Ssimon 859160814Ssimon *) Add support for building of engines under engine/ as shared libraries 860160814Ssimon under VC++ build system. 861160814Ssimon [Steve Henson] 862160814Ssimon 863160814Ssimon *) Corrected the numerous bugs in the Win32 path splitter in DSO. 864160814Ssimon Hopefully, we will not see any false combination of paths any more. 865160814Ssimon [Richard Levitte] 866160814Ssimon 867160814Ssimon Changes between 0.9.8 and 0.9.8a [11 Oct 2005] 868160814Ssimon 869160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 870160814Ssimon (part of SSL_OP_ALL). This option used to disable the 871160814Ssimon countermeasure against man-in-the-middle protocol-version 872160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 873160814Ssimon idea. (CVE-2005-2969) 874160814Ssimon 875160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 876160814Ssimon for Information Security, National Institute of Advanced Industrial 877160814Ssimon Science and Technology [AIST], Japan)] 878160814Ssimon 879160814Ssimon *) Add two function to clear and return the verify parameter flags. 880160814Ssimon [Steve Henson] 881160814Ssimon 882160814Ssimon *) Keep cipherlists sorted in the source instead of sorting them at 883160814Ssimon runtime, thus removing the need for a lock. 884160814Ssimon [Nils Larsch] 885160814Ssimon 886160814Ssimon *) Avoid some small subgroup attacks in Diffie-Hellman. 887160814Ssimon [Nick Mathewson and Ben Laurie] 888160814Ssimon 889160814Ssimon *) Add functions for well-known primes. 890160814Ssimon [Nick Mathewson] 891160814Ssimon 892160814Ssimon *) Extended Windows CE support. 893160814Ssimon [Satoshi Nakamura and Andy Polyakov] 894160814Ssimon 895160814Ssimon *) Initialize SSL_METHOD structures at compile time instead of during 896160814Ssimon runtime, thus removing the need for a lock. 897160814Ssimon [Steve Henson] 898160814Ssimon 899160814Ssimon *) Make PKCS7_decrypt() work even if no certificate is supplied by 900160814Ssimon attempting to decrypt each encrypted key in turn. Add support to 901160814Ssimon smime utility. 902160814Ssimon [Steve Henson] 903160814Ssimon 904160814Ssimon Changes between 0.9.7h and 0.9.8 [05 Jul 2005] 905160814Ssimon 906162911Ssimon [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after 907162911Ssimon OpenSSL 0.9.8.] 908162911Ssimon 909160814Ssimon *) Add libcrypto.pc and libssl.pc for those who feel they need them. 910160814Ssimon [Richard Levitte] 911160814Ssimon 912160814Ssimon *) Change CA.sh and CA.pl so they don't bundle the CSR and the private 913160814Ssimon key into the same file any more. 914160814Ssimon [Richard Levitte] 915160814Ssimon 916160814Ssimon *) Add initial support for Win64, both IA64 and AMD64/x64 flavors. 917160814Ssimon [Andy Polyakov] 918160814Ssimon 919160814Ssimon *) Add -utf8 command line and config file option to 'ca'. 920160814Ssimon [Stefan <stf@udoma.org] 921160814Ssimon 922160814Ssimon *) Removed the macro des_crypt(), as it seems to conflict with some 923160814Ssimon libraries. Use DES_crypt(). 924160814Ssimon [Richard Levitte] 925160814Ssimon 926160814Ssimon *) Correct naming of the 'chil' and '4758cca' ENGINEs. This 927160814Ssimon involves renaming the source and generated shared-libs for 928160814Ssimon both. The engines will accept the corrected or legacy ids 929160814Ssimon ('ncipher' and '4758_cca' respectively) when binding. NB, 930160814Ssimon this only applies when building 'shared'. 931160814Ssimon [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe] 932160814Ssimon 933160814Ssimon *) Add attribute functions to EVP_PKEY structure. Modify 934160814Ssimon PKCS12_create() to recognize a CSP name attribute and 935160814Ssimon use it. Make -CSP option work again in pkcs12 utility. 936160814Ssimon [Steve Henson] 937160814Ssimon 938160814Ssimon *) Add new functionality to the bn blinding code: 939160814Ssimon - automatic re-creation of the BN_BLINDING parameters after 940160814Ssimon a fixed number of uses (currently 32) 941160814Ssimon - add new function for parameter creation 942160814Ssimon - introduce flags to control the update behaviour of the 943160814Ssimon BN_BLINDING parameters 944160814Ssimon - hide BN_BLINDING structure 945160814Ssimon Add a second BN_BLINDING slot to the RSA structure to improve 946160814Ssimon performance when a single RSA object is shared among several 947160814Ssimon threads. 948160814Ssimon [Nils Larsch] 949160814Ssimon 950160814Ssimon *) Add support for DTLS. 951160814Ssimon [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie] 952160814Ssimon 953160814Ssimon *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1) 954160814Ssimon to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file() 955160814Ssimon [Walter Goulet] 956160814Ssimon 957160814Ssimon *) Remove buggy and incompletet DH cert support from 958160814Ssimon ssl/ssl_rsa.c and ssl/s3_both.c 959160814Ssimon [Nils Larsch] 960160814Ssimon 961160814Ssimon *) Use SHA-1 instead of MD5 as the default digest algorithm for 962160814Ssimon the apps/openssl applications. 963160814Ssimon [Nils Larsch] 964160814Ssimon 965160814Ssimon *) Compile clean with "-Wall -Wmissing-prototypes 966160814Ssimon -Wstrict-prototypes -Wmissing-declarations -Werror". Currently 967160814Ssimon DEBUG_SAFESTACK must also be set. 968160814Ssimon [Ben Laurie] 969160814Ssimon 970160814Ssimon *) Change ./Configure so that certain algorithms can be disabled by default. 971160814Ssimon The new counterpiece to "no-xxx" is "enable-xxx". 972160814Ssimon 973160814Ssimon The patented RC5 and MDC2 algorithms will now be disabled unless 974160814Ssimon "enable-rc5" and "enable-mdc2", respectively, are specified. 975160814Ssimon 976160814Ssimon (IDEA remains enabled despite being patented. This is because IDEA 977160814Ssimon is frequently required for interoperability, and there is no license 978160814Ssimon fee for non-commercial use. As before, "no-idea" can be used to 979160814Ssimon avoid this algorithm.) 980160814Ssimon 981160814Ssimon [Bodo Moeller] 982160814Ssimon 983160814Ssimon *) Add processing of proxy certificates (see RFC 3820). This work was 984160814Ssimon sponsored by KTH (The Royal Institute of Technology in Stockholm) and 985160814Ssimon EGEE (Enabling Grids for E-science in Europe). 986160814Ssimon [Richard Levitte] 987160814Ssimon 988160814Ssimon *) RC4 performance overhaul on modern architectures/implementations, such 989160814Ssimon as Intel P4, IA-64 and AMD64. 990160814Ssimon [Andy Polyakov] 991160814Ssimon 992160814Ssimon *) New utility extract-section.pl. This can be used specify an alternative 993160814Ssimon section number in a pod file instead of having to treat each file as 994160814Ssimon a separate case in Makefile. This can be done by adding two lines to the 995160814Ssimon pod file: 996160814Ssimon 997160814Ssimon =for comment openssl_section:XXX 998160814Ssimon 999160814Ssimon The blank line is mandatory. 1000160814Ssimon 1001160814Ssimon [Steve Henson] 1002160814Ssimon 1003160814Ssimon *) New arguments -certform, -keyform and -pass for s_client and s_server 1004160814Ssimon to allow alternative format key and certificate files and passphrase 1005160814Ssimon sources. 1006160814Ssimon [Steve Henson] 1007160814Ssimon 1008160814Ssimon *) New structure X509_VERIFY_PARAM which combines current verify parameters, 1009160814Ssimon update associated structures and add various utility functions. 1010160814Ssimon 1011160814Ssimon Add new policy related verify parameters, include policy checking in 1012160814Ssimon standard verify code. Enhance 'smime' application with extra parameters 1013160814Ssimon to support policy checking and print out. 1014160814Ssimon [Steve Henson] 1015160814Ssimon 1016160814Ssimon *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3 1017160814Ssimon Nehemiah processors. These extensions support AES encryption in hardware 1018160814Ssimon as well as RNG (though RNG support is currently disabled). 1019160814Ssimon [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov] 1020160814Ssimon 1021160814Ssimon *) Deprecate BN_[get|set]_params() functions (they were ignored internally). 1022160814Ssimon [Geoff Thorpe] 1023160814Ssimon 1024160814Ssimon *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented. 1025160814Ssimon [Andy Polyakov and a number of other people] 1026160814Ssimon 1027160814Ssimon *) Improved PowerPC platform support. Most notably BIGNUM assembler 1028160814Ssimon implementation contributed by IBM. 1029160814Ssimon [Suresh Chari, Peter Waltenberg, Andy Polyakov] 1030160814Ssimon 1031160814Ssimon *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public 1032160814Ssimon exponent rather than 'unsigned long'. There is a corresponding change to 1033160814Ssimon the new 'rsa_keygen' element of the RSA_METHOD structure. 1034160814Ssimon [Jelte Jansen, Geoff Thorpe] 1035160814Ssimon 1036160814Ssimon *) Functionality for creating the initial serial number file is now 1037160814Ssimon moved from CA.pl to the 'ca' utility with a new option -create_serial. 1038160814Ssimon 1039160814Ssimon (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial 1040160814Ssimon number file to 1, which is bound to cause problems. To avoid 1041160814Ssimon the problems while respecting compatibility between different 0.9.7 1042160814Ssimon patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in 1043160814Ssimon CA.pl for serial number initialization. With the new release 0.9.8, 1044160814Ssimon we can fix the problem directly in the 'ca' utility.) 1045160814Ssimon [Steve Henson] 1046160814Ssimon 1047160814Ssimon *) Reduced header interdepencies by declaring more opaque objects in 1048160814Ssimon ossl_typ.h. As a consequence, including some headers (eg. engine.h) will 1049160814Ssimon give fewer recursive includes, which could break lazy source code - so 1050160814Ssimon this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always, 1051160814Ssimon developers should define this symbol when building and using openssl to 1052160814Ssimon ensure they track the recommended behaviour, interfaces, [etc], but 1053160814Ssimon backwards-compatible behaviour prevails when this isn't defined. 1054160814Ssimon [Geoff Thorpe] 1055160814Ssimon 1056160814Ssimon *) New function X509_POLICY_NODE_print() which prints out policy nodes. 1057160814Ssimon [Steve Henson] 1058160814Ssimon 1059160814Ssimon *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality. 1060160814Ssimon This will generate a random key of the appropriate length based on the 1061160814Ssimon cipher context. The EVP_CIPHER can provide its own random key generation 1062160814Ssimon routine to support keys of a specific form. This is used in the des and 1063160814Ssimon 3des routines to generate a key of the correct parity. Update S/MIME 1064160814Ssimon code to use new functions and hence generate correct parity DES keys. 1065160814Ssimon Add EVP_CHECK_DES_KEY #define to return an error if the key is not 1066160814Ssimon valid (weak or incorrect parity). 1067160814Ssimon [Steve Henson] 1068160814Ssimon 1069160814Ssimon *) Add a local set of CRLs that can be used by X509_verify_cert() as well 1070160814Ssimon as looking them up. This is useful when the verified structure may contain 1071160814Ssimon CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs 1072160814Ssimon present unless the new PKCS7_NO_CRL flag is asserted. 1073160814Ssimon [Steve Henson] 1074160814Ssimon 1075160814Ssimon *) Extend ASN1 oid configuration module. It now additionally accepts the 1076160814Ssimon syntax: 1077160814Ssimon 1078160814Ssimon shortName = some long name, 1.2.3.4 1079160814Ssimon [Steve Henson] 1080160814Ssimon 1081160814Ssimon *) Reimplemented the BN_CTX implementation. There is now no more static 1082160814Ssimon limitation on the number of variables it can handle nor the depth of the 1083160814Ssimon "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack 1084160814Ssimon information can now expand as required, and rather than having a single 1085160814Ssimon static array of bignums, BN_CTX now uses a linked-list of such arrays 1086160814Ssimon allowing it to expand on demand whilst maintaining the usefulness of 1087160814Ssimon BN_CTX's "bundling". 1088160814Ssimon [Geoff Thorpe] 1089160814Ssimon 1090160814Ssimon *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD 1091160814Ssimon to allow all RSA operations to function using a single BN_CTX. 1092160814Ssimon [Geoff Thorpe] 1093160814Ssimon 1094160814Ssimon *) Preliminary support for certificate policy evaluation and checking. This 1095160814Ssimon is initially intended to pass the tests outlined in "Conformance Testing 1096160814Ssimon of Relying Party Client Certificate Path Processing Logic" v1.07. 1097160814Ssimon [Steve Henson] 1098160814Ssimon 1099160814Ssimon *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and 1100160814Ssimon remained unused and not that useful. A variety of other little bignum 1101160814Ssimon tweaks and fixes have also been made continuing on from the audit (see 1102160814Ssimon below). 1103160814Ssimon [Geoff Thorpe] 1104160814Ssimon 1105160814Ssimon *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with 1106160814Ssimon associated ASN1, EVP and SSL functions and old ASN1 macros. 1107160814Ssimon [Richard Levitte] 1108160814Ssimon 1109160814Ssimon *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results, 1110160814Ssimon and this should never fail. So the return value from the use of 1111160814Ssimon BN_set_word() (which can fail due to needless expansion) is now deprecated; 1112160814Ssimon if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro. 1113160814Ssimon [Geoff Thorpe] 1114160814Ssimon 1115160814Ssimon *) BN_CTX_get() should return zero-valued bignums, providing the same 1116160814Ssimon initialised value as BN_new(). 1117160814Ssimon [Geoff Thorpe, suggested by Ulf M�ller] 1118160814Ssimon 1119160814Ssimon *) Support for inhibitAnyPolicy certificate extension. 1120160814Ssimon [Steve Henson] 1121160814Ssimon 1122160814Ssimon *) An audit of the BIGNUM code is underway, for which debugging code is 1123160814Ssimon enabled when BN_DEBUG is defined. This makes stricter enforcements on what 1124160814Ssimon is considered valid when processing BIGNUMs, and causes execution to 1125160814Ssimon assert() when a problem is discovered. If BN_DEBUG_RAND is defined, 1126160814Ssimon further steps are taken to deliberately pollute unused data in BIGNUM 1127160814Ssimon structures to try and expose faulty code further on. For now, openssl will 1128160814Ssimon (in its default mode of operation) continue to tolerate the inconsistent 1129160814Ssimon forms that it has tolerated in the past, but authors and packagers should 1130160814Ssimon consider trying openssl and their own applications when compiled with 1131160814Ssimon these debugging symbols defined. It will help highlight potential bugs in 1132160814Ssimon their own code, and will improve the test coverage for OpenSSL itself. At 1133160814Ssimon some point, these tighter rules will become openssl's default to improve 1134160814Ssimon maintainability, though the assert()s and other overheads will remain only 1135160814Ssimon in debugging configurations. See bn.h for more details. 1136160814Ssimon [Geoff Thorpe, Nils Larsch, Ulf M�ller] 1137160814Ssimon 1138160814Ssimon *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure 1139160814Ssimon that can only be obtained through BN_CTX_new() (which implicitly 1140160814Ssimon initialises it). The presence of this function only made it possible 1141160814Ssimon to overwrite an existing structure (and cause memory leaks). 1142160814Ssimon [Geoff Thorpe] 1143160814Ssimon 1144160814Ssimon *) Because of the callback-based approach for implementing LHASH as a 1145160814Ssimon template type, lh_insert() adds opaque objects to hash-tables and 1146160814Ssimon lh_doall() or lh_doall_arg() are typically used with a destructor callback 1147160814Ssimon to clean up those corresponding objects before destroying the hash table 1148160814Ssimon (and losing the object pointers). So some over-zealous constifications in 1149160814Ssimon LHASH have been relaxed so that lh_insert() does not take (nor store) the 1150160814Ssimon objects as "const" and the lh_doall[_arg] callback wrappers are not 1151160814Ssimon prototyped to have "const" restrictions on the object pointers they are 1152160814Ssimon given (and so aren't required to cast them away any more). 1153160814Ssimon [Geoff Thorpe] 1154160814Ssimon 1155160814Ssimon *) The tmdiff.h API was so ugly and minimal that our own timing utility 1156160814Ssimon (speed) prefers to use its own implementation. The two implementations 1157160814Ssimon haven't been consolidated as yet (volunteers?) but the tmdiff API has had 1158160814Ssimon its object type properly exposed (MS_TM) instead of casting to/from "char 1159160814Ssimon *". This may still change yet if someone realises MS_TM and "ms_time_***" 1160160814Ssimon aren't necessarily the greatest nomenclatures - but this is what was used 1161160814Ssimon internally to the implementation so I've used that for now. 1162160814Ssimon [Geoff Thorpe] 1163160814Ssimon 1164160814Ssimon *) Ensure that deprecated functions do not get compiled when 1165160814Ssimon OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of 1166160814Ssimon the self-tests were still using deprecated key-generation functions so 1167160814Ssimon these have been updated also. 1168160814Ssimon [Geoff Thorpe] 1169160814Ssimon 1170160814Ssimon *) Reorganise PKCS#7 code to separate the digest location functionality 1171160814Ssimon into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest(). 1172160814Ssimon New function PKCS7_set_digest() to set the digest type for PKCS#7 1173160814Ssimon digestedData type. Add additional code to correctly generate the 1174160814Ssimon digestedData type and add support for this type in PKCS7 initialization 1175160814Ssimon functions. 1176160814Ssimon [Steve Henson] 1177160814Ssimon 1178160814Ssimon *) New function PKCS7_set0_type_other() this initializes a PKCS7 1179160814Ssimon structure of type "other". 1180160814Ssimon [Steve Henson] 1181160814Ssimon 1182160814Ssimon *) Fix prime generation loop in crypto/bn/bn_prime.pl by making 1183160814Ssimon sure the loop does correctly stop and breaking ("division by zero") 1184160814Ssimon modulus operations are not performed. The (pre-generated) prime 1185160814Ssimon table crypto/bn/bn_prime.h was already correct, but it could not be 1186160814Ssimon re-generated on some platforms because of the "division by zero" 1187160814Ssimon situation in the script. 1188160814Ssimon [Ralf S. Engelschall] 1189160814Ssimon 1190160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 1191160814Ssimon draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with 1192160814Ssimon SHA-1 now is only used for "small" curves (where the 1193160814Ssimon representation of a field element takes up to 24 bytes); for 1194160814Ssimon larger curves, the field element resulting from ECDH is directly 1195160814Ssimon used as premaster secret. 1196160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 1197160814Ssimon 1198160814Ssimon *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2 1199160814Ssimon curve secp160r1 to the tests. 1200160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 1201160814Ssimon 1202160814Ssimon *) Add the possibility to load symbols globally with DSO. 1203160814Ssimon [G�tz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte] 1204160814Ssimon 1205160814Ssimon *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better 1206160814Ssimon control of the error stack. 1207160814Ssimon [Richard Levitte] 1208160814Ssimon 1209160814Ssimon *) Add support for STORE in ENGINE. 1210160814Ssimon [Richard Levitte] 1211160814Ssimon 1212160814Ssimon *) Add the STORE type. The intention is to provide a common interface 1213160814Ssimon to certificate and key stores, be they simple file-based stores, or 1214160814Ssimon HSM-type store, or LDAP stores, or... 1215160814Ssimon NOTE: The code is currently UNTESTED and isn't really used anywhere. 1216160814Ssimon [Richard Levitte] 1217160814Ssimon 1218160814Ssimon *) Add a generic structure called OPENSSL_ITEM. This can be used to 1219160814Ssimon pass a list of arguments to any function as well as provide a way 1220160814Ssimon for a function to pass data back to the caller. 1221160814Ssimon [Richard Levitte] 1222160814Ssimon 1223160814Ssimon *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup() 1224160814Ssimon works like BUF_strdup() but can be used to duplicate a portion of 1225160814Ssimon a string. The copy gets NUL-terminated. BUF_memdup() duplicates 1226160814Ssimon a memory area. 1227160814Ssimon [Richard Levitte] 1228160814Ssimon 1229160814Ssimon *) Add the function sk_find_ex() which works like sk_find(), but will 1230160814Ssimon return an index to an element even if an exact match couldn't be 1231160814Ssimon found. The index is guaranteed to point at the element where the 1232160814Ssimon searched-for key would be inserted to preserve sorting order. 1233160814Ssimon [Richard Levitte] 1234160814Ssimon 1235160814Ssimon *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but 1236160814Ssimon takes an extra flags argument for optional functionality. Currently, 1237160814Ssimon the following flags are defined: 1238160814Ssimon 1239160814Ssimon OBJ_BSEARCH_VALUE_ON_NOMATCH 1240160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 1241160814Ssimon element where the comparing function returns a negative or zero 1242160814Ssimon number. 1243160814Ssimon 1244160814Ssimon OBJ_BSEARCH_FIRST_VALUE_ON_MATCH 1245160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 1246160814Ssimon element where the comparing function returns zero. This is useful 1247160814Ssimon if there are more than one element where the comparing function 1248160814Ssimon returns zero. 1249160814Ssimon [Richard Levitte] 1250160814Ssimon 1251160814Ssimon *) Make it possible to create self-signed certificates with 'openssl ca' 1252160814Ssimon in such a way that the self-signed certificate becomes part of the 1253160814Ssimon CA database and uses the same mechanisms for serial number generation 1254160814Ssimon as all other certificate signing. The new flag '-selfsign' enables 1255160814Ssimon this functionality. Adapt CA.sh and CA.pl.in. 1256160814Ssimon [Richard Levitte] 1257160814Ssimon 1258160814Ssimon *) Add functionality to check the public key of a certificate request 1259160814Ssimon against a given private. This is useful to check that a certificate 1260160814Ssimon request can be signed by that key (self-signing). 1261160814Ssimon [Richard Levitte] 1262160814Ssimon 1263160814Ssimon *) Make it possible to have multiple active certificates with the same 1264160814Ssimon subject in the CA index file. This is done only if the keyword 1265160814Ssimon 'unique_subject' is set to 'no' in the main CA section (default 1266160814Ssimon if 'CA_default') of the configuration file. The value is saved 1267160814Ssimon with the database itself in a separate index attribute file, 1268160814Ssimon named like the index file with '.attr' appended to the name. 1269160814Ssimon [Richard Levitte] 1270160814Ssimon 1271160814Ssimon *) Generate muti valued AVAs using '+' notation in config files for 1272160814Ssimon req and dirName. 1273160814Ssimon [Steve Henson] 1274160814Ssimon 1275160814Ssimon *) Support for nameConstraints certificate extension. 1276160814Ssimon [Steve Henson] 1277160814Ssimon 1278160814Ssimon *) Support for policyConstraints certificate extension. 1279160814Ssimon [Steve Henson] 1280160814Ssimon 1281160814Ssimon *) Support for policyMappings certificate extension. 1282160814Ssimon [Steve Henson] 1283160814Ssimon 1284160814Ssimon *) Make sure the default DSA_METHOD implementation only uses its 1285160814Ssimon dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL, 1286160814Ssimon and change its own handlers to be NULL so as to remove unnecessary 1287160814Ssimon indirection. This lets alternative implementations fallback to the 1288160814Ssimon default implementation more easily. 1289160814Ssimon [Geoff Thorpe] 1290160814Ssimon 1291160814Ssimon *) Support for directoryName in GeneralName related extensions 1292160814Ssimon in config files. 1293160814Ssimon [Steve Henson] 1294160814Ssimon 1295160814Ssimon *) Make it possible to link applications using Makefile.shared. 1296160814Ssimon Make that possible even when linking against static libraries! 1297160814Ssimon [Richard Levitte] 1298160814Ssimon 1299160814Ssimon *) Support for single pass processing for S/MIME signing. This now 1300160814Ssimon means that S/MIME signing can be done from a pipe, in addition 1301160814Ssimon cleartext signing (multipart/signed type) is effectively streaming 1302160814Ssimon and the signed data does not need to be all held in memory. 1303160814Ssimon 1304160814Ssimon This is done with a new flag PKCS7_STREAM. When this flag is set 1305160814Ssimon PKCS7_sign() only initializes the PKCS7 structure and the actual signing 1306160814Ssimon is done after the data is output (and digests calculated) in 1307160814Ssimon SMIME_write_PKCS7(). 1308160814Ssimon [Steve Henson] 1309160814Ssimon 1310160814Ssimon *) Add full support for -rpath/-R, both in shared libraries and 1311160814Ssimon applications, at least on the platforms where it's known how 1312160814Ssimon to do it. 1313160814Ssimon [Richard Levitte] 1314160814Ssimon 1315160814Ssimon *) In crypto/ec/ec_mult.c, implement fast point multiplication with 1316160814Ssimon precomputation, based on wNAF splitting: EC_GROUP_precompute_mult() 1317160814Ssimon will now compute a table of multiples of the generator that 1318160814Ssimon makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul() 1319160814Ssimon faster (notably in the case of a single point multiplication, 1320160814Ssimon scalar * generator). 1321160814Ssimon [Nils Larsch, Bodo Moeller] 1322160814Ssimon 1323160814Ssimon *) IPv6 support for certificate extensions. The various extensions 1324160814Ssimon which use the IP:a.b.c.d can now take IPv6 addresses using the 1325160814Ssimon formats of RFC1884 2.2 . IPv6 addresses are now also displayed 1326160814Ssimon correctly. 1327160814Ssimon [Steve Henson] 1328160814Ssimon 1329160814Ssimon *) Added an ENGINE that implements RSA by performing private key 1330160814Ssimon exponentiations with the GMP library. The conversions to and from 1331160814Ssimon GMP's mpz_t format aren't optimised nor are any montgomery forms 1332160814Ssimon cached, and on x86 it appears OpenSSL's own performance has caught up. 1333160814Ssimon However there are likely to be other architectures where GMP could 1334160814Ssimon provide a boost. This ENGINE is not built in by default, but it can be 1335160814Ssimon specified at Configure time and should be accompanied by the necessary 1336160814Ssimon linker additions, eg; 1337160814Ssimon ./config -DOPENSSL_USE_GMP -lgmp 1338160814Ssimon [Geoff Thorpe] 1339160814Ssimon 1340160814Ssimon *) "openssl engine" will not display ENGINE/DSO load failure errors when 1341160814Ssimon testing availability of engines with "-t" - the old behaviour is 1342160814Ssimon produced by increasing the feature's verbosity with "-tt". 1343160814Ssimon [Geoff Thorpe] 1344160814Ssimon 1345160814Ssimon *) ECDSA routines: under certain error conditions uninitialized BN objects 1346160814Ssimon could be freed. Solution: make sure initialization is performed early 1347160814Ssimon enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de> 1348160814Ssimon via PR#459) 1349160814Ssimon [Lutz Jaenicke] 1350160814Ssimon 1351160814Ssimon *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD 1352160814Ssimon and DH_METHOD (eg. by ENGINE implementations) to override the normal 1353160814Ssimon software implementations. For DSA and DH, parameter generation can 1354160814Ssimon also be overriden by providing the appropriate method callbacks. 1355160814Ssimon [Geoff Thorpe] 1356160814Ssimon 1357160814Ssimon *) Change the "progress" mechanism used in key-generation and 1358160814Ssimon primality testing to functions that take a new BN_GENCB pointer in 1359160814Ssimon place of callback/argument pairs. The new API functions have "_ex" 1360160814Ssimon postfixes and the older functions are reimplemented as wrappers for 1361160814Ssimon the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide 1362160814Ssimon declarations of the old functions to help (graceful) attempts to 1363160814Ssimon migrate to the new functions. Also, the new key-generation API 1364160814Ssimon functions operate on a caller-supplied key-structure and return 1365160814Ssimon success/failure rather than returning a key or NULL - this is to 1366160814Ssimon help make "keygen" another member function of RSA_METHOD etc. 1367160814Ssimon 1368160814Ssimon Example for using the new callback interface: 1369160814Ssimon 1370160814Ssimon int (*my_callback)(int a, int b, BN_GENCB *cb) = ...; 1371160814Ssimon void *my_arg = ...; 1372160814Ssimon BN_GENCB my_cb; 1373160814Ssimon 1374160814Ssimon BN_GENCB_set(&my_cb, my_callback, my_arg); 1375160814Ssimon 1376160814Ssimon return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb); 1377160814Ssimon /* For the meaning of a, b in calls to my_callback(), see the 1378160814Ssimon * documentation of the function that calls the callback. 1379160814Ssimon * cb will point to my_cb; my_arg can be retrieved as cb->arg. 1380160814Ssimon * my_callback should return 1 if it wants BN_is_prime_ex() 1381160814Ssimon * to continue, or 0 to stop. 1382160814Ssimon */ 1383160814Ssimon 1384160814Ssimon [Geoff Thorpe] 1385160814Ssimon 1386160814Ssimon *) Change the ZLIB compression method to be stateful, and make it 1387160814Ssimon available to TLS with the number defined in 1388160814Ssimon draft-ietf-tls-compression-04.txt. 1389160814Ssimon [Richard Levitte] 1390160814Ssimon 1391160814Ssimon *) Add the ASN.1 structures and functions for CertificatePair, which 1392160814Ssimon is defined as follows (according to X.509_4thEditionDraftV6.pdf): 1393160814Ssimon 1394160814Ssimon CertificatePair ::= SEQUENCE { 1395160814Ssimon forward [0] Certificate OPTIONAL, 1396160814Ssimon reverse [1] Certificate OPTIONAL, 1397160814Ssimon -- at least one of the pair shall be present -- } 1398160814Ssimon 1399160814Ssimon Also implement the PEM functions to read and write certificate 1400160814Ssimon pairs, and defined the PEM tag as "CERTIFICATE PAIR". 1401160814Ssimon 1402160814Ssimon This needed to be defined, mostly for the sake of the LDAP 1403160814Ssimon attribute crossCertificatePair, but may prove useful elsewhere as 1404160814Ssimon well. 1405160814Ssimon [Richard Levitte] 1406160814Ssimon 1407160814Ssimon *) Make it possible to inhibit symlinking of shared libraries in 1408160814Ssimon Makefile.shared, for Cygwin's sake. 1409160814Ssimon [Richard Levitte] 1410160814Ssimon 1411160814Ssimon *) Extend the BIGNUM API by creating a function 1412160814Ssimon void BN_set_negative(BIGNUM *a, int neg); 1413160814Ssimon and a macro that behave like 1414160814Ssimon int BN_is_negative(const BIGNUM *a); 1415160814Ssimon 1416160814Ssimon to avoid the need to access 'a->neg' directly in applications. 1417160814Ssimon [Nils Larsch] 1418160814Ssimon 1419160814Ssimon *) Implement fast modular reduction for pseudo-Mersenne primes 1420160814Ssimon used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c). 1421160814Ssimon EC_GROUP_new_curve_GFp() will now automatically use this 1422160814Ssimon if applicable. 1423160814Ssimon [Nils Larsch <nla@trustcenter.de>] 1424160814Ssimon 1425160814Ssimon *) Add new lock type (CRYPTO_LOCK_BN). 1426160814Ssimon [Bodo Moeller] 1427160814Ssimon 1428160814Ssimon *) Change the ENGINE framework to automatically load engines 1429160814Ssimon dynamically from specific directories unless they could be 1430160814Ssimon found to already be built in or loaded. Move all the 1431160814Ssimon current engines except for the cryptodev one to a new 1432160814Ssimon directory engines/. 1433160814Ssimon The engines in engines/ are built as shared libraries if 1434160814Ssimon the "shared" options was given to ./Configure or ./config. 1435160814Ssimon Otherwise, they are inserted in libcrypto.a. 1436160814Ssimon /usr/local/ssl/engines is the default directory for dynamic 1437160814Ssimon engines, but that can be overriden at configure time through 1438160814Ssimon the usual use of --prefix and/or --openssldir, and at run 1439160814Ssimon time with the environment variable OPENSSL_ENGINES. 1440160814Ssimon [Geoff Thorpe and Richard Levitte] 1441160814Ssimon 1442160814Ssimon *) Add Makefile.shared, a helper makefile to build shared 1443160814Ssimon libraries. Addapt Makefile.org. 1444160814Ssimon [Richard Levitte] 1445160814Ssimon 1446160814Ssimon *) Add version info to Win32 DLLs. 1447160814Ssimon [Peter 'Luna' Runestig" <peter@runestig.com>] 1448160814Ssimon 1449160814Ssimon *) Add new 'medium level' PKCS#12 API. Certificates and keys 1450160814Ssimon can be added using this API to created arbitrary PKCS#12 1451160814Ssimon files while avoiding the low level API. 1452160814Ssimon 1453160814Ssimon New options to PKCS12_create(), key or cert can be NULL and 1454160814Ssimon will then be omitted from the output file. The encryption 1455160814Ssimon algorithm NIDs can be set to -1 for no encryption, the mac 1456160814Ssimon iteration count can be set to 0 to omit the mac. 1457160814Ssimon 1458160814Ssimon Enhance pkcs12 utility by making the -nokeys and -nocerts 1459160814Ssimon options work when creating a PKCS#12 file. New option -nomac 1460160814Ssimon to omit the mac, NONE can be set for an encryption algorithm. 1461160814Ssimon New code is modified to use the enhanced PKCS12_create() 1462160814Ssimon instead of the low level API. 1463160814Ssimon [Steve Henson] 1464160814Ssimon 1465160814Ssimon *) Extend ASN1 encoder to support indefinite length constructed 1466160814Ssimon encoding. This can output sequences tags and octet strings in 1467160814Ssimon this form. Modify pk7_asn1.c to support indefinite length 1468160814Ssimon encoding. This is experimental and needs additional code to 1469160814Ssimon be useful, such as an ASN1 bio and some enhanced streaming 1470160814Ssimon PKCS#7 code. 1471160814Ssimon 1472160814Ssimon Extend template encode functionality so that tagging is passed 1473160814Ssimon down to the template encoder. 1474160814Ssimon [Steve Henson] 1475160814Ssimon 1476160814Ssimon *) Let 'openssl req' fail if an argument to '-newkey' is not 1477160814Ssimon recognized instead of using RSA as a default. 1478160814Ssimon [Bodo Moeller] 1479160814Ssimon 1480160814Ssimon *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt. 1481160814Ssimon As these are not official, they are not included in "ALL"; 1482160814Ssimon the "ECCdraft" ciphersuite group alias can be used to select them. 1483160814Ssimon [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)] 1484160814Ssimon 1485160814Ssimon *) Add ECDH engine support. 1486160814Ssimon [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)] 1487160814Ssimon 1488160814Ssimon *) Add ECDH in new directory crypto/ecdh/. 1489160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 1490160814Ssimon 1491160814Ssimon *) Let BN_rand_range() abort with an error after 100 iterations 1492160814Ssimon without success (which indicates a broken PRNG). 1493160814Ssimon [Bodo Moeller] 1494160814Ssimon 1495160814Ssimon *) Change BN_mod_sqrt() so that it verifies that the input value 1496160814Ssimon is really the square of the return value. (Previously, 1497160814Ssimon BN_mod_sqrt would show GIGO behaviour.) 1498160814Ssimon [Bodo Moeller] 1499160814Ssimon 1500160814Ssimon *) Add named elliptic curves over binary fields from X9.62, SECG, 1501160814Ssimon and WAP/WTLS; add OIDs that were still missing. 1502160814Ssimon 1503160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 1504160814Ssimon (Sun Microsystems Laboratories)] 1505160814Ssimon 1506160814Ssimon *) Extend the EC library for elliptic curves over binary fields 1507160814Ssimon (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/). 1508160814Ssimon New EC_METHOD: 1509160814Ssimon 1510160814Ssimon EC_GF2m_simple_method 1511160814Ssimon 1512160814Ssimon New API functions: 1513160814Ssimon 1514160814Ssimon EC_GROUP_new_curve_GF2m 1515160814Ssimon EC_GROUP_set_curve_GF2m 1516160814Ssimon EC_GROUP_get_curve_GF2m 1517160814Ssimon EC_POINT_set_affine_coordinates_GF2m 1518160814Ssimon EC_POINT_get_affine_coordinates_GF2m 1519160814Ssimon EC_POINT_set_compressed_coordinates_GF2m 1520160814Ssimon 1521160814Ssimon Point compression for binary fields is disabled by default for 1522160814Ssimon patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to 1523160814Ssimon enable it). 1524160814Ssimon 1525160814Ssimon As binary polynomials are represented as BIGNUMs, various members 1526160814Ssimon of the EC_GROUP and EC_POINT data structures can be shared 1527160814Ssimon between the implementations for prime fields and binary fields; 1528160814Ssimon the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m) 1529160814Ssimon are essentially identical to their ..._GFp counterparts. 1530160814Ssimon (For simplicity, the '..._GFp' prefix has been dropped from 1531160814Ssimon various internal method names.) 1532160814Ssimon 1533160814Ssimon An internal 'field_div' method (similar to 'field_mul' and 1534160814Ssimon 'field_sqr') has been added; this is used only for binary fields. 1535160814Ssimon 1536160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 1537160814Ssimon (Sun Microsystems Laboratories)] 1538160814Ssimon 1539160814Ssimon *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult() 1540160814Ssimon through methods ('mul', 'precompute_mult'). 1541160814Ssimon 1542160814Ssimon The generic implementations (now internally called 'ec_wNAF_mul' 1543160814Ssimon and 'ec_wNAF_precomputed_mult') remain the default if these 1544160814Ssimon methods are undefined. 1545160814Ssimon 1546160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 1547160814Ssimon (Sun Microsystems Laboratories)] 1548160814Ssimon 1549160814Ssimon *) New function EC_GROUP_get_degree, which is defined through 1550160814Ssimon EC_METHOD. For curves over prime fields, this returns the bit 1551160814Ssimon length of the modulus. 1552160814Ssimon 1553160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 1554160814Ssimon (Sun Microsystems Laboratories)] 1555160814Ssimon 1556160814Ssimon *) New functions EC_GROUP_dup, EC_POINT_dup. 1557160814Ssimon (These simply call ..._new and ..._copy). 1558160814Ssimon 1559160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 1560160814Ssimon (Sun Microsystems Laboratories)] 1561160814Ssimon 1562160814Ssimon *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c. 1563160814Ssimon Polynomials are represented as BIGNUMs (where the sign bit is not 1564160814Ssimon used) in the following functions [macros]: 1565160814Ssimon 1566160814Ssimon BN_GF2m_add 1567160814Ssimon BN_GF2m_sub [= BN_GF2m_add] 1568160814Ssimon BN_GF2m_mod [wrapper for BN_GF2m_mod_arr] 1569160814Ssimon BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr] 1570160814Ssimon BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr] 1571160814Ssimon BN_GF2m_mod_inv 1572160814Ssimon BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr] 1573160814Ssimon BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr] 1574160814Ssimon BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr] 1575160814Ssimon BN_GF2m_cmp [= BN_ucmp] 1576160814Ssimon 1577160814Ssimon (Note that only the 'mod' functions are actually for fields GF(2^m). 1578160814Ssimon BN_GF2m_add() is misnomer, but this is for the sake of consistency.) 1579160814Ssimon 1580160814Ssimon For some functions, an the irreducible polynomial defining a 1581160814Ssimon field can be given as an 'unsigned int[]' with strictly 1582160814Ssimon decreasing elements giving the indices of those bits that are set; 1583160814Ssimon i.e., p[] represents the polynomial 1584160814Ssimon f(t) = t^p[0] + t^p[1] + ... + t^p[k] 1585160814Ssimon where 1586160814Ssimon p[0] > p[1] > ... > p[k] = 0. 1587160814Ssimon This applies to the following functions: 1588160814Ssimon 1589160814Ssimon BN_GF2m_mod_arr 1590160814Ssimon BN_GF2m_mod_mul_arr 1591160814Ssimon BN_GF2m_mod_sqr_arr 1592160814Ssimon BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv] 1593160814Ssimon BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div] 1594160814Ssimon BN_GF2m_mod_exp_arr 1595160814Ssimon BN_GF2m_mod_sqrt_arr 1596160814Ssimon BN_GF2m_mod_solve_quad_arr 1597160814Ssimon BN_GF2m_poly2arr 1598160814Ssimon BN_GF2m_arr2poly 1599160814Ssimon 1600160814Ssimon Conversion can be performed by the following functions: 1601160814Ssimon 1602160814Ssimon BN_GF2m_poly2arr 1603160814Ssimon BN_GF2m_arr2poly 1604160814Ssimon 1605160814Ssimon bntest.c has additional tests for binary polynomial arithmetic. 1606160814Ssimon 1607160814Ssimon Two implementations for BN_GF2m_mod_div() are available. 1608160814Ssimon The default algorithm simply uses BN_GF2m_mod_inv() and 1609160814Ssimon BN_GF2m_mod_mul(). The alternative algorithm is compiled in only 1610160814Ssimon if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the 1611160814Ssimon copyright notice in crypto/bn/bn_gf2m.c before enabling it). 1612160814Ssimon 1613160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 1614160814Ssimon (Sun Microsystems Laboratories)] 1615160814Ssimon 1616160814Ssimon *) Add new error code 'ERR_R_DISABLED' that can be used when some 1617160814Ssimon functionality is disabled at compile-time. 1618160814Ssimon [Douglas Stebila <douglas.stebila@sun.com>] 1619160814Ssimon 1620160814Ssimon *) Change default behaviour of 'openssl asn1parse' so that more 1621160814Ssimon information is visible when viewing, e.g., a certificate: 1622160814Ssimon 1623160814Ssimon Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump' 1624160814Ssimon mode the content of non-printable OCTET STRINGs is output in a 1625160814Ssimon style similar to INTEGERs, but with '[HEX DUMP]' prepended to 1626160814Ssimon avoid the appearance of a printable string. 1627160814Ssimon [Nils Larsch <nla@trustcenter.de>] 1628160814Ssimon 1629160814Ssimon *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access 1630160814Ssimon functions 1631160814Ssimon EC_GROUP_set_asn1_flag() 1632160814Ssimon EC_GROUP_get_asn1_flag() 1633160814Ssimon EC_GROUP_set_point_conversion_form() 1634160814Ssimon EC_GROUP_get_point_conversion_form() 1635160814Ssimon These control ASN1 encoding details: 1636160814Ssimon - Curves (i.e., groups) are encoded explicitly unless asn1_flag 1637160814Ssimon has been set to OPENSSL_EC_NAMED_CURVE. 1638160814Ssimon - Points are encoded in uncompressed form by default; options for 1639160814Ssimon asn1_for are as for point2oct, namely 1640160814Ssimon POINT_CONVERSION_COMPRESSED 1641160814Ssimon POINT_CONVERSION_UNCOMPRESSED 1642160814Ssimon POINT_CONVERSION_HYBRID 1643160814Ssimon 1644160814Ssimon Also add 'seed' and 'seed_len' members to EC_GROUP with access 1645160814Ssimon functions 1646160814Ssimon EC_GROUP_set_seed() 1647160814Ssimon EC_GROUP_get0_seed() 1648160814Ssimon EC_GROUP_get_seed_len() 1649160814Ssimon This is used only for ASN1 purposes (so far). 1650160814Ssimon [Nils Larsch <nla@trustcenter.de>] 1651160814Ssimon 1652160814Ssimon *) Add 'field_type' member to EC_METHOD, which holds the NID 1653160814Ssimon of the appropriate field type OID. The new function 1654160814Ssimon EC_METHOD_get_field_type() returns this value. 1655160814Ssimon [Nils Larsch <nla@trustcenter.de>] 1656160814Ssimon 1657160814Ssimon *) Add functions 1658160814Ssimon EC_POINT_point2bn() 1659160814Ssimon EC_POINT_bn2point() 1660160814Ssimon EC_POINT_point2hex() 1661160814Ssimon EC_POINT_hex2point() 1662160814Ssimon providing useful interfaces to EC_POINT_point2oct() and 1663160814Ssimon EC_POINT_oct2point(). 1664160814Ssimon [Nils Larsch <nla@trustcenter.de>] 1665160814Ssimon 1666160814Ssimon *) Change internals of the EC library so that the functions 1667160814Ssimon EC_GROUP_set_generator() 1668160814Ssimon EC_GROUP_get_generator() 1669160814Ssimon EC_GROUP_get_order() 1670160814Ssimon EC_GROUP_get_cofactor() 1671160814Ssimon are implemented directly in crypto/ec/ec_lib.c and not dispatched 1672160814Ssimon to methods, which would lead to unnecessary code duplication when 1673160814Ssimon adding different types of curves. 1674160814Ssimon [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller] 1675160814Ssimon 1676160814Ssimon *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM 1677160814Ssimon arithmetic, and such that modified wNAFs are generated 1678160814Ssimon (which avoid length expansion in many cases). 1679160814Ssimon [Bodo Moeller] 1680160814Ssimon 1681160814Ssimon *) Add a function EC_GROUP_check_discriminant() (defined via 1682160814Ssimon EC_METHOD) that verifies that the curve discriminant is non-zero. 1683160814Ssimon 1684160814Ssimon Add a function EC_GROUP_check() that makes some sanity tests 1685160814Ssimon on a EC_GROUP, its generator and order. This includes 1686160814Ssimon EC_GROUP_check_discriminant(). 1687160814Ssimon [Nils Larsch <nla@trustcenter.de>] 1688160814Ssimon 1689160814Ssimon *) Add ECDSA in new directory crypto/ecdsa/. 1690160814Ssimon 1691160814Ssimon Add applications 'openssl ecparam' and 'openssl ecdsa' 1692160814Ssimon (these are based on 'openssl dsaparam' and 'openssl dsa'). 1693160814Ssimon 1694160814Ssimon ECDSA support is also included in various other files across the 1695160814Ssimon library. Most notably, 1696160814Ssimon - 'openssl req' now has a '-newkey ecdsa:file' option; 1697160814Ssimon - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA; 1698160814Ssimon - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and 1699160814Ssimon d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make 1700160814Ssimon them suitable for ECDSA where domain parameters must be 1701160814Ssimon extracted before the specific public key; 1702160814Ssimon - ECDSA engine support has been added. 1703160814Ssimon [Nils Larsch <nla@trustcenter.de>] 1704160814Ssimon 1705160814Ssimon *) Include some named elliptic curves, and add OIDs from X9.62, 1706160814Ssimon SECG, and WAP/WTLS. Each curve can be obtained from the new 1707160814Ssimon function 1708160814Ssimon EC_GROUP_new_by_curve_name(), 1709160814Ssimon and the list of available named curves can be obtained with 1710160814Ssimon EC_get_builtin_curves(). 1711160814Ssimon Also add a 'curve_name' member to EC_GROUP objects, which can be 1712160814Ssimon accessed via 1713160814Ssimon EC_GROUP_set_curve_name() 1714160814Ssimon EC_GROUP_get_curve_name() 1715160814Ssimon [Nils Larsch <larsch@trustcenter.de, Bodo Moeller] 1716160814Ssimon 1717160814Ssimon *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 1718160814Ssimon was actually never needed) and in BN_mul(). The removal in BN_mul() 1719160814Ssimon required a small change in bn_mul_part_recursive() and the addition 1720160814Ssimon of the functions bn_cmp_part_words(), bn_sub_part_words() and 1721160814Ssimon bn_add_part_words(), which do the same thing as bn_cmp_words(), 1722160814Ssimon bn_sub_words() and bn_add_words() except they take arrays with 1723160814Ssimon differing sizes. 1724160814Ssimon [Richard Levitte] 1725160814Ssimon 1726194206Ssimon Changes between 0.9.7l and 0.9.7m [23 Feb 2007] 1727194206Ssimon 1728167612Ssimon *) Cleanse PEM buffers before freeing them since they may contain 1729167612Ssimon sensitive data. 1730167612Ssimon [Benjamin Bennett <ben@psc.edu>] 1731167612Ssimon 1732167612Ssimon *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 1733167612Ssimon a ciphersuite string such as "DEFAULT:RSA" cannot enable 1734167612Ssimon authentication-only ciphersuites. 1735167612Ssimon [Bodo Moeller] 1736167612Ssimon 1737167612Ssimon *) Since AES128 and AES256 share a single mask bit in the logic of 1738167612Ssimon ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 1739167612Ssimon kludge to work properly if AES128 is available and AES256 isn't. 1740167612Ssimon [Victor Duchovni] 1741167612Ssimon 1742194206Ssimon *) Expand security boundary to match 1.1.1 module. 1743194206Ssimon [Steve Henson] 1744194206Ssimon 1745194206Ssimon *) Remove redundant features: hash file source, editing of test vectors 1746194206Ssimon modify fipsld to use external fips_premain.c signature. 1747194206Ssimon [Steve Henson] 1748194206Ssimon 1749194206Ssimon *) New perl script mkfipsscr.pl to create shell scripts or batch files to 1750194206Ssimon run algorithm test programs. 1751194206Ssimon [Steve Henson] 1752194206Ssimon 1753194206Ssimon *) Make algorithm test programs more tolerant of whitespace. 1754194206Ssimon [Steve Henson] 1755194206Ssimon 1756167612Ssimon *) Have SSL/TLS server implementation tolerate "mismatched" record 1757167612Ssimon protocol version while receiving ClientHello even if the 1758167612Ssimon ClientHello is fragmented. (The server can't insist on the 1759167612Ssimon particular protocol version it has chosen before the ServerHello 1760167612Ssimon message has informed the client about his choice.) 1761167612Ssimon [Bodo Moeller] 1762167612Ssimon 1763167612Ssimon *) Load error codes if they are not already present instead of using a 1764167612Ssimon static variable. This allows them to be cleanly unloaded and reloaded. 1765167612Ssimon [Steve Henson] 1766167612Ssimon 1767167612Ssimon Changes between 0.9.7k and 0.9.7l [28 Sep 2006] 1768167612Ssimon 1769167612Ssimon *) Introduce limits to prevent malicious keys being able to 1770167612Ssimon cause a denial of service. (CVE-2006-2940) 1771167612Ssimon [Steve Henson, Bodo Moeller] 1772167612Ssimon 1773167612Ssimon *) Fix ASN.1 parsing of certain invalid structures that can result 1774167612Ssimon in a denial of service. (CVE-2006-2937) [Steve Henson] 1775167612Ssimon 1776167612Ssimon *) Fix buffer overflow in SSL_get_shared_ciphers() function. 1777167612Ssimon (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 1778167612Ssimon 1779167612Ssimon *) Fix SSL client code which could crash if connecting to a 1780167612Ssimon malicious SSLv2 server. (CVE-2006-4343) 1781167612Ssimon [Tavis Ormandy and Will Drewry, Google Security Team] 1782167612Ssimon 1783162911Ssimon *) Change ciphersuite string processing so that an explicit 1784162911Ssimon ciphersuite selects this one ciphersuite (so that "AES256-SHA" 1785162911Ssimon will no longer include "AES128-SHA"), and any other similar 1786162911Ssimon ciphersuite (same bitmap) from *other* protocol versions (so that 1787162911Ssimon "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the 1788162911Ssimon SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining 1789162911Ssimon changes from 0.9.8b and 0.9.8d. 1790162911Ssimon [Bodo Moeller] 1791162911Ssimon 1792162911Ssimon Changes between 0.9.7j and 0.9.7k [05 Sep 2006] 1793162911Ssimon 1794162911Ssimon *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 1795162911Ssimon (CVE-2006-4339) [Ben Laurie and Google Security Team] 1796162911Ssimon 1797162911Ssimon *) Change the Unix randomness entropy gathering to use poll() when 1798162911Ssimon possible instead of select(), since the latter has some 1799162911Ssimon undesirable limitations. 1800162911Ssimon [Darryl Miles via Richard Levitte and Bodo Moeller] 1801162911Ssimon 1802162911Ssimon *) Disable rogue ciphersuites: 1803162911Ssimon 1804162911Ssimon - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 1805162911Ssimon - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 1806162911Ssimon - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 1807162911Ssimon 1808162911Ssimon The latter two were purportedly from 1809162911Ssimon draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 1810162911Ssimon appear there. 1811162911Ssimon 1812162911Ssimon Also deactive the remaining ciphersuites from 1813162911Ssimon draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 1814162911Ssimon unofficial, and the ID has long expired. 1815162911Ssimon [Bodo Moeller] 1816162911Ssimon 1817162911Ssimon *) Fix RSA blinding Heisenbug (problems sometimes occured on 1818162911Ssimon dual-core machines) and other potential thread-safety issues. 1819162911Ssimon [Bodo Moeller] 1820162911Ssimon 1821162911Ssimon Changes between 0.9.7i and 0.9.7j [04 May 2006] 1822162911Ssimon 1823162911Ssimon *) Adapt fipsld and the build system to link against the validated FIPS 1824162911Ssimon module in FIPS mode. 1825162911Ssimon [Steve Henson] 1826162911Ssimon 1827162911Ssimon *) Fixes for VC++ 2005 build under Windows. 1828162911Ssimon [Steve Henson] 1829162911Ssimon 1830162911Ssimon *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make 1831162911Ssimon from a Windows bash shell such as MSYS. It is autodetected from the 1832162911Ssimon "config" script when run from a VC++ environment. Modify standard VC++ 1833162911Ssimon build to use fipscanister.o from the GNU make build. 1834162911Ssimon [Steve Henson] 1835162911Ssimon 1836160814Ssimon Changes between 0.9.7h and 0.9.7i [14 Oct 2005] 1837160814Ssimon 1838160814Ssimon *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS. 1839160814Ssimon The value now differs depending on if you build for FIPS or not. 1840160814Ssimon BEWARE! A program linked with a shared FIPSed libcrypto can't be 1841160814Ssimon safely run with a non-FIPSed libcrypto, as it may crash because of 1842160814Ssimon the difference induced by this change. 1843160814Ssimon [Andy Polyakov] 1844160814Ssimon 1845160814Ssimon Changes between 0.9.7g and 0.9.7h [11 Oct 2005] 1846160814Ssimon 1847160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 1848160814Ssimon (part of SSL_OP_ALL). This option used to disable the 1849160814Ssimon countermeasure against man-in-the-middle protocol-version 1850160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 1851160814Ssimon idea. (CVE-2005-2969) 1852160814Ssimon 1853160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 1854160814Ssimon for Information Security, National Institute of Advanced Industrial 1855160814Ssimon Science and Technology [AIST], Japan)] 1856160814Ssimon 1857160814Ssimon *) Minimal support for X9.31 signatures and PSS padding modes. This is 1858160814Ssimon mainly for FIPS compliance and not fully integrated at this stage. 1859160814Ssimon [Steve Henson] 1860160814Ssimon 1861160814Ssimon *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform 1862160814Ssimon the exponentiation using a fixed-length exponent. (Otherwise, 1863160814Ssimon the information leaked through timing could expose the secret key 1864160814Ssimon after many signatures; cf. Bleichenbacher's attack on DSA with 1865160814Ssimon biased k.) 1866160814Ssimon [Bodo Moeller] 1867160814Ssimon 1868160814Ssimon *) Make a new fixed-window mod_exp implementation the default for 1869160814Ssimon RSA, DSA, and DH private-key operations so that the sequence of 1870160814Ssimon squares and multiplies and the memory access pattern are 1871160814Ssimon independent of the particular secret key. This will mitigate 1872160814Ssimon cache-timing and potential related attacks. 1873160814Ssimon 1874160814Ssimon BN_mod_exp_mont_consttime() is the new exponentiation implementation, 1875160814Ssimon and this is automatically used by BN_mod_exp_mont() if the new flag 1876160814Ssimon BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH 1877160814Ssimon will use this BN flag for private exponents unless the flag 1878160814Ssimon RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or 1879160814Ssimon DH_FLAG_NO_EXP_CONSTTIME, respectively, is set. 1880160814Ssimon 1881160814Ssimon [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller] 1882160814Ssimon 1883160814Ssimon *) Change the client implementation for SSLv23_method() and 1884160814Ssimon SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0 1885160814Ssimon Client Hello message format if the SSL_OP_NO_SSLv2 option is set. 1886160814Ssimon (Previously, the SSL 2.0 backwards compatible Client Hello 1887160814Ssimon message format would be used even with SSL_OP_NO_SSLv2.) 1888160814Ssimon [Bodo Moeller] 1889160814Ssimon 1890160814Ssimon *) Add support for smime-type MIME parameter in S/MIME messages which some 1891160814Ssimon clients need. 1892160814Ssimon [Steve Henson] 1893160814Ssimon 1894160814Ssimon *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in 1895160814Ssimon a threadsafe manner. Modify rsa code to use new function and add calls 1896160814Ssimon to dsa and dh code (which had race conditions before). 1897160814Ssimon [Steve Henson] 1898160814Ssimon 1899160814Ssimon *) Include the fixed error library code in the C error file definitions 1900160814Ssimon instead of fixing them up at runtime. This keeps the error code 1901160814Ssimon structures constant. 1902160814Ssimon [Steve Henson] 1903160814Ssimon 1904160814Ssimon Changes between 0.9.7f and 0.9.7g [11 Apr 2005] 1905160814Ssimon 1906160814Ssimon [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after 1907160814Ssimon OpenSSL 0.9.8.] 1908160814Ssimon 1909160814Ssimon *) Fixes for newer kerberos headers. NB: the casts are needed because 1910160814Ssimon the 'length' field is signed on one version and unsigned on another 1911160814Ssimon with no (?) obvious way to tell the difference, without these VC++ 1912160814Ssimon complains. Also the "definition" of FAR (blank) is no longer included 1913160814Ssimon nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up 1914160814Ssimon some needed definitions. 1915160814Ssimon [Steve Henson] 1916160814Ssimon 1917160814Ssimon *) Undo Cygwin change. 1918160814Ssimon [Ulf M�ller] 1919160814Ssimon 1920160814Ssimon *) Added support for proxy certificates according to RFC 3820. 1921160814Ssimon Because they may be a security thread to unaware applications, 1922160814Ssimon they must be explicitely allowed in run-time. See 1923160814Ssimon docs/HOWTO/proxy_certificates.txt for further information. 1924160814Ssimon [Richard Levitte] 1925160814Ssimon 1926160814Ssimon Changes between 0.9.7e and 0.9.7f [22 Mar 2005] 1927160814Ssimon 1928160814Ssimon *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating 1929160814Ssimon server and client random values. Previously 1930160814Ssimon (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in 1931160814Ssimon less random data when sizeof(time_t) > 4 (some 64 bit platforms). 1932160814Ssimon 1933160814Ssimon This change has negligible security impact because: 1934160814Ssimon 1935160814Ssimon 1. Server and client random values still have 24 bytes of pseudo random 1936160814Ssimon data. 1937160814Ssimon 1938160814Ssimon 2. Server and client random values are sent in the clear in the initial 1939160814Ssimon handshake. 1940160814Ssimon 1941160814Ssimon 3. The master secret is derived using the premaster secret (48 bytes in 1942160814Ssimon size for static RSA ciphersuites) as well as client server and random 1943160814Ssimon values. 1944160814Ssimon 1945160814Ssimon The OpenSSL team would like to thank the UK NISCC for bringing this issue 1946160814Ssimon to our attention. 1947160814Ssimon 1948160814Ssimon [Stephen Henson, reported by UK NISCC] 1949160814Ssimon 1950160814Ssimon *) Use Windows randomness collection on Cygwin. 1951160814Ssimon [Ulf M�ller] 1952160814Ssimon 1953160814Ssimon *) Fix hang in EGD/PRNGD query when communication socket is closed 1954160814Ssimon prematurely by EGD/PRNGD. 1955160814Ssimon [Darren Tucker <dtucker@zip.com.au> via Lutz J�nicke, resolves #1014] 1956160814Ssimon 1957160814Ssimon *) Prompt for pass phrases when appropriate for PKCS12 input format. 1958160814Ssimon [Steve Henson] 1959160814Ssimon 1960160814Ssimon *) Back-port of selected performance improvements from development 1961160814Ssimon branch, as well as improved support for PowerPC platforms. 1962160814Ssimon [Andy Polyakov] 1963160814Ssimon 1964160814Ssimon *) Add lots of checks for memory allocation failure, error codes to indicate 1965160814Ssimon failure and freeing up memory if a failure occurs. 1966160814Ssimon [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson] 1967160814Ssimon 1968160814Ssimon *) Add new -passin argument to dgst. 1969160814Ssimon [Steve Henson] 1970160814Ssimon 1971160814Ssimon *) Perform some character comparisons of different types in X509_NAME_cmp: 1972160814Ssimon this is needed for some certificates that reencode DNs into UTF8Strings 1973160814Ssimon (in violation of RFC3280) and can't or wont issue name rollover 1974160814Ssimon certificates. 1975160814Ssimon [Steve Henson] 1976160814Ssimon 1977160814Ssimon *) Make an explicit check during certificate validation to see that 1978160814Ssimon the CA setting in each certificate on the chain is correct. As a 1979160814Ssimon side effect always do the following basic checks on extensions, 1980160814Ssimon not just when there's an associated purpose to the check: 1981160814Ssimon 1982160814Ssimon - if there is an unhandled critical extension (unless the user 1983160814Ssimon has chosen to ignore this fault) 1984160814Ssimon - if the path length has been exceeded (if one is set at all) 1985160814Ssimon - that certain extensions fit the associated purpose (if one has 1986160814Ssimon been given) 1987160814Ssimon [Richard Levitte] 1988160814Ssimon 1989142425Snectar Changes between 0.9.7d and 0.9.7e [25 Oct 2004] 1990142425Snectar 1991142425Snectar *) Avoid a race condition when CRLs are checked in a multi threaded 1992142425Snectar environment. This would happen due to the reordering of the revoked 1993142425Snectar entries during signature checking and serial number lookup. Now the 1994142425Snectar encoding is cached and the serial number sort performed under a lock. 1995142425Snectar Add new STACK function sk_is_sorted(). 1996142425Snectar [Steve Henson] 1997142425Snectar 1998142425Snectar *) Add Delta CRL to the extension code. 1999142425Snectar [Steve Henson] 2000142425Snectar 2001142425Snectar *) Various fixes to s3_pkt.c so alerts are sent properly. 2002142425Snectar [David Holmes <d.holmes@f5.com>] 2003142425Snectar 2004142425Snectar *) Reduce the chances of duplicate issuer name and serial numbers (in 2005142425Snectar violation of RFC3280) using the OpenSSL certificate creation utilities. 2006142425Snectar This is done by creating a random 64 bit value for the initial serial 2007142425Snectar number when a serial number file is created or when a self signed 2008142425Snectar certificate is created using 'openssl req -x509'. The initial serial 2009142425Snectar number file is created using 'openssl x509 -next_serial' in CA.pl 2010142425Snectar rather than being initialized to 1. 2011142425Snectar [Steve Henson] 2012142425Snectar 2013127128Snectar Changes between 0.9.7c and 0.9.7d [17 Mar 2004] 2014127128Snectar 2015127128Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 2016160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 2017127128Snectar [Joe Orton, Steve Henson] 2018127128Snectar 2019127128Snectar *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites 2020160814Ssimon (CVE-2004-0112) 2021127128Snectar [Joe Orton, Steve Henson] 2022127128Snectar 2023127128Snectar *) Make it possible to have multiple active certificates with the same 2024127128Snectar subject in the CA index file. This is done only if the keyword 2025127128Snectar 'unique_subject' is set to 'no' in the main CA section (default 2026127128Snectar if 'CA_default') of the configuration file. The value is saved 2027127128Snectar with the database itself in a separate index attribute file, 2028127128Snectar named like the index file with '.attr' appended to the name. 2029127128Snectar [Richard Levitte] 2030127128Snectar 2031127128Snectar *) X509 verify fixes. Disable broken certificate workarounds when 2032127128Snectar X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if 2033127128Snectar keyUsage extension present. Don't accept CRLs with unhandled critical 2034127128Snectar extensions: since verify currently doesn't process CRL extensions this 2035127128Snectar rejects a CRL with *any* critical extensions. Add new verify error codes 2036127128Snectar for these cases. 2037127128Snectar [Steve Henson] 2038127128Snectar 2039127128Snectar *) When creating an OCSP nonce use an OCTET STRING inside the extnValue. 2040127128Snectar A clarification of RFC2560 will require the use of OCTET STRINGs and 2041127128Snectar some implementations cannot handle the current raw format. Since OpenSSL 2042127128Snectar copies and compares OCSP nonces as opaque blobs without any attempt at 2043127128Snectar parsing them this should not create any compatibility issues. 2044127128Snectar [Steve Henson] 2045127128Snectar 2046127128Snectar *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when 2047127128Snectar calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without 2048127128Snectar this HMAC (and other) operations are several times slower than OpenSSL 2049127128Snectar < 0.9.7. 2050127128Snectar [Steve Henson] 2051127128Snectar 2052127128Snectar *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex(). 2053127128Snectar [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 2054127128Snectar 2055127128Snectar *) Use the correct content when signing type "other". 2056127128Snectar [Steve Henson] 2057127128Snectar 2058120631Snectar Changes between 0.9.7b and 0.9.7c [30 Sep 2003] 2059120631Snectar 2060120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 2061120631Snectar 2062120631Snectar Stop out of bounds reads in the ASN1 code when presented with 2063160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 2064120631Snectar 2065160814Ssimon Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545). 2066120631Snectar 2067120631Snectar If verify callback ignores invalid public key errors don't try to check 2068120631Snectar certificate signature with the NULL public key. 2069120631Snectar 2070120631Snectar [Steve Henson] 2071120631Snectar 2072120631Snectar *) New -ignore_err option in ocsp application to stop the server 2073120631Snectar exiting on the first error in a request. 2074120631Snectar [Steve Henson] 2075120631Snectar 2076120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 2077120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 2078120631Snectar specifications. 2079120631Snectar [Steve Henson] 2080120631Snectar 2081120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 2082120631Snectar extra data after the compression methods not only for TLS 1.0 2083120631Snectar but also for SSL 3.0 (as required by the specification). 2084120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 2085120631Snectar 2086120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 2087120631Snectar when it's 512 *bits* long, not 512 bytes. 2088120631Snectar [Richard Levitte] 2089120631Snectar 2090120631Snectar *) Change AES_cbc_encrypt() so it outputs exact multiple of 2091120631Snectar blocks during encryption. 2092120631Snectar [Richard Levitte] 2093120631Snectar 2094120631Snectar *) Various fixes to base64 BIO and non blocking I/O. On write 2095120631Snectar flushes were not handled properly if the BIO retried. On read 2096120631Snectar data was not being buffered properly and had various logic bugs. 2097120631Snectar This also affects blocking I/O when the data being decoded is a 2098120631Snectar certain size. 2099120631Snectar [Steve Henson] 2100120631Snectar 2101120631Snectar *) Various S/MIME bugfixes and compatibility changes: 2102120631Snectar output correct application/pkcs7 MIME type if 2103120631Snectar PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures. 2104120631Snectar Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening 2105120631Snectar of files as .eml work). Correctly handle very long lines in MIME 2106120631Snectar parser. 2107120631Snectar [Steve Henson] 2108120631Snectar 2109120631Snectar Changes between 0.9.7a and 0.9.7b [10 Apr 2003] 2110120631Snectar 2111120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 2112120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 2113120631Snectar a protocol version number mismatch like a decryption error 2114120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 2115120631Snectar [Bodo Moeller] 2116120631Snectar 2117120631Snectar *) Turn on RSA blinding by default in the default implementation 2118120631Snectar to avoid a timing attack. Applications that don't want it can call 2119120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 2120120631Snectar They would be ill-advised to do so in most cases. 2121120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 2122120631Snectar 2123120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 2124120631Snectar seeded (in this case, the secret RSA exponent is abused as 2125120631Snectar an unpredictable seed -- if it is not unpredictable, there 2126120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 2127120631Snectar by remembering the creator's thread ID in rsa->blinding and 2128120631Snectar having all other threads use local one-time blinding factors 2129120631Snectar (this requires more computation than sharing rsa->blinding, but 2130120631Snectar avoids excessive locking; and if an RSA object is not shared 2131120631Snectar between threads, blinding will still be very fast). 2132120631Snectar [Bodo Moeller] 2133120631Snectar 2134120631Snectar *) Fixed a typo bug that would cause ENGINE_set_default() to set an 2135120631Snectar ENGINE as defaults for all supported algorithms irrespective of 2136120631Snectar the 'flags' parameter. 'flags' is now honoured, so applications 2137120631Snectar should make sure they are passing it correctly. 2138120631Snectar [Geoff Thorpe] 2139120631Snectar 2140120631Snectar *) Target "mingw" now allows native Windows code to be generated in 2141120631Snectar the Cygwin environment as well as with the MinGW compiler. 2142120631Snectar [Ulf Moeller] 2143120631Snectar 2144111147Snectar Changes between 0.9.7 and 0.9.7a [19 Feb 2003] 2145111147Snectar 2146111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 2147111147Snectar via timing by performing a MAC computation even if incorrrect 2148111147Snectar block cipher padding has been found. This is a countermeasure 2149111147Snectar against active attacks where the attacker has to distinguish 2150160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 2151111147Snectar 2152111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 2153111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 2154111147Snectar Martin Vuagnoux (EPFL, Ilion)] 2155111147Snectar 2156111147Snectar *) Make the no-err option work as intended. The intention with no-err 2157111147Snectar is not to have the whole error stack handling routines removed from 2158111147Snectar libcrypto, it's only intended to remove all the function name and 2159111147Snectar reason texts, thereby removing some of the footprint that may not 2160111147Snectar be interesting if those errors aren't displayed anyway. 2161111147Snectar 2162111147Snectar NOTE: it's still possible for any application or module to have it's 2163111147Snectar own set of error texts inserted. The routines are there, just not 2164111147Snectar used by default when no-err is given. 2165111147Snectar [Richard Levitte] 2166111147Snectar 2167111147Snectar *) Add support for FreeBSD on IA64. 2168111147Snectar [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454] 2169111147Snectar 2170111147Snectar *) Adjust DES_cbc_cksum() so it returns the same value as the MIT 2171111147Snectar Kerberos function mit_des_cbc_cksum(). Before this change, 2172111147Snectar the value returned by DES_cbc_cksum() was like the one from 2173111147Snectar mit_des_cbc_cksum(), except the bytes were swapped. 2174111147Snectar [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte] 2175111147Snectar 2176111147Snectar *) Allow an application to disable the automatic SSL chain building. 2177111147Snectar Before this a rather primitive chain build was always performed in 2178111147Snectar ssl3_output_cert_chain(): an application had no way to send the 2179111147Snectar correct chain if the automatic operation produced an incorrect result. 2180111147Snectar 2181111147Snectar Now the chain builder is disabled if either: 2182111147Snectar 2183111147Snectar 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert(). 2184111147Snectar 2185111147Snectar 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set. 2186111147Snectar 2187111147Snectar The reasoning behind this is that an application would not want the 2188111147Snectar auto chain building to take place if extra chain certificates are 2189111147Snectar present and it might also want a means of sending no additional 2190111147Snectar certificates (for example the chain has two certificates and the 2191111147Snectar root is omitted). 2192111147Snectar [Steve Henson] 2193111147Snectar 2194111147Snectar *) Add the possibility to build without the ENGINE framework. 2195111147Snectar [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 2196111147Snectar 2197111147Snectar *) Under Win32 gmtime() can return NULL: check return value in 2198111147Snectar OPENSSL_gmtime(). Add error code for case where gmtime() fails. 2199111147Snectar [Steve Henson] 2200111147Snectar 2201111147Snectar *) DSA routines: under certain error conditions uninitialized BN objects 2202111147Snectar could be freed. Solution: make sure initialization is performed early 2203111147Snectar enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>, 2204111147Snectar Nils Larsch <nla@trustcenter.de> via PR#459) 2205111147Snectar [Lutz Jaenicke] 2206111147Snectar 2207111147Snectar *) Another fix for SSLv2 session ID handling: the session ID was incorrectly 2208111147Snectar checked on reconnect on the client side, therefore session resumption 2209111147Snectar could still fail with a "ssl session id is different" error. This 2210111147Snectar behaviour is masked when SSL_OP_ALL is used due to 2211111147Snectar SSL_OP_MICROSOFT_SESS_ID_BUG being set. 2212111147Snectar Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 2213111147Snectar followup to PR #377. 2214111147Snectar [Lutz Jaenicke] 2215111147Snectar 2216111147Snectar *) IA-32 assembler support enhancements: unified ELF targets, support 2217111147Snectar for SCO/Caldera platforms, fix for Cygwin shared build. 2218111147Snectar [Andy Polyakov] 2219111147Snectar 2220111147Snectar *) Add support for FreeBSD on sparc64. As a consequence, support for 2221111147Snectar FreeBSD on non-x86 processors is separate from x86 processors on 2222111147Snectar the config script, much like the NetBSD support. 2223111147Snectar [Richard Levitte & Kris Kennaway <kris@obsecurity.org>] 2224111147Snectar 2225109998Smarkm Changes between 0.9.6h and 0.9.7 [31 Dec 2002] 2226109998Smarkm 2227120631Snectar [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after 2228120631Snectar OpenSSL 0.9.7.] 2229120631Snectar 2230109998Smarkm *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED 2231109998Smarkm code (06) was taken as the first octet of the session ID and the last 2232109998Smarkm octet was ignored consequently. As a result SSLv2 client side session 2233109998Smarkm caching could not have worked due to the session ID mismatch between 2234109998Smarkm client and server. 2235109998Smarkm Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 2236109998Smarkm PR #377. 2237109998Smarkm [Lutz Jaenicke] 2238109998Smarkm 2239109998Smarkm *) Change the declaration of needed Kerberos libraries to use EX_LIBS 2240109998Smarkm instead of the special (and badly supported) LIBKRB5. LIBKRB5 is 2241109998Smarkm removed entirely. 2242109998Smarkm [Richard Levitte] 2243109998Smarkm 2244109998Smarkm *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it 2245109998Smarkm seems that in spite of existing for more than a year, many application 2246109998Smarkm author have done nothing to provide the necessary callbacks, which 2247109998Smarkm means that this particular engine will not work properly anywhere. 2248109998Smarkm This is a very unfortunate situation which forces us, in the name 2249109998Smarkm of usability, to give the hw_ncipher.c a static lock, which is part 2250109998Smarkm of libcrypto. 2251109998Smarkm NOTE: This is for the 0.9.7 series ONLY. This hack will never 2252109998Smarkm appear in 0.9.8 or later. We EXPECT application authors to have 2253109998Smarkm dealt properly with this when 0.9.8 is released (unless we actually 2254109998Smarkm make such changes in the libcrypto locking code that changes will 2255109998Smarkm have to be made anyway). 2256109998Smarkm [Richard Levitte] 2257109998Smarkm 2258109998Smarkm *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content 2259109998Smarkm octets have been read, EOF or an error occurs. Without this change 2260109998Smarkm some truncated ASN1 structures will not produce an error. 2261109998Smarkm [Steve Henson] 2262109998Smarkm 2263109998Smarkm *) Disable Heimdal support, since it hasn't been fully implemented. 2264109998Smarkm Still give the possibility to force the use of Heimdal, but with 2265109998Smarkm warnings and a request that patches get sent to openssl-dev. 2266109998Smarkm [Richard Levitte] 2267109998Smarkm 2268109998Smarkm *) Add the VC-CE target, introduce the WINCE sysname, and add 2269109998Smarkm INSTALL.WCE and appropriate conditionals to make it build. 2270109998Smarkm [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 2271109998Smarkm 2272109998Smarkm *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and 2273109998Smarkm cygssl-x.y.z.dll, where x, y and z are the major, minor and 2274109998Smarkm edit numbers of the version. 2275109998Smarkm [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 2276109998Smarkm 2277109998Smarkm *) Introduce safe string copy and catenation functions 2278109998Smarkm (BUF_strlcpy() and BUF_strlcat()). 2279109998Smarkm [Ben Laurie (CHATS) and Richard Levitte] 2280109998Smarkm 2281109998Smarkm *) Avoid using fixed-size buffers for one-line DNs. 2282109998Smarkm [Ben Laurie (CHATS)] 2283109998Smarkm 2284109998Smarkm *) Add BUF_MEM_grow_clean() to avoid information leakage when 2285109998Smarkm resizing buffers containing secrets, and use where appropriate. 2286109998Smarkm [Ben Laurie (CHATS)] 2287109998Smarkm 2288109998Smarkm *) Avoid using fixed size buffers for configuration file location. 2289109998Smarkm [Ben Laurie (CHATS)] 2290109998Smarkm 2291109998Smarkm *) Avoid filename truncation for various CA files. 2292109998Smarkm [Ben Laurie (CHATS)] 2293109998Smarkm 2294109998Smarkm *) Use sizeof in preference to magic numbers. 2295109998Smarkm [Ben Laurie (CHATS)] 2296109998Smarkm 2297109998Smarkm *) Avoid filename truncation in cert requests. 2298109998Smarkm [Ben Laurie (CHATS)] 2299109998Smarkm 2300109998Smarkm *) Add assertions to check for (supposedly impossible) buffer 2301109998Smarkm overflows. 2302109998Smarkm [Ben Laurie (CHATS)] 2303109998Smarkm 2304109998Smarkm *) Don't cache truncated DNS entries in the local cache (this could 2305109998Smarkm potentially lead to a spoofing attack). 2306109998Smarkm [Ben Laurie (CHATS)] 2307109998Smarkm 2308109998Smarkm *) Fix various buffers to be large enough for hex/decimal 2309109998Smarkm representations in a platform independent manner. 2310109998Smarkm [Ben Laurie (CHATS)] 2311109998Smarkm 2312109998Smarkm *) Add CRYPTO_realloc_clean() to avoid information leakage when 2313109998Smarkm resizing buffers containing secrets, and use where appropriate. 2314109998Smarkm [Ben Laurie (CHATS)] 2315109998Smarkm 2316109998Smarkm *) Add BIO_indent() to avoid much slightly worrying code to do 2317109998Smarkm indents. 2318109998Smarkm [Ben Laurie (CHATS)] 2319109998Smarkm 2320109998Smarkm *) Convert sprintf()/BIO_puts() to BIO_printf(). 2321109998Smarkm [Ben Laurie (CHATS)] 2322109998Smarkm 2323109998Smarkm *) buffer_gets() could terminate with the buffer only half 2324109998Smarkm full. Fixed. 2325109998Smarkm [Ben Laurie (CHATS)] 2326109998Smarkm 2327109998Smarkm *) Add assertions to prevent user-supplied crypto functions from 2328109998Smarkm overflowing internal buffers by having large block sizes, etc. 2329109998Smarkm [Ben Laurie (CHATS)] 2330109998Smarkm 2331109998Smarkm *) New OPENSSL_assert() macro (similar to assert(), but enabled 2332109998Smarkm unconditionally). 2333109998Smarkm [Ben Laurie (CHATS)] 2334109998Smarkm 2335109998Smarkm *) Eliminate unused copy of key in RC4. 2336109998Smarkm [Ben Laurie (CHATS)] 2337109998Smarkm 2338109998Smarkm *) Eliminate unused and incorrectly sized buffers for IV in pem.h. 2339109998Smarkm [Ben Laurie (CHATS)] 2340109998Smarkm 2341109998Smarkm *) Fix off-by-one error in EGD path. 2342109998Smarkm [Ben Laurie (CHATS)] 2343109998Smarkm 2344109998Smarkm *) If RANDFILE path is too long, ignore instead of truncating. 2345109998Smarkm [Ben Laurie (CHATS)] 2346109998Smarkm 2347109998Smarkm *) Eliminate unused and incorrectly sized X.509 structure 2348109998Smarkm CBCParameter. 2349109998Smarkm [Ben Laurie (CHATS)] 2350109998Smarkm 2351109998Smarkm *) Eliminate unused and dangerous function knumber(). 2352109998Smarkm [Ben Laurie (CHATS)] 2353109998Smarkm 2354109998Smarkm *) Eliminate unused and dangerous structure, KSSL_ERR. 2355109998Smarkm [Ben Laurie (CHATS)] 2356109998Smarkm 2357109998Smarkm *) Protect against overlong session ID context length in an encoded 2358109998Smarkm session object. Since these are local, this does not appear to be 2359109998Smarkm exploitable. 2360109998Smarkm [Ben Laurie (CHATS)] 2361109998Smarkm 2362109998Smarkm *) Change from security patch (see 0.9.6e below) that did not affect 2363109998Smarkm the 0.9.6 release series: 2364109998Smarkm 2365109998Smarkm Remote buffer overflow in SSL3 protocol - an attacker could 2366109998Smarkm supply an oversized master key in Kerberos-enabled versions. 2367160814Ssimon (CVE-2002-0657) 2368109998Smarkm [Ben Laurie (CHATS)] 2369109998Smarkm 2370109998Smarkm *) Change the SSL kerb5 codes to match RFC 2712. 2371109998Smarkm [Richard Levitte] 2372109998Smarkm 2373109998Smarkm *) Make -nameopt work fully for req and add -reqopt switch. 2374109998Smarkm [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson] 2375109998Smarkm 2376109998Smarkm *) The "block size" for block ciphers in CFB and OFB mode should be 1. 2377109998Smarkm [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>] 2378109998Smarkm 2379109998Smarkm *) Make sure tests can be performed even if the corresponding algorithms 2380109998Smarkm have been removed entirely. This was also the last step to make 2381109998Smarkm OpenSSL compilable with DJGPP under all reasonable conditions. 2382109998Smarkm [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>] 2383109998Smarkm 2384109998Smarkm *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT 2385109998Smarkm to allow version independent disabling of normally unselected ciphers, 2386109998Smarkm which may be activated as a side-effect of selecting a single cipher. 2387109998Smarkm 2388109998Smarkm (E.g., cipher list string "RSA" enables ciphersuites that are left 2389109998Smarkm out of "ALL" because they do not provide symmetric encryption. 2390109998Smarkm "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) 2391109998Smarkm [Lutz Jaenicke, Bodo Moeller] 2392109998Smarkm 2393109998Smarkm *) Add appropriate support for separate platform-dependent build 2394109998Smarkm directories. The recommended way to make a platform-dependent 2395109998Smarkm build directory is the following (tested on Linux), maybe with 2396109998Smarkm some local tweaks: 2397109998Smarkm 2398109998Smarkm # Place yourself outside of the OpenSSL source tree. In 2399109998Smarkm # this example, the environment variable OPENSSL_SOURCE 2400109998Smarkm # is assumed to contain the absolute OpenSSL source directory. 2401109998Smarkm mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" 2402109998Smarkm cd objtree/"`uname -s`-`uname -r`-`uname -m`" 2403111147Snectar (cd $OPENSSL_SOURCE; find . -type f) | while read F; do 2404109998Smarkm mkdir -p `dirname $F` 2405109998Smarkm ln -s $OPENSSL_SOURCE/$F $F 2406109998Smarkm done 2407109998Smarkm 2408109998Smarkm To be absolutely sure not to disturb the source tree, a "make clean" 2409109998Smarkm is a good thing. If it isn't successfull, don't worry about it, 2410109998Smarkm it probably means the source directory is very clean. 2411109998Smarkm [Richard Levitte] 2412109998Smarkm 2413109998Smarkm *) Make sure any ENGINE control commands make local copies of string 2414109998Smarkm pointers passed to them whenever necessary. Otherwise it is possible 2415109998Smarkm the caller may have overwritten (or deallocated) the original string 2416109998Smarkm data when a later ENGINE operation tries to use the stored values. 2417109998Smarkm [G�tz Babin-Ebell <babinebell@trustcenter.de>] 2418109998Smarkm 2419109998Smarkm *) Improve diagnostics in file reading and command-line digests. 2420109998Smarkm [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>] 2421109998Smarkm 2422109998Smarkm *) Add AES modes CFB and OFB to the object database. Correct an 2423109998Smarkm error in AES-CFB decryption. 2424109998Smarkm [Richard Levitte] 2425109998Smarkm 2426109998Smarkm *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this 2427109998Smarkm allows existing EVP_CIPHER_CTX structures to be reused after 2428109998Smarkm calling EVP_*Final(). This behaviour is used by encryption 2429109998Smarkm BIOs and some applications. This has the side effect that 2430109998Smarkm applications must explicitly clean up cipher contexts with 2431109998Smarkm EVP_CIPHER_CTX_cleanup() or they will leak memory. 2432109998Smarkm [Steve Henson] 2433109998Smarkm 2434109998Smarkm *) Check the values of dna and dnb in bn_mul_recursive before calling 2435109998Smarkm bn_mul_comba (a non zero value means the a or b arrays do not contain 2436109998Smarkm n2 elements) and fallback to bn_mul_normal if either is not zero. 2437109998Smarkm [Steve Henson] 2438109998Smarkm 2439109998Smarkm *) Fix escaping of non-ASCII characters when using the -subj option 2440109998Smarkm of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) 2441109998Smarkm [Lutz Jaenicke] 2442109998Smarkm 2443109998Smarkm *) Make object definitions compliant to LDAP (RFC2256): SN is the short 2444109998Smarkm form for "surname", serialNumber has no short form. 2445109998Smarkm Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; 2446109998Smarkm therefore remove "mail" short name for "internet 7". 2447109998Smarkm The OID for unique identifiers in X509 certificates is 2448109998Smarkm x500UniqueIdentifier, not uniqueIdentifier. 2449109998Smarkm Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) 2450109998Smarkm [Lutz Jaenicke] 2451109998Smarkm 2452109998Smarkm *) Add an "init" command to the ENGINE config module and auto initialize 2453109998Smarkm ENGINEs. Without any "init" command the ENGINE will be initialized 2454109998Smarkm after all ctrl commands have been executed on it. If init=1 the 2455109998Smarkm ENGINE is initailized at that point (ctrls before that point are run 2456109998Smarkm on the uninitialized ENGINE and after on the initialized one). If 2457109998Smarkm init=0 then the ENGINE will not be iniatialized at all. 2458109998Smarkm [Steve Henson] 2459109998Smarkm 2460109998Smarkm *) Fix the 'app_verify_callback' interface so that the user-defined 2461109998Smarkm argument is actually passed to the callback: In the 2462109998Smarkm SSL_CTX_set_cert_verify_callback() prototype, the callback 2463109998Smarkm declaration has been changed from 2464109998Smarkm int (*cb)() 2465109998Smarkm into 2466109998Smarkm int (*cb)(X509_STORE_CTX *,void *); 2467109998Smarkm in ssl_verify_cert_chain (ssl/ssl_cert.c), the call 2468109998Smarkm i=s->ctx->app_verify_callback(&ctx) 2469109998Smarkm has been changed into 2470109998Smarkm i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). 2471109998Smarkm 2472109998Smarkm To update applications using SSL_CTX_set_cert_verify_callback(), 2473109998Smarkm a dummy argument can be added to their callback functions. 2474109998Smarkm [D. K. Smetters <smetters@parc.xerox.com>] 2475109998Smarkm 2476109998Smarkm *) Added the '4758cca' ENGINE to support IBM 4758 cards. 2477109998Smarkm [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe] 2478109998Smarkm 2479109998Smarkm *) Add and OPENSSL_LOAD_CONF define which will cause 2480109998Smarkm OpenSSL_add_all_algorithms() to load the openssl.cnf config file. 2481109998Smarkm This allows older applications to transparently support certain 2482109998Smarkm OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. 2483109998Smarkm Two new functions OPENSSL_add_all_algorithms_noconf() which will never 2484109998Smarkm load the config file and OPENSSL_add_all_algorithms_conf() which will 2485109998Smarkm always load it have also been added. 2486109998Smarkm [Steve Henson] 2487109998Smarkm 2488109998Smarkm *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. 2489109998Smarkm Adjust NIDs and EVP layer. 2490109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 2491109998Smarkm 2492109998Smarkm *) Config modules support in openssl utility. 2493109998Smarkm 2494109998Smarkm Most commands now load modules from the config file, 2495109998Smarkm though in a few (such as version) this isn't done 2496109998Smarkm because it couldn't be used for anything. 2497109998Smarkm 2498109998Smarkm In the case of ca and req the config file used is 2499109998Smarkm the same as the utility itself: that is the -config 2500109998Smarkm command line option can be used to specify an 2501109998Smarkm alternative file. 2502109998Smarkm [Steve Henson] 2503109998Smarkm 2504109998Smarkm *) Move default behaviour from OPENSSL_config(). If appname is NULL 2505109998Smarkm use "openssl_conf" if filename is NULL use default openssl config file. 2506109998Smarkm [Steve Henson] 2507109998Smarkm 2508109998Smarkm *) Add an argument to OPENSSL_config() to allow the use of an alternative 2509109998Smarkm config section name. Add a new flag to tolerate a missing config file 2510109998Smarkm and move code to CONF_modules_load_file(). 2511109998Smarkm [Steve Henson] 2512109998Smarkm 2513109998Smarkm *) Support for crypto accelerator cards from Accelerated Encryption 2514109998Smarkm Processing, www.aep.ie. (Use engine 'aep') 2515109998Smarkm The support was copied from 0.9.6c [engine] and adapted/corrected 2516109998Smarkm to work with the new engine framework. 2517109998Smarkm [AEP Inc. and Richard Levitte] 2518109998Smarkm 2519109998Smarkm *) Support for SureWare crypto accelerator cards from Baltimore 2520109998Smarkm Technologies. (Use engine 'sureware') 2521109998Smarkm The support was copied from 0.9.6c [engine] and adapted 2522109998Smarkm to work with the new engine framework. 2523109998Smarkm [Richard Levitte] 2524109998Smarkm 2525109998Smarkm *) Have the CHIL engine fork-safe (as defined by nCipher) and actually 2526109998Smarkm make the newer ENGINE framework commands for the CHIL engine work. 2527109998Smarkm [Toomas Kiisk <vix@cyber.ee> and Richard Levitte] 2528109998Smarkm 2529109998Smarkm *) Make it possible to produce shared libraries on ReliantUNIX. 2530109998Smarkm [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte] 2531109998Smarkm 2532109998Smarkm *) Add the configuration target debug-linux-ppro. 2533109998Smarkm Make 'openssl rsa' use the general key loading routines 2534109998Smarkm implemented in apps.c, and make those routines able to 2535109998Smarkm handle the key format FORMAT_NETSCAPE and the variant 2536109998Smarkm FORMAT_IISSGC. 2537109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 2538109998Smarkm 2539109998Smarkm *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 2540109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 2541109998Smarkm 2542109998Smarkm *) Add -keyform to rsautl, and document -engine. 2543109998Smarkm [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>] 2544109998Smarkm 2545109998Smarkm *) Change BIO_new_file (crypto/bio/bss_file.c) to use new 2546109998Smarkm BIO_R_NO_SUCH_FILE error code rather than the generic 2547109998Smarkm ERR_R_SYS_LIB error code if fopen() fails with ENOENT. 2548109998Smarkm [Ben Laurie] 2549109998Smarkm 2550109998Smarkm *) Add new functions 2551109998Smarkm ERR_peek_last_error 2552109998Smarkm ERR_peek_last_error_line 2553109998Smarkm ERR_peek_last_error_line_data. 2554109998Smarkm These are similar to 2555109998Smarkm ERR_peek_error 2556109998Smarkm ERR_peek_error_line 2557109998Smarkm ERR_peek_error_line_data, 2558109998Smarkm but report on the latest error recorded rather than the first one 2559109998Smarkm still in the error queue. 2560109998Smarkm [Ben Laurie, Bodo Moeller] 2561109998Smarkm 2562109998Smarkm *) default_algorithms option in ENGINE config module. This allows things 2563109998Smarkm like: 2564109998Smarkm default_algorithms = ALL 2565109998Smarkm default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS 2566109998Smarkm [Steve Henson] 2567109998Smarkm 2568109998Smarkm *) Prelminary ENGINE config module. 2569109998Smarkm [Steve Henson] 2570109998Smarkm 2571109998Smarkm *) New experimental application configuration code. 2572109998Smarkm [Steve Henson] 2573109998Smarkm 2574109998Smarkm *) Change the AES code to follow the same name structure as all other 2575109998Smarkm symmetric ciphers, and behave the same way. Move everything to 2576109998Smarkm the directory crypto/aes, thereby obsoleting crypto/rijndael. 2577109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 2578109998Smarkm 2579109998Smarkm *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. 2580109998Smarkm [Ben Laurie and Theo de Raadt] 2581109998Smarkm 2582109998Smarkm *) Add option to output public keys in req command. 2583109998Smarkm [Massimiliano Pala madwolf@openca.org] 2584109998Smarkm 2585109998Smarkm *) Use wNAFs in EC_POINTs_mul() for improved efficiency 2586109998Smarkm (up to about 10% better than before for P-192 and P-224). 2587109998Smarkm [Bodo Moeller] 2588109998Smarkm 2589109998Smarkm *) New functions/macros 2590109998Smarkm 2591109998Smarkm SSL_CTX_set_msg_callback(ctx, cb) 2592109998Smarkm SSL_CTX_set_msg_callback_arg(ctx, arg) 2593109998Smarkm SSL_set_msg_callback(ssl, cb) 2594109998Smarkm SSL_set_msg_callback_arg(ssl, arg) 2595109998Smarkm 2596109998Smarkm to request calling a callback function 2597109998Smarkm 2598109998Smarkm void cb(int write_p, int version, int content_type, 2599109998Smarkm const void *buf, size_t len, SSL *ssl, void *arg) 2600109998Smarkm 2601109998Smarkm whenever a protocol message has been completely received 2602109998Smarkm (write_p == 0) or sent (write_p == 1). Here 'version' is the 2603109998Smarkm protocol version according to which the SSL library interprets 2604109998Smarkm the current protocol message (SSL2_VERSION, SSL3_VERSION, or 2605109998Smarkm TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or 2606109998Smarkm the content type as defined in the SSL 3.0/TLS 1.0 protocol 2607109998Smarkm specification (change_cipher_spec(20), alert(21), handshake(22)). 2608109998Smarkm 'buf' and 'len' point to the actual message, 'ssl' to the 2609109998Smarkm SSL object, and 'arg' is the application-defined value set by 2610109998Smarkm SSL[_CTX]_set_msg_callback_arg(). 2611109998Smarkm 2612109998Smarkm 'openssl s_client' and 'openssl s_server' have new '-msg' options 2613109998Smarkm to enable a callback that displays all protocol messages. 2614109998Smarkm [Bodo Moeller] 2615109998Smarkm 2616109998Smarkm *) Change the shared library support so shared libraries are built as 2617109998Smarkm soon as the corresponding static library is finished, and thereby get 2618109998Smarkm openssl and the test programs linked against the shared library. 2619109998Smarkm This still only happens when the keyword "shard" has been given to 2620109998Smarkm the configuration scripts. 2621109998Smarkm 2622109998Smarkm NOTE: shared library support is still an experimental thing, and 2623109998Smarkm backward binary compatibility is still not guaranteed. 2624109998Smarkm ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte] 2625109998Smarkm 2626109998Smarkm *) Add support for Subject Information Access extension. 2627109998Smarkm [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 2628109998Smarkm 2629109998Smarkm *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero 2630109998Smarkm additional bytes when new memory had to be allocated, not just 2631109998Smarkm when reusing an existing buffer. 2632109998Smarkm [Bodo Moeller] 2633109998Smarkm 2634109998Smarkm *) New command line and configuration option 'utf8' for the req command. 2635109998Smarkm This allows field values to be specified as UTF8 strings. 2636109998Smarkm [Steve Henson] 2637109998Smarkm 2638109998Smarkm *) Add -multi and -mr options to "openssl speed" - giving multiple parallel 2639109998Smarkm runs for the former and machine-readable output for the latter. 2640109998Smarkm [Ben Laurie] 2641109998Smarkm 2642109998Smarkm *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion 2643109998Smarkm of the e-mail address in the DN (i.e., it will go into a certificate 2644109998Smarkm extension only). The new configuration file option 'email_in_dn = no' 2645109998Smarkm has the same effect. 2646109998Smarkm [Massimiliano Pala madwolf@openca.org] 2647109998Smarkm 2648109998Smarkm *) Change all functions with names starting with des_ to be starting 2649109998Smarkm with DES_ instead. Add wrappers that are compatible with libdes, 2650109998Smarkm but are named _ossl_old_des_*. Finally, add macros that map the 2651109998Smarkm des_* symbols to the corresponding _ossl_old_des_* if libdes 2652109998Smarkm compatibility is desired. If OpenSSL 0.9.6c compatibility is 2653109998Smarkm desired, the des_* symbols will be mapped to DES_*, with one 2654109998Smarkm exception. 2655109998Smarkm 2656109998Smarkm Since we provide two compatibility mappings, the user needs to 2657109998Smarkm define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes 2658109998Smarkm compatibility is desired. The default (i.e., when that macro 2659109998Smarkm isn't defined) is OpenSSL 0.9.6c compatibility. 2660109998Smarkm 2661109998Smarkm There are also macros that enable and disable the support of old 2662109998Smarkm des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT 2663109998Smarkm and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those 2664109998Smarkm are defined, the default will apply: to support the old des routines. 2665109998Smarkm 2666109998Smarkm In either case, one must include openssl/des.h to get the correct 2667109998Smarkm definitions. Do not try to just include openssl/des_old.h, that 2668109998Smarkm won't work. 2669109998Smarkm 2670109998Smarkm NOTE: This is a major break of an old API into a new one. Software 2671109998Smarkm authors are encouraged to switch to the DES_ style functions. Some 2672109998Smarkm time in the future, des_old.h and the libdes compatibility functions 2673109998Smarkm will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the 2674109998Smarkm default), and then completely removed. 2675109998Smarkm [Richard Levitte] 2676109998Smarkm 2677109998Smarkm *) Test for certificates which contain unsupported critical extensions. 2678109998Smarkm If such a certificate is found during a verify operation it is 2679109998Smarkm rejected by default: this behaviour can be overridden by either 2680109998Smarkm handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or 2681109998Smarkm by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function 2682109998Smarkm X509_supported_extension() has also been added which returns 1 if a 2683109998Smarkm particular extension is supported. 2684109998Smarkm [Steve Henson] 2685109998Smarkm 2686109998Smarkm *) Modify the behaviour of EVP cipher functions in similar way to digests 2687109998Smarkm to retain compatibility with existing code. 2688109998Smarkm [Steve Henson] 2689109998Smarkm 2690109998Smarkm *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain 2691109998Smarkm compatibility with existing code. In particular the 'ctx' parameter does 2692109998Smarkm not have to be to be initialized before the call to EVP_DigestInit() and 2693109998Smarkm it is tidied up after a call to EVP_DigestFinal(). New function 2694109998Smarkm EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function 2695109998Smarkm EVP_MD_CTX_copy() changed to not require the destination to be 2696109998Smarkm initialized valid and new function EVP_MD_CTX_copy_ex() added which 2697109998Smarkm requires the destination to be valid. 2698109998Smarkm 2699109998Smarkm Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), 2700109998Smarkm EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). 2701109998Smarkm [Steve Henson] 2702109998Smarkm 2703109998Smarkm *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it 2704109998Smarkm so that complete 'Handshake' protocol structures are kept in memory 2705109998Smarkm instead of overwriting 'msg_type' and 'length' with 'body' data. 2706109998Smarkm [Bodo Moeller] 2707109998Smarkm 2708109998Smarkm *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. 2709109998Smarkm [Massimo Santin via Richard Levitte] 2710109998Smarkm 2711109998Smarkm *) Major restructuring to the underlying ENGINE code. This includes 2712109998Smarkm reduction of linker bloat, separation of pure "ENGINE" manipulation 2713109998Smarkm (initialisation, etc) from functionality dealing with implementations 2714109998Smarkm of specific crypto iterfaces. This change also introduces integrated 2715109998Smarkm support for symmetric ciphers and digest implementations - so ENGINEs 2716109998Smarkm can now accelerate these by providing EVP_CIPHER and EVP_MD 2717109998Smarkm implementations of their own. This is detailed in crypto/engine/README 2718109998Smarkm as it couldn't be adequately described here. However, there are a few 2719109998Smarkm API changes worth noting - some RSA, DSA, DH, and RAND functions that 2720109998Smarkm were changed in the original introduction of ENGINE code have now 2721109998Smarkm reverted back - the hooking from this code to ENGINE is now a good 2722109998Smarkm deal more passive and at run-time, operations deal directly with 2723109998Smarkm RSA_METHODs, DSA_METHODs (etc) as they did before, rather than 2724109998Smarkm dereferencing through an ENGINE pointer any more. Also, the ENGINE 2725109998Smarkm functions dealing with BN_MOD_EXP[_CRT] handlers have been removed - 2726109998Smarkm they were not being used by the framework as there is no concept of a 2727109998Smarkm BIGNUM_METHOD and they could not be generalised to the new 2728109998Smarkm 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, 2729109998Smarkm ENGINE_cpy() has been removed as it cannot be consistently defined in 2730109998Smarkm the new code. 2731109998Smarkm [Geoff Thorpe] 2732109998Smarkm 2733109998Smarkm *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. 2734109998Smarkm [Steve Henson] 2735109998Smarkm 2736109998Smarkm *) Change mkdef.pl to sort symbols that get the same entry number, 2737109998Smarkm and make sure the automatically generated functions ERR_load_* 2738109998Smarkm become part of libeay.num as well. 2739109998Smarkm [Richard Levitte] 2740109998Smarkm 2741109998Smarkm *) New function SSL_renegotiate_pending(). This returns true once 2742109998Smarkm renegotiation has been requested (either SSL_renegotiate() call 2743109998Smarkm or HelloRequest/ClientHello receveived from the peer) and becomes 2744109998Smarkm false once a handshake has been completed. 2745109998Smarkm (For servers, SSL_renegotiate() followed by SSL_do_handshake() 2746109998Smarkm sends a HelloRequest, but does not ensure that a handshake takes 2747109998Smarkm place. SSL_renegotiate_pending() is useful for checking if the 2748109998Smarkm client has followed the request.) 2749109998Smarkm [Bodo Moeller] 2750109998Smarkm 2751109998Smarkm *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. 2752109998Smarkm By default, clients may request session resumption even during 2753109998Smarkm renegotiation (if session ID contexts permit); with this option, 2754109998Smarkm session resumption is possible only in the first handshake. 2755109998Smarkm 2756109998Smarkm SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes 2757109998Smarkm more bits available for options that should not be part of 2758109998Smarkm SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). 2759109998Smarkm [Bodo Moeller] 2760109998Smarkm 2761109998Smarkm *) Add some demos for certificate and certificate request creation. 2762109998Smarkm [Steve Henson] 2763109998Smarkm 2764109998Smarkm *) Make maximum certificate chain size accepted from the peer application 2765109998Smarkm settable (SSL*_get/set_max_cert_list()), as proposed by 2766109998Smarkm "Douglas E. Engert" <deengert@anl.gov>. 2767109998Smarkm [Lutz Jaenicke] 2768109998Smarkm 2769109998Smarkm *) Add support for shared libraries for Unixware-7 2770109998Smarkm (Boyd Lynn Gerber <gerberb@zenez.com>). 2771109998Smarkm [Lutz Jaenicke] 2772109998Smarkm 2773109998Smarkm *) Add a "destroy" handler to ENGINEs that allows structural cleanup to 2774109998Smarkm be done prior to destruction. Use this to unload error strings from 2775109998Smarkm ENGINEs that load their own error strings. NB: This adds two new API 2776109998Smarkm functions to "get" and "set" this destroy handler in an ENGINE. 2777109998Smarkm [Geoff Thorpe] 2778109998Smarkm 2779109998Smarkm *) Alter all existing ENGINE implementations (except "openssl" and 2780109998Smarkm "openbsd") to dynamically instantiate their own error strings. This 2781109998Smarkm makes them more flexible to be built both as statically-linked ENGINEs 2782109998Smarkm and self-contained shared-libraries loadable via the "dynamic" ENGINE. 2783109998Smarkm Also, add stub code to each that makes building them as self-contained 2784109998Smarkm shared-libraries easier (see README.ENGINE). 2785109998Smarkm [Geoff Thorpe] 2786109998Smarkm 2787109998Smarkm *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE 2788109998Smarkm implementations into applications that are completely implemented in 2789109998Smarkm self-contained shared-libraries. The "dynamic" ENGINE exposes control 2790109998Smarkm commands that can be used to configure what shared-library to load and 2791109998Smarkm to control aspects of the way it is handled. Also, made an update to 2792109998Smarkm the README.ENGINE file that brings its information up-to-date and 2793109998Smarkm provides some information and instructions on the "dynamic" ENGINE 2794109998Smarkm (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). 2795109998Smarkm [Geoff Thorpe] 2796109998Smarkm 2797109998Smarkm *) Make it possible to unload ranges of ERR strings with a new 2798109998Smarkm "ERR_unload_strings" function. 2799109998Smarkm [Geoff Thorpe] 2800109998Smarkm 2801109998Smarkm *) Add a copy() function to EVP_MD. 2802109998Smarkm [Ben Laurie] 2803109998Smarkm 2804109998Smarkm *) Make EVP_MD routines take a context pointer instead of just the 2805109998Smarkm md_data void pointer. 2806109998Smarkm [Ben Laurie] 2807109998Smarkm 2808109998Smarkm *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates 2809109998Smarkm that the digest can only process a single chunk of data 2810109998Smarkm (typically because it is provided by a piece of 2811109998Smarkm hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application 2812109998Smarkm is only going to provide a single chunk of data, and hence the 2813109998Smarkm framework needn't accumulate the data for oneshot drivers. 2814109998Smarkm [Ben Laurie] 2815109998Smarkm 2816109998Smarkm *) As with "ERR", make it possible to replace the underlying "ex_data" 2817109998Smarkm functions. This change also alters the storage and management of global 2818109998Smarkm ex_data state - it's now all inside ex_data.c and all "class" code (eg. 2819109998Smarkm RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class 2820109998Smarkm index counters. The API functions that use this state have been changed 2821109998Smarkm to take a "class_index" rather than pointers to the class's local STACK 2822109998Smarkm and counter, and there is now an API function to dynamically create new 2823109998Smarkm classes. This centralisation allows us to (a) plug a lot of the 2824109998Smarkm thread-safety problems that existed, and (b) makes it possible to clean 2825109998Smarkm up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) 2826109998Smarkm such data would previously have always leaked in application code and 2827109998Smarkm workarounds were in place to make the memory debugging turn a blind eye 2828109998Smarkm to it. Application code that doesn't use this new function will still 2829109998Smarkm leak as before, but their memory debugging output will announce it now 2830109998Smarkm rather than letting it slide. 2831109998Smarkm 2832109998Smarkm Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change 2833109998Smarkm induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now 2834109998Smarkm has a return value to indicate success or failure. 2835109998Smarkm [Geoff Thorpe] 2836109998Smarkm 2837109998Smarkm *) Make it possible to replace the underlying "ERR" functions such that the 2838109998Smarkm global state (2 LHASH tables and 2 locks) is only used by the "default" 2839109998Smarkm implementation. This change also adds two functions to "get" and "set" 2840109998Smarkm the implementation prior to it being automatically set the first time 2841109998Smarkm any other ERR function takes place. Ie. an application can call "get", 2842109998Smarkm pass the return value to a module it has just loaded, and that module 2843109998Smarkm can call its own "set" function using that value. This means the 2844109998Smarkm module's "ERR" operations will use (and modify) the error state in the 2845109998Smarkm application and not in its own statically linked copy of OpenSSL code. 2846109998Smarkm [Geoff Thorpe] 2847109998Smarkm 2848109998Smarkm *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment 2849109998Smarkm reference counts. This performs normal REF_PRINT/REF_CHECK macros on 2850109998Smarkm the operation, and provides a more encapsulated way for external code 2851109998Smarkm (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code 2852109998Smarkm to use these functions rather than manually incrementing the counts. 2853109998Smarkm 2854109998Smarkm Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". 2855109998Smarkm [Geoff Thorpe] 2856109998Smarkm 2857109998Smarkm *) Add EVP test program. 2858109998Smarkm [Ben Laurie] 2859109998Smarkm 2860109998Smarkm *) Add symmetric cipher support to ENGINE. Expect the API to change! 2861109998Smarkm [Ben Laurie] 2862109998Smarkm 2863109998Smarkm *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() 2864109998Smarkm X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), 2865109998Smarkm X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). 2866109998Smarkm These allow a CRL to be built without having to access X509_CRL fields 2867109998Smarkm directly. Modify 'ca' application to use new functions. 2868109998Smarkm [Steve Henson] 2869109998Smarkm 2870109998Smarkm *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended 2871109998Smarkm bug workarounds. Rollback attack detection is a security feature. 2872109998Smarkm The problem will only arise on OpenSSL servers when TLSv1 is not 2873109998Smarkm available (sslv3_server_method() or SSL_OP_NO_TLSv1). 2874109998Smarkm Software authors not wanting to support TLSv1 will have special reasons 2875109998Smarkm for their choice and can explicitly enable this option. 2876109998Smarkm [Bodo Moeller, Lutz Jaenicke] 2877109998Smarkm 2878109998Smarkm *) Rationalise EVP so it can be extended: don't include a union of 2879109998Smarkm cipher/digest structures, add init/cleanup functions for EVP_MD_CTX 2880109998Smarkm (similar to those existing for EVP_CIPHER_CTX). 2881109998Smarkm Usage example: 2882109998Smarkm 2883109998Smarkm EVP_MD_CTX md; 2884109998Smarkm 2885109998Smarkm EVP_MD_CTX_init(&md); /* new function call */ 2886109998Smarkm EVP_DigestInit(&md, EVP_sha1()); 2887109998Smarkm EVP_DigestUpdate(&md, in, len); 2888109998Smarkm EVP_DigestFinal(&md, out, NULL); 2889109998Smarkm EVP_MD_CTX_cleanup(&md); /* new function call */ 2890109998Smarkm 2891109998Smarkm [Ben Laurie] 2892109998Smarkm 2893109998Smarkm *) Make DES key schedule conform to the usual scheme, as well as 2894109998Smarkm correcting its structure. This means that calls to DES functions 2895109998Smarkm now have to pass a pointer to a des_key_schedule instead of a 2896109998Smarkm plain des_key_schedule (which was actually always a pointer 2897109998Smarkm anyway): E.g., 2898109998Smarkm 2899109998Smarkm des_key_schedule ks; 2900109998Smarkm 2901109998Smarkm des_set_key_checked(..., &ks); 2902109998Smarkm des_ncbc_encrypt(..., &ks, ...); 2903109998Smarkm 2904109998Smarkm (Note that a later change renames 'des_...' into 'DES_...'.) 2905109998Smarkm [Ben Laurie] 2906109998Smarkm 2907109998Smarkm *) Initial reduction of linker bloat: the use of some functions, such as 2908109998Smarkm PEM causes large amounts of unused functions to be linked in due to 2909109998Smarkm poor organisation. For example pem_all.c contains every PEM function 2910109998Smarkm which has a knock on effect of linking in large amounts of (unused) 2911109998Smarkm ASN1 code. Grouping together similar functions and splitting unrelated 2912109998Smarkm functions prevents this. 2913109998Smarkm [Steve Henson] 2914109998Smarkm 2915109998Smarkm *) Cleanup of EVP macros. 2916109998Smarkm [Ben Laurie] 2917109998Smarkm 2918109998Smarkm *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the 2919109998Smarkm correct _ecb suffix. 2920109998Smarkm [Ben Laurie] 2921109998Smarkm 2922109998Smarkm *) Add initial OCSP responder support to ocsp application. The 2923109998Smarkm revocation information is handled using the text based index 2924109998Smarkm use by the ca application. The responder can either handle 2925109998Smarkm requests generated internally, supplied in files (for example 2926109998Smarkm via a CGI script) or using an internal minimal server. 2927109998Smarkm [Steve Henson] 2928109998Smarkm 2929109998Smarkm *) Add configuration choices to get zlib compression for TLS. 2930109998Smarkm [Richard Levitte] 2931109998Smarkm 2932109998Smarkm *) Changes to Kerberos SSL for RFC 2712 compliance: 2933109998Smarkm 1. Implemented real KerberosWrapper, instead of just using 2934109998Smarkm KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] 2935109998Smarkm 2. Implemented optional authenticator field of KerberosWrapper. 2936109998Smarkm 2937109998Smarkm Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, 2938109998Smarkm and authenticator structs; see crypto/krb5/. 2939109998Smarkm 2940109998Smarkm Generalized Kerberos calls to support multiple Kerberos libraries. 2941109998Smarkm [Vern Staats <staatsvr@asc.hpc.mil>, 2942109998Smarkm Jeffrey Altman <jaltman@columbia.edu> 2943109998Smarkm via Richard Levitte] 2944109998Smarkm 2945109998Smarkm *) Cause 'openssl speed' to use fully hard-coded DSA keys as it 2946109998Smarkm already does with RSA. testdsa.h now has 'priv_key/pub_key' 2947109998Smarkm values for each of the key sizes rather than having just 2948109998Smarkm parameters (and 'speed' generating keys each time). 2949109998Smarkm [Geoff Thorpe] 2950109998Smarkm 2951109998Smarkm *) Speed up EVP routines. 2952109998Smarkm Before: 2953109998Smarkmencrypt 2954109998Smarkmtype 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 2955109998Smarkmdes-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k 2956109998Smarkmdes-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k 2957109998Smarkmdes-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k 2958109998Smarkmdecrypt 2959109998Smarkmdes-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k 2960109998Smarkmdes-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k 2961109998Smarkmdes-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k 2962109998Smarkm After: 2963109998Smarkmencrypt 2964109998Smarkmdes-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k 2965109998Smarkmdecrypt 2966109998Smarkmdes-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k 2967109998Smarkm [Ben Laurie] 2968109998Smarkm 2969109998Smarkm *) Added the OS2-EMX target. 2970109998Smarkm ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte] 2971109998Smarkm 2972109998Smarkm *) Rewrite apps to use NCONF routines instead of the old CONF. New functions 2973109998Smarkm to support NCONF routines in extension code. New function CONF_set_nconf() 2974109998Smarkm to allow functions which take an NCONF to also handle the old LHASH 2975109998Smarkm structure: this means that the old CONF compatible routines can be 2976109998Smarkm retained (in particular wrt extensions) without having to duplicate the 2977109998Smarkm code. New function X509V3_add_ext_nconf_sk to add extensions to a stack. 2978109998Smarkm [Steve Henson] 2979109998Smarkm 2980109998Smarkm *) Enhance the general user interface with mechanisms for inner control 2981109998Smarkm and with possibilities to have yes/no kind of prompts. 2982109998Smarkm [Richard Levitte] 2983109998Smarkm 2984109998Smarkm *) Change all calls to low level digest routines in the library and 2985109998Smarkm applications to use EVP. Add missing calls to HMAC_cleanup() and 2986109998Smarkm don't assume HMAC_CTX can be copied using memcpy(). 2987109998Smarkm [Verdon Walker <VWalker@novell.com>, Steve Henson] 2988109998Smarkm 2989109998Smarkm *) Add the possibility to control engines through control names but with 2990109998Smarkm arbitrary arguments instead of just a string. 2991109998Smarkm Change the key loaders to take a UI_METHOD instead of a callback 2992109998Smarkm function pointer. NOTE: this breaks binary compatibility with earlier 2993109998Smarkm versions of OpenSSL [engine]. 2994109998Smarkm Adapt the nCipher code for these new conditions and add a card insertion 2995109998Smarkm callback. 2996109998Smarkm [Richard Levitte] 2997109998Smarkm 2998109998Smarkm *) Enhance the general user interface with mechanisms to better support 2999109998Smarkm dialog box interfaces, application-defined prompts, the possibility 3000109998Smarkm to use defaults (for example default passwords from somewhere else) 3001109998Smarkm and interrupts/cancellations. 3002109998Smarkm [Richard Levitte] 3003109998Smarkm 3004109998Smarkm *) Tidy up PKCS#12 attribute handling. Add support for the CSP name 3005109998Smarkm attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. 3006109998Smarkm [Steve Henson] 3007109998Smarkm 3008109998Smarkm *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also 3009109998Smarkm tidy up some unnecessarily weird code in 'sk_new()'). 3010109998Smarkm [Geoff, reported by Diego Tartara <dtartara@novamens.com>] 3011109998Smarkm 3012109998Smarkm *) Change the key loading routines for ENGINEs to use the same kind 3013109998Smarkm callback (pem_password_cb) as all other routines that need this 3014109998Smarkm kind of callback. 3015109998Smarkm [Richard Levitte] 3016109998Smarkm 3017109998Smarkm *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with 3018109998Smarkm 256 bit (=32 byte) keys. Of course seeding with more entropy bytes 3019109998Smarkm than this minimum value is recommended. 3020109998Smarkm [Lutz Jaenicke] 3021109998Smarkm 3022109998Smarkm *) New random seeder for OpenVMS, using the system process statistics 3023109998Smarkm that are easily reachable. 3024109998Smarkm [Richard Levitte] 3025109998Smarkm 3026109998Smarkm *) Windows apparently can't transparently handle global 3027109998Smarkm variables defined in DLLs. Initialisations such as: 3028109998Smarkm 3029109998Smarkm const ASN1_ITEM *it = &ASN1_INTEGER_it; 3030109998Smarkm 3031109998Smarkm wont compile. This is used by the any applications that need to 3032109998Smarkm declare their own ASN1 modules. This was fixed by adding the option 3033109998Smarkm EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly 3034109998Smarkm needed for static libraries under Win32. 3035109998Smarkm [Steve Henson] 3036109998Smarkm 3037109998Smarkm *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle 3038109998Smarkm setting of purpose and trust fields. New X509_STORE trust and 3039109998Smarkm purpose functions and tidy up setting in other SSL functions. 3040109998Smarkm [Steve Henson] 3041109998Smarkm 3042109998Smarkm *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE 3043109998Smarkm structure. These are inherited by X509_STORE_CTX when it is 3044109998Smarkm initialised. This allows various defaults to be set in the 3045109998Smarkm X509_STORE structure (such as flags for CRL checking and custom 3046109998Smarkm purpose or trust settings) for functions which only use X509_STORE_CTX 3047109998Smarkm internally such as S/MIME. 3048109998Smarkm 3049109998Smarkm Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and 3050109998Smarkm trust settings if they are not set in X509_STORE. This allows X509_STORE 3051109998Smarkm purposes and trust (in S/MIME for example) to override any set by default. 3052109998Smarkm 3053109998Smarkm Add command line options for CRL checking to smime, s_client and s_server 3054109998Smarkm applications. 3055109998Smarkm [Steve Henson] 3056109998Smarkm 3057109998Smarkm *) Initial CRL based revocation checking. If the CRL checking flag(s) 3058109998Smarkm are set then the CRL is looked up in the X509_STORE structure and 3059109998Smarkm its validity and signature checked, then if the certificate is found 3060109998Smarkm in the CRL the verify fails with a revoked error. 3061109998Smarkm 3062109998Smarkm Various new CRL related callbacks added to X509_STORE_CTX structure. 3063109998Smarkm 3064109998Smarkm Command line options added to 'verify' application to support this. 3065109998Smarkm 3066109998Smarkm This needs some additional work, such as being able to handle multiple 3067109998Smarkm CRLs with different times, extension based lookup (rather than just 3068109998Smarkm by subject name) and ultimately more complete V2 CRL extension 3069109998Smarkm handling. 3070109998Smarkm [Steve Henson] 3071109998Smarkm 3072109998Smarkm *) Add a general user interface API (crypto/ui/). This is designed 3073109998Smarkm to replace things like des_read_password and friends (backward 3074109998Smarkm compatibility functions using this new API are provided). 3075109998Smarkm The purpose is to remove prompting functions from the DES code 3076109998Smarkm section as well as provide for prompting through dialog boxes in 3077109998Smarkm a window system and the like. 3078109998Smarkm [Richard Levitte] 3079109998Smarkm 3080109998Smarkm *) Add "ex_data" support to ENGINE so implementations can add state at a 3081109998Smarkm per-structure level rather than having to store it globally. 3082109998Smarkm [Geoff] 3083109998Smarkm 3084109998Smarkm *) Make it possible for ENGINE structures to be copied when retrieved by 3085109998Smarkm ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. 3086109998Smarkm This causes the "original" ENGINE structure to act like a template, 3087109998Smarkm analogous to the RSA vs. RSA_METHOD type of separation. Because of this 3088109998Smarkm operational state can be localised to each ENGINE structure, despite the 3089109998Smarkm fact they all share the same "methods". New ENGINE structures returned in 3090109998Smarkm this case have no functional references and the return value is the single 3091109998Smarkm structural reference. This matches the single structural reference returned 3092109998Smarkm by ENGINE_by_id() normally, when it is incremented on the pre-existing 3093109998Smarkm ENGINE structure. 3094109998Smarkm [Geoff] 3095109998Smarkm 3096109998Smarkm *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this 3097109998Smarkm needs to match any other type at all we need to manually clear the 3098109998Smarkm tag cache. 3099109998Smarkm [Steve Henson] 3100109998Smarkm 3101109998Smarkm *) Changes to the "openssl engine" utility to include; 3102109998Smarkm - verbosity levels ('-v', '-vv', and '-vvv') that provide information 3103109998Smarkm about an ENGINE's available control commands. 3104109998Smarkm - executing control commands from command line arguments using the 3105109998Smarkm '-pre' and '-post' switches. '-post' is only used if '-t' is 3106109998Smarkm specified and the ENGINE is successfully initialised. The syntax for 3107109998Smarkm the individual commands are colon-separated, for example; 3108109998Smarkm openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so 3109109998Smarkm [Geoff] 3110109998Smarkm 3111109998Smarkm *) New dynamic control command support for ENGINEs. ENGINEs can now 3112109998Smarkm declare their own commands (numbers), names (strings), descriptions, 3113109998Smarkm and input types for run-time discovery by calling applications. A 3114109998Smarkm subset of these commands are implicitly classed as "executable" 3115109998Smarkm depending on their input type, and only these can be invoked through 3116109998Smarkm the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this 3117109998Smarkm can be based on user input, config files, etc). The distinction is 3118109998Smarkm that "executable" commands cannot return anything other than a boolean 3119109998Smarkm result and can only support numeric or string input, whereas some 3120109998Smarkm discoverable commands may only be for direct use through 3121109998Smarkm ENGINE_ctrl(), eg. supporting the exchange of binary data, function 3122109998Smarkm pointers, or other custom uses. The "executable" commands are to 3123109998Smarkm support parameterisations of ENGINE behaviour that can be 3124109998Smarkm unambiguously defined by ENGINEs and used consistently across any 3125109998Smarkm OpenSSL-based application. Commands have been added to all the 3126109998Smarkm existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow 3127109998Smarkm control over shared-library paths without source code alterations. 3128109998Smarkm [Geoff] 3129109998Smarkm 3130109998Smarkm *) Changed all ENGINE implementations to dynamically allocate their 3131109998Smarkm ENGINEs rather than declaring them statically. Apart from this being 3132109998Smarkm necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, 3133109998Smarkm this also allows the implementations to compile without using the 3134109998Smarkm internal engine_int.h header. 3135109998Smarkm [Geoff] 3136109998Smarkm 3137109998Smarkm *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a 3138109998Smarkm 'const' value. Any code that should be able to modify a RAND_METHOD 3139109998Smarkm should already have non-const pointers to it (ie. they should only 3140109998Smarkm modify their own ones). 3141109998Smarkm [Geoff] 3142109998Smarkm 3143109998Smarkm *) Made a variety of little tweaks to the ENGINE code. 3144109998Smarkm - "atalla" and "ubsec" string definitions were moved from header files 3145109998Smarkm to C code. "nuron" string definitions were placed in variables 3146109998Smarkm rather than hard-coded - allowing parameterisation of these values 3147109998Smarkm later on via ctrl() commands. 3148109998Smarkm - Removed unused "#if 0"'d code. 3149109998Smarkm - Fixed engine list iteration code so it uses ENGINE_free() to release 3150109998Smarkm structural references. 3151109998Smarkm - Constified the RAND_METHOD element of ENGINE structures. 3152109998Smarkm - Constified various get/set functions as appropriate and added 3153109998Smarkm missing functions (including a catch-all ENGINE_cpy that duplicates 3154109998Smarkm all ENGINE values onto a new ENGINE except reference counts/state). 3155109998Smarkm - Removed NULL parameter checks in get/set functions. Setting a method 3156109998Smarkm or function to NULL is a way of cancelling out a previously set 3157109998Smarkm value. Passing a NULL ENGINE parameter is just plain stupid anyway 3158109998Smarkm and doesn't justify the extra error symbols and code. 3159109998Smarkm - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for 3160109998Smarkm flags from engine_int.h to engine.h. 3161109998Smarkm - Changed prototypes for ENGINE handler functions (init(), finish(), 3162109998Smarkm ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. 3163109998Smarkm [Geoff] 3164109998Smarkm 3165109998Smarkm *) Implement binary inversion algorithm for BN_mod_inverse in addition 3166109998Smarkm to the algorithm using long division. The binary algorithm can be 3167109998Smarkm used only if the modulus is odd. On 32-bit systems, it is faster 3168109998Smarkm only for relatively small moduli (roughly 20-30% for 128-bit moduli, 3169109998Smarkm roughly 5-15% for 256-bit moduli), so we use it only for moduli 3170109998Smarkm up to 450 bits. In 64-bit environments, the binary algorithm 3171109998Smarkm appears to be advantageous for much longer moduli; here we use it 3172109998Smarkm for moduli up to 2048 bits. 3173109998Smarkm [Bodo Moeller] 3174109998Smarkm 3175109998Smarkm *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code 3176109998Smarkm could not support the combine flag in choice fields. 3177109998Smarkm [Steve Henson] 3178109998Smarkm 3179109998Smarkm *) Add a 'copy_extensions' option to the 'ca' utility. This copies 3180109998Smarkm extensions from a certificate request to the certificate. 3181109998Smarkm [Steve Henson] 3182109998Smarkm 3183109998Smarkm *) Allow multiple 'certopt' and 'nameopt' options to be separated 3184109998Smarkm by commas. Add 'namopt' and 'certopt' options to the 'ca' config 3185109998Smarkm file: this allows the display of the certificate about to be 3186109998Smarkm signed to be customised, to allow certain fields to be included 3187109998Smarkm or excluded and extension details. The old system didn't display 3188109998Smarkm multicharacter strings properly, omitted fields not in the policy 3189109998Smarkm and couldn't display additional details such as extensions. 3190109998Smarkm [Steve Henson] 3191109998Smarkm 3192109998Smarkm *) Function EC_POINTs_mul for multiple scalar multiplication 3193109998Smarkm of an arbitrary number of elliptic curve points 3194109998Smarkm \sum scalars[i]*points[i], 3195109998Smarkm optionally including the generator defined for the EC_GROUP: 3196109998Smarkm scalar*generator + \sum scalars[i]*points[i]. 3197109998Smarkm 3198109998Smarkm EC_POINT_mul is a simple wrapper function for the typical case 3199109998Smarkm that the point list has just one item (besides the optional 3200109998Smarkm generator). 3201109998Smarkm [Bodo Moeller] 3202109998Smarkm 3203109998Smarkm *) First EC_METHODs for curves over GF(p): 3204109998Smarkm 3205109998Smarkm EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr 3206109998Smarkm operations and provides various method functions that can also 3207109998Smarkm operate with faster implementations of modular arithmetic. 3208109998Smarkm 3209109998Smarkm EC_GFp_mont_method() reuses most functions that are part of 3210109998Smarkm EC_GFp_simple_method, but uses Montgomery arithmetic. 3211109998Smarkm 3212109998Smarkm [Bodo Moeller; point addition and point doubling 3213109998Smarkm implementation directly derived from source code provided by 3214109998Smarkm Lenka Fibikova <fibikova@exp-math.uni-essen.de>] 3215109998Smarkm 3216109998Smarkm *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, 3217109998Smarkm crypto/ec/ec_lib.c): 3218109998Smarkm 3219109998Smarkm Curves are EC_GROUP objects (with an optional group generator) 3220109998Smarkm based on EC_METHODs that are built into the library. 3221109998Smarkm 3222109998Smarkm Points are EC_POINT objects based on EC_GROUP objects. 3223109998Smarkm 3224109998Smarkm Most of the framework would be able to handle curves over arbitrary 3225109998Smarkm finite fields, but as there are no obvious types for fields other 3226109998Smarkm than GF(p), some functions are limited to that for now. 3227109998Smarkm [Bodo Moeller] 3228109998Smarkm 3229109998Smarkm *) Add the -HTTP option to s_server. It is similar to -WWW, but requires 3230109998Smarkm that the file contains a complete HTTP response. 3231109998Smarkm [Richard Levitte] 3232109998Smarkm 3233109998Smarkm *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl 3234109998Smarkm change the def and num file printf format specifier from "%-40sXXX" 3235109998Smarkm to "%-39s XXX". The latter will always guarantee a space after the 3236109998Smarkm field while the former will cause them to run together if the field 3237109998Smarkm is 40 of more characters long. 3238109998Smarkm [Steve Henson] 3239109998Smarkm 3240109998Smarkm *) Constify the cipher and digest 'method' functions and structures 3241109998Smarkm and modify related functions to take constant EVP_MD and EVP_CIPHER 3242109998Smarkm pointers. 3243109998Smarkm [Steve Henson] 3244109998Smarkm 3245109998Smarkm *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them 3246109998Smarkm in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. 3247109998Smarkm [Bodo Moeller] 3248109998Smarkm 3249109998Smarkm *) Modify EVP_Digest*() routines so they now return values. Although the 3250109998Smarkm internal software routines can never fail additional hardware versions 3251109998Smarkm might. 3252109998Smarkm [Steve Henson] 3253109998Smarkm 3254109998Smarkm *) Clean up crypto/err/err.h and change some error codes to avoid conflicts: 3255109998Smarkm 3256109998Smarkm Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 3257109998Smarkm (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. 3258109998Smarkm 3259109998Smarkm ASN1 error codes 3260109998Smarkm ERR_R_NESTED_ASN1_ERROR 3261109998Smarkm ... 3262109998Smarkm ERR_R_MISSING_ASN1_EOS 3263109998Smarkm were 4 .. 9, conflicting with 3264109998Smarkm ERR_LIB_RSA (= ERR_R_RSA_LIB) 3265109998Smarkm ... 3266109998Smarkm ERR_LIB_PEM (= ERR_R_PEM_LIB). 3267109998Smarkm They are now 58 .. 63 (i.e., just below ERR_R_FATAL). 3268109998Smarkm 3269109998Smarkm Add new error code 'ERR_R_INTERNAL_ERROR'. 3270109998Smarkm [Bodo Moeller] 3271109998Smarkm 3272109998Smarkm *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock 3273109998Smarkm suffices. 3274109998Smarkm [Bodo Moeller] 3275109998Smarkm 3276109998Smarkm *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This 3277109998Smarkm sets the subject name for a new request or supersedes the 3278109998Smarkm subject name in a given request. Formats that can be parsed are 3279109998Smarkm 'CN=Some Name, OU=myOU, C=IT' 3280109998Smarkm and 3281109998Smarkm 'CN=Some Name/OU=myOU/C=IT'. 3282109998Smarkm 3283109998Smarkm Add options '-batch' and '-verbose' to 'openssl req'. 3284109998Smarkm [Massimiliano Pala <madwolf@hackmasters.net>] 3285109998Smarkm 3286109998Smarkm *) Introduce the possibility to access global variables through 3287109998Smarkm functions on platform were that's the best way to handle exporting 3288109998Smarkm global variables in shared libraries. To enable this functionality, 3289109998Smarkm one must configure with "EXPORT_VAR_AS_FN" or defined the C macro 3290109998Smarkm "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter 3291109998Smarkm is normally done by Configure or something similar). 3292109998Smarkm 3293109998Smarkm To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL 3294109998Smarkm in the source file (foo.c) like this: 3295109998Smarkm 3296109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; 3297109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(double,bar); 3298109998Smarkm 3299109998Smarkm To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL 3300109998Smarkm and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: 3301109998Smarkm 3302109998Smarkm OPENSSL_DECLARE_GLOBAL(int,foo); 3303109998Smarkm #define foo OPENSSL_GLOBAL_REF(foo) 3304109998Smarkm OPENSSL_DECLARE_GLOBAL(double,bar); 3305109998Smarkm #define bar OPENSSL_GLOBAL_REF(bar) 3306109998Smarkm 3307109998Smarkm The #defines are very important, and therefore so is including the 3308109998Smarkm header file everywhere where the defined globals are used. 3309109998Smarkm 3310109998Smarkm The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition 3311109998Smarkm of ASN.1 items, but that structure is a bit different. 3312109998Smarkm 3313109998Smarkm The largest change is in util/mkdef.pl which has been enhanced with 3314109998Smarkm better and easier to understand logic to choose which symbols should 3315109998Smarkm go into the Windows .def files as well as a number of fixes and code 3316109998Smarkm cleanup (among others, algorithm keywords are now sorted 3317109998Smarkm lexicographically to avoid constant rewrites). 3318109998Smarkm [Richard Levitte] 3319109998Smarkm 3320109998Smarkm *) In BN_div() keep a copy of the sign of 'num' before writing the 3321109998Smarkm result to 'rm' because if rm==num the value will be overwritten 3322109998Smarkm and produce the wrong result if 'num' is negative: this caused 3323109998Smarkm problems with BN_mod() and BN_nnmod(). 3324109998Smarkm [Steve Henson] 3325109998Smarkm 3326109998Smarkm *) Function OCSP_request_verify(). This checks the signature on an 3327109998Smarkm OCSP request and verifies the signer certificate. The signer 3328109998Smarkm certificate is just checked for a generic purpose and OCSP request 3329109998Smarkm trust settings. 3330109998Smarkm [Steve Henson] 3331109998Smarkm 3332109998Smarkm *) Add OCSP_check_validity() function to check the validity of OCSP 3333109998Smarkm responses. OCSP responses are prepared in real time and may only 3334109998Smarkm be a few seconds old. Simply checking that the current time lies 3335109998Smarkm between thisUpdate and nextUpdate max reject otherwise valid responses 3336109998Smarkm caused by either OCSP responder or client clock inaccuracy. Instead 3337109998Smarkm we allow thisUpdate and nextUpdate to fall within a certain period of 3338109998Smarkm the current time. The age of the response can also optionally be 3339109998Smarkm checked. Two new options -validity_period and -status_age added to 3340109998Smarkm ocsp utility. 3341109998Smarkm [Steve Henson] 3342109998Smarkm 3343109998Smarkm *) If signature or public key algorithm is unrecognized print out its 3344109998Smarkm OID rather that just UNKNOWN. 3345109998Smarkm [Steve Henson] 3346109998Smarkm 3347109998Smarkm *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and 3348109998Smarkm OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate 3349109998Smarkm ID to be generated from the issuer certificate alone which can then be 3350109998Smarkm passed to OCSP_id_issuer_cmp(). 3351109998Smarkm [Steve Henson] 3352109998Smarkm 3353109998Smarkm *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new 3354109998Smarkm ASN1 modules to export functions returning ASN1_ITEM pointers 3355109998Smarkm instead of the ASN1_ITEM structures themselves. This adds several 3356109998Smarkm new macros which allow the underlying ASN1 function/structure to 3357109998Smarkm be accessed transparently. As a result code should not use ASN1_ITEM 3358109998Smarkm references directly (such as &X509_it) but instead use the relevant 3359109998Smarkm macros (such as ASN1_ITEM_rptr(X509)). This option is to allow 3360109998Smarkm use of the new ASN1 code on platforms where exporting structures 3361109998Smarkm is problematical (for example in shared libraries) but exporting 3362109998Smarkm functions returning pointers to structures is not. 3363109998Smarkm [Steve Henson] 3364109998Smarkm 3365109998Smarkm *) Add support for overriding the generation of SSL/TLS session IDs. 3366109998Smarkm These callbacks can be registered either in an SSL_CTX or per SSL. 3367109998Smarkm The purpose of this is to allow applications to control, if they wish, 3368109998Smarkm the arbitrary values chosen for use as session IDs, particularly as it 3369109998Smarkm can be useful for session caching in multiple-server environments. A 3370109998Smarkm command-line switch for testing this (and any client code that wishes 3371109998Smarkm to use such a feature) has been added to "s_server". 3372109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 3373109998Smarkm 3374109998Smarkm *) Modify mkdef.pl to recognise and parse preprocessor conditionals 3375109998Smarkm of the form '#if defined(...) || defined(...) || ...' and 3376109998Smarkm '#if !defined(...) && !defined(...) && ...'. This also avoids 3377109998Smarkm the growing number of special cases it was previously handling. 3378109998Smarkm [Richard Levitte] 3379109998Smarkm 3380109998Smarkm *) Make all configuration macros available for application by making 3381109998Smarkm sure they are available in opensslconf.h, by giving them names starting 3382109998Smarkm with "OPENSSL_" to avoid conflicts with other packages and by making 3383109998Smarkm sure e_os2.h will cover all platform-specific cases together with 3384109998Smarkm opensslconf.h. 3385109998Smarkm Additionally, it is now possible to define configuration/platform- 3386109998Smarkm specific names (called "system identities"). In the C code, these 3387109998Smarkm are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another 3388109998Smarkm macro with the name beginning with "OPENSSL_SYS_", which is determined 3389109998Smarkm from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on 3390109998Smarkm what is available. 3391109998Smarkm [Richard Levitte] 3392109998Smarkm 3393109998Smarkm *) New option -set_serial to 'req' and 'x509' this allows the serial 3394109998Smarkm number to use to be specified on the command line. Previously self 3395109998Smarkm signed certificates were hard coded with serial number 0 and the 3396109998Smarkm CA options of 'x509' had to use a serial number in a file which was 3397109998Smarkm auto incremented. 3398109998Smarkm [Steve Henson] 3399109998Smarkm 3400109998Smarkm *) New options to 'ca' utility to support V2 CRL entry extensions. 3401109998Smarkm Currently CRL reason, invalidity date and hold instruction are 3402109998Smarkm supported. Add new CRL extensions to V3 code and some new objects. 3403109998Smarkm [Steve Henson] 3404109998Smarkm 3405109998Smarkm *) New function EVP_CIPHER_CTX_set_padding() this is used to 3406109998Smarkm disable standard block padding (aka PKCS#5 padding) in the EVP 3407109998Smarkm API, which was previously mandatory. This means that the data is 3408109998Smarkm not padded in any way and so the total length much be a multiple 3409109998Smarkm of the block size, otherwise an error occurs. 3410109998Smarkm [Steve Henson] 3411109998Smarkm 3412109998Smarkm *) Initial (incomplete) OCSP SSL support. 3413109998Smarkm [Steve Henson] 3414109998Smarkm 3415109998Smarkm *) New function OCSP_parse_url(). This splits up a URL into its host, 3416109998Smarkm port and path components: primarily to parse OCSP URLs. New -url 3417109998Smarkm option to ocsp utility. 3418109998Smarkm [Steve Henson] 3419109998Smarkm 3420109998Smarkm *) New nonce behavior. The return value of OCSP_check_nonce() now 3421109998Smarkm reflects the various checks performed. Applications can decide 3422109998Smarkm whether to tolerate certain situations such as an absent nonce 3423109998Smarkm in a response when one was present in a request: the ocsp application 3424109998Smarkm just prints out a warning. New function OCSP_add1_basic_nonce() 3425109998Smarkm this is to allow responders to include a nonce in a response even if 3426109998Smarkm the request is nonce-less. 3427109998Smarkm [Steve Henson] 3428109998Smarkm 3429109998Smarkm *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are 3430109998Smarkm skipped when using openssl x509 multiple times on a single input file, 3431109998Smarkm e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs". 3432109998Smarkm [Bodo Moeller] 3433109998Smarkm 3434109998Smarkm *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() 3435109998Smarkm set string type: to handle setting ASN1_TIME structures. Fix ca 3436109998Smarkm utility to correctly initialize revocation date of CRLs. 3437109998Smarkm [Steve Henson] 3438109998Smarkm 3439109998Smarkm *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override 3440109998Smarkm the clients preferred ciphersuites and rather use its own preferences. 3441109998Smarkm Should help to work around M$ SGC (Server Gated Cryptography) bug in 3442109998Smarkm Internet Explorer by ensuring unchanged hash method during stepup. 3443109998Smarkm (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) 3444109998Smarkm [Lutz Jaenicke] 3445109998Smarkm 3446109998Smarkm *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael 3447109998Smarkm to aes and add a new 'exist' option to print out symbols that don't 3448109998Smarkm appear to exist. 3449109998Smarkm [Steve Henson] 3450109998Smarkm 3451109998Smarkm *) Additional options to ocsp utility to allow flags to be set and 3452109998Smarkm additional certificates supplied. 3453109998Smarkm [Steve Henson] 3454109998Smarkm 3455109998Smarkm *) Add the option -VAfile to 'openssl ocsp', so the user can give the 3456109998Smarkm OCSP client a number of certificate to only verify the response 3457109998Smarkm signature against. 3458109998Smarkm [Richard Levitte] 3459109998Smarkm 3460109998Smarkm *) Update Rijndael code to version 3.0 and change EVP AES ciphers to 3461109998Smarkm handle the new API. Currently only ECB, CBC modes supported. Add new 3462109998Smarkm AES OIDs. 3463109998Smarkm 3464109998Smarkm Add TLS AES ciphersuites as described in RFC3268, "Advanced 3465109998Smarkm Encryption Standard (AES) Ciphersuites for Transport Layer 3466109998Smarkm Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were 3467109998Smarkm not enabled by default and were not part of the "ALL" ciphersuite 3468109998Smarkm alias because they were not yet official; they could be 3469109998Smarkm explicitly requested by specifying the "AESdraft" ciphersuite 3470109998Smarkm group alias. In the final release of OpenSSL 0.9.7, the group 3471109998Smarkm alias is called "AES" and is part of "ALL".) 3472109998Smarkm [Ben Laurie, Steve Henson, Bodo Moeller] 3473109998Smarkm 3474109998Smarkm *) New function OCSP_copy_nonce() to copy nonce value (if present) from 3475109998Smarkm request to response. 3476109998Smarkm [Steve Henson] 3477109998Smarkm 3478109998Smarkm *) Functions for OCSP responders. OCSP_request_onereq_count(), 3479109998Smarkm OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() 3480109998Smarkm extract information from a certificate request. OCSP_response_create() 3481109998Smarkm creates a response and optionally adds a basic response structure. 3482109998Smarkm OCSP_basic_add1_status() adds a complete single response to a basic 3483109998Smarkm response and returns the OCSP_SINGLERESP structure just added (to allow 3484109998Smarkm extensions to be included for example). OCSP_basic_add1_cert() adds a 3485109998Smarkm certificate to a basic response and OCSP_basic_sign() signs a basic 3486109998Smarkm response with various flags. New helper functions ASN1_TIME_check() 3487109998Smarkm (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() 3488109998Smarkm (converts ASN1_TIME to GeneralizedTime). 3489109998Smarkm [Steve Henson] 3490109998Smarkm 3491109998Smarkm *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() 3492109998Smarkm in a single operation. X509_get0_pubkey_bitstr() extracts the public_key 3493109998Smarkm structure from a certificate. X509_pubkey_digest() digests the public_key 3494109998Smarkm contents: this is used in various key identifiers. 3495109998Smarkm [Steve Henson] 3496109998Smarkm 3497109998Smarkm *) Make sk_sort() tolerate a NULL argument. 3498109998Smarkm [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>] 3499109998Smarkm 3500109998Smarkm *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates 3501109998Smarkm passed by the function are trusted implicitly. If any of them signed the 3502109998Smarkm response then it is assumed to be valid and is not verified. 3503109998Smarkm [Steve Henson] 3504109998Smarkm 3505109998Smarkm *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT 3506109998Smarkm to data. This was previously part of the PKCS7 ASN1 code. This 3507109998Smarkm was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. 3508109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 3509109998Smarkm <support@securenetterm.com>] 3510109998Smarkm 3511109998Smarkm *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 3512109998Smarkm routines: without these tracing memory leaks is very painful. 3513109998Smarkm Fix leaks in PKCS12 and PKCS7 routines. 3514109998Smarkm [Steve Henson] 3515109998Smarkm 3516109998Smarkm *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). 3517109998Smarkm Previously it initialised the 'type' argument to V_ASN1_UTCTIME which 3518109998Smarkm effectively meant GeneralizedTime would never be used. Now it 3519109998Smarkm is initialised to -1 but X509_time_adj() now has to check the value 3520109998Smarkm and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or 3521109998Smarkm V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. 3522109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 3523109998Smarkm <support@securenetterm.com>] 3524109998Smarkm 3525109998Smarkm *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously 3526109998Smarkm result in a zero length in the ASN1_INTEGER structure which was 3527109998Smarkm not consistent with the structure when d2i_ASN1_INTEGER() was used 3528109998Smarkm and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() 3529109998Smarkm to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() 3530109998Smarkm where it did not print out a minus for negative ASN1_INTEGER. 3531109998Smarkm [Steve Henson] 3532109998Smarkm 3533109998Smarkm *) Add summary printout to ocsp utility. The various functions which 3534109998Smarkm convert status values to strings have been renamed to: 3535109998Smarkm OCSP_response_status_str(), OCSP_cert_status_str() and 3536109998Smarkm OCSP_crl_reason_str() and are no longer static. New options 3537109998Smarkm to verify nonce values and to disable verification. OCSP response 3538109998Smarkm printout format cleaned up. 3539109998Smarkm [Steve Henson] 3540109998Smarkm 3541109998Smarkm *) Add additional OCSP certificate checks. These are those specified 3542109998Smarkm in RFC2560. This consists of two separate checks: the CA of the 3543109998Smarkm certificate being checked must either be the OCSP signer certificate 3544109998Smarkm or the issuer of the OCSP signer certificate. In the latter case the 3545109998Smarkm OCSP signer certificate must contain the OCSP signing extended key 3546109998Smarkm usage. This check is performed by attempting to match the OCSP 3547109998Smarkm signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash 3548109998Smarkm in the OCSP_CERTID structures of the response. 3549109998Smarkm [Steve Henson] 3550109998Smarkm 3551109998Smarkm *) Initial OCSP certificate verification added to OCSP_basic_verify() 3552109998Smarkm and related routines. This uses the standard OpenSSL certificate 3553109998Smarkm verify routines to perform initial checks (just CA validity) and 3554109998Smarkm to obtain the certificate chain. Then additional checks will be 3555109998Smarkm performed on the chain. Currently the root CA is checked to see 3556109998Smarkm if it is explicitly trusted for OCSP signing. This is used to set 3557109998Smarkm a root CA as a global signing root: that is any certificate that 3558109998Smarkm chains to that CA is an acceptable OCSP signing certificate. 3559109998Smarkm [Steve Henson] 3560109998Smarkm 3561109998Smarkm *) New '-extfile ...' option to 'openssl ca' for reading X.509v3 3562109998Smarkm extensions from a separate configuration file. 3563109998Smarkm As when reading extensions from the main configuration file, 3564109998Smarkm the '-extensions ...' option may be used for specifying the 3565109998Smarkm section to use. 3566109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 3567109998Smarkm 3568109998Smarkm *) New OCSP utility. Allows OCSP requests to be generated or 3569109998Smarkm read. The request can be sent to a responder and the output 3570109998Smarkm parsed, outputed or printed in text form. Not complete yet: 3571109998Smarkm still needs to check the OCSP response validity. 3572109998Smarkm [Steve Henson] 3573109998Smarkm 3574109998Smarkm *) New subcommands for 'openssl ca': 3575109998Smarkm 'openssl ca -status <serial>' prints the status of the cert with 3576109998Smarkm the given serial number (according to the index file). 3577109998Smarkm 'openssl ca -updatedb' updates the expiry status of certificates 3578109998Smarkm in the index file. 3579109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 3580109998Smarkm 3581109998Smarkm *) New '-newreq-nodes' command option to CA.pl. This is like 3582109998Smarkm '-newreq', but calls 'openssl req' with the '-nodes' option 3583109998Smarkm so that the resulting key is not encrypted. 3584109998Smarkm [Damien Miller <djm@mindrot.org>] 3585109998Smarkm 3586109998Smarkm *) New configuration for the GNU Hurd. 3587109998Smarkm [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte] 3588109998Smarkm 3589109998Smarkm *) Initial code to implement OCSP basic response verify. This 3590109998Smarkm is currently incomplete. Currently just finds the signer's 3591109998Smarkm certificate and verifies the signature on the response. 3592109998Smarkm [Steve Henson] 3593109998Smarkm 3594109998Smarkm *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in 3595109998Smarkm value of OPENSSLDIR. This is available via the new '-d' option 3596109998Smarkm to 'openssl version', and is also included in 'openssl version -a'. 3597109998Smarkm [Bodo Moeller] 3598109998Smarkm 3599109998Smarkm *) Allowing defining memory allocation callbacks that will be given 3600109998Smarkm file name and line number information in additional arguments 3601109998Smarkm (a const char* and an int). The basic functionality remains, as 3602109998Smarkm well as the original possibility to just replace malloc(), 3603109998Smarkm realloc() and free() by functions that do not know about these 3604109998Smarkm additional arguments. To register and find out the current 3605109998Smarkm settings for extended allocation functions, the following 3606109998Smarkm functions are provided: 3607109998Smarkm 3608109998Smarkm CRYPTO_set_mem_ex_functions 3609109998Smarkm CRYPTO_set_locked_mem_ex_functions 3610109998Smarkm CRYPTO_get_mem_ex_functions 3611109998Smarkm CRYPTO_get_locked_mem_ex_functions 3612109998Smarkm 3613109998Smarkm These work the same way as CRYPTO_set_mem_functions and friends. 3614109998Smarkm CRYPTO_get_[locked_]mem_functions now writes 0 where such an 3615109998Smarkm extended allocation function is enabled. 3616109998Smarkm Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where 3617109998Smarkm a conventional allocation function is enabled. 3618109998Smarkm [Richard Levitte, Bodo Moeller] 3619109998Smarkm 3620109998Smarkm *) Finish off removing the remaining LHASH function pointer casts. 3621109998Smarkm There should no longer be any prototype-casting required when using 3622109998Smarkm the LHASH abstraction, and any casts that remain are "bugs". See 3623109998Smarkm the callback types and macros at the head of lhash.h for details 3624109998Smarkm (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). 3625109998Smarkm [Geoff Thorpe] 3626109998Smarkm 3627109998Smarkm *) Add automatic query of EGD sockets in RAND_poll() for the unix variant. 3628109998Smarkm If /dev/[u]random devices are not available or do not return enough 3629109998Smarkm entropy, EGD style sockets (served by EGD or PRNGD) will automatically 3630109998Smarkm be queried. 3631109998Smarkm The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and 3632109998Smarkm /etc/entropy will be queried once each in this sequence, quering stops 3633109998Smarkm when enough entropy was collected without querying more sockets. 3634109998Smarkm [Lutz Jaenicke] 3635109998Smarkm 3636109998Smarkm *) Change the Unix RAND_poll() variant to be able to poll several 3637109998Smarkm random devices, as specified by DEVRANDOM, until a sufficient amount 3638109998Smarkm of data has been collected. We spend at most 10 ms on each file 3639109998Smarkm (select timeout) and read in non-blocking mode. DEVRANDOM now 3640109998Smarkm defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" 3641109998Smarkm (previously it was just the string "/dev/urandom"), so on typical 3642109998Smarkm platforms the 10 ms delay will never occur. 3643109998Smarkm Also separate out the Unix variant to its own file, rand_unix.c. 3644109998Smarkm For VMS, there's a currently-empty rand_vms.c. 3645109998Smarkm [Richard Levitte] 3646109998Smarkm 3647109998Smarkm *) Move OCSP client related routines to ocsp_cl.c. These 3648109998Smarkm provide utility functions which an application needing 3649109998Smarkm to issue a request to an OCSP responder and analyse the 3650109998Smarkm response will typically need: as opposed to those which an 3651109998Smarkm OCSP responder itself would need which will be added later. 3652109998Smarkm 3653109998Smarkm OCSP_request_sign() signs an OCSP request with an API similar 3654109998Smarkm to PKCS7_sign(). OCSP_response_status() returns status of OCSP 3655109998Smarkm response. OCSP_response_get1_basic() extracts basic response 3656109998Smarkm from response. OCSP_resp_find_status(): finds and extracts status 3657109998Smarkm information from an OCSP_CERTID structure (which will be created 3658109998Smarkm when the request structure is built). These are built from lower 3659109998Smarkm level functions which work on OCSP_SINGLERESP structures but 3660109998Smarkm wont normally be used unless the application wishes to examine 3661109998Smarkm extensions in the OCSP response for example. 3662109998Smarkm 3663109998Smarkm Replace nonce routines with a pair of functions. 3664109998Smarkm OCSP_request_add1_nonce() adds a nonce value and optionally 3665109998Smarkm generates a random value. OCSP_check_nonce() checks the 3666109998Smarkm validity of the nonce in an OCSP response. 3667109998Smarkm [Steve Henson] 3668109998Smarkm 3669109998Smarkm *) Change function OCSP_request_add() to OCSP_request_add0_id(). 3670109998Smarkm This doesn't copy the supplied OCSP_CERTID and avoids the 3671109998Smarkm need to free up the newly created id. Change return type 3672109998Smarkm to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. 3673109998Smarkm This can then be used to add extensions to the request. 3674109998Smarkm Deleted OCSP_request_new(), since most of its functionality 3675109998Smarkm is now in OCSP_REQUEST_new() (and the case insensitive name 3676109998Smarkm clash) apart from the ability to set the request name which 3677109998Smarkm will be added elsewhere. 3678109998Smarkm [Steve Henson] 3679109998Smarkm 3680109998Smarkm *) Update OCSP API. Remove obsolete extensions argument from 3681109998Smarkm various functions. Extensions are now handled using the new 3682109998Smarkm OCSP extension code. New simple OCSP HTTP function which 3683109998Smarkm can be used to send requests and parse the response. 3684109998Smarkm [Steve Henson] 3685109998Smarkm 3686109998Smarkm *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new 3687109998Smarkm ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN 3688109998Smarkm uses the special reorder version of SET OF to sort the attributes 3689109998Smarkm and reorder them to match the encoded order. This resolves a long 3690109998Smarkm standing problem: a verify on a PKCS7 structure just after signing 3691109998Smarkm it used to fail because the attribute order did not match the 3692109998Smarkm encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: 3693109998Smarkm it uses the received order. This is necessary to tolerate some broken 3694109998Smarkm software that does not order SET OF. This is handled by encoding 3695109998Smarkm as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) 3696109998Smarkm to produce the required SET OF. 3697109998Smarkm [Steve Henson] 3698109998Smarkm 3699109998Smarkm *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and 3700109998Smarkm OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header 3701109998Smarkm files to get correct declarations of the ASN.1 item variables. 3702109998Smarkm [Richard Levitte] 3703109998Smarkm 3704109998Smarkm *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many 3705109998Smarkm PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: 3706109998Smarkm asn1_check_tlen() would sometimes attempt to use 'ctx' when it was 3707109998Smarkm NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). 3708109998Smarkm New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant 3709109998Smarkm ASN1_ITEM and no wrapper functions. 3710109998Smarkm [Steve Henson] 3711109998Smarkm 3712109998Smarkm *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These 3713109998Smarkm replace the old function pointer based I/O routines. Change most of 3714109998Smarkm the *_d2i_bio() and *_d2i_fp() functions to use these. 3715109998Smarkm [Steve Henson] 3716109998Smarkm 3717109998Smarkm *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor 3718109998Smarkm lines, recognice more "algorithms" that can be deselected, and make 3719109998Smarkm it complain about algorithm deselection that isn't recognised. 3720109998Smarkm [Richard Levitte] 3721109998Smarkm 3722109998Smarkm *) New ASN1 functions to handle dup, sign, verify, digest, pack and 3723109998Smarkm unpack operations in terms of ASN1_ITEM. Modify existing wrappers 3724109998Smarkm to use new functions. Add NO_ASN1_OLD which can be set to remove 3725109998Smarkm some old style ASN1 functions: this can be used to determine if old 3726109998Smarkm code will still work when these eventually go away. 3727109998Smarkm [Steve Henson] 3728109998Smarkm 3729109998Smarkm *) New extension functions for OCSP structures, these follow the 3730109998Smarkm same conventions as certificates and CRLs. 3731109998Smarkm [Steve Henson] 3732109998Smarkm 3733109998Smarkm *) New function X509V3_add1_i2d(). This automatically encodes and 3734109998Smarkm adds an extension. Its behaviour can be customised with various 3735109998Smarkm flags to append, replace or delete. Various wrappers added for 3736109998Smarkm certifcates and CRLs. 3737109998Smarkm [Steve Henson] 3738109998Smarkm 3739109998Smarkm *) Fix to avoid calling the underlying ASN1 print routine when 3740109998Smarkm an extension cannot be parsed. Correct a typo in the 3741109998Smarkm OCSP_SERVICELOC extension. Tidy up print OCSP format. 3742109998Smarkm [Steve Henson] 3743109998Smarkm 3744109998Smarkm *) Make mkdef.pl parse some of the ASN1 macros and add apropriate 3745109998Smarkm entries for variables. 3746109998Smarkm [Steve Henson] 3747109998Smarkm 3748109998Smarkm *) Add functionality to apps/openssl.c for detecting locking 3749109998Smarkm problems: As the program is single-threaded, all we have 3750109998Smarkm to do is register a locking callback using an array for 3751109998Smarkm storing which locks are currently held by the program. 3752109998Smarkm [Bodo Moeller] 3753109998Smarkm 3754109998Smarkm *) Use a lock around the call to CRYPTO_get_ex_new_index() in 3755109998Smarkm SSL_get_ex_data_X509_STORE_idx(), which is used in 3756109998Smarkm ssl_verify_cert_chain() and thus can be called at any time 3757109998Smarkm during TLS/SSL handshakes so that thread-safety is essential. 3758109998Smarkm Unfortunately, the ex_data design is not at all suited 3759109998Smarkm for multi-threaded use, so it probably should be abolished. 3760109998Smarkm [Bodo Moeller] 3761109998Smarkm 3762109998Smarkm *) Added Broadcom "ubsec" ENGINE to OpenSSL. 3763109998Smarkm [Broadcom, tweaked and integrated by Geoff Thorpe] 3764109998Smarkm 3765109998Smarkm *) Move common extension printing code to new function 3766109998Smarkm X509V3_print_extensions(). Reorganise OCSP print routines and 3767109998Smarkm implement some needed OCSP ASN1 functions. Add OCSP extensions. 3768109998Smarkm [Steve Henson] 3769109998Smarkm 3770109998Smarkm *) New function X509_signature_print() to remove duplication in some 3771109998Smarkm print routines. 3772109998Smarkm [Steve Henson] 3773109998Smarkm 3774109998Smarkm *) Add a special meaning when SET OF and SEQUENCE OF flags are both 3775109998Smarkm set (this was treated exactly the same as SET OF previously). This 3776109998Smarkm is used to reorder the STACK representing the structure to match the 3777109998Smarkm encoding. This will be used to get round a problem where a PKCS7 3778109998Smarkm structure which was signed could not be verified because the STACK 3779109998Smarkm order did not reflect the encoded order. 3780109998Smarkm [Steve Henson] 3781109998Smarkm 3782109998Smarkm *) Reimplement the OCSP ASN1 module using the new code. 3783109998Smarkm [Steve Henson] 3784109998Smarkm 3785109998Smarkm *) Update the X509V3 code to permit the use of an ASN1_ITEM structure 3786109998Smarkm for its ASN1 operations. The old style function pointers still exist 3787109998Smarkm for now but they will eventually go away. 3788109998Smarkm [Steve Henson] 3789109998Smarkm 3790109998Smarkm *) Merge in replacement ASN1 code from the ASN1 branch. This almost 3791109998Smarkm completely replaces the old ASN1 functionality with a table driven 3792109998Smarkm encoder and decoder which interprets an ASN1_ITEM structure describing 3793109998Smarkm the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is 3794109998Smarkm largely maintained. Almost all of the old asn1_mac.h macro based ASN1 3795109998Smarkm has also been converted to the new form. 3796109998Smarkm [Steve Henson] 3797109998Smarkm 3798109998Smarkm *) Change BN_mod_exp_recp so that negative moduli are tolerated 3799109998Smarkm (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set 3800109998Smarkm so that BN_mod_exp_mont and BN_mod_exp_mont_word work 3801109998Smarkm for negative moduli. 3802109998Smarkm [Bodo Moeller] 3803109998Smarkm 3804109998Smarkm *) Fix BN_uadd and BN_usub: Always return non-negative results instead 3805109998Smarkm of not touching the result's sign bit. 3806109998Smarkm [Bodo Moeller] 3807109998Smarkm 3808109998Smarkm *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be 3809109998Smarkm set. 3810109998Smarkm [Bodo Moeller] 3811109998Smarkm 3812109998Smarkm *) Changed the LHASH code to use prototypes for callbacks, and created 3813109998Smarkm macros to declare and implement thin (optionally static) functions 3814109998Smarkm that provide type-safety and avoid function pointer casting for the 3815109998Smarkm type-specific callbacks. 3816109998Smarkm [Geoff Thorpe] 3817109998Smarkm 3818109998Smarkm *) Added Kerberos Cipher Suites to be used with TLS, as written in 3819109998Smarkm RFC 2712. 3820109998Smarkm [Veers Staats <staatsvr@asc.hpc.mil>, 3821109998Smarkm Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte] 3822109998Smarkm 3823109998Smarkm *) Reformat the FAQ so the different questions and answers can be divided 3824109998Smarkm in sections depending on the subject. 3825109998Smarkm [Richard Levitte] 3826109998Smarkm 3827109998Smarkm *) Have the zlib compression code load ZLIB.DLL dynamically under 3828109998Smarkm Windows. 3829109998Smarkm [Richard Levitte] 3830109998Smarkm 3831109998Smarkm *) New function BN_mod_sqrt for computing square roots modulo a prime 3832109998Smarkm (using the probabilistic Tonelli-Shanks algorithm unless 3833109998Smarkm p == 3 (mod 4) or p == 5 (mod 8), which are cases that can 3834109998Smarkm be handled deterministically). 3835109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 3836109998Smarkm 3837109998Smarkm *) Make BN_mod_inverse faster by explicitly handling small quotients 3838109998Smarkm in the Euclid loop. (Speed gain about 20% for small moduli [256 or 3839109998Smarkm 512 bits], about 30% for larger ones [1024 or 2048 bits].) 3840109998Smarkm [Bodo Moeller] 3841109998Smarkm 3842109998Smarkm *) New function BN_kronecker. 3843109998Smarkm [Bodo Moeller] 3844109998Smarkm 3845109998Smarkm *) Fix BN_gcd so that it works on negative inputs; the result is 3846109998Smarkm positive unless both parameters are zero. 3847109998Smarkm Previously something reasonably close to an infinite loop was 3848109998Smarkm possible because numbers could be growing instead of shrinking 3849109998Smarkm in the implementation of Euclid's algorithm. 3850109998Smarkm [Bodo Moeller] 3851109998Smarkm 3852109998Smarkm *) Fix BN_is_word() and BN_is_one() macros to take into account the 3853109998Smarkm sign of the number in question. 3854109998Smarkm 3855109998Smarkm Fix BN_is_word(a,w) to work correctly for w == 0. 3856109998Smarkm 3857109998Smarkm The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) 3858109998Smarkm because its test if the absolute value of 'a' equals 'w'. 3859109998Smarkm Note that BN_abs_is_word does *not* handle w == 0 reliably; 3860109998Smarkm it exists mostly for use in the implementations of BN_is_zero(), 3861109998Smarkm BN_is_one(), and BN_is_word(). 3862109998Smarkm [Bodo Moeller] 3863109998Smarkm 3864109998Smarkm *) New function BN_swap. 3865109998Smarkm [Bodo Moeller] 3866109998Smarkm 3867109998Smarkm *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that 3868109998Smarkm the exponentiation functions are more likely to produce reasonable 3869109998Smarkm results on negative inputs. 3870109998Smarkm [Bodo Moeller] 3871109998Smarkm 3872109998Smarkm *) Change BN_mod_mul so that the result is always non-negative. 3873109998Smarkm Previously, it could be negative if one of the factors was negative; 3874109998Smarkm I don't think anyone really wanted that behaviour. 3875109998Smarkm [Bodo Moeller] 3876109998Smarkm 3877109998Smarkm *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c 3878109998Smarkm (except for exponentiation, which stays in crypto/bn/bn_exp.c, 3879109998Smarkm and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c) 3880109998Smarkm and add new functions: 3881109998Smarkm 3882109998Smarkm BN_nnmod 3883109998Smarkm BN_mod_sqr 3884109998Smarkm BN_mod_add 3885109998Smarkm BN_mod_add_quick 3886109998Smarkm BN_mod_sub 3887109998Smarkm BN_mod_sub_quick 3888109998Smarkm BN_mod_lshift1 3889109998Smarkm BN_mod_lshift1_quick 3890109998Smarkm BN_mod_lshift 3891109998Smarkm BN_mod_lshift_quick 3892109998Smarkm 3893109998Smarkm These functions always generate non-negative results. 3894109998Smarkm 3895109998Smarkm BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r 3896109998Smarkm such that |m| < r < 0, BN_nnmod will output rem + |m| instead). 3897109998Smarkm 3898109998Smarkm BN_mod_XXX_quick(r, a, [b,] m) generates the same result as 3899109998Smarkm BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b] 3900109998Smarkm be reduced modulo m. 3901109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 3902109998Smarkm 3903111147Snectar#if 0 3904111147Snectar The following entry accidentily appeared in the CHANGES file 3905111147Snectar distributed with OpenSSL 0.9.7. The modifications described in 3906111147Snectar it do *not* apply to OpenSSL 0.9.7. 3907111147Snectar 3908109998Smarkm *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 3909109998Smarkm was actually never needed) and in BN_mul(). The removal in BN_mul() 3910109998Smarkm required a small change in bn_mul_part_recursive() and the addition 3911109998Smarkm of the functions bn_cmp_part_words(), bn_sub_part_words() and 3912109998Smarkm bn_add_part_words(), which do the same thing as bn_cmp_words(), 3913109998Smarkm bn_sub_words() and bn_add_words() except they take arrays with 3914109998Smarkm differing sizes. 3915109998Smarkm [Richard Levitte] 3916111147Snectar#endif 3917109998Smarkm 3918109998Smarkm *) In 'openssl passwd', verify passwords read from the terminal 3919109998Smarkm unless the '-salt' option is used (which usually means that 3920109998Smarkm verification would just waste user's time since the resulting 3921109998Smarkm hash is going to be compared with some given password hash) 3922109998Smarkm or the new '-noverify' option is used. 3923109998Smarkm 3924109998Smarkm This is an incompatible change, but it does not affect 3925109998Smarkm non-interactive use of 'openssl passwd' (passwords on the command 3926109998Smarkm line, '-stdin' option, '-in ...' option) and thus should not 3927109998Smarkm cause any problems. 3928109998Smarkm [Bodo Moeller] 3929109998Smarkm 3930109998Smarkm *) Remove all references to RSAref, since there's no more need for it. 3931109998Smarkm [Richard Levitte] 3932109998Smarkm 3933109998Smarkm *) Make DSO load along a path given through an environment variable 3934109998Smarkm (SHLIB_PATH) with shl_load(). 3935109998Smarkm [Richard Levitte] 3936109998Smarkm 3937109998Smarkm *) Constify the ENGINE code as a result of BIGNUM constification. 3938109998Smarkm Also constify the RSA code and most things related to it. In a 3939109998Smarkm few places, most notable in the depth of the ASN.1 code, ugly 3940109998Smarkm casts back to non-const were required (to be solved at a later 3941109998Smarkm time) 3942109998Smarkm [Richard Levitte] 3943109998Smarkm 3944109998Smarkm *) Make it so the openssl application has all engines loaded by default. 3945109998Smarkm [Richard Levitte] 3946109998Smarkm 3947109998Smarkm *) Constify the BIGNUM routines a little more. 3948109998Smarkm [Richard Levitte] 3949109998Smarkm 3950109998Smarkm *) Add the following functions: 3951109998Smarkm 3952109998Smarkm ENGINE_load_cswift() 3953109998Smarkm ENGINE_load_chil() 3954109998Smarkm ENGINE_load_atalla() 3955109998Smarkm ENGINE_load_nuron() 3956109998Smarkm ENGINE_load_builtin_engines() 3957109998Smarkm 3958109998Smarkm That way, an application can itself choose if external engines that 3959109998Smarkm are built-in in OpenSSL shall ever be used or not. The benefit is 3960109998Smarkm that applications won't have to be linked with libdl or other dso 3961109998Smarkm libraries unless it's really needed. 3962109998Smarkm 3963109998Smarkm Changed 'openssl engine' to load all engines on demand. 3964109998Smarkm Changed the engine header files to avoid the duplication of some 3965109998Smarkm declarations (they differed!). 3966109998Smarkm [Richard Levitte] 3967109998Smarkm 3968109998Smarkm *) 'openssl engine' can now list capabilities. 3969109998Smarkm [Richard Levitte] 3970109998Smarkm 3971109998Smarkm *) Better error reporting in 'openssl engine'. 3972109998Smarkm [Richard Levitte] 3973109998Smarkm 3974109998Smarkm *) Never call load_dh_param(NULL) in s_server. 3975109998Smarkm [Bodo Moeller] 3976109998Smarkm 3977109998Smarkm *) Add engine application. It can currently list engines by name and 3978109998Smarkm identity, and test if they are actually available. 3979109998Smarkm [Richard Levitte] 3980109998Smarkm 3981109998Smarkm *) Improve RPM specification file by forcing symbolic linking and making 3982109998Smarkm sure the installed documentation is also owned by root.root. 3983109998Smarkm [Damien Miller <djm@mindrot.org>] 3984109998Smarkm 3985109998Smarkm *) Give the OpenSSL applications more possibilities to make use of 3986109998Smarkm keys (public as well as private) handled by engines. 3987109998Smarkm [Richard Levitte] 3988109998Smarkm 3989109998Smarkm *) Add OCSP code that comes from CertCo. 3990109998Smarkm [Richard Levitte] 3991109998Smarkm 3992109998Smarkm *) Add VMS support for the Rijndael code. 3993109998Smarkm [Richard Levitte] 3994109998Smarkm 3995109998Smarkm *) Added untested support for Nuron crypto accelerator. 3996109998Smarkm [Ben Laurie] 3997109998Smarkm 3998109998Smarkm *) Add support for external cryptographic devices. This code was 3999109998Smarkm previously distributed separately as the "engine" branch. 4000109998Smarkm [Geoff Thorpe, Richard Levitte] 4001109998Smarkm 4002109998Smarkm *) Rework the filename-translation in the DSO code. It is now possible to 4003109998Smarkm have far greater control over how a "name" is turned into a filename 4004109998Smarkm depending on the operating environment and any oddities about the 4005109998Smarkm different shared library filenames on each system. 4006109998Smarkm [Geoff Thorpe] 4007109998Smarkm 4008109998Smarkm *) Support threads on FreeBSD-elf in Configure. 4009109998Smarkm [Richard Levitte] 4010109998Smarkm 4011109998Smarkm *) Fix for SHA1 assembly problem with MASM: it produces 4012109998Smarkm warnings about corrupt line number information when assembling 4013109998Smarkm with debugging information. This is caused by the overlapping 4014109998Smarkm of two sections. 4015109998Smarkm [Bernd Matthes <mainbug@celocom.de>, Steve Henson] 4016109998Smarkm 4017109998Smarkm *) NCONF changes. 4018109998Smarkm NCONF_get_number() has no error checking at all. As a replacement, 4019109998Smarkm NCONF_get_number_e() is defined (_e for "error checking") and is 4020109998Smarkm promoted strongly. The old NCONF_get_number is kept around for 4021109998Smarkm binary backward compatibility. 4022109998Smarkm Make it possible for methods to load from something other than a BIO, 4023109998Smarkm by providing a function pointer that is given a name instead of a BIO. 4024109998Smarkm For example, this could be used to load configuration data from an 4025109998Smarkm LDAP server. 4026109998Smarkm [Richard Levitte] 4027109998Smarkm 4028109998Smarkm *) Fix for non blocking accept BIOs. Added new I/O special reason 4029109998Smarkm BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs 4030109998Smarkm with non blocking I/O was not possible because no retry code was 4031109998Smarkm implemented. Also added new SSL code SSL_WANT_ACCEPT to cover 4032109998Smarkm this case. 4033109998Smarkm [Steve Henson] 4034109998Smarkm 4035109998Smarkm *) Added the beginnings of Rijndael support. 4036109998Smarkm [Ben Laurie] 4037109998Smarkm 4038109998Smarkm *) Fix for bug in DirectoryString mask setting. Add support for 4039109998Smarkm X509_NAME_print_ex() in 'req' and X509_print_ex() function 4040109998Smarkm to allow certificate printing to more controllable, additional 4041109998Smarkm 'certopt' option to 'x509' to allow new printing options to be 4042109998Smarkm set. 4043109998Smarkm [Steve Henson] 4044109998Smarkm 4045109998Smarkm *) Clean old EAY MD5 hack from e_os.h. 4046109998Smarkm [Richard Levitte] 4047109998Smarkm 4048142425Snectar Changes between 0.9.6l and 0.9.6m [17 Mar 2004] 4049142425Snectar 4050142425Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 4051160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 4052142425Snectar [Joe Orton, Steve Henson] 4053142425Snectar 4054142425Snectar Changes between 0.9.6k and 0.9.6l [04 Nov 2003] 4055142425Snectar 4056142425Snectar *) Fix additional bug revealed by the NISCC test suite: 4057142425Snectar 4058142425Snectar Stop bug triggering large recursion when presented with 4059160814Ssimon certain ASN.1 tags (CVE-2003-0851) 4060142425Snectar [Steve Henson] 4061142425Snectar 4062120631Snectar Changes between 0.9.6j and 0.9.6k [30 Sep 2003] 4063120631Snectar 4064120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 4065120631Snectar 4066120631Snectar Stop out of bounds reads in the ASN1 code when presented with 4067160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 4068120631Snectar 4069120631Snectar If verify callback ignores invalid public key errors don't try to check 4070120631Snectar certificate signature with the NULL public key. 4071120631Snectar 4072120631Snectar [Steve Henson] 4073120631Snectar 4074120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 4075120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 4076120631Snectar specifications. 4077120631Snectar [Steve Henson] 4078120631Snectar 4079120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 4080120631Snectar extra data after the compression methods not only for TLS 1.0 4081120631Snectar but also for SSL 3.0 (as required by the specification). 4082120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 4083120631Snectar 4084120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 4085120631Snectar when it's 512 *bits* long, not 512 bytes. 4086120631Snectar [Richard Levitte] 4087120631Snectar 4088120631Snectar Changes between 0.9.6i and 0.9.6j [10 Apr 2003] 4089120631Snectar 4090120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 4091120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 4092120631Snectar a protocol version number mismatch like a decryption error 4093120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 4094120631Snectar [Bodo Moeller] 4095120631Snectar 4096120631Snectar *) Turn on RSA blinding by default in the default implementation 4097120631Snectar to avoid a timing attack. Applications that don't want it can call 4098120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 4099120631Snectar They would be ill-advised to do so in most cases. 4100120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 4101120631Snectar 4102120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 4103120631Snectar seeded (in this case, the secret RSA exponent is abused as 4104120631Snectar an unpredictable seed -- if it is not unpredictable, there 4105120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 4106120631Snectar by remembering the creator's thread ID in rsa->blinding and 4107120631Snectar having all other threads use local one-time blinding factors 4108120631Snectar (this requires more computation than sharing rsa->blinding, but 4109120631Snectar avoids excessive locking; and if an RSA object is not shared 4110120631Snectar between threads, blinding will still be very fast). 4111120631Snectar [Bodo Moeller] 4112120631Snectar 4113111147Snectar Changes between 0.9.6h and 0.9.6i [19 Feb 2003] 4114111147Snectar 4115111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 4116111147Snectar via timing by performing a MAC computation even if incorrrect 4117111147Snectar block cipher padding has been found. This is a countermeasure 4118111147Snectar against active attacks where the attacker has to distinguish 4119160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 4120111147Snectar 4121111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 4122111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 4123111147Snectar Martin Vuagnoux (EPFL, Ilion)] 4124111147Snectar 4125109998Smarkm Changes between 0.9.6g and 0.9.6h [5 Dec 2002] 4126109998Smarkm 4127109998Smarkm *) New function OPENSSL_cleanse(), which is used to cleanse a section of 4128109998Smarkm memory from it's contents. This is done with a counter that will 4129109998Smarkm place alternating values in each byte. This can be used to solve 4130109998Smarkm two issues: 1) the removal of calls to memset() by highly optimizing 4131109998Smarkm compilers, and 2) cleansing with other values than 0, since those can 4132109998Smarkm be read through on certain media, for example a swap space on disk. 4133109998Smarkm [Geoff Thorpe] 4134109998Smarkm 4135109998Smarkm *) Bugfix: client side session caching did not work with external caching, 4136109998Smarkm because the session->cipher setting was not restored when reloading 4137109998Smarkm from the external cache. This problem was masked, when 4138109998Smarkm SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. 4139109998Smarkm (Found by Steve Haslam <steve@araqnid.ddts.net>.) 4140109998Smarkm [Lutz Jaenicke] 4141109998Smarkm 4142109998Smarkm *) Fix client_certificate (ssl/s2_clnt.c): The permissible total 4143109998Smarkm length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. 4144109998Smarkm [Zeev Lieber <zeev-l@yahoo.com>] 4145109998Smarkm 4146109998Smarkm *) Undo an undocumented change introduced in 0.9.6e which caused 4147109998Smarkm repeated calls to OpenSSL_add_all_ciphers() and 4148109998Smarkm OpenSSL_add_all_digests() to be ignored, even after calling 4149109998Smarkm EVP_cleanup(). 4150109998Smarkm [Richard Levitte] 4151109998Smarkm 4152109998Smarkm *) Change the default configuration reader to deal with last line not 4153109998Smarkm being properly terminated. 4154109998Smarkm [Richard Levitte] 4155109998Smarkm 4156109998Smarkm *) Change X509_NAME_cmp() so it applies the special rules on handling 4157109998Smarkm DN values that are of type PrintableString, as well as RDNs of type 4158109998Smarkm emailAddress where the value has the type ia5String. 4159109998Smarkm [stefank@valicert.com via Richard Levitte] 4160109998Smarkm 4161109998Smarkm *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half 4162109998Smarkm the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently 4163109998Smarkm doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be 4164109998Smarkm the bitwise-OR of the two for use by the majority of applications 4165109998Smarkm wanting this behaviour, and update the docs. The documented 4166109998Smarkm behaviour and actual behaviour were inconsistent and had been 4167109998Smarkm changing anyway, so this is more a bug-fix than a behavioural 4168109998Smarkm change. 4169109998Smarkm [Geoff Thorpe, diagnosed by Nadav Har'El] 4170109998Smarkm 4171109998Smarkm *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c 4172109998Smarkm (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). 4173109998Smarkm [Bodo Moeller] 4174109998Smarkm 4175109998Smarkm *) Fix initialization code race conditions in 4176109998Smarkm SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), 4177109998Smarkm SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), 4178109998Smarkm SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), 4179109998Smarkm TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), 4180109998Smarkm ssl2_get_cipher_by_char(), 4181109998Smarkm ssl3_get_cipher_by_char(). 4182109998Smarkm [Patrick McCormick <patrick@tellme.com>, Bodo Moeller] 4183109998Smarkm 4184109998Smarkm *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after 4185109998Smarkm the cached sessions are flushed, as the remove_cb() might use ex_data 4186109998Smarkm contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> 4187109998Smarkm (see [openssl.org #212]). 4188109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 4189109998Smarkm 4190109998Smarkm *) Fix typo in OBJ_txt2obj which incorrectly passed the content 4191109998Smarkm length, instead of the encoding length to d2i_ASN1_OBJECT. 4192109998Smarkm [Steve Henson] 4193109998Smarkm 4194101618Snectar Changes between 0.9.6f and 0.9.6g [9 Aug 2002] 4195101618Snectar 4196101618Snectar *) [In 0.9.6g-engine release:] 4197101618Snectar Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall'). 4198101618Snectar [Lynn Gazis <lgazis@rainbow.com>] 4199101618Snectar 4200101613Snectar Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 4201101613Snectar 4202101613Snectar *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 4203101613Snectar and get fix the header length calculation. 4204101613Snectar [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 4205101613Snectar Alon Kantor <alonk@checkpoint.com> (and others), 4206101613Snectar Steve Henson] 4207101613Snectar 4208101613Snectar *) Use proper error handling instead of 'assertions' in buffer 4209101613Snectar overflow checks added in 0.9.6e. This prevents DoS (the 4210101613Snectar assertions could call abort()). 4211101613Snectar [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] 4212101613Snectar 4213100936Snectar Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 4214100936Snectar 4215109998Smarkm *) Add various sanity checks to asn1_get_length() to reject 4216109998Smarkm the ASN1 length bytes if they exceed sizeof(long), will appear 4217109998Smarkm negative or the content length exceeds the length of the 4218109998Smarkm supplied buffer. 4219109998Smarkm [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 4220109998Smarkm 4221100936Snectar *) Fix cipher selection routines: ciphers without encryption had no flags 4222100936Snectar for the cipher strength set and where therefore not handled correctly 4223100936Snectar by the selection routines (PR #130). 4224100936Snectar [Lutz Jaenicke] 4225100936Snectar 4226100936Snectar *) Fix EVP_dsa_sha macro. 4227100936Snectar [Nils Larsch] 4228100936Snectar 4229100936Snectar *) New option 4230100936Snectar SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 4231100936Snectar for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 4232100936Snectar that was added in OpenSSL 0.9.6d. 4233100936Snectar 4234100936Snectar As the countermeasure turned out to be incompatible with some 4235100936Snectar broken SSL implementations, the new option is part of SSL_OP_ALL. 4236100936Snectar SSL_OP_ALL is usually employed when compatibility with weird SSL 4237100936Snectar implementations is desired (e.g. '-bugs' option to 's_client' and 4238100936Snectar 's_server'), so the new option is automatically set in many 4239100936Snectar applications. 4240100936Snectar [Bodo Moeller] 4241100936Snectar 4242100936Snectar *) Changes in security patch: 4243100936Snectar 4244100936Snectar Changes marked "(CHATS)" were sponsored by the Defense Advanced 4245100936Snectar Research Projects Agency (DARPA) and Air Force Research Laboratory, 4246100936Snectar Air Force Materiel Command, USAF, under agreement number 4247100936Snectar F30602-01-2-0537. 4248100936Snectar 4249100936Snectar *) Add various sanity checks to asn1_get_length() to reject 4250100936Snectar the ASN1 length bytes if they exceed sizeof(long), will appear 4251100936Snectar negative or the content length exceeds the length of the 4252160814Ssimon supplied buffer. (CVE-2002-0659) 4253100936Snectar [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 4254100936Snectar 4255100936Snectar *) Assertions for various potential buffer overflows, not known to 4256100936Snectar happen in practice. 4257100936Snectar [Ben Laurie (CHATS)] 4258100936Snectar 4259100936Snectar *) Various temporary buffers to hold ASCII versions of integers were 4260160814Ssimon too small for 64 bit platforms. (CVE-2002-0655) 4261100936Snectar [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 4262100936Snectar 4263100936Snectar *) Remote buffer overflow in SSL3 protocol - an attacker could 4264160814Ssimon supply an oversized session ID to a client. (CVE-2002-0656) 4265100936Snectar [Ben Laurie (CHATS)] 4266100936Snectar 4267100936Snectar *) Remote buffer overflow in SSL2 protocol - an attacker could 4268160814Ssimon supply an oversized client master key. (CVE-2002-0656) 4269100936Snectar [Ben Laurie (CHATS)] 4270100936Snectar 4271100928Snectar Changes between 0.9.6c and 0.9.6d [9 May 2002] 4272100928Snectar 4273100928Snectar *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 4274100928Snectar encoded as NULL) with id-dsa-with-sha1. 4275100928Snectar [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 4276100928Snectar 4277100928Snectar *) Check various X509_...() return values in apps/req.c. 4278100928Snectar [Nils Larsch <nla@trustcenter.de>] 4279100928Snectar 4280100928Snectar *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 4281100928Snectar an end-of-file condition would erronously be flagged, when the CRLF 4282100928Snectar was just at the end of a processed block. The bug was discovered when 4283100928Snectar processing data through a buffering memory BIO handing the data to a 4284100928Snectar BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 4285100928Snectar <ptsekov@syntrex.com> and Nedelcho Stanev. 4286100928Snectar [Lutz Jaenicke] 4287100928Snectar 4288100928Snectar *) Implement a countermeasure against a vulnerability recently found 4289100928Snectar in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 4290100928Snectar before application data chunks to avoid the use of known IVs 4291100928Snectar with data potentially chosen by the attacker. 4292100928Snectar [Bodo Moeller] 4293100928Snectar 4294100928Snectar *) Fix length checks in ssl3_get_client_hello(). 4295100928Snectar [Bodo Moeller] 4296100928Snectar 4297100928Snectar *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 4298100928Snectar to prevent ssl3_read_internal() from incorrectly assuming that 4299100928Snectar ssl3_read_bytes() found application data while handshake 4300100928Snectar processing was enabled when in fact s->s3->in_read_app_data was 4301100928Snectar merely automatically cleared during the initial handshake. 4302100928Snectar [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 4303100928Snectar 4304100928Snectar *) Fix object definitions for Private and Enterprise: they were not 4305100928Snectar recognized in their shortname (=lowercase) representation. Extend 4306100928Snectar obj_dat.pl to issue an error when using undefined keywords instead 4307100928Snectar of silently ignoring the problem (Svenning Sorensen 4308100928Snectar <sss@sss.dnsalias.net>). 4309100928Snectar [Lutz Jaenicke] 4310100928Snectar 4311100928Snectar *) Fix DH_generate_parameters() so that it works for 'non-standard' 4312100928Snectar generators, i.e. generators other than 2 and 5. (Previously, the 4313100928Snectar code did not properly initialise the 'add' and 'rem' values to 4314100928Snectar BN_generate_prime().) 4315100928Snectar 4316100928Snectar In the new general case, we do not insist that 'generator' is 4317100928Snectar actually a primitive root: This requirement is rather pointless; 4318100928Snectar a generator of the order-q subgroup is just as good, if not 4319100928Snectar better. 4320100928Snectar [Bodo Moeller] 4321100928Snectar 4322100928Snectar *) Map new X509 verification errors to alerts. Discovered and submitted by 4323100928Snectar Tom Wu <tom@arcot.com>. 4324100928Snectar [Lutz Jaenicke] 4325100928Snectar 4326100928Snectar *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 4327100928Snectar returning non-zero before the data has been completely received 4328100928Snectar when using non-blocking I/O. 4329100928Snectar [Bodo Moeller; problem pointed out by John Hughes] 4330100928Snectar 4331100928Snectar *) Some of the ciphers missed the strength entry (SSL_LOW etc). 4332100928Snectar [Ben Laurie, Lutz Jaenicke] 4333100928Snectar 4334100928Snectar *) Fix bug in SSL_clear(): bad sessions were not removed (found by 4335100928Snectar Yoram Zahavi <YoramZ@gilian.com>). 4336100928Snectar [Lutz Jaenicke] 4337100928Snectar 4338100928Snectar *) Add information about CygWin 1.3 and on, and preserve proper 4339100928Snectar configuration for the versions before that. 4340100928Snectar [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 4341100928Snectar 4342100928Snectar *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 4343100928Snectar check whether we deal with a copy of a session and do not delete from 4344100928Snectar the cache in this case. Problem reported by "Izhar Shoshani Levi" 4345100928Snectar <izhar@checkpoint.com>. 4346100928Snectar [Lutz Jaenicke] 4347100928Snectar 4348100928Snectar *) Do not store session data into the internal session cache, if it 4349100928Snectar is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 4350100928Snectar flag is set). Proposed by Aslam <aslam@funk.com>. 4351100928Snectar [Lutz Jaenicke] 4352100928Snectar 4353100928Snectar *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 4354100928Snectar value is 0. 4355100928Snectar [Richard Levitte] 4356100928Snectar 4357109998Smarkm *) [In 0.9.6d-engine release:] 4358109998Smarkm Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 4359100928Snectar [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 4360100928Snectar 4361100928Snectar *) Add the configuration target linux-s390x. 4362100928Snectar [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 4363100928Snectar 4364100928Snectar *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 4365100928Snectar ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 4366100928Snectar variable as an indication that a ClientHello message has been 4367100928Snectar received. As the flag value will be lost between multiple 4368100928Snectar invocations of ssl3_accept when using non-blocking I/O, the 4369100928Snectar function may not be aware that a handshake has actually taken 4370100928Snectar place, thus preventing a new session from being added to the 4371100928Snectar session cache. 4372100928Snectar 4373100928Snectar To avoid this problem, we now set s->new_session to 2 instead of 4374100928Snectar using a local variable. 4375100928Snectar [Lutz Jaenicke, Bodo Moeller] 4376100928Snectar 4377100928Snectar *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 4378100928Snectar if the SSL_R_LENGTH_MISMATCH error is detected. 4379100928Snectar [Geoff Thorpe, Bodo Moeller] 4380100928Snectar 4381100928Snectar *) New 'shared_ldflag' column in Configure platform table. 4382100928Snectar [Richard Levitte] 4383100928Snectar 4384100928Snectar *) Fix EVP_CIPHER_mode macro. 4385100928Snectar ["Dan S. Camper" <dan@bti.net>] 4386100928Snectar 4387100928Snectar *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 4388100928Snectar type, we must throw them away by setting rr->length to 0. 4389100928Snectar [D P Chang <dpc@qualys.com>] 4390100928Snectar 439189837Skris Changes between 0.9.6b and 0.9.6c [21 dec 2001] 439289837Skris 439389837Skris *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 439489837Skris <Dominikus.Scherkl@biodata.com>. (The previous implementation 439589837Skris worked incorrectly for those cases where range = 10..._2 and 439689837Skris 3*range is two bits longer than range.) 439789837Skris [Bodo Moeller] 439889837Skris 439989837Skris *) Only add signing time to PKCS7 structures if it is not already 440089837Skris present. 440189837Skris [Steve Henson] 440289837Skris 440389837Skris *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 440489837Skris OBJ_ld_ce should be OBJ_id_ce. 440589837Skris Also some ip-pda OIDs in crypto/objects/objects.txt were 440689837Skris incorrect (cf. RFC 3039). 440789837Skris [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 440889837Skris 440989837Skris *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 441089837Skris returns early because it has nothing to do. 441189837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 441289837Skris 441389837Skris *) [In 0.9.6c-engine release:] 441489837Skris Fix mutex callback return values in crypto/engine/hw_ncipher.c. 441589837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 441689837Skris 441789837Skris *) [In 0.9.6c-engine release:] 441889837Skris Add support for Cryptographic Appliance's keyserver technology. 441989837Skris (Use engine 'keyclient') 442089837Skris [Cryptographic Appliances and Geoff Thorpe] 442189837Skris 442289837Skris *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 442389837Skris is called via tools/c89.sh because arguments have to be 442489837Skris rearranged (all '-L' options must appear before the first object 442589837Skris modules). 442689837Skris [Richard Shapiro <rshapiro@abinitio.com>] 442789837Skris 442889837Skris *) [In 0.9.6c-engine release:] 442989837Skris Add support for Broadcom crypto accelerator cards, backported 443089837Skris from 0.9.7. 443189837Skris [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 443289837Skris 443389837Skris *) [In 0.9.6c-engine release:] 443489837Skris Add support for SureWare crypto accelerator cards from 443589837Skris Baltimore Technologies. (Use engine 'sureware') 443689837Skris [Baltimore Technologies and Mark Cox] 443789837Skris 443889837Skris *) [In 0.9.6c-engine release:] 443989837Skris Add support for crypto accelerator cards from Accelerated 444089837Skris Encryption Processing, www.aep.ie. (Use engine 'aep') 444189837Skris [AEP Inc. and Mark Cox] 444289837Skris 444389837Skris *) Add a configuration entry for gcc on UnixWare. 444489837Skris [Gary Benson <gbenson@redhat.com>] 444589837Skris 444689837Skris *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 444789837Skris messages are stored in a single piece (fixed-length part and 444889837Skris variable-length part combined) and fix various bugs found on the way. 444989837Skris [Bodo Moeller] 445089837Skris 445189837Skris *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 445289837Skris instead. BIO_gethostbyname() does not know what timeouts are 445389837Skris appropriate, so entries would stay in cache even when they have 445489837Skris become invalid. 445589837Skris [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 445689837Skris 445789837Skris *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 445889837Skris faced with a pathologically small ClientHello fragment that does 445989837Skris not contain client_version: Instead of aborting with an error, 446089837Skris simply choose the highest available protocol version (i.e., 446189837Skris TLS 1.0 unless it is disabled). In practice, ClientHello 446289837Skris messages are never sent like this, but this change gives us 446389837Skris strictly correct behaviour at least for TLS. 446489837Skris [Bodo Moeller] 446589837Skris 446689837Skris *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 446789837Skris never resets s->method to s->ctx->method when called from within 446889837Skris one of the SSL handshake functions. 446989837Skris [Bodo Moeller; problem pointed out by Niko Baric] 447089837Skris 447189837Skris *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 447289837Skris (sent using the client's version number) if client_version is 447389837Skris smaller than the protocol version in use. Also change 447489837Skris ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 447589837Skris the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 447689837Skris the client will at least see that alert. 447789837Skris [Bodo Moeller] 447889837Skris 447989837Skris *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 448089837Skris correctly. 448189837Skris [Bodo Moeller] 448289837Skris 448389837Skris *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 448489837Skris client receives HelloRequest while in a handshake. 448589837Skris [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 448689837Skris 448789837Skris *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 448889837Skris should end in 'break', not 'goto end' which circuments various 448989837Skris cleanups done in state SSL_ST_OK. But session related stuff 449089837Skris must be disabled for SSL_ST_OK in the case that we just sent a 449189837Skris HelloRequest. 449289837Skris 449389837Skris Also avoid some overhead by not calling ssl_init_wbio_buffer() 449489837Skris before just sending a HelloRequest. 449589837Skris [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 449689837Skris 449789837Skris *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 449889837Skris reveal whether illegal block cipher padding was found or a MAC 449989837Skris verification error occured. (Neither SSLerr() codes nor alerts 450089837Skris are directly visible to potential attackers, but the information 450189837Skris may leak via logfiles.) 450289837Skris 450389837Skris Similar changes are not required for the SSL 2.0 implementation 450489837Skris because the number of padding bytes is sent in clear for SSL 2.0, 450589837Skris and the extra bytes are just ignored. However ssl/s2_pkt.c 450689837Skris failed to verify that the purported number of padding bytes is in 450789837Skris the legal range. 450889837Skris [Bodo Moeller] 450989837Skris 451089837Skris *) Add OpenUNIX-8 support including shared libraries 451189837Skris (Boyd Lynn Gerber <gerberb@zenez.com>). 451289837Skris [Lutz Jaenicke] 451389837Skris 451489837Skris *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 451589837Skris 'wristwatch attack' using huge encoding parameters (cf. 451689837Skris James H. Manger's CRYPTO 2001 paper). Note that the 451789837Skris RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 451889837Skris encoding parameters and hence was not vulnerable. 451989837Skris [Bodo Moeller] 452089837Skris 452189837Skris *) BN_sqr() bug fix. 452289837Skris [Ulf M�ller, reported by Jim Ellis <jim.ellis@cavium.com>] 452389837Skris 452489837Skris *) Rabin-Miller test analyses assume uniformly distributed witnesses, 452589837Skris so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 452689837Skris followed by modular reduction. 452789837Skris [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 452889837Skris 452989837Skris *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 453089837Skris equivalent based on BN_pseudo_rand() instead of BN_rand(). 453189837Skris [Bodo Moeller] 453289837Skris 453389837Skris *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 453489837Skris This function was broken, as the check for a new client hello message 453589837Skris to handle SGC did not allow these large messages. 453689837Skris (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 453789837Skris [Lutz Jaenicke] 453889837Skris 453989837Skris *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 454089837Skris [Lutz Jaenicke] 454189837Skris 454289837Skris *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 454389837Skris for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 454489837Skris [Lutz Jaenicke] 454589837Skris 454689837Skris *) Rework the configuration and shared library support for Tru64 Unix. 454789837Skris The configuration part makes use of modern compiler features and 454889837Skris still retains old compiler behavior for those that run older versions 454989837Skris of the OS. The shared library support part includes a variant that 455089837Skris uses the RPATH feature, and is available through the special 455189837Skris configuration target "alpha-cc-rpath", which will never be selected 455289837Skris automatically. 455389837Skris [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 455489837Skris 455589837Skris *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 455689837Skris with the same message size as in ssl3_get_certificate_request(). 455789837Skris Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 455889837Skris messages might inadvertently be reject as too long. 455989837Skris [Petr Lampa <lampa@fee.vutbr.cz>] 456089837Skris 456189837Skris *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 456289837Skris [Andy Polyakov] 456389837Skris 456489837Skris *) Modified SSL library such that the verify_callback that has been set 456589837Skris specificly for an SSL object with SSL_set_verify() is actually being 456689837Skris used. Before the change, a verify_callback set with this function was 456789837Skris ignored and the verify_callback() set in the SSL_CTX at the time of 456889837Skris the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 456989837Skris to allow the necessary settings. 457089837Skris [Lutz Jaenicke] 457189837Skris 457289837Skris *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 457389837Skris explicitly to NULL, as at least on Solaris 8 this seems not always to be 457489837Skris done automatically (in contradiction to the requirements of the C 457589837Skris standard). This made problems when used from OpenSSH. 457689837Skris [Lutz Jaenicke] 457789837Skris 457889837Skris *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 457989837Skris dh->length and always used 458089837Skris 458189837Skris BN_rand_range(priv_key, dh->p). 458289837Skris 458389837Skris BN_rand_range() is not necessary for Diffie-Hellman, and this 458489837Skris specific range makes Diffie-Hellman unnecessarily inefficient if 458589837Skris dh->length (recommended exponent length) is much smaller than the 458689837Skris length of dh->p. We could use BN_rand_range() if the order of 458789837Skris the subgroup was stored in the DH structure, but we only have 458889837Skris dh->length. 458989837Skris 459089837Skris So switch back to 459189837Skris 459289837Skris BN_rand(priv_key, l, ...) 459389837Skris 459489837Skris where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 459589837Skris otherwise. 459689837Skris [Bodo Moeller] 459789837Skris 459889837Skris *) In 459989837Skris 460089837Skris RSA_eay_public_encrypt 460189837Skris RSA_eay_private_decrypt 460289837Skris RSA_eay_private_encrypt (signing) 460389837Skris RSA_eay_public_decrypt (signature verification) 460489837Skris 460589837Skris (default implementations for RSA_public_encrypt, 460689837Skris RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 460789837Skris always reject numbers >= n. 460889837Skris [Bodo Moeller] 460989837Skris 461089837Skris *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 461189837Skris to synchronize access to 'locking_thread'. This is necessary on 461289837Skris systems where access to 'locking_thread' (an 'unsigned long' 461389837Skris variable) is not atomic. 461489837Skris [Bodo Moeller] 461589837Skris 461689837Skris *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 461789837Skris *before* setting the 'crypto_lock_rand' flag. The previous code had 461889837Skris a race condition if 0 is a valid thread ID. 461989837Skris [Travis Vitek <vitek@roguewave.com>] 462089837Skris 462189837Skris *) Add support for shared libraries under Irix. 462289837Skris [Albert Chin-A-Young <china@thewrittenword.com>] 462389837Skris 462489837Skris *) Add configuration option to build on Linux on both big-endian and 462589837Skris little-endian MIPS. 462689837Skris [Ralf Baechle <ralf@uni-koblenz.de>] 462789837Skris 462889837Skris *) Add the possibility to create shared libraries on HP-UX. 462989837Skris [Richard Levitte] 463089837Skris 463179998Skris Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 463279998Skris 463379998Skris *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 463479998Skris to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 463579998Skris Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 463679998Skris PRNG state recovery was possible based on the output of 463779998Skris one PRNG request appropriately sized to gain knowledge on 463879998Skris 'md' followed by enough consecutive 1-byte PRNG requests 463979998Skris to traverse all of 'state'. 464079998Skris 464179998Skris 1. When updating 'md_local' (the current thread's copy of 'md') 464279998Skris during PRNG output generation, hash all of the previous 464379998Skris 'md_local' value, not just the half used for PRNG output. 464479998Skris 464579998Skris 2. Make the number of bytes from 'state' included into the hash 464679998Skris independent from the number of PRNG bytes requested. 464779998Skris 464879998Skris The first measure alone would be sufficient to avoid 464979998Skris Markku-Juhani's attack. (Actually it had never occurred 465079998Skris to me that the half of 'md_local' used for chaining was the 465179998Skris half from which PRNG output bytes were taken -- I had always 465279998Skris assumed that the secret half would be used.) The second 465379998Skris measure makes sure that additional data from 'state' is never 465479998Skris mixed into 'md_local' in small portions; this heuristically 465579998Skris further strengthens the PRNG. 465679998Skris [Bodo Moeller] 465779998Skris 465879998Skris *) Fix crypto/bn/asm/mips3.s. 465979998Skris [Andy Polyakov] 466079998Skris 466179998Skris *) When only the key is given to "enc", the IV is undefined. Print out 466279998Skris an error message in this case. 466379998Skris [Lutz Jaenicke] 466479998Skris 466579998Skris *) Handle special case when X509_NAME is empty in X509 printing routines. 466679998Skris [Steve Henson] 466779998Skris 466879998Skris *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 466979998Skris positive and less than q. 467079998Skris [Bodo Moeller] 467179998Skris 467279998Skris *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 467379998Skris used: it isn't thread safe and the add_lock_callback should handle 467479998Skris that itself. 467579998Skris [Paul Rose <Paul.Rose@bridge.com>] 467679998Skris 467779998Skris *) Verify that incoming data obeys the block size in 467879998Skris ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 467979998Skris [Bodo Moeller] 468079998Skris 468179998Skris *) Fix OAEP check. 468279998Skris [Ulf M�ller, Bodo M�ller] 468379998Skris 468479998Skris *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 468589837Skris RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 468679998Skris when fixing the server behaviour for backwards-compatible 'client 468779998Skris hello' messages. (Note that the attack is impractical against 468879998Skris SSL 3.0 and TLS 1.0 anyway because length and version checking 468979998Skris means that the probability of guessing a valid ciphertext is 469079998Skris around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 469179998Skris paper.) 469279998Skris 469379998Skris Before 0.9.5, the countermeasure (hide the error by generating a 469479998Skris random 'decryption result') did not work properly because 469579998Skris ERR_clear_error() was missing, meaning that SSL_get_error() would 469679998Skris detect the supposedly ignored error. 469779998Skris 469879998Skris Both problems are now fixed. 469979998Skris [Bodo Moeller] 470079998Skris 470179998Skris *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 470279998Skris (previously it was 1024). 470379998Skris [Bodo Moeller] 470479998Skris 470579998Skris *) Fix for compatibility mode trust settings: ignore trust settings 470679998Skris unless some valid trust or reject settings are present. 470779998Skris [Steve Henson] 470879998Skris 470979998Skris *) Fix for blowfish EVP: its a variable length cipher. 471079998Skris [Steve Henson] 471179998Skris 471279998Skris *) Fix various bugs related to DSA S/MIME verification. Handle missing 471379998Skris parameters in DSA public key structures and return an error in the 471479998Skris DSA routines if parameters are absent. 471579998Skris [Steve Henson] 471679998Skris 471779998Skris *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 471879998Skris in the current directory if neither $RANDFILE nor $HOME was set. 471979998Skris RAND_file_name() in 0.9.6a returned NULL in this case. This has 472079998Skris caused some confusion to Windows users who haven't defined $HOME. 472179998Skris Thus RAND_file_name() is changed again: e_os.h can define a 472279998Skris DEFAULT_HOME, which will be used if $HOME is not set. 472379998Skris For Windows, we use "C:"; on other platforms, we still require 472479998Skris environment variables. 472579998Skris 472679998Skris *) Move 'if (!initialized) RAND_poll()' into regions protected by 472779998Skris CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 472879998Skris having multiple threads call RAND_poll() concurrently. 472979998Skris [Bodo Moeller] 473079998Skris 473179998Skris *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 473279998Skris combination of a flag and a thread ID variable. 473379998Skris Otherwise while one thread is in ssleay_rand_bytes (which sets the 473479998Skris flag), *other* threads can enter ssleay_add_bytes without obeying 473589837Skris the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 473679998Skris that they do not hold after the first thread unsets add_do_not_lock). 473779998Skris [Bodo Moeller] 473879998Skris 473979998Skris *) Change bctest again: '-x' expressions are not available in all 474079998Skris versions of 'test'. 474179998Skris [Bodo Moeller] 474279998Skris 474376866Skris Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 474472613Skris 474576866Skris *) Fix a couple of memory leaks in PKCS7_dataDecode() 474676866Skris [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 474776866Skris 474876866Skris *) Change Configure and Makefiles to provide EXE_EXT, which will contain 474976866Skris the default extension for executables, if any. Also, make the perl 475076866Skris scripts that use symlink() to test if it really exists and use "cp" 475176866Skris if it doesn't. All this made OpenSSL compilable and installable in 475276866Skris CygWin. 475376866Skris [Richard Levitte] 475476866Skris 475576866Skris *) Fix for asn1_GetSequence() for indefinite length constructed data. 475676866Skris If SEQUENCE is length is indefinite just set c->slen to the total 475776866Skris amount of data available. 475876866Skris [Steve Henson, reported by shige@FreeBSD.org] 475976866Skris [This change does not apply to 0.9.7.] 476076866Skris 476176866Skris *) Change bctest to avoid here-documents inside command substitution 476276866Skris (workaround for FreeBSD /bin/sh bug). 476376866Skris For compatibility with Ultrix, avoid shell functions (introduced 476476866Skris in the bctest version that searches along $PATH). 476576866Skris [Bodo Moeller] 476676866Skris 476776866Skris *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 476876866Skris with des_encrypt() defined on some operating systems, like Solaris 476976866Skris and UnixWare. 477076866Skris [Richard Levitte] 477176866Skris 477276866Skris *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 477376866Skris On the Importance of Eliminating Errors in Cryptographic 477476866Skris Computations, J. Cryptology 14 (2001) 2, 101-119, 477576866Skris http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 477676866Skris [Ulf Moeller] 477776866Skris 477876866Skris *) MIPS assembler BIGNUM division bug fix. 477976866Skris [Andy Polyakov] 478076866Skris 478176866Skris *) Disabled incorrect Alpha assembler code. 478276866Skris [Richard Levitte] 478376866Skris 478476866Skris *) Fix PKCS#7 decode routines so they correctly update the length 478576866Skris after reading an EOC for the EXPLICIT tag. 478676866Skris [Steve Henson] 478776866Skris [This change does not apply to 0.9.7.] 478876866Skris 478976866Skris *) Fix bug in PKCS#12 key generation routines. This was triggered 479076866Skris if a 3DES key was generated with a 0 initial byte. Include 479176866Skris PKCS12_BROKEN_KEYGEN compilation option to retain the old 479276866Skris (but broken) behaviour. 479376866Skris [Steve Henson] 479476866Skris 479576866Skris *) Enhance bctest to search for a working bc along $PATH and print 479676866Skris it when found. 479776866Skris [Tim Rice <tim@multitalents.net> via Richard Levitte] 479876866Skris 479976866Skris *) Fix memory leaks in err.c: free err_data string if necessary; 480076866Skris don't write to the wrong index in ERR_set_error_data. 480176866Skris [Bodo Moeller] 480276866Skris 480376866Skris *) Implement ssl23_peek (analogous to ssl23_read), which previously 480476866Skris did not exist. 480576866Skris [Bodo Moeller] 480676866Skris 480776866Skris *) Replace rdtsc with _emit statements for VC++ version 5. 480876866Skris [Jeremy Cooper <jeremy@baymoo.org>] 480976866Skris 481076866Skris *) Make it possible to reuse SSLv2 sessions. 481176866Skris [Richard Levitte] 481276866Skris 481376866Skris *) In copy_email() check for >= 0 as a return value for 481476866Skris X509_NAME_get_index_by_NID() since 0 is a valid index. 481576866Skris [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 481676866Skris 481776866Skris *) Avoid coredump with unsupported or invalid public keys by checking if 481876866Skris X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 481976866Skris PKCS7_verify() fails with non detached data. 482076866Skris [Steve Henson] 482176866Skris 482276866Skris *) Don't use getenv in library functions when run as setuid/setgid. 482376866Skris New function OPENSSL_issetugid(). 482476866Skris [Ulf Moeller] 482576866Skris 482676866Skris *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 482776866Skris due to incorrect handling of multi-threading: 482876866Skris 482976866Skris 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 483076866Skris 483176866Skris 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 483276866Skris 483376866Skris 3. Count how many times MemCheck_off() has been called so that 483476866Skris nested use can be treated correctly. This also avoids 483576866Skris inband-signalling in the previous code (which relied on the 483676866Skris assumption that thread ID 0 is impossible). 483776866Skris [Bodo Moeller] 483876866Skris 483976866Skris *) Add "-rand" option also to s_client and s_server. 484076866Skris [Lutz Jaenicke] 484176866Skris 484276866Skris *) Fix CPU detection on Irix 6.x. 484376866Skris [Kurt Hockenbury <khockenb@stevens-tech.edu> and 484476866Skris "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 484576866Skris 484676866Skris *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 484776866Skris was empty. 484876866Skris [Steve Henson] 484976866Skris [This change does not apply to 0.9.7.] 485076866Skris 485176866Skris *) Use the cached encoding of an X509_NAME structure rather than 485276866Skris copying it. This is apparently the reason for the libsafe "errors" 485376866Skris but the code is actually correct. 485476866Skris [Steve Henson] 485576866Skris 485672613Skris *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 485772613Skris Bleichenbacher's DSA attack. 485876866Skris Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 485976866Skris to be set and top=0 forces the highest bit to be set; top=-1 is new 486076866Skris and leaves the highest bit random. 486176866Skris [Ulf Moeller, Bodo Moeller] 486272613Skris 486372613Skris *) In the NCONF_...-based implementations for CONF_... queries 486472613Skris (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 486572613Skris a temporary CONF structure with the data component set to NULL 486672613Skris (which gives segmentation faults in lh_retrieve). 486772613Skris Instead, use NULL for the CONF pointer in CONF_get_string and 486872613Skris CONF_get_number (which may use environment variables) and directly 486972613Skris return NULL from CONF_get_section. 487072613Skris [Bodo Moeller] 487172613Skris 487272613Skris *) Fix potential buffer overrun for EBCDIC. 487372613Skris [Ulf Moeller] 487472613Skris 487572613Skris *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 487672613Skris keyUsage if basicConstraints absent for a CA. 487772613Skris [Steve Henson] 487872613Skris 487972613Skris *) Make SMIME_write_PKCS7() write mail header values with a format that 488072613Skris is more generally accepted (no spaces before the semicolon), since 488172613Skris some programs can't parse those values properly otherwise. Also make 488272613Skris sure BIO's that break lines after each write do not create invalid 488372613Skris headers. 488472613Skris [Richard Levitte] 488572613Skris 488672613Skris *) Make the CRL encoding routines work with empty SEQUENCE OF. The 488772613Skris macros previously used would not encode an empty SEQUENCE OF 488872613Skris and break the signature. 488972613Skris [Steve Henson] 489076866Skris [This change does not apply to 0.9.7.] 489172613Skris 489272613Skris *) Zero the premaster secret after deriving the master secret in 489372613Skris DH ciphersuites. 489472613Skris [Steve Henson] 489572613Skris 489672613Skris *) Add some EVP_add_digest_alias registrations (as found in 489772613Skris OpenSSL_add_all_digests()) to SSL_library_init() 489872613Skris aka OpenSSL_add_ssl_algorithms(). This provides improved 489972613Skris compatibility with peers using X.509 certificates 490072613Skris with unconventional AlgorithmIdentifier OIDs. 490172613Skris [Bodo Moeller] 490272613Skris 490372613Skris *) Fix for Irix with NO_ASM. 490472613Skris ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 490572613Skris 490672613Skris *) ./config script fixes. 490772613Skris [Ulf Moeller, Richard Levitte] 490872613Skris 490972613Skris *) Fix 'openssl passwd -1'. 491072613Skris [Bodo Moeller] 491172613Skris 491272613Skris *) Change PKCS12_key_gen_asc() so it can cope with non null 491372613Skris terminated strings whose length is passed in the passlen 491472613Skris parameter, for example from PEM callbacks. This was done 491572613Skris by adding an extra length parameter to asc2uni(). 491672613Skris [Steve Henson, reported by <oddissey@samsung.co.kr>] 491772613Skris 491872613Skris *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 491972613Skris call failed, free the DSA structure. 492072613Skris [Bodo Moeller] 492172613Skris 492272613Skris *) Fix to uni2asc() to cope with zero length Unicode strings. 492372613Skris These are present in some PKCS#12 files. 492472613Skris [Steve Henson] 492572613Skris 492672613Skris *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 492772613Skris Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 492872613Skris when writing a 32767 byte record. 492972613Skris [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 493072613Skris 493172613Skris *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 493272613Skris obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 493372613Skris 493472613Skris (RSA objects have a reference count access to which is protected 493572613Skris by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 493672613Skris so they are meant to be shared between threads.) 493772613Skris [Bodo Moeller, Geoff Thorpe; original patch submitted by 493872613Skris "Reddie, Steven" <Steven.Reddie@ca.com>] 493972613Skris 494072613Skris *) Fix a deadlock in CRYPTO_mem_leaks(). 494172613Skris [Bodo Moeller] 494272613Skris 494376866Skris *) Use better test patterns in bntest. 494476866Skris [Ulf M�ller] 494576866Skris 494672613Skris *) rand_win.c fix for Borland C. 494772613Skris [Ulf M�ller] 494872613Skris 494972613Skris *) BN_rshift bugfix for n == 0. 495072613Skris [Bodo Moeller] 495172613Skris 495276866Skris *) Add a 'bctest' script that checks for some known 'bc' bugs 495376866Skris so that 'make test' does not abort just because 'bc' is broken. 495476866Skris [Bodo Moeller] 495576866Skris 495672613Skris *) Store verify_result within SSL_SESSION also for client side to 495772613Skris avoid potential security hole. (Re-used sessions on the client side 495872613Skris always resulted in verify_result==X509_V_OK, not using the original 495972613Skris result of the server certificate verification.) 496072613Skris [Lutz Jaenicke] 496172613Skris 496272613Skris *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 496372613Skris SSL3_RT_APPLICATION_DATA, return 0. 496472613Skris Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 496572613Skris [Bodo Moeller] 496672613Skris 496772613Skris *) Fix SSL_peek: 496872613Skris Both ssl2_peek and ssl3_peek, which were totally broken in earlier 496972613Skris releases, have been re-implemented by renaming the previous 497072613Skris implementations of ssl2_read and ssl3_read to ssl2_read_internal 497172613Skris and ssl3_read_internal, respectively, and adding 'peek' parameters 497272613Skris to them. The new ssl[23]_{read,peek} functions are calls to 497372613Skris ssl[23]_read_internal with the 'peek' flag set appropriately. 497472613Skris A 'peek' parameter has also been added to ssl3_read_bytes, which 497572613Skris does the actual work for ssl3_read_internal. 497672613Skris [Bodo Moeller] 497772613Skris 497876866Skris *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 497976866Skris the method-specific "init()" handler. Also clean up ex_data after 498076866Skris calling the method-specific "finish()" handler. Previously, this was 498176866Skris happening the other way round. 498276866Skris [Geoff Thorpe] 498376866Skris 498472613Skris *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 498572613Skris The previous value, 12, was not always sufficient for BN_mod_exp(). 498672613Skris [Bodo Moeller] 498772613Skris 498876866Skris *) Make sure that shared libraries get the internal name engine with 498976866Skris the full version number and not just 0. This should mark the 499076866Skris shared libraries as not backward compatible. Of course, this should 499176866Skris be changed again when we can guarantee backward binary compatibility. 499276866Skris [Richard Levitte] 499376866Skris 499472613Skris *) Fix typo in get_cert_by_subject() in by_dir.c 499572613Skris [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 499672613Skris 499776866Skris *) Rework the system to generate shared libraries: 499876866Skris 499976866Skris - Make note of the expected extension for the shared libraries and 500076866Skris if there is a need for symbolic links from for example libcrypto.so.0 500176866Skris to libcrypto.so.0.9.7. There is extended info in Configure for 500276866Skris that. 500376866Skris 500476866Skris - Make as few rebuilds of the shared libraries as possible. 500576866Skris 500676866Skris - Still avoid linking the OpenSSL programs with the shared libraries. 500776866Skris 500876866Skris - When installing, install the shared libraries separately from the 500976866Skris static ones. 501076866Skris [Richard Levitte] 501176866Skris 501272613Skris *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 501372613Skris 501472613Skris Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 501572613Skris and not in SSL_clear because the latter is also used by the 501672613Skris accept/connect functions; previously, the settings made by 501772613Skris SSL_set_read_ahead would be lost during the handshake. 501872613Skris [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 501972613Skris 502072613Skris *) Correct util/mkdef.pl to be selective about disabled algorithms. 502172613Skris Previously, it would create entries for disableed algorithms no 502272613Skris matter what. 502372613Skris [Richard Levitte] 502472613Skris 502572613Skris *) Added several new manual pages for SSL_* function. 502672613Skris [Lutz Jaenicke] 502772613Skris 502868651Skris Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 502968651Skris 503068651Skris *) In ssl23_get_client_hello, generate an error message when faced 503168651Skris with an initial SSL 3.0/TLS record that is too small to contain the 503268651Skris first two bytes of the ClientHello message, i.e. client_version. 503368651Skris (Note that this is a pathologic case that probably has never happened 503468651Skris in real life.) The previous approach was to use the version number 503568651Skris from the record header as a substitute; but our protocol choice 503668651Skris should not depend on that one because it is not authenticated 503768651Skris by the Finished messages. 503868651Skris [Bodo Moeller] 503968651Skris 504068651Skris *) More robust randomness gathering functions for Windows. 504168651Skris [Jeffrey Altman <jaltman@columbia.edu>] 504268651Skris 504368651Skris *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 504468651Skris not set then we don't setup the error code for issuer check errors 504568651Skris to avoid possibly overwriting other errors which the callback does 504668651Skris handle. If an application does set the flag then we assume it knows 504768651Skris what it is doing and can handle the new informational codes 504868651Skris appropriately. 504968651Skris [Steve Henson] 505068651Skris 505168651Skris *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 505268651Skris a general "ANY" type, as such it should be able to decode anything 505368651Skris including tagged types. However it didn't check the class so it would 505468651Skris wrongly interpret tagged types in the same way as their universal 505568651Skris counterpart and unknown types were just rejected. Changed so that the 505668651Skris tagged and unknown types are handled in the same way as a SEQUENCE: 505768651Skris that is the encoding is stored intact. There is also a new type 505868651Skris "V_ASN1_OTHER" which is used when the class is not universal, in this 505968651Skris case we have no idea what the actual type is so we just lump them all 506068651Skris together. 506168651Skris [Steve Henson] 506268651Skris 506368651Skris *) On VMS, stdout may very well lead to a file that is written to 506468651Skris in a record-oriented fashion. That means that every write() will 506568651Skris write a separate record, which will be read separately by the 506668651Skris programs trying to read from it. This can be very confusing. 506768651Skris 506868651Skris The solution is to put a BIO filter in the way that will buffer 506968651Skris text until a linefeed is reached, and then write everything a 507068651Skris line at a time, so every record written will be an actual line, 507168651Skris not chunks of lines and not (usually doesn't happen, but I've 507268651Skris seen it once) several lines in one record. BIO_f_linebuffer() is 507368651Skris the answer. 507468651Skris 507568651Skris Currently, it's a VMS-only method, because that's where it has 507668651Skris been tested well enough. 507768651Skris [Richard Levitte] 507868651Skris 507968651Skris *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 508068651Skris it can return incorrect results. 508168651Skris (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 508268651Skris but it was in 0.9.6-beta[12].) 508368651Skris [Bodo Moeller] 508468651Skris 508568651Skris *) Disable the check for content being present when verifying detached 508668651Skris signatures in pk7_smime.c. Some versions of Netscape (wrongly) 508768651Skris include zero length content when signing messages. 508868651Skris [Steve Henson] 508968651Skris 509068651Skris *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 509168651Skris BIO_ctrl (for BIO pairs). 509268651Skris [Bodo M�ller] 509368651Skris 509468651Skris *) Add DSO method for VMS. 509568651Skris [Richard Levitte] 509668651Skris 509768651Skris *) Bug fix: Montgomery multiplication could produce results with the 509868651Skris wrong sign. 509968651Skris [Ulf M�ller] 510068651Skris 510168651Skris *) Add RPM specification openssl.spec and modify it to build three 510268651Skris packages. The default package contains applications, application 510368651Skris documentation and run-time libraries. The devel package contains 510468651Skris include files, static libraries and function documentation. The 510568651Skris doc package contains the contents of the doc directory. The original 510668651Skris openssl.spec was provided by Damien Miller <djm@mindrot.org>. 510768651Skris [Richard Levitte] 510868651Skris 510968651Skris *) Add a large number of documentation files for many SSL routines. 511068651Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 511168651Skris 511268651Skris *) Add a configuration entry for Sony News 4. 511368651Skris [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 511468651Skris 511568651Skris *) Don't set the two most significant bits to one when generating a 511668651Skris random number < q in the DSA library. 511768651Skris [Ulf M�ller] 511868651Skris 511968651Skris *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 512068651Skris behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 512168651Skris the underlying transport is blocking) if a handshake took place. 512268651Skris (The default behaviour is needed by applications such as s_client 512368651Skris and s_server that use select() to determine when to use SSL_read; 512468651Skris but for applications that know in advance when to expect data, it 512568651Skris just makes things more complicated.) 512668651Skris [Bodo Moeller] 512768651Skris 512868651Skris *) Add RAND_egd_bytes(), which gives control over the number of bytes read 512968651Skris from EGD. 513068651Skris [Ben Laurie] 513168651Skris 513268651Skris *) Add a few more EBCDIC conditionals that make `req' and `x509' 513368651Skris work better on such systems. 513468651Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 513568651Skris 513668651Skris *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 513768651Skris Update PKCS12_parse() so it copies the friendlyName and the 513868651Skris keyid to the certificates aux info. 513968651Skris [Steve Henson] 514068651Skris 514168651Skris *) Fix bug in PKCS7_verify() which caused an infinite loop 514268651Skris if there was more than one signature. 514368651Skris [Sven Uszpelkat <su@celocom.de>] 514468651Skris 514568651Skris *) Major change in util/mkdef.pl to include extra information 514668651Skris about each symbol, as well as presentig variables as well 514768651Skris as functions. This change means that there's n more need 514868651Skris to rebuild the .num files when some algorithms are excluded. 514968651Skris [Richard Levitte] 515068651Skris 515168651Skris *) Allow the verify time to be set by an application, 515268651Skris rather than always using the current time. 515368651Skris [Steve Henson] 515468651Skris 515568651Skris *) Phase 2 verify code reorganisation. The certificate 515668651Skris verify code now looks up an issuer certificate by a 515768651Skris number of criteria: subject name, authority key id 515868651Skris and key usage. It also verifies self signed certificates 515968651Skris by the same criteria. The main comparison function is 516068651Skris X509_check_issued() which performs these checks. 516168651Skris 516268651Skris Lot of changes were necessary in order to support this 516368651Skris without completely rewriting the lookup code. 516468651Skris 516568651Skris Authority and subject key identifier are now cached. 516668651Skris 516768651Skris The LHASH 'certs' is X509_STORE has now been replaced 516868651Skris by a STACK_OF(X509_OBJECT). This is mainly because an 516968651Skris LHASH can't store or retrieve multiple objects with 517068651Skris the same hash value. 517168651Skris 517268651Skris As a result various functions (which were all internal 517368651Skris use only) have changed to handle the new X509_STORE 517468651Skris structure. This will break anything that messed round 517568651Skris with X509_STORE internally. 517668651Skris 517768651Skris The functions X509_STORE_add_cert() now checks for an 517868651Skris exact match, rather than just subject name. 517968651Skris 518068651Skris The X509_STORE API doesn't directly support the retrieval 518168651Skris of multiple certificates matching a given criteria, however 518268651Skris this can be worked round by performing a lookup first 518368651Skris (which will fill the cache with candidate certificates) 518468651Skris and then examining the cache for matches. This is probably 518568651Skris the best we can do without throwing out X509_LOOKUP 518668651Skris entirely (maybe later...). 518768651Skris 518868651Skris The X509_VERIFY_CTX structure has been enhanced considerably. 518968651Skris 519068651Skris All certificate lookup operations now go via a get_issuer() 519168651Skris callback. Although this currently uses an X509_STORE it 519268651Skris can be replaced by custom lookups. This is a simple way 519368651Skris to bypass the X509_STORE hackery necessary to make this 519468651Skris work and makes it possible to use more efficient techniques 519568651Skris in future. A very simple version which uses a simple 519668651Skris STACK for its trusted certificate store is also provided 519768651Skris using X509_STORE_CTX_trusted_stack(). 519868651Skris 519968651Skris The verify_cb() and verify() callbacks now have equivalents 520068651Skris in the X509_STORE_CTX structure. 520168651Skris 520268651Skris X509_STORE_CTX also has a 'flags' field which can be used 520368651Skris to customise the verify behaviour. 520468651Skris [Steve Henson] 520568651Skris 520668651Skris *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 520768651Skris excludes S/MIME capabilities. 520868651Skris [Steve Henson] 520968651Skris 521068651Skris *) When a certificate request is read in keep a copy of the 521168651Skris original encoding of the signed data and use it when outputing 521268651Skris again. Signatures then use the original encoding rather than 521368651Skris a decoded, encoded version which may cause problems if the 521468651Skris request is improperly encoded. 521568651Skris [Steve Henson] 521668651Skris 521768651Skris *) For consistency with other BIO_puts implementations, call 521868651Skris buffer_write(b, ...) directly in buffer_puts instead of calling 521968651Skris BIO_write(b, ...). 522068651Skris 522168651Skris In BIO_puts, increment b->num_write as in BIO_write. 522268651Skris [Peter.Sylvester@EdelWeb.fr] 522368651Skris 522468651Skris *) Fix BN_mul_word for the case where the word is 0. (We have to use 522568651Skris BN_zero, we may not return a BIGNUM with an array consisting of 522668651Skris words set to zero.) 522768651Skris [Bodo Moeller] 522868651Skris 522968651Skris *) Avoid calling abort() from within the library when problems are 523068651Skris detected, except if preprocessor symbols have been defined 523168651Skris (such as REF_CHECK, BN_DEBUG etc.). 523268651Skris [Bodo Moeller] 523368651Skris 523468651Skris *) New openssl application 'rsautl'. This utility can be 523568651Skris used for low level RSA operations. DER public key 523668651Skris BIO/fp routines also added. 523768651Skris [Steve Henson] 523868651Skris 523968651Skris *) New Configure entry and patches for compiling on QNX 4. 524068651Skris [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 524168651Skris 524268651Skris *) A demo state-machine implementation was sponsored by 524368651Skris Nuron (http://www.nuron.com/) and is now available in 524468651Skris demos/state_machine. 524568651Skris [Ben Laurie] 524668651Skris 524768651Skris *) New options added to the 'dgst' utility for signature 524868651Skris generation and verification. 524968651Skris [Steve Henson] 525068651Skris 525168651Skris *) Unrecognized PKCS#7 content types are now handled via a 525268651Skris catch all ASN1_TYPE structure. This allows unsupported 525368651Skris types to be stored as a "blob" and an application can 525468651Skris encode and decode it manually. 525568651Skris [Steve Henson] 525668651Skris 525768651Skris *) Fix various signed/unsigned issues to make a_strex.c 525868651Skris compile under VC++. 525968651Skris [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 526068651Skris 526168651Skris *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 526268651Skris length if passed a buffer. ASN1_INTEGER_to_BN failed 526368651Skris if passed a NULL BN and its argument was negative. 526468651Skris [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 526568651Skris 526668651Skris *) Modification to PKCS#7 encoding routines to output definite 526768651Skris length encoding. Since currently the whole structures are in 526868651Skris memory there's not real point in using indefinite length 526968651Skris constructed encoding. However if OpenSSL is compiled with 527068651Skris the flag PKCS7_INDEFINITE_ENCODING the old form is used. 527168651Skris [Steve Henson] 527268651Skris 527368651Skris *) Added BIO_vprintf() and BIO_vsnprintf(). 527468651Skris [Richard Levitte] 527568651Skris 527668651Skris *) Added more prefixes to parse for in the the strings written 527768651Skris through a logging bio, to cover all the levels that are available 527868651Skris through syslog. The prefixes are now: 527968651Skris 528068651Skris PANIC, EMERG, EMR => LOG_EMERG 528168651Skris ALERT, ALR => LOG_ALERT 528268651Skris CRIT, CRI => LOG_CRIT 528368651Skris ERROR, ERR => LOG_ERR 528468651Skris WARNING, WARN, WAR => LOG_WARNING 528568651Skris NOTICE, NOTE, NOT => LOG_NOTICE 528668651Skris INFO, INF => LOG_INFO 528768651Skris DEBUG, DBG => LOG_DEBUG 528868651Skris 528968651Skris and as before, if none of those prefixes are present at the 529068651Skris beginning of the string, LOG_ERR is chosen. 529168651Skris 529268651Skris On Win32, the LOG_* levels are mapped according to this: 529368651Skris 529468651Skris LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 529568651Skris LOG_WARNING => EVENTLOG_WARNING_TYPE 529668651Skris LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 529768651Skris 529868651Skris [Richard Levitte] 529968651Skris 530068651Skris *) Made it possible to reconfigure with just the configuration 530168651Skris argument "reconf" or "reconfigure". The command line arguments 530268651Skris are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 530368651Skris and are retrieved from there when reconfiguring. 530468651Skris [Richard Levitte] 530568651Skris 530668651Skris *) MD4 implemented. 530768651Skris [Assar Westerlund <assar@sics.se>, Richard Levitte] 530868651Skris 530968651Skris *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 531068651Skris [Richard Levitte] 531168651Skris 531268651Skris *) The obj_dat.pl script was messing up the sorting of object 531368651Skris names. The reason was that it compared the quoted version 531468651Skris of strings as a result "OCSP" > "OCSP Signing" because 531568651Skris " > SPACE. Changed script to store unquoted versions of 531668651Skris names and add quotes on output. It was also omitting some 531768651Skris names from the lookup table if they were given a default 531868651Skris value (that is if SN is missing it is given the same 531968651Skris value as LN and vice versa), these are now added on the 532068651Skris grounds that if an object has a name we should be able to 532168651Skris look it up. Finally added warning output when duplicate 532268651Skris short or long names are found. 532368651Skris [Steve Henson] 532468651Skris 532568651Skris *) Changes needed for Tandem NSK. 532668651Skris [Scott Uroff <scott@xypro.com>] 532768651Skris 532868651Skris *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 532968651Skris RSA_padding_check_SSLv23(), special padding was never detected 533068651Skris and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 533168651Skris version rollback attacks was not effective. 533268651Skris 533368651Skris In s23_clnt.c, don't use special rollback-attack detection padding 533468651Skris (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 533568651Skris client; similarly, in s23_srvr.c, don't do the rollback check if 533668651Skris SSL 2.0 is the only protocol enabled in the server. 533768651Skris [Bodo Moeller] 533868651Skris 533968651Skris *) Make it possible to get hexdumps of unprintable data with 'openssl 534068651Skris asn1parse'. By implication, the functions ASN1_parse_dump() and 534168651Skris BIO_dump_indent() are added. 534268651Skris [Richard Levitte] 534368651Skris 534468651Skris *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 534568651Skris these print out strings and name structures based on various 534668651Skris flags including RFC2253 support and proper handling of 534768651Skris multibyte characters. Added options to the 'x509' utility 534868651Skris to allow the various flags to be set. 534968651Skris [Steve Henson] 535068651Skris 535168651Skris *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 535268651Skris Also change the functions X509_cmp_current_time() and 535368651Skris X509_gmtime_adj() work with an ASN1_TIME structure, 535468651Skris this will enable certificates using GeneralizedTime in validity 535568651Skris dates to be checked. 535668651Skris [Steve Henson] 535768651Skris 535868651Skris *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 535968651Skris negative public key encodings) on by default, 536068651Skris NO_NEG_PUBKEY_BUG can be set to disable it. 536168651Skris [Steve Henson] 536268651Skris 536368651Skris *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 536468651Skris content octets. An i2c_ASN1_OBJECT is unnecessary because 536568651Skris the encoding can be trivially obtained from the structure. 536668651Skris [Steve Henson] 536768651Skris 536868651Skris *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 536968651Skris not read locks (CRYPTO_r_[un]lock). 537068651Skris [Bodo Moeller] 537168651Skris 537268651Skris *) A first attempt at creating official support for shared 537368651Skris libraries through configuration. I've kept it so the 537468651Skris default is static libraries only, and the OpenSSL programs 537568651Skris are always statically linked for now, but there are 537668651Skris preparations for dynamic linking in place. 537789837Skris This has been tested on Linux and Tru64. 537868651Skris [Richard Levitte] 537968651Skris 538068651Skris *) Randomness polling function for Win9x, as described in: 538168651Skris Peter Gutmann, Software Generation of Practically Strong 538268651Skris Random Numbers. 538368651Skris [Ulf M�ller] 538468651Skris 538568651Skris *) Fix so PRNG is seeded in req if using an already existing 538668651Skris DSA key. 538768651Skris [Steve Henson] 538868651Skris 538968651Skris *) New options to smime application. -inform and -outform 539068651Skris allow alternative formats for the S/MIME message including 539168651Skris PEM and DER. The -content option allows the content to be 539268651Skris specified separately. This should allow things like Netscape 539368651Skris form signing output easier to verify. 539468651Skris [Steve Henson] 539568651Skris 539668651Skris *) Fix the ASN1 encoding of tags using the 'long form'. 539768651Skris [Steve Henson] 539868651Skris 539968651Skris *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 540068651Skris STRING types. These convert content octets to and from the 540168651Skris underlying type. The actual tag and length octets are 540268651Skris already assumed to have been read in and checked. These 540368651Skris are needed because all other string types have virtually 540468651Skris identical handling apart from the tag. By having versions 540568651Skris of the ASN1 functions that just operate on content octets 540668651Skris IMPLICIT tagging can be handled properly. It also allows 540768651Skris the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 540868651Skris and ASN1_INTEGER are identical apart from the tag. 540968651Skris [Steve Henson] 541068651Skris 541168651Skris *) Change the handling of OID objects as follows: 541268651Skris 541368651Skris - New object identifiers are inserted in objects.txt, following 541468651Skris the syntax given in objects.README. 541568651Skris - objects.pl is used to process obj_mac.num and create a new 541668651Skris obj_mac.h. 541768651Skris - obj_dat.pl is used to create a new obj_dat.h, using the data in 541868651Skris obj_mac.h. 541968651Skris 542068651Skris This is currently kind of a hack, and the perl code in objects.pl 542168651Skris isn't very elegant, but it works as I intended. The simplest way 542268651Skris to check that it worked correctly is to look in obj_dat.h and 542368651Skris check the array nid_objs and make sure the objects haven't moved 542468651Skris around (this is important!). Additions are OK, as well as 542568651Skris consistent name changes. 542668651Skris [Richard Levitte] 542768651Skris 542868651Skris *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 542968651Skris [Bodo Moeller] 543068651Skris 543168651Skris *) Addition of the command line parameter '-rand file' to 'openssl req'. 543268651Skris The given file adds to whatever has already been seeded into the 543368651Skris random pool through the RANDFILE configuration file option or 543468651Skris environment variable, or the default random state file. 543568651Skris [Richard Levitte] 543668651Skris 543768651Skris *) mkstack.pl now sorts each macro group into lexical order. 543868651Skris Previously the output order depended on the order the files 543968651Skris appeared in the directory, resulting in needless rewriting 544068651Skris of safestack.h . 544168651Skris [Steve Henson] 544268651Skris 544368651Skris *) Patches to make OpenSSL compile under Win32 again. Mostly 544468651Skris work arounds for the VC++ problem that it treats func() as 544568651Skris func(void). Also stripped out the parts of mkdef.pl that 544668651Skris added extra typesafe functions: these no longer exist. 544768651Skris [Steve Henson] 544868651Skris 544968651Skris *) Reorganisation of the stack code. The macros are now all 545068651Skris collected in safestack.h . Each macro is defined in terms of 545168651Skris a "stack macro" of the form SKM_<name>(type, a, b). The 545268651Skris DEBUG_SAFESTACK is now handled in terms of function casts, 545368651Skris this has the advantage of retaining type safety without the 545468651Skris use of additional functions. If DEBUG_SAFESTACK is not defined 545568651Skris then the non typesafe macros are used instead. Also modified the 545668651Skris mkstack.pl script to handle the new form. Needs testing to see 545768651Skris if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 545868651Skris the default if no major problems. Similar behaviour for ASN1_SET_OF 545968651Skris and PKCS12_STACK_OF. 546068651Skris [Steve Henson] 546168651Skris 546268651Skris *) When some versions of IIS use the 'NET' form of private key the 546368651Skris key derivation algorithm is different. Normally MD5(password) is 546468651Skris used as a 128 bit RC4 key. In the modified case 546568651Skris MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 546668651Skris new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 546768651Skris as the old Netscape_RSA functions except they have an additional 546868651Skris 'sgckey' parameter which uses the modified algorithm. Also added 546968651Skris an -sgckey command line option to the rsa utility. Thanks to 547068651Skris Adrian Peck <bertie@ncipher.com> for posting details of the modified 547168651Skris algorithm to openssl-dev. 547268651Skris [Steve Henson] 547368651Skris 547468651Skris *) The evp_local.h macros were using 'c.##kname' which resulted in 547568651Skris invalid expansion on some systems (SCO 5.0.5 for example). 547668651Skris Corrected to 'c.kname'. 547768651Skris [Phillip Porch <root@theporch.com>] 547868651Skris 547968651Skris *) New X509_get1_email() and X509_REQ_get1_email() functions that return 548068651Skris a STACK of email addresses from a certificate or request, these look 548168651Skris in the subject name and the subject alternative name extensions and 548268651Skris omit any duplicate addresses. 548368651Skris [Steve Henson] 548468651Skris 548568651Skris *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 548668651Skris This makes DSA verification about 2 % faster. 548768651Skris [Bodo Moeller] 548868651Skris 548968651Skris *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 549068651Skris (meaning that now 2^5 values will be precomputed, which is only 4 KB 549168651Skris plus overhead for 1024 bit moduli). 549268651Skris This makes exponentiations about 0.5 % faster for 1024 bit 549368651Skris exponents (as measured by "openssl speed rsa2048"). 549468651Skris [Bodo Moeller] 549568651Skris 549668651Skris *) Rename memory handling macros to avoid conflicts with other 549768651Skris software: 549868651Skris Malloc => OPENSSL_malloc 549968651Skris Malloc_locked => OPENSSL_malloc_locked 550068651Skris Realloc => OPENSSL_realloc 550168651Skris Free => OPENSSL_free 550268651Skris [Richard Levitte] 550368651Skris 550468651Skris *) New function BN_mod_exp_mont_word for small bases (roughly 15% 550568651Skris faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 550668651Skris [Bodo Moeller] 550768651Skris 550868651Skris *) CygWin32 support. 550968651Skris [John Jarvie <jjarvie@newsguy.com>] 551068651Skris 551168651Skris *) The type-safe stack code has been rejigged. It is now only compiled 551268651Skris in when OpenSSL is configured with the DEBUG_SAFESTACK option and 551368651Skris by default all type-specific stack functions are "#define"d back to 551468651Skris standard stack functions. This results in more streamlined output 551568651Skris but retains the type-safety checking possibilities of the original 551668651Skris approach. 551768651Skris [Geoff Thorpe] 551868651Skris 551968651Skris *) The STACK code has been cleaned up, and certain type declarations 552068651Skris that didn't make a lot of sense have been brought in line. This has 552168651Skris also involved a cleanup of sorts in safestack.h to more correctly 552268651Skris map type-safe stack functions onto their plain stack counterparts. 552368651Skris This work has also resulted in a variety of "const"ifications of 552468651Skris lots of the code, especially "_cmp" operations which should normally 552568651Skris be prototyped with "const" parameters anyway. 552668651Skris [Geoff Thorpe] 552768651Skris 552868651Skris *) When generating bytes for the first time in md_rand.c, 'stir the pool' 552968651Skris by seeding with STATE_SIZE dummy bytes (with zero entropy count). 553068651Skris (The PRNG state consists of two parts, the large pool 'state' and 'md', 553168651Skris where all of 'md' is used each time the PRNG is used, but 'state' 553268651Skris is used only indexed by a cyclic counter. As entropy may not be 553368651Skris well distributed from the beginning, 'md' is important as a 553468651Skris chaining variable. However, the output function chains only half 553568651Skris of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 553668651Skris all of 'md', and seeding with STATE_SIZE dummy bytes will result 553768651Skris in all of 'state' being rewritten, with the new values depending 553868651Skris on virtually all of 'md'. This overcomes the 80 bit limitation.) 553968651Skris [Bodo Moeller] 554068651Skris 554168651Skris *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 554268651Skris the handshake is continued after ssl_verify_cert_chain(); 554368651Skris otherwise, if SSL_VERIFY_NONE is set, remaining error codes 554468651Skris can lead to 'unexplainable' connection aborts later. 554568651Skris [Bodo Moeller; problem tracked down by Lutz Jaenicke] 554668651Skris 554768651Skris *) Major EVP API cipher revision. 554868651Skris Add hooks for extra EVP features. This allows various cipher 554968651Skris parameters to be set in the EVP interface. Support added for variable 555068651Skris key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 555168651Skris setting of RC2 and RC5 parameters. 555268651Skris 555368651Skris Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 555468651Skris ciphers. 555568651Skris 555668651Skris Remove lots of duplicated code from the EVP library. For example *every* 555768651Skris cipher init() function handles the 'iv' in the same way according to the 555868651Skris cipher mode. They also all do nothing if the 'key' parameter is NULL and 555968651Skris for CFB and OFB modes they zero ctx->num. 556068651Skris 556168651Skris New functionality allows removal of S/MIME code RC2 hack. 556268651Skris 556368651Skris Most of the routines have the same form and so can be declared in terms 556468651Skris of macros. 556568651Skris 556668651Skris By shifting this to the top level EVP_CipherInit() it can be removed from 556768651Skris all individual ciphers. If the cipher wants to handle IVs or keys 556868651Skris differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 556968651Skris flags. 557068651Skris 557168651Skris Change lots of functions like EVP_EncryptUpdate() to now return a 557268651Skris value: although software versions of the algorithms cannot fail 557368651Skris any installed hardware versions can. 557468651Skris [Steve Henson] 557568651Skris 557668651Skris *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 557768651Skris this option is set, tolerate broken clients that send the negotiated 557868651Skris protocol version number instead of the requested protocol version 557968651Skris number. 558068651Skris [Bodo Moeller] 558168651Skris 558268651Skris *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 558368651Skris i.e. non-zero for export ciphersuites, zero otherwise. 558468651Skris Previous versions had this flag inverted, inconsistent with 558568651Skris rsa_tmp_cb (..._TMP_RSA_CB). 558668651Skris [Bodo Moeller; problem reported by Amit Chopra] 558768651Skris 558868651Skris *) Add missing DSA library text string. Work around for some IIS 558968651Skris key files with invalid SEQUENCE encoding. 559068651Skris [Steve Henson] 559168651Skris 559268651Skris *) Add a document (doc/standards.txt) that list all kinds of standards 559368651Skris and so on that are implemented in OpenSSL. 559468651Skris [Richard Levitte] 559568651Skris 559668651Skris *) Enhance c_rehash script. Old version would mishandle certificates 559768651Skris with the same subject name hash and wouldn't handle CRLs at all. 559868651Skris Added -fingerprint option to crl utility, to support new c_rehash 559968651Skris features. 560068651Skris [Steve Henson] 560168651Skris 560268651Skris *) Eliminate non-ANSI declarations in crypto.h and stack.h. 560368651Skris [Ulf M�ller] 560468651Skris 560568651Skris *) Fix for SSL server purpose checking. Server checking was 560668651Skris rejecting certificates which had extended key usage present 560768651Skris but no ssl client purpose. 560868651Skris [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 560968651Skris 561068651Skris *) Make PKCS#12 code work with no password. The PKCS#12 spec 561168651Skris is a little unclear about how a blank password is handled. 561268651Skris Since the password in encoded as a BMPString with terminating 561368651Skris double NULL a zero length password would end up as just the 561468651Skris double NULL. However no password at all is different and is 561568651Skris handled differently in the PKCS#12 key generation code. NS 561668651Skris treats a blank password as zero length. MSIE treats it as no 561768651Skris password on export: but it will try both on import. We now do 561868651Skris the same: PKCS12_parse() tries zero length and no password if 561968651Skris the password is set to "" or NULL (NULL is now a valid password: 562068651Skris it wasn't before) as does the pkcs12 application. 562168651Skris [Steve Henson] 562268651Skris 562368651Skris *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 562468651Skris perror when PEM_read_bio_X509_REQ fails, the error message must 562568651Skris be obtained from the error queue. 562668651Skris [Bodo Moeller] 562768651Skris 562868651Skris *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 562968651Skris it in ERR_remove_state if appropriate, and change ERR_get_state 563068651Skris accordingly to avoid race conditions (this is necessary because 563168651Skris thread_hash is no longer constant once set). 563268651Skris [Bodo Moeller] 563368651Skris 563468651Skris *) Bugfix for linux-elf makefile.one. 563568651Skris [Ulf M�ller] 563668651Skris 563768651Skris *) RSA_get_default_method() will now cause a default 563868651Skris RSA_METHOD to be chosen if one doesn't exist already. 563968651Skris Previously this was only set during a call to RSA_new() 564068651Skris or RSA_new_method(NULL) meaning it was possible for 564168651Skris RSA_get_default_method() to return NULL. 564268651Skris [Geoff Thorpe] 564368651Skris 564468651Skris *) Added native name translation to the existing DSO code 564568651Skris that will convert (if the flag to do so is set) filenames 564668651Skris that are sufficiently small and have no path information 564768651Skris into a canonical native form. Eg. "blah" converted to 564868651Skris "libblah.so" or "blah.dll" etc. 564968651Skris [Geoff Thorpe] 565068651Skris 565168651Skris *) New function ERR_error_string_n(e, buf, len) which is like 565268651Skris ERR_error_string(e, buf), but writes at most 'len' bytes 565368651Skris including the 0 terminator. For ERR_error_string_n, 'buf' 565468651Skris may not be NULL. 565568651Skris [Damien Miller <djm@mindrot.org>, Bodo Moeller] 565668651Skris 565768651Skris *) CONF library reworked to become more general. A new CONF 565868651Skris configuration file reader "class" is implemented as well as a 565968651Skris new functions (NCONF_*, for "New CONF") to handle it. The now 566068651Skris old CONF_* functions are still there, but are reimplemented to 566168651Skris work in terms of the new functions. Also, a set of functions 566268651Skris to handle the internal storage of the configuration data is 566368651Skris provided to make it easier to write new configuration file 566468651Skris reader "classes" (I can definitely see something reading a 566568651Skris configuration file in XML format, for example), called _CONF_*, 566668651Skris or "the configuration storage API"... 566768651Skris 566868651Skris The new configuration file reading functions are: 566968651Skris 567068651Skris NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 567168651Skris NCONF_get_section, NCONF_get_string, NCONF_get_numbre 567268651Skris 567368651Skris NCONF_default, NCONF_WIN32 567468651Skris 567568651Skris NCONF_dump_fp, NCONF_dump_bio 567668651Skris 567768651Skris NCONF_default and NCONF_WIN32 are method (or "class") choosers, 567868651Skris NCONF_new creates a new CONF object. This works in the same way 567968651Skris as other interfaces in OpenSSL, like the BIO interface. 568068651Skris NCONF_dump_* dump the internal storage of the configuration file, 568168651Skris which is useful for debugging. All other functions take the same 568268651Skris arguments as the old CONF_* functions wth the exception of the 568368651Skris first that must be a `CONF *' instead of a `LHASH *'. 568468651Skris 568568651Skris To make it easer to use the new classes with the old CONF_* functions, 568668651Skris the function CONF_set_default_method is provided. 568768651Skris [Richard Levitte] 568868651Skris 568968651Skris *) Add '-tls1' option to 'openssl ciphers', which was already 569068651Skris mentioned in the documentation but had not been implemented. 569168651Skris (This option is not yet really useful because even the additional 569268651Skris experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 569368651Skris [Bodo Moeller] 569468651Skris 569568651Skris *) Initial DSO code added into libcrypto for letting OpenSSL (and 569668651Skris OpenSSL-based applications) load shared libraries and bind to 569768651Skris them in a portable way. 569868651Skris [Geoff Thorpe, with contributions from Richard Levitte] 569968651Skris 570059191Skris Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 570159191Skris 570259191Skris *) Make sure _lrotl and _lrotr are only used with MSVC. 570359191Skris 570459191Skris *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 570559191Skris (the default implementation of RAND_status). 570659191Skris 570759191Skris *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 570859191Skris to '-clrext' (= clear extensions), as intended and documented. 570959191Skris [Bodo Moeller; inconsistency pointed out by Michael Attili 571059191Skris <attili@amaxo.com>] 571159191Skris 571259191Skris *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 571359191Skris was larger than the MD block size. 571459191Skris [Steve Henson, pointed out by Yost William <YostW@tce.com>] 571559191Skris 571659191Skris *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 571759191Skris fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 571859191Skris using the passed key: if the passed key was a private key the result 571959191Skris of X509_print(), for example, would be to print out all the private key 572059191Skris components. 572159191Skris [Steve Henson] 572259191Skris 572359191Skris *) des_quad_cksum() byte order bug fix. 572459191Skris [Ulf M�ller, using the problem description in krb4-0.9.7, where 572559191Skris the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 572659191Skris 572759191Skris *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 572859191Skris discouraged. 572959191Skris [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 573059191Skris 573159191Skris *) For easily testing in shell scripts whether some command 573259191Skris 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 573359191Skris returns with exit code 0 iff no command of the given name is available. 573459191Skris 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 573559191Skris the output goes to stdout and nothing is printed to stderr. 573659191Skris Additional arguments are always ignored. 573759191Skris 573859191Skris Since for each cipher there is a command of the same name, 573959191Skris the 'no-cipher' compilation switches can be tested this way. 574059191Skris 574159191Skris ('openssl no-XXX' is not able to detect pseudo-commands such 574259191Skris as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 574359191Skris [Bodo Moeller] 574459191Skris 574559191Skris *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 574659191Skris [Bodo Moeller] 574759191Skris 574859191Skris *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 574959191Skris is set; it will be thrown away anyway because each handshake creates 575059191Skris its own key. 575159191Skris ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 575259191Skris to parameters -- in previous versions (since OpenSSL 0.9.3) the 575359191Skris 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 575459191Skris you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 575559191Skris [Bodo Moeller] 575659191Skris 575759191Skris *) New s_client option -ign_eof: EOF at stdin is ignored, and 575859191Skris 'Q' and 'R' lose their special meanings (quit/renegotiate). 575959191Skris This is part of what -quiet does; unlike -quiet, -ign_eof 576059191Skris does not suppress any output. 576159191Skris [Richard Levitte] 576259191Skris 576359191Skris *) Add compatibility options to the purpose and trust code. The 576459191Skris purpose X509_PURPOSE_ANY is "any purpose" which automatically 576559191Skris accepts a certificate or CA, this was the previous behaviour, 576659191Skris with all the associated security issues. 576759191Skris 576859191Skris X509_TRUST_COMPAT is the old trust behaviour: only and 576959191Skris automatically trust self signed roots in certificate store. A 577059191Skris new trust setting X509_TRUST_DEFAULT is used to specify that 577159191Skris a purpose has no associated trust setting and it should instead 577259191Skris use the value in the default purpose. 577359191Skris [Steve Henson] 577459191Skris 577559191Skris *) Fix the PKCS#8 DSA private key code so it decodes keys again 577659191Skris and fix a memory leak. 577759191Skris [Steve Henson] 577859191Skris 577959191Skris *) In util/mkerr.pl (which implements 'make errors'), preserve 578059191Skris reason strings from the previous version of the .c file, as 578159191Skris the default to have only downcase letters (and digits) in 578259191Skris automatically generated reasons codes is not always appropriate. 578359191Skris [Bodo Moeller] 578459191Skris 578559191Skris *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 578659191Skris using strerror. Previously, ERR_reason_error_string() returned 578759191Skris library names as reason strings for SYSerr; but SYSerr is a special 578859191Skris case where small numbers are errno values, not library numbers. 578959191Skris [Bodo Moeller] 579059191Skris 579159191Skris *) Add '-dsaparam' option to 'openssl dhparam' application. This 579259191Skris converts DSA parameters into DH parameters. (When creating parameters, 579359191Skris DSA_generate_parameters is used.) 579459191Skris [Bodo Moeller] 579559191Skris 579659191Skris *) Include 'length' (recommended exponent length) in C code generated 579759191Skris by 'openssl dhparam -C'. 579859191Skris [Bodo Moeller] 579959191Skris 580059191Skris *) The second argument to set_label in perlasm was already being used 580159191Skris so couldn't be used as a "file scope" flag. Moved to third argument 580259191Skris which was free. 580359191Skris [Steve Henson] 580459191Skris 580559191Skris *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 580659191Skris instead of RAND_bytes for encryption IVs and salts. 580759191Skris [Bodo Moeller] 580859191Skris 580959191Skris *) Include RAND_status() into RAND_METHOD instead of implementing 581059191Skris it only for md_rand.c Otherwise replacing the PRNG by calling 581159191Skris RAND_set_rand_method would be impossible. 581259191Skris [Bodo Moeller] 581359191Skris 581459191Skris *) Don't let DSA_generate_key() enter an infinite loop if the random 581559191Skris number generation fails. 581659191Skris [Bodo Moeller] 581759191Skris 581859191Skris *) New 'rand' application for creating pseudo-random output. 581959191Skris [Bodo Moeller] 582059191Skris 582159191Skris *) Added configuration support for Linux/IA64 582259191Skris [Rolf Haberrecker <rolf@suse.de>] 582359191Skris 582459191Skris *) Assembler module support for Mingw32. 582559191Skris [Ulf M�ller] 582659191Skris 582759191Skris *) Shared library support for HPUX (in shlib/). 582859191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 582959191Skris 583059191Skris *) Shared library support for Solaris gcc. 583159191Skris [Lutz Behnke <behnke@trustcenter.de>] 583259191Skris 583359191Skris Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 583459191Skris 583559191Skris *) PKCS7_encrypt() was adding text MIME headers twice because they 583659191Skris were added manually and by SMIME_crlf_copy(). 583759191Skris [Steve Henson] 583859191Skris 583959191Skris *) In bntest.c don't call BN_rand with zero bits argument. 584059191Skris [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 584159191Skris 584259191Skris *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 584359191Skris case was implemented. This caused BN_div_recp() to fail occasionally. 584459191Skris [Ulf M�ller] 584559191Skris 584659191Skris *) Add an optional second argument to the set_label() in the perl 584759191Skris assembly language builder. If this argument exists and is set 584859191Skris to 1 it signals that the assembler should use a symbol whose 584959191Skris scope is the entire file, not just the current function. This 585059191Skris is needed with MASM which uses the format label:: for this scope. 585159191Skris [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 585259191Skris 585359191Skris *) Change the ASN1 types so they are typedefs by default. Before 585459191Skris almost all types were #define'd to ASN1_STRING which was causing 585559191Skris STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 585659191Skris for example. 585759191Skris [Steve Henson] 585859191Skris 585959191Skris *) Change names of new functions to the new get1/get0 naming 586059191Skris convention: After 'get1', the caller owns a reference count 586159191Skris and has to call ..._free; 'get0' returns a pointer to some 586259191Skris data structure without incrementing reference counters. 586359191Skris (Some of the existing 'get' functions increment a reference 586459191Skris counter, some don't.) 586559191Skris Similarly, 'set1' and 'add1' functions increase reference 586659191Skris counters or duplicate objects. 586759191Skris [Steve Henson] 586859191Skris 586959191Skris *) Allow for the possibility of temp RSA key generation failure: 587059191Skris the code used to assume it always worked and crashed on failure. 587159191Skris [Steve Henson] 587259191Skris 587359191Skris *) Fix potential buffer overrun problem in BIO_printf(). 587459191Skris [Ulf M�ller, using public domain code by Patrick Powell; problem 587559191Skris pointed out by David Sacerdote <das33@cornell.edu>] 587659191Skris 587759191Skris *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 587859191Skris RAND_egd() and RAND_status(). In the command line application, 587959191Skris the EGD socket can be specified like a seed file using RANDFILE 588059191Skris or -rand. 588159191Skris [Ulf M�ller] 588259191Skris 588359191Skris *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 588459191Skris Some CAs (e.g. Verisign) distribute certificates in this form. 588559191Skris [Steve Henson] 588659191Skris 588759191Skris *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 588859191Skris list to exclude them. This means that no special compilation option 588959191Skris is needed to use anonymous DH: it just needs to be included in the 589059191Skris cipher list. 589159191Skris [Steve Henson] 589259191Skris 589359191Skris *) Change the EVP_MD_CTX_type macro so its meaning consistent with 589459191Skris EVP_MD_type. The old functionality is available in a new macro called 589559191Skris EVP_MD_md(). Change code that uses it and update docs. 589659191Skris [Steve Henson] 589759191Skris 589859191Skris *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 589959191Skris where the 'void *' argument is replaced by a function pointer argument. 590059191Skris Previously 'void *' was abused to point to functions, which works on 590159191Skris many platforms, but is not correct. As these functions are usually 590259191Skris called by macros defined in OpenSSL header files, most source code 590359191Skris should work without changes. 590459191Skris [Richard Levitte] 590559191Skris 590659191Skris *) <openssl/opensslconf.h> (which is created by Configure) now contains 590759191Skris sections with information on -D... compiler switches used for 590859191Skris compiling the library so that applications can see them. To enable 590959191Skris one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 591059191Skris must be defined. E.g., 591159191Skris #define OPENSSL_ALGORITHM_DEFINES 591259191Skris #include <openssl/opensslconf.h> 591359191Skris defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 591459191Skris [Richard Levitte, Ulf and Bodo M�ller] 591559191Skris 591659191Skris *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 591759191Skris record layer. 591859191Skris [Bodo Moeller] 591959191Skris 592059191Skris *) Change the 'other' type in certificate aux info to a STACK_OF 592159191Skris X509_ALGOR. Although not an AlgorithmIdentifier as such it has 592259191Skris the required ASN1 format: arbitrary types determined by an OID. 592359191Skris [Steve Henson] 592459191Skris 592559191Skris *) Add some PEM_write_X509_REQ_NEW() functions and a command line 592659191Skris argument to 'req'. This is not because the function is newer or 592759191Skris better than others it just uses the work 'NEW' in the certificate 592859191Skris request header lines. Some software needs this. 592959191Skris [Steve Henson] 593059191Skris 593159191Skris *) Reorganise password command line arguments: now passwords can be 593259191Skris obtained from various sources. Delete the PEM_cb function and make 593359191Skris it the default behaviour: i.e. if the callback is NULL and the 593459191Skris usrdata argument is not NULL interpret it as a null terminated pass 593559191Skris phrase. If usrdata and the callback are NULL then the pass phrase 593659191Skris is prompted for as usual. 593759191Skris [Steve Henson] 593859191Skris 593959191Skris *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 594059191Skris the support is automatically enabled. The resulting binaries will 594159191Skris autodetect the card and use it if present. 594259191Skris [Ben Laurie and Compaq Inc.] 594359191Skris 594459191Skris *) Work around for Netscape hang bug. This sends certificate request 594559191Skris and server done in one record. Since this is perfectly legal in the 594659191Skris SSL/TLS protocol it isn't a "bug" option and is on by default. See 594759191Skris the bugs/SSLv3 entry for more info. 594859191Skris [Steve Henson] 594959191Skris 595059191Skris *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 595159191Skris [Andy Polyakov] 595259191Skris 595359191Skris *) Add -rand argument to smime and pkcs12 applications and read/write 595459191Skris of seed file. 595559191Skris [Steve Henson] 595659191Skris 595759191Skris *) New 'passwd' tool for crypt(3) and apr1 password hashes. 595859191Skris [Bodo Moeller] 595959191Skris 596059191Skris *) Add command line password options to the remaining applications. 596159191Skris [Steve Henson] 596259191Skris 596359191Skris *) Bug fix for BN_div_recp() for numerators with an even number of 596459191Skris bits. 596559191Skris [Ulf M�ller] 596659191Skris 596759191Skris *) More tests in bntest.c, and changed test_bn output. 596859191Skris [Ulf M�ller] 596959191Skris 597059191Skris *) ./config recognizes MacOS X now. 597159191Skris [Andy Polyakov] 597259191Skris 597359191Skris *) Bug fix for BN_div() when the first words of num and divsor are 597459191Skris equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 597559191Skris [Ulf M�ller] 597659191Skris 597759191Skris *) Add support for various broken PKCS#8 formats, and command line 597859191Skris options to produce them. 597959191Skris [Steve Henson] 598059191Skris 598159191Skris *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 598259191Skris get temporary BIGNUMs from a BN_CTX. 598359191Skris [Ulf M�ller] 598459191Skris 598559191Skris *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 598659191Skris for p == 0. 598759191Skris [Ulf M�ller] 598859191Skris 598959191Skris *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 599059191Skris include a #define from the old name to the new. The original intent 599159191Skris was that statically linked binaries could for example just call 599259191Skris SSLeay_add_all_ciphers() to just add ciphers to the table and not 599359191Skris link with digests. This never worked becayse SSLeay_add_all_digests() 599459191Skris and SSLeay_add_all_ciphers() were in the same source file so calling 599559191Skris one would link with the other. They are now in separate source files. 599659191Skris [Steve Henson] 599759191Skris 599859191Skris *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 599959191Skris [Steve Henson] 600059191Skris 600159191Skris *) Use a less unusual form of the Miller-Rabin primality test (it used 600259191Skris a binary algorithm for exponentiation integrated into the Miller-Rabin 600359191Skris loop, our standard modexp algorithms are faster). 600459191Skris [Bodo Moeller] 600559191Skris 600659191Skris *) Support for the EBCDIC character set completed. 600759191Skris [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 600859191Skris 600959191Skris *) Source code cleanups: use const where appropriate, eliminate casts, 601059191Skris use void * instead of char * in lhash. 601159191Skris [Ulf M�ller] 601259191Skris 601359191Skris *) Bugfix: ssl3_send_server_key_exchange was not restartable 601459191Skris (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 601559191Skris this the server could overwrite ephemeral keys that the client 601659191Skris has already seen). 601759191Skris [Bodo Moeller] 601859191Skris 601959191Skris *) Turn DSA_is_prime into a macro that calls BN_is_prime, 602059191Skris using 50 iterations of the Rabin-Miller test. 602159191Skris 602259191Skris DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 602359191Skris iterations of the Rabin-Miller test as required by the appendix 602459191Skris to FIPS PUB 186[-1]) instead of DSA_is_prime. 602559191Skris As BN_is_prime_fasttest includes trial division, DSA parameter 602659191Skris generation becomes much faster. 602759191Skris 602859191Skris This implies a change for the callback functions in DSA_is_prime 602959191Skris and DSA_generate_parameters: The callback function is called once 603059191Skris for each positive witness in the Rabin-Miller test, not just 603159191Skris occasionally in the inner loop; and the parameters to the 603259191Skris callback function now provide an iteration count for the outer 603359191Skris loop rather than for the current invocation of the inner loop. 603459191Skris DSA_generate_parameters additionally can call the callback 603559191Skris function with an 'iteration count' of -1, meaning that a 603659191Skris candidate has passed the trial division test (when q is generated 603759191Skris from an application-provided seed, trial division is skipped). 603859191Skris [Bodo Moeller] 603959191Skris 604059191Skris *) New function BN_is_prime_fasttest that optionally does trial 604159191Skris division before starting the Rabin-Miller test and has 604259191Skris an additional BN_CTX * argument (whereas BN_is_prime always 604359191Skris has to allocate at least one BN_CTX). 604459191Skris 'callback(1, -1, cb_arg)' is called when a number has passed the 604559191Skris trial division stage. 604659191Skris [Bodo Moeller] 604759191Skris 604859191Skris *) Fix for bug in CRL encoding. The validity dates weren't being handled 604959191Skris as ASN1_TIME. 605059191Skris [Steve Henson] 605159191Skris 605259191Skris *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 605359191Skris [Steve Henson] 605459191Skris 605559191Skris *) New function BN_pseudo_rand(). 605659191Skris [Ulf M�ller] 605759191Skris 605859191Skris *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 605959191Skris bignum version of BN_from_montgomery() with the working code from 606059191Skris SSLeay 0.9.0 (the word based version is faster anyway), and clean up 606159191Skris the comments. 606259191Skris [Ulf M�ller] 606359191Skris 606459191Skris *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 606559191Skris made it impossible to use the same SSL_SESSION data structure in 606659191Skris SSL2 clients in multiple threads. 606759191Skris [Bodo Moeller] 606859191Skris 606959191Skris *) The return value of RAND_load_file() no longer counts bytes obtained 607059191Skris by stat(). RAND_load_file(..., -1) is new and uses the complete file 607159191Skris to seed the PRNG (previously an explicit byte count was required). 607259191Skris [Ulf M�ller, Bodo M�ller] 607359191Skris 607459191Skris *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 607559191Skris used (char *) instead of (void *) and had casts all over the place. 607659191Skris [Steve Henson] 607759191Skris 607859191Skris *) Make BN_generate_prime() return NULL on error if ret!=NULL. 607959191Skris [Ulf M�ller] 608059191Skris 608159191Skris *) Retain source code compatibility for BN_prime_checks macro: 608259191Skris BN_is_prime(..., BN_prime_checks, ...) now uses 608359191Skris BN_prime_checks_for_size to determine the appropriate number of 608459191Skris Rabin-Miller iterations. 608559191Skris [Ulf M�ller] 608659191Skris 608759191Skris *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 608859191Skris DH_CHECK_P_NOT_SAFE_PRIME. 608959191Skris (Check if this is true? OpenPGP calls them "strong".) 609059191Skris [Ulf M�ller] 609159191Skris 609259191Skris *) Merge the functionality of "dh" and "gendh" programs into a new program 609359191Skris "dhparam". The old programs are retained for now but will handle DH keys 609459191Skris (instead of parameters) in future. 609559191Skris [Steve Henson] 609659191Skris 609759191Skris *) Make the ciphers, s_server and s_client programs check the return values 609859191Skris when a new cipher list is set. 609959191Skris [Steve Henson] 610059191Skris 610159191Skris *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 610259191Skris ciphers. Before when the 56bit ciphers were enabled the sorting was 610359191Skris wrong. 610459191Skris 610559191Skris The syntax for the cipher sorting has been extended to support sorting by 610659191Skris cipher-strength (using the strength_bits hard coded in the tables). 610759191Skris The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 610859191Skris 610959191Skris Fix a bug in the cipher-command parser: when supplying a cipher command 611059191Skris string with an "undefined" symbol (neither command nor alphanumeric 611159191Skris [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 611259191Skris an error is flagged. 611359191Skris 611459191Skris Due to the strength-sorting extension, the code of the 611559191Skris ssl_create_cipher_list() function was completely rearranged. I hope that 611659191Skris the readability was also increased :-) 611759191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 611859191Skris 611959191Skris *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 612059191Skris for the first serial number and places 2 in the serial number file. This 612159191Skris avoids problems when the root CA is created with serial number zero and 612259191Skris the first user certificate has the same issuer name and serial number 612359191Skris as the root CA. 612459191Skris [Steve Henson] 612559191Skris 612659191Skris *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 612759191Skris the new code. Add documentation for this stuff. 612859191Skris [Steve Henson] 612959191Skris 613059191Skris *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 613159191Skris X509_*() to X509at_*() on the grounds that they don't handle X509 613259191Skris structures and behave in an analagous way to the X509v3 functions: 613359191Skris they shouldn't be called directly but wrapper functions should be used 613459191Skris instead. 613559191Skris 613659191Skris So we also now have some wrapper functions that call the X509at functions 613759191Skris when passed certificate requests. (TO DO: similar things can be done with 613859191Skris PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 613959191Skris things. Some of these need some d2i or i2d and print functionality 614059191Skris because they handle more complex structures.) 614159191Skris [Steve Henson] 614259191Skris 614359191Skris *) Add missing #ifndefs that caused missing symbols when building libssl 614459191Skris as a shared library without RSA. Use #ifndef NO_SSL2 instead of 614559191Skris NO_RSA in ssl/s2*.c. 614659191Skris [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 614759191Skris 614859191Skris *) Precautions against using the PRNG uninitialized: RAND_bytes() now 614959191Skris has a return value which indicates the quality of the random data 615059191Skris (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 615159191Skris error queue. New function RAND_pseudo_bytes() generates output that is 615259191Skris guaranteed to be unique but not unpredictable. RAND_add is like 615359191Skris RAND_seed, but takes an extra argument for an entropy estimate 615459191Skris (RAND_seed always assumes full entropy). 615559191Skris [Ulf M�ller] 615659191Skris 615759191Skris *) Do more iterations of Rabin-Miller probable prime test (specifically, 615859191Skris 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 615959191Skris instead of only 2 for all lengths; see BN_prime_checks_for_size definition 616059191Skris in crypto/bn/bn_prime.c for the complete table). This guarantees a 616159191Skris false-positive rate of at most 2^-80 for random input. 616259191Skris [Bodo Moeller] 616359191Skris 616459191Skris *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 616559191Skris [Bodo Moeller] 616659191Skris 616759191Skris *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 616859191Skris in the 0.9.5 release), this returns the chain 616959191Skris from an X509_CTX structure with a dup of the stack and all 617059191Skris the X509 reference counts upped: so the stack will exist 617159191Skris after X509_CTX_cleanup() has been called. Modify pkcs12.c 617259191Skris to use this. 617359191Skris 617459191Skris Also make SSL_SESSION_print() print out the verify return 617559191Skris code. 617659191Skris [Steve Henson] 617759191Skris 617859191Skris *) Add manpage for the pkcs12 command. Also change the default 617959191Skris behaviour so MAC iteration counts are used unless the new 618059191Skris -nomaciter option is used. This improves file security and 618159191Skris only older versions of MSIE (4.0 for example) need it. 618259191Skris [Steve Henson] 618359191Skris 618459191Skris *) Honor the no-xxx Configure options when creating .DEF files. 618559191Skris [Ulf M�ller] 618659191Skris 618759191Skris *) Add PKCS#10 attributes to field table: challengePassword, 618859191Skris unstructuredName and unstructuredAddress. These are taken from 618959191Skris draft PKCS#9 v2.0 but are compatible with v1.2 provided no 619059191Skris international characters are used. 619159191Skris 619259191Skris More changes to X509_ATTRIBUTE code: allow the setting of types 619359191Skris based on strings. Remove the 'loc' parameter when adding 619459191Skris attributes because these will be a SET OF encoding which is sorted 619559191Skris in ASN1 order. 619659191Skris [Steve Henson] 619759191Skris 619859191Skris *) Initial changes to the 'req' utility to allow request generation 619959191Skris automation. This will allow an application to just generate a template 620059191Skris file containing all the field values and have req construct the 620159191Skris request. 620259191Skris 620359191Skris Initial support for X509_ATTRIBUTE handling. Stacks of these are 620459191Skris used all over the place including certificate requests and PKCS#7 620559191Skris structures. They are currently handled manually where necessary with 620659191Skris some primitive wrappers for PKCS#7. The new functions behave in a 620759191Skris manner analogous to the X509 extension functions: they allow 620859191Skris attributes to be looked up by NID and added. 620959191Skris 621059191Skris Later something similar to the X509V3 code would be desirable to 621159191Skris automatically handle the encoding, decoding and printing of the 621259191Skris more complex types. The string types like challengePassword can 621359191Skris be handled by the string table functions. 621459191Skris 621559191Skris Also modified the multi byte string table handling. Now there is 621659191Skris a 'global mask' which masks out certain types. The table itself 621759191Skris can use the flag STABLE_NO_MASK to ignore the mask setting: this 621859191Skris is useful when for example there is only one permissible type 621959191Skris (as in countryName) and using the mask might result in no valid 622059191Skris types at all. 622159191Skris [Steve Henson] 622259191Skris 622359191Skris *) Clean up 'Finished' handling, and add functions SSL_get_finished and 622459191Skris SSL_get_peer_finished to allow applications to obtain the latest 622559191Skris Finished messages sent to the peer or expected from the peer, 622659191Skris respectively. (SSL_get_peer_finished is usually the Finished message 622759191Skris actually received from the peer, otherwise the protocol will be aborted.) 622859191Skris 622959191Skris As the Finished message are message digests of the complete handshake 623059191Skris (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 623159191Skris be used for external authentication procedures when the authentication 623259191Skris provided by SSL/TLS is not desired or is not enough. 623359191Skris [Bodo Moeller] 623459191Skris 623559191Skris *) Enhanced support for Alpha Linux is added. Now ./config checks if 623659191Skris the host supports BWX extension and if Compaq C is present on the 623759191Skris $PATH. Just exploiting of the BWX extension results in 20-30% 623859191Skris performance kick for some algorithms, e.g. DES and RC4 to mention 623959191Skris a couple. Compaq C in turn generates ~20% faster code for MD5 and 624059191Skris SHA1. 624159191Skris [Andy Polyakov] 624259191Skris 624359191Skris *) Add support for MS "fast SGC". This is arguably a violation of the 624459191Skris SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 624559191Skris weak crypto and after checking the certificate is SGC a second one 624659191Skris with strong crypto. MS SGC stops the first handshake after receiving 624759191Skris the server certificate message and sends a second client hello. Since 624859191Skris a server will typically do all the time consuming operations before 624959191Skris expecting any further messages from the client (server key exchange 625059191Skris is the most expensive) there is little difference between the two. 625159191Skris 625259191Skris To get OpenSSL to support MS SGC we have to permit a second client 625359191Skris hello message after we have sent server done. In addition we have to 625459191Skris reset the MAC if we do get this second client hello. 625559191Skris [Steve Henson] 625659191Skris 625759191Skris *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 625859191Skris if a DER encoded private key is RSA or DSA traditional format. Changed 625959191Skris d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 626059191Skris format DER encoded private key. Newer code should use PKCS#8 format which 626159191Skris has the key type encoded in the ASN1 structure. Added DER private key 626259191Skris support to pkcs8 application. 626359191Skris [Steve Henson] 626459191Skris 626559191Skris *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 626659191Skris ciphersuites has been selected (as required by the SSL 3/TLS 1 626759191Skris specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 626859191Skris is set, we interpret this as a request to violate the specification 626959191Skris (the worst that can happen is a handshake failure, and 'correct' 627059191Skris behaviour would result in a handshake failure anyway). 627159191Skris [Bodo Moeller] 627259191Skris 627359191Skris *) In SSL_CTX_add_session, take into account that there might be multiple 627459191Skris SSL_SESSION structures with the same session ID (e.g. when two threads 627559191Skris concurrently obtain them from an external cache). 627659191Skris The internal cache can handle only one SSL_SESSION with a given ID, 627759191Skris so if there's a conflict, we now throw out the old one to achieve 627859191Skris consistency. 627959191Skris [Bodo Moeller] 628059191Skris 628159191Skris *) Add OIDs for idea and blowfish in CBC mode. This will allow both 628259191Skris to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 628359191Skris some routines that use cipher OIDs: some ciphers do not have OIDs 628459191Skris defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 628559191Skris example. 628659191Skris [Steve Henson] 628759191Skris 628859191Skris *) Simplify the trust setting structure and code. Now we just have 628959191Skris two sequences of OIDs for trusted and rejected settings. These will 629059191Skris typically have values the same as the extended key usage extension 629159191Skris and any application specific purposes. 629259191Skris 629359191Skris The trust checking code now has a default behaviour: it will just 629459191Skris check for an object with the same NID as the passed id. Functions can 629559191Skris be provided to override either the default behaviour or the behaviour 629659191Skris for a given id. SSL client, server and email already have functions 629759191Skris in place for compatibility: they check the NID and also return "trusted" 629859191Skris if the certificate is self signed. 629959191Skris [Steve Henson] 630059191Skris 630159191Skris *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 630259191Skris traditional format into an EVP_PKEY structure. 630359191Skris [Steve Henson] 630459191Skris 630559191Skris *) Add a password callback function PEM_cb() which either prompts for 630659191Skris a password if usr_data is NULL or otherwise assumes it is a null 630759191Skris terminated password. Allow passwords to be passed on command line 630859191Skris environment or config files in a few more utilities. 630959191Skris [Steve Henson] 631059191Skris 631159191Skris *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 631259191Skris keys. Add some short names for PKCS#8 PBE algorithms and allow them 631359191Skris to be specified on the command line for the pkcs8 and pkcs12 utilities. 631459191Skris Update documentation. 631559191Skris [Steve Henson] 631659191Skris 631759191Skris *) Support for ASN1 "NULL" type. This could be handled before by using 631859191Skris ASN1_TYPE but there wasn't any function that would try to read a NULL 631959191Skris and produce an error if it couldn't. For compatibility we also have 632059191Skris ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 632159191Skris don't allocate anything because they don't need to. 632259191Skris [Steve Henson] 632359191Skris 632459191Skris *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 632559191Skris for details. 632659191Skris [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 632759191Skris 632859191Skris *) Rebuild of the memory allocation routines used by OpenSSL code and 632959191Skris possibly others as well. The purpose is to make an interface that 633059191Skris provide hooks so anyone can build a separate set of allocation and 633159191Skris deallocation routines to be used by OpenSSL, for example memory 633259191Skris pool implementations, or something else, which was previously hard 633359191Skris since Malloc(), Realloc() and Free() were defined as macros having 633459191Skris the values malloc, realloc and free, respectively (except for Win32 633559191Skris compilations). The same is provided for memory debugging code. 633659191Skris OpenSSL already comes with functionality to find memory leaks, but 633759191Skris this gives people a chance to debug other memory problems. 633859191Skris 633959191Skris With these changes, a new set of functions and macros have appeared: 634059191Skris 634168651Skris CRYPTO_set_mem_debug_functions() [F] 634268651Skris CRYPTO_get_mem_debug_functions() [F] 634368651Skris CRYPTO_dbg_set_options() [F] 634468651Skris CRYPTO_dbg_get_options() [F] 634568651Skris CRYPTO_malloc_debug_init() [M] 634659191Skris 634759191Skris The memory debug functions are NULL by default, unless the library 634859191Skris is compiled with CRYPTO_MDEBUG or friends is defined. If someone 634959191Skris wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 635059191Skris gives the standard debugging functions that come with OpenSSL) or 635159191Skris CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 635259191Skris provided by the library user) must be used. When the standard 635359191Skris debugging functions are used, CRYPTO_dbg_set_options can be used to 635459191Skris request additional information: 635559191Skris CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 635659191Skris the CRYPTO_MDEBUG_xxx macro when compiling the library. 635759191Skris 635859191Skris Also, things like CRYPTO_set_mem_functions will always give the 635959191Skris expected result (the new set of functions is used for allocation 636059191Skris and deallocation) at all times, regardless of platform and compiler 636159191Skris options. 636259191Skris 636359191Skris To finish it up, some functions that were never use in any other 636459191Skris way than through macros have a new API and new semantic: 636559191Skris 636659191Skris CRYPTO_dbg_malloc() 636759191Skris CRYPTO_dbg_realloc() 636859191Skris CRYPTO_dbg_free() 636959191Skris 637059191Skris All macros of value have retained their old syntax. 637159191Skris [Richard Levitte and Bodo Moeller] 637259191Skris 637359191Skris *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 637459191Skris ordering of SMIMECapabilities wasn't in "strength order" and there 637559191Skris was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 637659191Skris algorithm. 637759191Skris [Steve Henson] 637859191Skris 637959191Skris *) Some ASN1 types with illegal zero length encoding (INTEGER, 638059191Skris ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 638159191Skris [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 638259191Skris 638359191Skris *) Merge in my S/MIME library for OpenSSL. This provides a simple 638459191Skris S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 638559191Skris functionality to handle multipart/signed properly) and a utility 638659191Skris called 'smime' to call all this stuff. This is based on code I 638759191Skris originally wrote for Celo who have kindly allowed it to be 638859191Skris included in OpenSSL. 638959191Skris [Steve Henson] 639059191Skris 639159191Skris *) Add variants des_set_key_checked and des_set_key_unchecked of 639259191Skris des_set_key (aka des_key_sched). Global variable des_check_key 639359191Skris decides which of these is called by des_set_key; this way 639459191Skris des_check_key behaves as it always did, but applications and 639559191Skris the library itself, which was buggy for des_check_key == 1, 639659191Skris have a cleaner way to pick the version they need. 639759191Skris [Bodo Moeller] 639859191Skris 639959191Skris *) New function PKCS12_newpass() which changes the password of a 640059191Skris PKCS12 structure. 640159191Skris [Steve Henson] 640259191Skris 640359191Skris *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 640459191Skris dynamic mix. In both cases the ids can be used as an index into the 640559191Skris table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 640659191Skris functions so they accept a list of the field values and the 640759191Skris application doesn't need to directly manipulate the X509_TRUST 640859191Skris structure. 640959191Skris [Steve Henson] 641059191Skris 641159191Skris *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 641259191Skris need initialising. 641359191Skris [Steve Henson] 641459191Skris 641559191Skris *) Modify the way the V3 extension code looks up extensions. This now 641659191Skris works in a similar way to the object code: we have some "standard" 641759191Skris extensions in a static table which is searched with OBJ_bsearch() 641859191Skris and the application can add dynamic ones if needed. The file 641959191Skris crypto/x509v3/ext_dat.h now has the info: this file needs to be 642059191Skris updated whenever a new extension is added to the core code and kept 642159191Skris in ext_nid order. There is a simple program 'tabtest.c' which checks 642259191Skris this. New extensions are not added too often so this file can readily 642359191Skris be maintained manually. 642459191Skris 642559191Skris There are two big advantages in doing things this way. The extensions 642659191Skris can be looked up immediately and no longer need to be "added" using 642759191Skris X509V3_add_standard_extensions(): this function now does nothing. 642859191Skris [Side note: I get *lots* of email saying the extension code doesn't 642959191Skris work because people forget to call this function] 643059191Skris Also no dynamic allocation is done unless new extensions are added: 643159191Skris so if we don't add custom extensions there is no need to call 643259191Skris X509V3_EXT_cleanup(). 643359191Skris [Steve Henson] 643459191Skris 643559191Skris *) Modify enc utility's salting as follows: make salting the default. Add a 643659191Skris magic header, so unsalted files fail gracefully instead of just decrypting 643759191Skris to garbage. This is because not salting is a big security hole, so people 643859191Skris should be discouraged from doing it. 643959191Skris [Ben Laurie] 644059191Skris 644159191Skris *) Fixes and enhancements to the 'x509' utility. It allowed a message 644259191Skris digest to be passed on the command line but it only used this 644359191Skris parameter when signing a certificate. Modified so all relevant 644459191Skris operations are affected by the digest parameter including the 644559191Skris -fingerprint and -x509toreq options. Also -x509toreq choked if a 644659191Skris DSA key was used because it didn't fix the digest. 644759191Skris [Steve Henson] 644859191Skris 644959191Skris *) Initial certificate chain verify code. Currently tests the untrusted 645059191Skris certificates for consistency with the verify purpose (which is set 645159191Skris when the X509_STORE_CTX structure is set up) and checks the pathlength. 645259191Skris 645359191Skris There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 645459191Skris this is because it will reject chains with invalid extensions whereas 645559191Skris every previous version of OpenSSL and SSLeay made no checks at all. 645659191Skris 645759191Skris Trust code: checks the root CA for the relevant trust settings. Trust 645859191Skris settings have an initial value consistent with the verify purpose: e.g. 645959191Skris if the verify purpose is for SSL client use it expects the CA to be 646059191Skris trusted for SSL client use. However the default value can be changed to 646159191Skris permit custom trust settings: one example of this would be to only trust 646259191Skris certificates from a specific "secure" set of CAs. 646359191Skris 646459191Skris Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 646559191Skris which should be used for version portability: especially since the 646659191Skris verify structure is likely to change more often now. 646759191Skris 646859191Skris SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 646959191Skris to set them. If not set then assume SSL clients will verify SSL servers 647059191Skris and vice versa. 647159191Skris 647259191Skris Two new options to the verify program: -untrusted allows a set of 647359191Skris untrusted certificates to be passed in and -purpose which sets the 647459191Skris intended purpose of the certificate. If a purpose is set then the 647559191Skris new chain verify code is used to check extension consistency. 647659191Skris [Steve Henson] 647759191Skris 647859191Skris *) Support for the authority information access extension. 647959191Skris [Steve Henson] 648059191Skris 648159191Skris *) Modify RSA and DSA PEM read routines to transparently handle 648259191Skris PKCS#8 format private keys. New *_PUBKEY_* functions that handle 648359191Skris public keys in a format compatible with certificate 648459191Skris SubjectPublicKeyInfo structures. Unfortunately there were already 648559191Skris functions called *_PublicKey_* which used various odd formats so 648659191Skris these are retained for compatibility: however the DSA variants were 648759191Skris never in a public release so they have been deleted. Changed dsa/rsa 648859191Skris utilities to handle the new format: note no releases ever handled public 648959191Skris keys so we should be OK. 649059191Skris 649159191Skris The primary motivation for this change is to avoid the same fiasco 649259191Skris that dogs private keys: there are several incompatible private key 649359191Skris formats some of which are standard and some OpenSSL specific and 649459191Skris require various evil hacks to allow partial transparent handling and 649559191Skris even then it doesn't work with DER formats. Given the option anything 649659191Skris other than PKCS#8 should be dumped: but the other formats have to 649759191Skris stay in the name of compatibility. 649859191Skris 649959191Skris With public keys and the benefit of hindsight one standard format 650059191Skris is used which works with EVP_PKEY, RSA or DSA structures: though 650159191Skris it clearly returns an error if you try to read the wrong kind of key. 650259191Skris 650359191Skris Added a -pubkey option to the 'x509' utility to output the public key. 650459191Skris Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 650559191Skris (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 650659191Skris EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 650759191Skris that do the same as the EVP_PKEY_assign_*() except they up the 650859191Skris reference count of the added key (they don't "swallow" the 650959191Skris supplied key). 651059191Skris [Steve Henson] 651159191Skris 651259191Skris *) Fixes to crypto/x509/by_file.c the code to read in certificates and 651359191Skris CRLs would fail if the file contained no certificates or no CRLs: 651459191Skris added a new function to read in both types and return the number 651559191Skris read: this means that if none are read it will be an error. The 651659191Skris DER versions of the certificate and CRL reader would always fail 651759191Skris because it isn't possible to mix certificates and CRLs in DER format 651859191Skris without choking one or the other routine. Changed this to just read 651959191Skris a certificate: this is the best we can do. Also modified the code 652059191Skris in apps/verify.c to take notice of return codes: it was previously 652159191Skris attempting to read in certificates from NULL pointers and ignoring 652259191Skris any errors: this is one reason why the cert and CRL reader seemed 652359191Skris to work. It doesn't check return codes from the default certificate 652459191Skris routines: these may well fail if the certificates aren't installed. 652559191Skris [Steve Henson] 652659191Skris 652759191Skris *) Code to support otherName option in GeneralName. 652859191Skris [Steve Henson] 652959191Skris 653059191Skris *) First update to verify code. Change the verify utility 653159191Skris so it warns if it is passed a self signed certificate: 653259191Skris for consistency with the normal behaviour. X509_verify 653359191Skris has been modified to it will now verify a self signed 653459191Skris certificate if *exactly* the same certificate appears 653559191Skris in the store: it was previously impossible to trust a 653659191Skris single self signed certificate. This means that: 653759191Skris openssl verify ss.pem 653859191Skris now gives a warning about a self signed certificate but 653959191Skris openssl verify -CAfile ss.pem ss.pem 654059191Skris is OK. 654159191Skris [Steve Henson] 654259191Skris 654359191Skris *) For servers, store verify_result in SSL_SESSION data structure 654459191Skris (and add it to external session representation). 654559191Skris This is needed when client certificate verifications fails, 654659191Skris but an application-provided verification callback (set by 654759191Skris SSL_CTX_set_cert_verify_callback) allows accepting the session 654859191Skris anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 654959191Skris but returns 1): When the session is reused, we have to set 655059191Skris ssl->verify_result to the appropriate error code to avoid 655159191Skris security holes. 655259191Skris [Bodo Moeller, problem pointed out by Lutz Jaenicke] 655359191Skris 655459191Skris *) Fix a bug in the new PKCS#7 code: it didn't consider the 655559191Skris case in PKCS7_dataInit() where the signed PKCS7 structure 655659191Skris didn't contain any existing data because it was being created. 655759191Skris [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 655859191Skris 655959191Skris *) Add a salt to the key derivation routines in enc.c. This 656059191Skris forms the first 8 bytes of the encrypted file. Also add a 656159191Skris -S option to allow a salt to be input on the command line. 656259191Skris [Steve Henson] 656359191Skris 656459191Skris *) New function X509_cmp(). Oddly enough there wasn't a function 656559191Skris to compare two certificates. We do this by working out the SHA1 656659191Skris hash and comparing that. X509_cmp() will be needed by the trust 656759191Skris code. 656859191Skris [Steve Henson] 656959191Skris 657059191Skris *) SSL_get1_session() is like SSL_get_session(), but increments 657159191Skris the reference count in the SSL_SESSION returned. 657259191Skris [Geoff Thorpe <geoff@eu.c2.net>] 657359191Skris 657459191Skris *) Fix for 'req': it was adding a null to request attributes. 657559191Skris Also change the X509_LOOKUP and X509_INFO code to handle 657659191Skris certificate auxiliary information. 657759191Skris [Steve Henson] 657859191Skris 657959191Skris *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 658059191Skris the 'enc' command. 658159191Skris [Steve Henson] 658259191Skris 658359191Skris *) Add the possibility to add extra information to the memory leak 658459191Skris detecting output, to form tracebacks, showing from where each 658559191Skris allocation was originated: CRYPTO_push_info("constant string") adds 658659191Skris the string plus current file name and line number to a per-thread 658759191Skris stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 658859191Skris is like calling CYRPTO_pop_info() until the stack is empty. 658959191Skris Also updated memory leak detection code to be multi-thread-safe. 659059191Skris [Richard Levitte] 659159191Skris 659259191Skris *) Add options -text and -noout to pkcs7 utility and delete the 659359191Skris encryption options which never did anything. Update docs. 659459191Skris [Steve Henson] 659559191Skris 659659191Skris *) Add options to some of the utilities to allow the pass phrase 659759191Skris to be included on either the command line (not recommended on 659859191Skris OSes like Unix) or read from the environment. Update the 659959191Skris manpages and fix a few bugs. 660059191Skris [Steve Henson] 660159191Skris 660259191Skris *) Add a few manpages for some of the openssl commands. 660359191Skris [Steve Henson] 660459191Skris 660559191Skris *) Fix the -revoke option in ca. It was freeing up memory twice, 660659191Skris leaking and not finding already revoked certificates. 660759191Skris [Steve Henson] 660859191Skris 660959191Skris *) Extensive changes to support certificate auxiliary information. 661059191Skris This involves the use of X509_CERT_AUX structure and X509_AUX 661159191Skris functions. An X509_AUX function such as PEM_read_X509_AUX() 661259191Skris can still read in a certificate file in the usual way but it 661359191Skris will also read in any additional "auxiliary information". By 661459191Skris doing things this way a fair degree of compatibility can be 661559191Skris retained: existing certificates can have this information added 661659191Skris using the new 'x509' options. 661759191Skris 661859191Skris Current auxiliary information includes an "alias" and some trust 661959191Skris settings. The trust settings will ultimately be used in enhanced 662059191Skris certificate chain verification routines: currently a certificate 662159191Skris can only be trusted if it is self signed and then it is trusted 662259191Skris for all purposes. 662359191Skris [Steve Henson] 662459191Skris 662559191Skris *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 662659191Skris The problem was that one of the replacement routines had not been working 662759191Skris since SSLeay releases. For now the offending routine has been replaced 662859191Skris with non-optimised assembler. Even so, this now gives around 95% 662959191Skris performance improvement for 1024 bit RSA signs. 663059191Skris [Mark Cox] 663159191Skris 663259191Skris *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 663359191Skris handling. Most clients have the effective key size in bits equal to 663459191Skris the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 663559191Skris A few however don't do this and instead use the size of the decrypted key 663659191Skris to determine the RC2 key length and the AlgorithmIdentifier to determine 663759191Skris the effective key length. In this case the effective key length can still 663859191Skris be 40 bits but the key length can be 168 bits for example. This is fixed 663959191Skris by manually forcing an RC2 key into the EVP_PKEY structure because the 664059191Skris EVP code can't currently handle unusual RC2 key sizes: it always assumes 664159191Skris the key length and effective key length are equal. 664259191Skris [Steve Henson] 664359191Skris 664459191Skris *) Add a bunch of functions that should simplify the creation of 664559191Skris X509_NAME structures. Now you should be able to do: 664659191Skris X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 664759191Skris and have it automatically work out the correct field type and fill in 664859191Skris the structures. The more adventurous can try: 664959191Skris X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 665059191Skris and it will (hopefully) work out the correct multibyte encoding. 665159191Skris [Steve Henson] 665259191Skris 665359191Skris *) Change the 'req' utility to use the new field handling and multibyte 665459191Skris copy routines. Before the DN field creation was handled in an ad hoc 665559191Skris way in req, ca, and x509 which was rather broken and didn't support 665659191Skris BMPStrings or UTF8Strings. Since some software doesn't implement 665759191Skris BMPStrings or UTF8Strings yet, they can be enabled using the config file 665859191Skris using the dirstring_type option. See the new comment in the default 665959191Skris openssl.cnf for more info. 666059191Skris [Steve Henson] 666159191Skris 666259191Skris *) Make crypto/rand/md_rand.c more robust: 666359191Skris - Assure unique random numbers after fork(). 666459191Skris - Make sure that concurrent threads access the global counter and 666559191Skris md serializably so that we never lose entropy in them 666659191Skris or use exactly the same state in multiple threads. 666759191Skris Access to the large state is not always serializable because 666859191Skris the additional locking could be a performance killer, and 666959191Skris md should be large enough anyway. 667059191Skris [Bodo Moeller] 667159191Skris 667259191Skris *) New file apps/app_rand.c with commonly needed functionality 667359191Skris for handling the random seed file. 667459191Skris 667559191Skris Use the random seed file in some applications that previously did not: 667659191Skris ca, 667759191Skris dsaparam -genkey (which also ignored its '-rand' option), 667859191Skris s_client, 667959191Skris s_server, 668059191Skris x509 (when signing). 668159191Skris Except on systems with /dev/urandom, it is crucial to have a random 668259191Skris seed file at least for key creation, DSA signing, and for DH exchanges; 668359191Skris for RSA signatures we could do without one. 668459191Skris 668559191Skris gendh and gendsa (unlike genrsa) used to read only the first byte 668659191Skris of each file listed in the '-rand' option. The function as previously 668759191Skris found in genrsa is now in app_rand.c and is used by all programs 668859191Skris that support '-rand'. 668959191Skris [Bodo Moeller] 669059191Skris 669159191Skris *) In RAND_write_file, use mode 0600 for creating files; 669259191Skris don't just chmod when it may be too late. 669359191Skris [Bodo Moeller] 669459191Skris 669559191Skris *) Report an error from X509_STORE_load_locations 669659191Skris when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 669759191Skris [Bill Perry] 669859191Skris 669959191Skris *) New function ASN1_mbstring_copy() this copies a string in either 670059191Skris ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 670159191Skris into an ASN1_STRING type. A mask of permissible types is passed 670259191Skris and it chooses the "minimal" type to use or an error if not type 670359191Skris is suitable. 670459191Skris [Steve Henson] 670559191Skris 670659191Skris *) Add function equivalents to the various macros in asn1.h. The old 670759191Skris macros are retained with an M_ prefix. Code inside the library can 670859191Skris use the M_ macros. External code (including the openssl utility) 670959191Skris should *NOT* in order to be "shared library friendly". 671059191Skris [Steve Henson] 671159191Skris 671259191Skris *) Add various functions that can check a certificate's extensions 671359191Skris to see if it usable for various purposes such as SSL client, 671459191Skris server or S/MIME and CAs of these types. This is currently 671559191Skris VERY EXPERIMENTAL but will ultimately be used for certificate chain 671659191Skris verification. Also added a -purpose flag to x509 utility to 671759191Skris print out all the purposes. 671859191Skris [Steve Henson] 671959191Skris 672059191Skris *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 672159191Skris functions. 672259191Skris [Steve Henson] 672359191Skris 672459191Skris *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 672559191Skris for, obtain and decode and extension and obtain its critical flag. 672659191Skris This allows all the necessary extension code to be handled in a 672759191Skris single function call. 672859191Skris [Steve Henson] 672959191Skris 673059191Skris *) RC4 tune-up featuring 30-40% performance improvement on most RISC 673159191Skris platforms. See crypto/rc4/rc4_enc.c for further details. 673259191Skris [Andy Polyakov] 673359191Skris 673459191Skris *) New -noout option to asn1parse. This causes no output to be produced 673559191Skris its main use is when combined with -strparse and -out to extract data 673659191Skris from a file (which may not be in ASN.1 format). 673759191Skris [Steve Henson] 673859191Skris 673959191Skris *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 674059191Skris when producing the local key id. 674159191Skris [Richard Levitte <levitte@stacken.kth.se>] 674259191Skris 674359191Skris *) New option -dhparam in s_server. This allows a DH parameter file to be 674459191Skris stated explicitly. If it is not stated then it tries the first server 674559191Skris certificate file. The previous behaviour hard coded the filename 674659191Skris "server.pem". 674759191Skris [Steve Henson] 674859191Skris 674959191Skris *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 675059191Skris a public key to be input or output. For example: 675159191Skris openssl rsa -in key.pem -pubout -out pubkey.pem 675259191Skris Also added necessary DSA public key functions to handle this. 675359191Skris [Steve Henson] 675459191Skris 675559191Skris *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 675659191Skris in the message. This was handled by allowing 675759191Skris X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 675859191Skris [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 675959191Skris 676059191Skris *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 676159191Skris to the end of the strings whereas this didn't. This would cause problems 676259191Skris if strings read with d2i_ASN1_bytes() were later modified. 676359191Skris [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 676459191Skris 676559191Skris *) Fix for base64 decode bug. When a base64 bio reads only one line of 676659191Skris data and it contains EOF it will end up returning an error. This is 676759191Skris caused by input 46 bytes long. The cause is due to the way base64 676859191Skris BIOs find the start of base64 encoded data. They do this by trying a 676959191Skris trial decode on each line until they find one that works. When they 677059191Skris do a flag is set and it starts again knowing it can pass all the 677159191Skris data directly through the decoder. Unfortunately it doesn't reset 677259191Skris the context it uses. This means that if EOF is reached an attempt 677359191Skris is made to pass two EOFs through the context and this causes the 677459191Skris resulting error. This can also cause other problems as well. As is 677559191Skris usual with these problems it takes *ages* to find and the fix is 677659191Skris trivial: move one line. 677759191Skris [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 677859191Skris 677959191Skris *) Ugly workaround to get s_client and s_server working under Windows. The 678059191Skris old code wouldn't work because it needed to select() on sockets and the 678159191Skris tty (for keypresses and to see if data could be written). Win32 only 678259191Skris supports select() on sockets so we select() with a 1s timeout on the 678359191Skris sockets and then see if any characters are waiting to be read, if none 678459191Skris are present then we retry, we also assume we can always write data to 678559191Skris the tty. This isn't nice because the code then blocks until we've 678659191Skris received a complete line of data and it is effectively polling the 678759191Skris keyboard at 1s intervals: however it's quite a bit better than not 678859191Skris working at all :-) A dedicated Windows application might handle this 678959191Skris with an event loop for example. 679059191Skris [Steve Henson] 679159191Skris 679259191Skris *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 679359191Skris and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 679459191Skris will be called when RSA_sign() and RSA_verify() are used. This is useful 679559191Skris if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 679659191Skris For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 679759191Skris should *not* be used: RSA_sign() and RSA_verify() must be used instead. 679859191Skris This necessitated the support of an extra signature type NID_md5_sha1 679959191Skris for SSL signatures and modifications to the SSL library to use it instead 680059191Skris of calling RSA_public_decrypt() and RSA_private_encrypt(). 680159191Skris [Steve Henson] 680259191Skris 680359191Skris *) Add new -verify -CAfile and -CApath options to the crl program, these 680459191Skris will lookup a CRL issuers certificate and verify the signature in a 680559191Skris similar way to the verify program. Tidy up the crl program so it 680659191Skris no longer accesses structures directly. Make the ASN1 CRL parsing a bit 680759191Skris less strict. It will now permit CRL extensions even if it is not 680859191Skris a V2 CRL: this will allow it to tolerate some broken CRLs. 680959191Skris [Steve Henson] 681059191Skris 681159191Skris *) Initialize all non-automatic variables each time one of the openssl 681259191Skris sub-programs is started (this is necessary as they may be started 681359191Skris multiple times from the "OpenSSL>" prompt). 681459191Skris [Lennart Bang, Bodo Moeller] 681559191Skris 681659191Skris *) Preliminary compilation option RSA_NULL which disables RSA crypto without 681759191Skris removing all other RSA functionality (this is what NO_RSA does). This 681859191Skris is so (for example) those in the US can disable those operations covered 681959191Skris by the RSA patent while allowing storage and parsing of RSA keys and RSA 682059191Skris key generation. 682159191Skris [Steve Henson] 682259191Skris 682359191Skris *) Non-copying interface to BIO pairs. 682459191Skris (still largely untested) 682559191Skris [Bodo Moeller] 682659191Skris 682759191Skris *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 682859191Skris ASCII string. This was handled independently in various places before. 682959191Skris [Steve Henson] 683059191Skris 683159191Skris *) New functions UTF8_getc() and UTF8_putc() that parse and generate 683259191Skris UTF8 strings a character at a time. 683359191Skris [Steve Henson] 683459191Skris 683559191Skris *) Use client_version from client hello to select the protocol 683659191Skris (s23_srvr.c) and for RSA client key exchange verification 683759191Skris (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 683859191Skris [Bodo Moeller] 683959191Skris 684059191Skris *) Add various utility functions to handle SPKACs, these were previously 684159191Skris handled by poking round in the structure internals. Added new function 684259191Skris NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 684359191Skris print, verify and generate SPKACs. Based on an original idea from 684459191Skris Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 684559191Skris [Steve Henson] 684659191Skris 684759191Skris *) RIPEMD160 is operational on all platforms and is back in 'make test'. 684859191Skris [Andy Polyakov] 684959191Skris 685059191Skris *) Allow the config file extension section to be overwritten on the 685159191Skris command line. Based on an original idea from Massimiliano Pala 685259191Skris <madwolf@comune.modena.it>. The new option is called -extensions 685359191Skris and can be applied to ca, req and x509. Also -reqexts to override 685459191Skris the request extensions in req and -crlexts to override the crl extensions 685559191Skris in ca. 685659191Skris [Steve Henson] 685759191Skris 685859191Skris *) Add new feature to the SPKAC handling in ca. Now you can include 685959191Skris the same field multiple times by preceding it by "XXXX." for example: 686059191Skris 1.OU="Unit name 1" 686159191Skris 2.OU="Unit name 2" 686259191Skris this is the same syntax as used in the req config file. 686359191Skris [Steve Henson] 686459191Skris 686559191Skris *) Allow certificate extensions to be added to certificate requests. These 686659191Skris are specified in a 'req_extensions' option of the req section of the 686759191Skris config file. They can be printed out with the -text option to req but 686859191Skris are otherwise ignored at present. 686959191Skris [Steve Henson] 687059191Skris 687159191Skris *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 687259191Skris data read consists of only the final block it would not decrypted because 687359191Skris EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 687459191Skris A misplaced 'break' also meant the decrypted final block might not be 687559191Skris copied until the next read. 687659191Skris [Steve Henson] 687759191Skris 687859191Skris *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 687959191Skris a few extra parameters to the DH structure: these will be useful if 688059191Skris for example we want the value of 'q' or implement X9.42 DH. 688159191Skris [Steve Henson] 688259191Skris 688359191Skris *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 688459191Skris provides hooks that allow the default DSA functions or functions on a 688559191Skris "per key" basis to be replaced. This allows hardware acceleration and 688659191Skris hardware key storage to be handled without major modification to the 688759191Skris library. Also added low level modexp hooks and CRYPTO_EX structure and 688859191Skris associated functions. 688959191Skris [Steve Henson] 689059191Skris 689159191Skris *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 689259191Skris as "read only": it can't be written to and the buffer it points to will 689359191Skris not be freed. Reading from a read only BIO is much more efficient than 689459191Skris a normal memory BIO. This was added because there are several times when 689559191Skris an area of memory needs to be read from a BIO. The previous method was 689659191Skris to create a memory BIO and write the data to it, this results in two 689759191Skris copies of the data and an O(n^2) reading algorithm. There is a new 689859191Skris function BIO_new_mem_buf() which creates a read only memory BIO from 689959191Skris an area of memory. Also modified the PKCS#7 routines to use read only 690059191Skris memory BIOs. 690159191Skris [Steve Henson] 690259191Skris 690359191Skris *) Bugfix: ssl23_get_client_hello did not work properly when called in 690459191Skris state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 690559191Skris a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 690659191Skris but a retry condition occured while trying to read the rest. 690759191Skris [Bodo Moeller] 690859191Skris 690959191Skris *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 691059191Skris NID_pkcs7_encrypted by default: this was wrong since this should almost 691159191Skris always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 691259191Skris the encrypted data type: this is a more sensible place to put it and it 691359191Skris allows the PKCS#12 code to be tidied up that duplicated this 691459191Skris functionality. 691559191Skris [Steve Henson] 691659191Skris 691759191Skris *) Changed obj_dat.pl script so it takes its input and output files on 691859191Skris the command line. This should avoid shell escape redirection problems 691959191Skris under Win32. 692059191Skris [Steve Henson] 692159191Skris 692259191Skris *) Initial support for certificate extension requests, these are included 692359191Skris in things like Xenroll certificate requests. Included functions to allow 692459191Skris extensions to be obtained and added. 692559191Skris [Steve Henson] 692659191Skris 692759191Skris *) -crlf option to s_client and s_server for sending newlines as 692859191Skris CRLF (as required by many protocols). 692959191Skris [Bodo Moeller] 693059191Skris 693155714Skris Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 693255714Skris 693355714Skris *) Install libRSAglue.a when OpenSSL is built with RSAref. 693455714Skris [Ralf S. Engelschall] 693555714Skris 693655714Skris *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 693755714Skris [Andrija Antonijevic <TheAntony2@bigfoot.com>] 693855714Skris 693955714Skris *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 694055714Skris program. 694155714Skris [Steve Henson] 694255714Skris 694355714Skris *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 694455714Skris DH parameters/keys (q is lost during that conversion, but the resulting 694555714Skris DH parameters contain its length). 694655714Skris 694755714Skris For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 694855714Skris much faster than DH_generate_parameters (which creates parameters 694955714Skris where p = 2*q + 1), and also the smaller q makes DH computations 695055714Skris much more efficient (160-bit exponentiation instead of 1024-bit 695155714Skris exponentiation); so this provides a convenient way to support DHE 695255714Skris ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 695355714Skris utter importance to use 695455714Skris SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 695555714Skris or 695655714Skris SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 695755714Skris when such DH parameters are used, because otherwise small subgroup 695855714Skris attacks may become possible! 695955714Skris [Bodo Moeller] 696055714Skris 696155714Skris *) Avoid memory leak in i2d_DHparams. 696255714Skris [Bodo Moeller] 696355714Skris 696455714Skris *) Allow the -k option to be used more than once in the enc program: 696555714Skris this allows the same encrypted message to be read by multiple recipients. 696655714Skris [Steve Henson] 696755714Skris 696855714Skris *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 696955714Skris an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 697055714Skris it will always use the numerical form of the OID, even if it has a short 697155714Skris or long name. 697255714Skris [Steve Henson] 697355714Skris 697455714Skris *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 697555714Skris method only got called if p,q,dmp1,dmq1,iqmp components were present, 697655714Skris otherwise bn_mod_exp was called. In the case of hardware keys for example 697755714Skris no private key components need be present and it might store extra data 697859191Skris in the RSA structure, which cannot be accessed from bn_mod_exp. 697959191Skris By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 698059191Skris private key operations. 698155714Skris [Steve Henson] 698255714Skris 698355714Skris *) Added support for SPARC Linux. 698455714Skris [Andy Polyakov] 698555714Skris 698655714Skris *) pem_password_cb function type incompatibly changed from 698755714Skris typedef int pem_password_cb(char *buf, int size, int rwflag); 698855714Skris to 698955714Skris ....(char *buf, int size, int rwflag, void *userdata); 699055714Skris so that applications can pass data to their callbacks: 699155714Skris The PEM[_ASN1]_{read,write}... functions and macros now take an 699255714Skris additional void * argument, which is just handed through whenever 699355714Skris the password callback is called. 699459191Skris [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 699555714Skris 699655714Skris New function SSL_CTX_set_default_passwd_cb_userdata. 699755714Skris 699855714Skris Compatibility note: As many C implementations push function arguments 699955714Skris onto the stack in reverse order, the new library version is likely to 700055714Skris interoperate with programs that have been compiled with the old 700155714Skris pem_password_cb definition (PEM_whatever takes some data that 700255714Skris happens to be on the stack as its last argument, and the callback 700355714Skris just ignores this garbage); but there is no guarantee whatsoever that 700455714Skris this will work. 700555714Skris 700655714Skris *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 700755714Skris (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 700855714Skris problems not only on Windows, but also on some Unix platforms. 700955714Skris To avoid problematic command lines, these definitions are now in an 701055714Skris auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 701155714Skris for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 701255714Skris [Bodo Moeller] 701355714Skris 701455714Skris *) MIPS III/IV assembler module is reimplemented. 701555714Skris [Andy Polyakov] 701655714Skris 701755714Skris *) More DES library cleanups: remove references to srand/rand and 701855714Skris delete an unused file. 701955714Skris [Ulf M�ller] 702055714Skris 702155714Skris *) Add support for the the free Netwide assembler (NASM) under Win32, 702255714Skris since not many people have MASM (ml) and it can be hard to obtain. 702355714Skris This is currently experimental but it seems to work OK and pass all 702455714Skris the tests. Check out INSTALL.W32 for info. 702555714Skris [Steve Henson] 702655714Skris 702755714Skris *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 702855714Skris without temporary keys kept an extra copy of the server key, 702955714Skris and connections with temporary keys did not free everything in case 703055714Skris of an error. 703155714Skris [Bodo Moeller] 703255714Skris 703355714Skris *) New function RSA_check_key and new openssl rsa option -check 703455714Skris for verifying the consistency of RSA keys. 703555714Skris [Ulf Moeller, Bodo Moeller] 703655714Skris 703755714Skris *) Various changes to make Win32 compile work: 703855714Skris 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 703955714Skris 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 704055714Skris comparison" warnings. 704155714Skris 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 704255714Skris [Steve Henson] 704355714Skris 704455714Skris *) Add a debugging option to PKCS#5 v2 key generation function: when 704555714Skris you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 704655714Skris derived keys are printed to stderr. 704755714Skris [Steve Henson] 704855714Skris 704955714Skris *) Copy the flags in ASN1_STRING_dup(). 705055714Skris [Roman E. Pavlov <pre@mo.msk.ru>] 705155714Skris 705255714Skris *) The x509 application mishandled signing requests containing DSA 705355714Skris keys when the signing key was also DSA and the parameters didn't match. 705455714Skris 705555714Skris It was supposed to omit the parameters when they matched the signing key: 705655714Skris the verifying software was then supposed to automatically use the CA's 705755714Skris parameters if they were absent from the end user certificate. 705855714Skris 705955714Skris Omitting parameters is no longer recommended. The test was also 706055714Skris the wrong way round! This was probably due to unusual behaviour in 706155714Skris EVP_cmp_parameters() which returns 1 if the parameters match. 706255714Skris This meant that parameters were omitted when they *didn't* match and 706355714Skris the certificate was useless. Certificates signed with 'ca' didn't have 706455714Skris this bug. 706555714Skris [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 706655714Skris 706755714Skris *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 706855714Skris The interface is as follows: 706955714Skris Applications can use 707055714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 707155714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 707255714Skris "off" is now the default. 707355714Skris The library internally uses 707455714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 707555714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 707655714Skris to disable memory-checking temporarily. 707755714Skris 707855714Skris Some inconsistent states that previously were possible (and were 707955714Skris even the default) are now avoided. 708055714Skris 708155714Skris -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 708255714Skris with each memory chunk allocated; this is occasionally more helpful 708355714Skris than just having a counter. 708455714Skris 708555714Skris -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 708655714Skris 708755714Skris -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 708855714Skris extensions. 708955714Skris [Bodo Moeller] 709055714Skris 709155714Skris *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 709255714Skris which largely parallels "options", but is for changing API behaviour, 709355714Skris whereas "options" are about protocol behaviour. 709455714Skris Initial "mode" flags are: 709555714Skris 709655714Skris SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 709755714Skris a single record has been written. 709855714Skris SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 709955714Skris retries use the same buffer location. 710055714Skris (But all of the contents must be 710155714Skris copied!) 710255714Skris [Bodo Moeller] 710355714Skris 710479998Skris *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 710555714Skris worked. 710655714Skris 710755714Skris *) Fix problems with no-hmac etc. 710855714Skris [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 710955714Skris 711055714Skris *) New functions RSA_get_default_method(), RSA_set_method() and 711155714Skris RSA_get_method(). These allows replacement of RSA_METHODs without having 711255714Skris to mess around with the internals of an RSA structure. 711355714Skris [Steve Henson] 711455714Skris 711555714Skris *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 711655714Skris Also really enable memory leak checks in openssl.c and in some 711755714Skris test programs. 711855714Skris [Chad C. Mulligan, Bodo Moeller] 711955714Skris 712055714Skris *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 712155714Skris up the length of negative integers. This has now been simplified to just 712255714Skris store the length when it is first determined and use it later, rather 712355714Skris than trying to keep track of where data is copied and updating it to 712455714Skris point to the end. 712555714Skris [Steve Henson, reported by Brien Wheeler 712655714Skris <bwheeler@authentica-security.com>] 712755714Skris 712855714Skris *) Add a new function PKCS7_signatureVerify. This allows the verification 712955714Skris of a PKCS#7 signature but with the signing certificate passed to the 713055714Skris function itself. This contrasts with PKCS7_dataVerify which assumes the 713155714Skris certificate is present in the PKCS#7 structure. This isn't always the 713255714Skris case: certificates can be omitted from a PKCS#7 structure and be 713355714Skris distributed by "out of band" means (such as a certificate database). 713455714Skris [Steve Henson] 713555714Skris 713655714Skris *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 713755714Skris function prototypes in pem.h, also change util/mkdef.pl to add the 713855714Skris necessary function names. 713955714Skris [Steve Henson] 714055714Skris 714155714Skris *) mk1mf.pl (used by Windows builds) did not properly read the 714255714Skris options set by Configure in the top level Makefile, and Configure 714355714Skris was not even able to write more than one option correctly. 714455714Skris Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 714555714Skris [Bodo Moeller] 714655714Skris 714755714Skris *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 714855714Skris file to be loaded from a BIO or FILE pointer. The BIO version will 714955714Skris for example allow memory BIOs to contain config info. 715055714Skris [Steve Henson] 715155714Skris 715255714Skris *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 715355714Skris Whoever hopes to achieve shared-library compatibility across versions 715455714Skris must use this, not the compile-time macro. 715555714Skris (Exercise 0.9.4: Which is the minimum library version required by 715655714Skris such programs?) 715755714Skris Note: All this applies only to multi-threaded programs, others don't 715855714Skris need locks. 715955714Skris [Bodo Moeller] 716055714Skris 716155714Skris *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 716255714Skris through a BIO pair triggered the default case, i.e. 716355714Skris SSLerr(...,SSL_R_UNKNOWN_STATE). 716455714Skris [Bodo Moeller] 716555714Skris 716655714Skris *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 716755714Skris can use the SSL library even if none of the specific BIOs is 716855714Skris appropriate. 716955714Skris [Bodo Moeller] 717055714Skris 717155714Skris *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 717255714Skris for the encoded length. 717355714Skris [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 717455714Skris 717555714Skris *) Add initial documentation of the X509V3 functions. 717655714Skris [Steve Henson] 717755714Skris 717855714Skris *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 717955714Skris PEM_write_bio_PKCS8PrivateKey() that are equivalent to 718055714Skris PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 718155714Skris secure PKCS#8 private key format with a high iteration count. 718255714Skris [Steve Henson] 718355714Skris 718455714Skris *) Fix determination of Perl interpreter: A perl or perl5 718555714Skris _directory_ in $PATH was also accepted as the interpreter. 718655714Skris [Ralf S. Engelschall] 718755714Skris 718855714Skris *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 718955714Skris wrong with it but it was very old and did things like calling 719055714Skris PEM_ASN1_read() directly and used MD5 for the hash not to mention some 719155714Skris unusual formatting. 719255714Skris [Steve Henson] 719355714Skris 719455714Skris *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 719555714Skris to use the new extension code. 719655714Skris [Steve Henson] 719755714Skris 719855714Skris *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 719955714Skris with macros. This should make it easier to change their form, add extra 720055714Skris arguments etc. Fix a few PEM prototypes which didn't have cipher as a 720155714Skris constant. 720255714Skris [Steve Henson] 720355714Skris 720455714Skris *) Add to configuration table a new entry that can specify an alternative 720555714Skris name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 720655714Skris according to Mark Crispin <MRC@Panda.COM>. 720755714Skris [Bodo Moeller] 720855714Skris 720955714Skris#if 0 721055714Skris *) DES CBC did not update the IV. Weird. 721155714Skris [Ben Laurie] 721255714Skris#else 721355714Skris des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 721455714Skris Changing the behaviour of the former might break existing programs -- 721555714Skris where IV updating is needed, des_ncbc_encrypt can be used. 721655714Skris#endif 721755714Skris 721855714Skris *) When bntest is run from "make test" it drives bc to check its 721955714Skris calculations, as well as internally checking them. If an internal check 722055714Skris fails, it needs to cause bc to give a non-zero result or make test carries 722155714Skris on without noticing the failure. Fixed. 722255714Skris [Ben Laurie] 722355714Skris 722455714Skris *) DES library cleanups. 722555714Skris [Ulf M�ller] 722655714Skris 722755714Skris *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 722855714Skris used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 722955714Skris ciphers. NOTE: although the key derivation function has been verified 723055714Skris against some published test vectors it has not been extensively tested 723155714Skris yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 723255714Skris of v2.0. 723355714Skris [Steve Henson] 723455714Skris 723555714Skris *) Instead of "mkdir -p", which is not fully portable, use new 723655714Skris Perl script "util/mkdir-p.pl". 723755714Skris [Bodo Moeller] 723855714Skris 723955714Skris *) Rewrite the way password based encryption (PBE) is handled. It used to 724055714Skris assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 724155714Skris structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 724255714Skris but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 724355714Skris the 'parameter' field of the AlgorithmIdentifier is passed to the 724455714Skris underlying key generation function so it must do its own ASN1 parsing. 724555714Skris This has also changed the EVP_PBE_CipherInit() function which now has a 724655714Skris 'parameter' argument instead of literal salt and iteration count values 724755714Skris and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 724855714Skris [Steve Henson] 724955714Skris 725055714Skris *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 725155714Skris and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 725255714Skris Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 725355714Skris KEY" because this clashed with PKCS#8 unencrypted string. Since this 725455714Skris value was just used as a "magic string" and not used directly its 725555714Skris value doesn't matter. 725655714Skris [Steve Henson] 725755714Skris 725855714Skris *) Introduce some semblance of const correctness to BN. Shame C doesn't 725955714Skris support mutable. 726055714Skris [Ben Laurie] 726155714Skris 726255714Skris *) "linux-sparc64" configuration (ultrapenguin). 726355714Skris [Ray Miller <ray.miller@oucs.ox.ac.uk>] 726455714Skris "linux-sparc" configuration. 726555714Skris [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 726655714Skris 726755714Skris *) config now generates no-xxx options for missing ciphers. 726855714Skris [Ulf M�ller] 726955714Skris 727055714Skris *) Support the EBCDIC character set (work in progress). 727155714Skris File ebcdic.c not yet included because it has a different license. 727255714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 727355714Skris 727455714Skris *) Support BS2000/OSD-POSIX. 727555714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 727655714Skris 727755714Skris *) Make callbacks for key generation use void * instead of char *. 727855714Skris [Ben Laurie] 727955714Skris 728055714Skris *) Make S/MIME samples compile (not yet tested). 728155714Skris [Ben Laurie] 728255714Skris 728355714Skris *) Additional typesafe stacks. 728455714Skris [Ben Laurie] 728555714Skris 728655714Skris *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 728755714Skris [Bodo Moeller] 728855714Skris 728955714Skris 729055714Skris Changes between 0.9.3 and 0.9.3a [29 May 1999] 729155714Skris 729255714Skris *) New configuration variant "sco5-gcc". 729355714Skris 729455714Skris *) Updated some demos. 729555714Skris [Sean O Riordain, Wade Scholine] 729655714Skris 729755714Skris *) Add missing BIO_free at exit of pkcs12 application. 729855714Skris [Wu Zhigang] 729955714Skris 730055714Skris *) Fix memory leak in conf.c. 730155714Skris [Steve Henson] 730255714Skris 730355714Skris *) Updates for Win32 to assembler version of MD5. 730455714Skris [Steve Henson] 730555714Skris 730655714Skris *) Set #! path to perl in apps/der_chop to where we found it 730755714Skris instead of using a fixed path. 730855714Skris [Bodo Moeller] 730955714Skris 731055714Skris *) SHA library changes for irix64-mips4-cc. 731155714Skris [Andy Polyakov] 731255714Skris 731355714Skris *) Improvements for VMS support. 731455714Skris [Richard Levitte] 731555714Skris 731655714Skris 731755714Skris Changes between 0.9.2b and 0.9.3 [24 May 1999] 731855714Skris 731955714Skris *) Bignum library bug fix. IRIX 6 passes "make test" now! 732055714Skris This also avoids the problems with SC4.2 and unpatched SC5. 732155714Skris [Andy Polyakov <appro@fy.chalmers.se>] 732255714Skris 732355714Skris *) New functions sk_num, sk_value and sk_set to replace the previous macros. 732455714Skris These are required because of the typesafe stack would otherwise break 732555714Skris existing code. If old code used a structure member which used to be STACK 732655714Skris and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 732755714Skris sk_num or sk_value it would produce an error because the num, data members 732855714Skris are not present in STACK_OF. Now it just produces a warning. sk_set 732955714Skris replaces the old method of assigning a value to sk_value 733055714Skris (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 733155714Skris that does this will no longer work (and should use sk_set instead) but 733255714Skris this could be regarded as a "questionable" behaviour anyway. 733355714Skris [Steve Henson] 733455714Skris 733555714Skris *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 733655714Skris correctly handle encrypted S/MIME data. 733755714Skris [Steve Henson] 733855714Skris 733955714Skris *) Change type of various DES function arguments from des_cblock 734055714Skris (which means, in function argument declarations, pointer to char) 734155714Skris to des_cblock * (meaning pointer to array with 8 char elements), 734255714Skris which allows the compiler to do more typechecking; it was like 734355714Skris that back in SSLeay, but with lots of ugly casts. 734455714Skris 734555714Skris Introduce new type const_des_cblock. 734655714Skris [Bodo Moeller] 734755714Skris 734855714Skris *) Reorganise the PKCS#7 library and get rid of some of the more obvious 734955714Skris problems: find RecipientInfo structure that matches recipient certificate 735055714Skris and initialise the ASN1 structures properly based on passed cipher. 735155714Skris [Steve Henson] 735255714Skris 735355714Skris *) Belatedly make the BN tests actually check the results. 735455714Skris [Ben Laurie] 735555714Skris 735655714Skris *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 735755714Skris to and from BNs: it was completely broken. New compilation option 735855714Skris NEG_PUBKEY_BUG to allow for some broken certificates that encode public 735955714Skris key elements as negative integers. 736055714Skris [Steve Henson] 736155714Skris 736255714Skris *) Reorganize and speed up MD5. 736355714Skris [Andy Polyakov <appro@fy.chalmers.se>] 736455714Skris 736555714Skris *) VMS support. 736655714Skris [Richard Levitte <richard@levitte.org>] 736755714Skris 736855714Skris *) New option -out to asn1parse to allow the parsed structure to be 736955714Skris output to a file. This is most useful when combined with the -strparse 737055714Skris option to examine the output of things like OCTET STRINGS. 737155714Skris [Steve Henson] 737255714Skris 737355714Skris *) Make SSL library a little more fool-proof by not requiring any longer 737455714Skris that SSL_set_{accept,connect}_state be called before 737555714Skris SSL_{accept,connect} may be used (SSL_set_..._state is omitted 737655714Skris in many applications because usually everything *appeared* to work as 737755714Skris intended anyway -- now it really works as intended). 737855714Skris [Bodo Moeller] 737955714Skris 738055714Skris *) Move openssl.cnf out of lib/. 738155714Skris [Ulf M�ller] 738255714Skris 738355714Skris *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 738455714Skris -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 738555714Skris -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 738655714Skris [Ralf S. Engelschall] 738755714Skris 738855714Skris *) Various fixes to the EVP and PKCS#7 code. It may now be able to 738955714Skris handle PKCS#7 enveloped data properly. 739055714Skris [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 739155714Skris 739255714Skris *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 739355714Skris copying pointers. The cert_st handling is changed by this in 739455714Skris various ways (and thus what used to be known as ctx->default_cert 739555714Skris is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 739655714Skris any longer when s->cert does not give us what we need). 739755714Skris ssl_cert_instantiate becomes obsolete by this change. 739855714Skris As soon as we've got the new code right (possibly it already is?), 739955714Skris we have solved a couple of bugs of the earlier code where s->cert 740055714Skris was used as if it could not have been shared with other SSL structures. 740155714Skris 740255714Skris Note that using the SSL API in certain dirty ways now will result 740355714Skris in different behaviour than observed with earlier library versions: 740455714Skris Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 740555714Skris does not influence s as it used to. 740655714Skris 740755714Skris In order to clean up things more thoroughly, inside SSL_SESSION 740855714Skris we don't use CERT any longer, but a new structure SESS_CERT 740955714Skris that holds per-session data (if available); currently, this is 741055714Skris the peer's certificate chain and, for clients, the server's certificate 741155714Skris and temporary key. CERT holds only those values that can have 741255714Skris meaningful defaults in an SSL_CTX. 741355714Skris [Bodo Moeller] 741455714Skris 741555714Skris *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 741655714Skris from the internal representation. Various PKCS#7 fixes: remove some 741755714Skris evil casts and set the enc_dig_alg field properly based on the signing 741855714Skris key type. 741955714Skris [Steve Henson] 742055714Skris 742155714Skris *) Allow PKCS#12 password to be set from the command line or the 742255714Skris environment. Let 'ca' get its config file name from the environment 742355714Skris variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 742455714Skris and 'x509'). 742555714Skris [Steve Henson] 742655714Skris 742755714Skris *) Allow certificate policies extension to use an IA5STRING for the 742855714Skris organization field. This is contrary to the PKIX definition but 742955714Skris VeriSign uses it and IE5 only recognises this form. Document 'x509' 743055714Skris extension option. 743155714Skris [Steve Henson] 743255714Skris 743355714Skris *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 743455714Skris without disallowing inline assembler and the like for non-pedantic builds. 743555714Skris [Ben Laurie] 743655714Skris 743755714Skris *) Support Borland C++ builder. 743855714Skris [Janez Jere <jj@void.si>, modified by Ulf M�ller] 743955714Skris 744055714Skris *) Support Mingw32. 744155714Skris [Ulf M�ller] 744255714Skris 744355714Skris *) SHA-1 cleanups and performance enhancements. 744455714Skris [Andy Polyakov <appro@fy.chalmers.se>] 744555714Skris 744655714Skris *) Sparc v8plus assembler for the bignum library. 744755714Skris [Andy Polyakov <appro@fy.chalmers.se>] 744855714Skris 744955714Skris *) Accept any -xxx and +xxx compiler options in Configure. 745055714Skris [Ulf M�ller] 745155714Skris 745255714Skris *) Update HPUX configuration. 745355714Skris [Anonymous] 745455714Skris 745555714Skris *) Add missing sk_<type>_unshift() function to safestack.h 745655714Skris [Ralf S. Engelschall] 745755714Skris 745855714Skris *) New function SSL_CTX_use_certificate_chain_file that sets the 745955714Skris "extra_cert"s in addition to the certificate. (This makes sense 746055714Skris only for "PEM" format files, as chains as a whole are not 746155714Skris DER-encoded.) 746255714Skris [Bodo Moeller] 746355714Skris 746455714Skris *) Support verify_depth from the SSL API. 746555714Skris x509_vfy.c had what can be considered an off-by-one-error: 746655714Skris Its depth (which was not part of the external interface) 746755714Skris was actually counting the number of certificates in a chain; 746855714Skris now it really counts the depth. 746955714Skris [Bodo Moeller] 747055714Skris 747155714Skris *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 747255714Skris instead of X509err, which often resulted in confusing error 747355714Skris messages since the error codes are not globally unique 747455714Skris (e.g. an alleged error in ssl3_accept when a certificate 747555714Skris didn't match the private key). 747655714Skris 747755714Skris *) New function SSL_CTX_set_session_id_context that allows to set a default 747855714Skris value (so that you don't need SSL_set_session_id_context for each 747955714Skris connection using the SSL_CTX). 748055714Skris [Bodo Moeller] 748155714Skris 748255714Skris *) OAEP decoding bug fix. 748355714Skris [Ulf M�ller] 748455714Skris 748555714Skris *) Support INSTALL_PREFIX for package builders, as proposed by 748655714Skris David Harris. 748755714Skris [Bodo Moeller] 748855714Skris 748955714Skris *) New Configure options "threads" and "no-threads". For systems 749055714Skris where the proper compiler options are known (currently Solaris 749155714Skris and Linux), "threads" is the default. 749255714Skris [Bodo Moeller] 749355714Skris 749455714Skris *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 749555714Skris [Bodo Moeller] 749655714Skris 749755714Skris *) Install various scripts to $(OPENSSLDIR)/misc, not to 749855714Skris $(INSTALLTOP)/bin -- they shouldn't clutter directories 749955714Skris such as /usr/local/bin. 750055714Skris [Bodo Moeller] 750155714Skris 750255714Skris *) "make linux-shared" to build shared libraries. 750355714Skris [Niels Poppe <niels@netbox.org>] 750455714Skris 750555714Skris *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 750655714Skris [Ulf M�ller] 750755714Skris 750855714Skris *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 750955714Skris extension adding in x509 utility. 751055714Skris [Steve Henson] 751155714Skris 751255714Skris *) Remove NOPROTO sections and error code comments. 751355714Skris [Ulf M�ller] 751455714Skris 751555714Skris *) Partial rewrite of the DEF file generator to now parse the ANSI 751655714Skris prototypes. 751755714Skris [Steve Henson] 751855714Skris 751955714Skris *) New Configure options --prefix=DIR and --openssldir=DIR. 752055714Skris [Ulf M�ller] 752155714Skris 752255714Skris *) Complete rewrite of the error code script(s). It is all now handled 752355714Skris by one script at the top level which handles error code gathering, 752455714Skris header rewriting and C source file generation. It should be much better 752555714Skris than the old method: it now uses a modified version of Ulf's parser to 752655714Skris read the ANSI prototypes in all header files (thus the old K&R definitions 752755714Skris aren't needed for error creation any more) and do a better job of 752855714Skris translating function codes into names. The old 'ASN1 error code imbedded 752955714Skris in a comment' is no longer necessary and it doesn't use .err files which 753055714Skris have now been deleted. Also the error code call doesn't have to appear all 753155714Skris on one line (which resulted in some large lines...). 753255714Skris [Steve Henson] 753355714Skris 753455714Skris *) Change #include filenames from <foo.h> to <openssl/foo.h>. 753555714Skris [Bodo Moeller] 753655714Skris 753755714Skris *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 753855714Skris 0 (which usually indicates a closed connection), but continue reading. 753955714Skris [Bodo Moeller] 754055714Skris 754155714Skris *) Fix some race conditions. 754255714Skris [Bodo Moeller] 754355714Skris 754455714Skris *) Add support for CRL distribution points extension. Add Certificate 754555714Skris Policies and CRL distribution points documentation. 754655714Skris [Steve Henson] 754755714Skris 754855714Skris *) Move the autogenerated header file parts to crypto/opensslconf.h. 754955714Skris [Ulf M�ller] 755055714Skris 755155714Skris *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 755255714Skris 8 of keying material. Merlin has also confirmed interop with this fix 755355714Skris between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 755455714Skris [Merlin Hughes <merlin@baltimore.ie>] 755555714Skris 755655714Skris *) Fix lots of warnings. 755755714Skris [Richard Levitte <levitte@stacken.kth.se>] 755855714Skris 755955714Skris *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 756055714Skris the directory spec didn't end with a LIST_SEPARATOR_CHAR. 756155714Skris [Richard Levitte <levitte@stacken.kth.se>] 756255714Skris 756355714Skris *) Fix problems with sizeof(long) == 8. 756455714Skris [Andy Polyakov <appro@fy.chalmers.se>] 756555714Skris 756655714Skris *) Change functions to ANSI C. 756755714Skris [Ulf M�ller] 756855714Skris 756955714Skris *) Fix typos in error codes. 757055714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 757155714Skris 757255714Skris *) Remove defunct assembler files from Configure. 757355714Skris [Ulf M�ller] 757455714Skris 757555714Skris *) SPARC v8 assembler BIGNUM implementation. 757655714Skris [Andy Polyakov <appro@fy.chalmers.se>] 757755714Skris 757855714Skris *) Support for Certificate Policies extension: both print and set. 757955714Skris Various additions to support the r2i method this uses. 758055714Skris [Steve Henson] 758155714Skris 758255714Skris *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 758355714Skris return a const string when you are expecting an allocated buffer. 758455714Skris [Ben Laurie] 758555714Skris 758655714Skris *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 758755714Skris types DirectoryString and DisplayText. 758855714Skris [Steve Henson] 758955714Skris 759055714Skris *) Add code to allow r2i extensions to access the configuration database, 759155714Skris add an LHASH database driver and add several ctx helper functions. 759255714Skris [Steve Henson] 759355714Skris 759455714Skris *) Fix an evil bug in bn_expand2() which caused various BN functions to 759555714Skris fail when they extended the size of a BIGNUM. 759655714Skris [Steve Henson] 759755714Skris 759855714Skris *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 759955714Skris support typesafe stack. 760055714Skris [Steve Henson] 760155714Skris 760255714Skris *) Fix typo in SSL_[gs]et_options(). 760355714Skris [Nils Frostberg <nils@medcom.se>] 760455714Skris 760555714Skris *) Delete various functions and files that belonged to the (now obsolete) 760655714Skris old X509V3 handling code. 760755714Skris [Steve Henson] 760855714Skris 760955714Skris *) New Configure option "rsaref". 761055714Skris [Ulf M�ller] 761155714Skris 761255714Skris *) Don't auto-generate pem.h. 761355714Skris [Bodo Moeller] 761455714Skris 761555714Skris *) Introduce type-safe ASN.1 SETs. 761655714Skris [Ben Laurie] 761755714Skris 761855714Skris *) Convert various additional casted stacks to type-safe STACK_OF() variants. 761955714Skris [Ben Laurie, Ralf S. Engelschall, Steve Henson] 762055714Skris 762155714Skris *) Introduce type-safe STACKs. This will almost certainly break lots of code 762255714Skris that links with OpenSSL (well at least cause lots of warnings), but fear 762355714Skris not: the conversion is trivial, and it eliminates loads of evil casts. A 762455714Skris few STACKed things have been converted already. Feel free to convert more. 762555714Skris In the fullness of time, I'll do away with the STACK type altogether. 762655714Skris [Ben Laurie] 762755714Skris 762855714Skris *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 762955714Skris specified in <certfile> by updating the entry in the index.txt file. 763055714Skris This way one no longer has to edit the index.txt file manually for 763155714Skris revoking a certificate. The -revoke option does the gory details now. 763255714Skris [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 763355714Skris 763455714Skris *) Fix `openssl crl -noout -text' combination where `-noout' killed the 763555714Skris `-text' option at all and this way the `-noout -text' combination was 763655714Skris inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 763755714Skris [Ralf S. Engelschall] 763855714Skris 763955714Skris *) Make sure a corresponding plain text error message exists for the 764055714Skris X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 764155714Skris verify callback function determined that a certificate was revoked. 764255714Skris [Ralf S. Engelschall] 764355714Skris 764455714Skris *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 764555714Skris ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 764655714Skris all available cipers including rc5, which was forgotten until now. 764755714Skris In order to let the testing shell script know which algorithms 764855714Skris are available, a new (up to now undocumented) command 764955714Skris "openssl list-cipher-commands" is used. 765055714Skris [Bodo Moeller] 765155714Skris 765255714Skris *) Bugfix: s_client occasionally would sleep in select() when 765355714Skris it should have checked SSL_pending() first. 765455714Skris [Bodo Moeller] 765555714Skris 765655714Skris *) New functions DSA_do_sign and DSA_do_verify to provide access to 765755714Skris the raw DSA values prior to ASN.1 encoding. 765855714Skris [Ulf M�ller] 765955714Skris 766055714Skris *) Tweaks to Configure 766155714Skris [Niels Poppe <niels@netbox.org>] 766255714Skris 766355714Skris *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 766455714Skris yet... 766555714Skris [Steve Henson] 766655714Skris 766755714Skris *) New variables $(RANLIB) and $(PERL) in the Makefiles. 766855714Skris [Ulf M�ller] 766955714Skris 767055714Skris *) New config option to avoid instructions that are illegal on the 80386. 767155714Skris The default code is faster, but requires at least a 486. 767255714Skris [Ulf M�ller] 767355714Skris 767455714Skris *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 767555714Skris SSL2_SERVER_VERSION (not used at all) macros, which are now the 767655714Skris same as SSL2_VERSION anyway. 767755714Skris [Bodo Moeller] 767855714Skris 767955714Skris *) New "-showcerts" option for s_client. 768055714Skris [Bodo Moeller] 768155714Skris 768255714Skris *) Still more PKCS#12 integration. Add pkcs12 application to openssl 768355714Skris application. Various cleanups and fixes. 768455714Skris [Steve Henson] 768555714Skris 768655714Skris *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 768755714Skris modify error routines to work internally. Add error codes and PBE init 768855714Skris to library startup routines. 768955714Skris [Steve Henson] 769055714Skris 769155714Skris *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 769255714Skris packing functions to asn1 and evp. Changed function names and error 769355714Skris codes along the way. 769455714Skris [Steve Henson] 769555714Skris 769655714Skris *) PKCS12 integration: and so it begins... First of several patches to 769755714Skris slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 769855714Skris objects to objects.h 769955714Skris [Steve Henson] 770055714Skris 770155714Skris *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 770255714Skris and display support for Thawte strong extranet extension. 770355714Skris [Steve Henson] 770455714Skris 770555714Skris *) Add LinuxPPC support. 770655714Skris [Jeff Dubrule <igor@pobox.org>] 770755714Skris 770855714Skris *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 770955714Skris bn_div_words in alpha.s. 771055714Skris [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 771155714Skris 771255714Skris *) Make sure the RSA OAEP test is skipped under -DRSAref because 771355714Skris OAEP isn't supported when OpenSSL is built with RSAref. 771455714Skris [Ulf Moeller <ulf@fitug.de>] 771555714Skris 771655714Skris *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 771755714Skris so they no longer are missing under -DNOPROTO. 771855714Skris [Soren S. Jorvang <soren@t.dk>] 771955714Skris 772055714Skris 772155714Skris Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 772255714Skris 772355714Skris *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 772455714Skris doesn't work when the session is reused. Coming soon! 772555714Skris [Ben Laurie] 772655714Skris 772755714Skris *) Fix a security hole, that allows sessions to be reused in the wrong 772855714Skris context thus bypassing client cert protection! All software that uses 772955714Skris client certs and session caches in multiple contexts NEEDS PATCHING to 773055714Skris allow session reuse! A fuller solution is in the works. 773155714Skris [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 773255714Skris 773355714Skris *) Some more source tree cleanups (removed obsolete files 773455714Skris crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 773555714Skris permission on "config" script to be executable) and a fix for the INSTALL 773655714Skris document. 773755714Skris [Ulf Moeller <ulf@fitug.de>] 773855714Skris 773955714Skris *) Remove some legacy and erroneous uses of malloc, free instead of 774055714Skris Malloc, Free. 774155714Skris [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 774255714Skris 774355714Skris *) Make rsa_oaep_test return non-zero on error. 774455714Skris [Ulf Moeller <ulf@fitug.de>] 774555714Skris 774655714Skris *) Add support for native Solaris shared libraries. Configure 774755714Skris solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 774855714Skris if someone would make that last step automatic. 774955714Skris [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 775055714Skris 775155714Skris *) ctx_size was not built with the right compiler during "make links". Fixed. 775255714Skris [Ben Laurie] 775355714Skris 775455714Skris *) Change the meaning of 'ALL' in the cipher list. It now means "everything 775555714Skris except NULL ciphers". This means the default cipher list will no longer 775655714Skris enable NULL ciphers. They need to be specifically enabled e.g. with 775755714Skris the string "DEFAULT:eNULL". 775855714Skris [Steve Henson] 775955714Skris 776055714Skris *) Fix to RSA private encryption routines: if p < q then it would 776155714Skris occasionally produce an invalid result. This will only happen with 776255714Skris externally generated keys because OpenSSL (and SSLeay) ensure p > q. 776355714Skris [Steve Henson] 776455714Skris 776555714Skris *) Be less restrictive and allow also `perl util/perlpath.pl 776655714Skris /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 776755714Skris because this way one can also use an interpreter named `perl5' (which is 776855714Skris usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 776955714Skris installed as `perl'). 777055714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 777155714Skris 777255714Skris *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 777355714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 777455714Skris 777555714Skris *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 777655714Skris advapi32.lib to Win32 build and change the pem test comparision 777755714Skris to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 777855714Skris suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 777955714Skris and crypto/des/ede_cbcm_enc.c. 778055714Skris [Steve Henson] 778155714Skris 778255714Skris *) DES quad checksum was broken on big-endian architectures. Fixed. 778355714Skris [Ben Laurie] 778455714Skris 778555714Skris *) Comment out two functions in bio.h that aren't implemented. Fix up the 778655714Skris Win32 test batch file so it (might) work again. The Win32 test batch file 778755714Skris is horrible: I feel ill.... 778855714Skris [Steve Henson] 778955714Skris 779055714Skris *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 779155714Skris in e_os.h. Audit of header files to check ANSI and non ANSI 779255714Skris sections: 10 functions were absent from non ANSI section and not exported 779355714Skris from Windows DLLs. Fixed up libeay.num for new functions. 779455714Skris [Steve Henson] 779555714Skris 779655714Skris *) Make `openssl version' output lines consistent. 779755714Skris [Ralf S. Engelschall] 779855714Skris 779955714Skris *) Fix Win32 symbol export lists for BIO functions: Added 780055714Skris BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 780155714Skris to ms/libeay{16,32}.def. 780255714Skris [Ralf S. Engelschall] 780355714Skris 780455714Skris *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 780555714Skris fine under Unix and passes some trivial tests I've now added. But the 780655714Skris whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 780755714Skris added to make sure no one expects that this stuff really works in the 780855714Skris OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 780955714Skris up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 781055714Skris openssl_bio.xs. 781155714Skris [Ralf S. Engelschall] 781255714Skris 781355714Skris *) Fix the generation of two part addresses in perl. 781455714Skris [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 781555714Skris 781655714Skris *) Add config entry for Linux on MIPS. 781755714Skris [John Tobey <jtobey@channel1.com>] 781855714Skris 781955714Skris *) Make links whenever Configure is run, unless we are on Windoze. 782055714Skris [Ben Laurie] 782155714Skris 782255714Skris *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 782355714Skris Currently only issuerAltName and AuthorityKeyIdentifier make any sense 782455714Skris in CRLs. 782555714Skris [Steve Henson] 782655714Skris 782755714Skris *) Add a useful kludge to allow package maintainers to specify compiler and 782855714Skris other platforms details on the command line without having to patch the 782955714Skris Configure script everytime: One now can use ``perl Configure 783055714Skris <id>:<details>'', i.e. platform ids are allowed to have details appended 783155714Skris to them (seperated by colons). This is treated as there would be a static 783255714Skris pre-configured entry in Configure's %table under key <id> with value 783355714Skris <details> and ``perl Configure <id>'' is called. So, when you want to 783455714Skris perform a quick test-compile under FreeBSD 3.1 with pgcc and without 783555714Skris assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 783655714Skris now, which overrides the FreeBSD-elf entry on-the-fly. 783755714Skris [Ralf S. Engelschall] 783855714Skris 783955714Skris *) Disable new TLS1 ciphersuites by default: they aren't official yet. 784055714Skris [Ben Laurie] 784155714Skris 784255714Skris *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 784355714Skris on the `perl Configure ...' command line. This way one can compile 784455714Skris OpenSSL libraries with Position Independent Code (PIC) which is needed 784555714Skris for linking it into DSOs. 784655714Skris [Ralf S. Engelschall] 784755714Skris 784855714Skris *) Remarkably, export ciphers were totally broken and no-one had noticed! 784955714Skris Fixed. 785055714Skris [Ben Laurie] 785155714Skris 785255714Skris *) Cleaned up the LICENSE document: The official contact for any license 785355714Skris questions now is the OpenSSL core team under openssl-core@openssl.org. 785455714Skris And add a paragraph about the dual-license situation to make sure people 785555714Skris recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 785655714Skris to the OpenSSL toolkit. 785755714Skris [Ralf S. Engelschall] 785855714Skris 785955714Skris *) General source tree makefile cleanups: Made `making xxx in yyy...' 786055714Skris display consistent in the source tree and replaced `/bin/rm' by `rm'. 786155714Skris Additonally cleaned up the `make links' target: Remove unnecessary 786255714Skris semicolons, subsequent redundant removes, inline point.sh into mklink.sh 786355714Skris to speed processing and no longer clutter the display with confusing 786455714Skris stuff. Instead only the actually done links are displayed. 786555714Skris [Ralf S. Engelschall] 786655714Skris 786755714Skris *) Permit null encryption ciphersuites, used for authentication only. It used 786855714Skris to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 786955714Skris It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 787055714Skris encryption. 787155714Skris [Ben Laurie] 787255714Skris 787355714Skris *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 787455714Skris signed attributes when verifying signatures (this would break them), 787555714Skris the detached data encoding was wrong and public keys obtained using 787655714Skris X509_get_pubkey() weren't freed. 787755714Skris [Steve Henson] 787855714Skris 787955714Skris *) Add text documentation for the BUFFER functions. Also added a work around 788055714Skris to a Win95 console bug. This was triggered by the password read stuff: the 788155714Skris last character typed gets carried over to the next fread(). If you were 788255714Skris generating a new cert request using 'req' for example then the last 788355714Skris character of the passphrase would be CR which would then enter the first 788455714Skris field as blank. 788555714Skris [Steve Henson] 788655714Skris 788755714Skris *) Added the new `Includes OpenSSL Cryptography Software' button as 788855714Skris doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 788955714Skris button and can be used by applications based on OpenSSL to show the 789055714Skris relationship to the OpenSSL project. 789155714Skris [Ralf S. Engelschall] 789255714Skris 789355714Skris *) Remove confusing variables in function signatures in files 789455714Skris ssl/ssl_lib.c and ssl/ssl.h. 789555714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 789655714Skris 789755714Skris *) Don't install bss_file.c under PREFIX/include/ 789855714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 789955714Skris 790055714Skris *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 790155714Skris functions that return function pointers and has support for NT specific 790255714Skris stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 790355714Skris #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 790455714Skris unsigned to signed types: this was killing the Win32 compile. 790555714Skris [Steve Henson] 790655714Skris 790755714Skris *) Add new certificate file to stack functions, 790855714Skris SSL_add_dir_cert_subjects_to_stack() and 790955714Skris SSL_add_file_cert_subjects_to_stack(). These largely supplant 791055714Skris SSL_load_client_CA_file(), and can be used to add multiple certs easily 791155714Skris to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 791255714Skris This means that Apache-SSL and similar packages don't have to mess around 791355714Skris to add as many CAs as they want to the preferred list. 791455714Skris [Ben Laurie] 791555714Skris 791655714Skris *) Experiment with doxygen documentation. Currently only partially applied to 791755714Skris ssl/ssl_lib.c. 791855714Skris See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 791955714Skris openssl.doxy as the configuration file. 792055714Skris [Ben Laurie] 792155714Skris 792255714Skris *) Get rid of remaining C++-style comments which strict C compilers hate. 792355714Skris [Ralf S. Engelschall, pointed out by Carlos Amengual] 792455714Skris 792555714Skris *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 792655714Skris compiled in by default: it has problems with large keys. 792755714Skris [Steve Henson] 792855714Skris 792955714Skris *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 793055714Skris DH private keys and/or callback functions which directly correspond to 793155714Skris their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 793255714Skris is needed for applications which have to configure certificates on a 793355714Skris per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 793455714Skris (e.g. s_server). 793555714Skris For the RSA certificate situation is makes no difference, but 793655714Skris for the DSA certificate situation this fixes the "no shared cipher" 793755714Skris problem where the OpenSSL cipher selection procedure failed because the 793855714Skris temporary keys were not overtaken from the context and the API provided 793955714Skris no way to reconfigure them. 794055714Skris The new functions now let applications reconfigure the stuff and they 794155714Skris are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 794255714Skris SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 794355714Skris non-public-API function ssl_cert_instantiate() is used as a helper 794455714Skris function and also to reduce code redundancy inside ssl_rsa.c. 794555714Skris [Ralf S. Engelschall] 794655714Skris 794755714Skris *) Move s_server -dcert and -dkey options out of the undocumented feature 794855714Skris area because they are useful for the DSA situation and should be 794955714Skris recognized by the users. 795055714Skris [Ralf S. Engelschall] 795155714Skris 795255714Skris *) Fix the cipher decision scheme for export ciphers: the export bits are 795355714Skris *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 795455714Skris SSL_EXP_MASK. So, the original variable has to be used instead of the 795555714Skris already masked variable. 795655714Skris [Richard Levitte <levitte@stacken.kth.se>] 795755714Skris 795855714Skris *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 795955714Skris [Richard Levitte <levitte@stacken.kth.se>] 796055714Skris 796155714Skris *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 796255714Skris from `int' to `unsigned int' because it's a length and initialized by 796355714Skris EVP_DigestFinal() which expects an `unsigned int *'. 796455714Skris [Richard Levitte <levitte@stacken.kth.se>] 796555714Skris 796655714Skris *) Don't hard-code path to Perl interpreter on shebang line of Configure 796755714Skris script. Instead use the usual Shell->Perl transition trick. 796855714Skris [Ralf S. Engelschall] 796955714Skris 797055714Skris *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 797155714Skris (in addition to RSA certificates) to match the behaviour of `openssl dsa 797255714Skris -noout -modulus' as it's already the case for `openssl rsa -noout 797355714Skris -modulus'. For RSA the -modulus is the real "modulus" while for DSA 797455714Skris currently the public key is printed (a decision which was already done by 797555714Skris `openssl dsa -modulus' in the past) which serves a similar purpose. 797655714Skris Additionally the NO_RSA no longer completely removes the whole -modulus 797755714Skris option; it now only avoids using the RSA stuff. Same applies to NO_DSA 797855714Skris now, too. 797955714Skris [Ralf S. Engelschall] 798055714Skris 798155714Skris *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 798255714Skris BIO. See the source (crypto/evp/bio_ok.c) for more info. 798355714Skris [Arne Ansper <arne@ats.cyber.ee>] 798455714Skris 798555714Skris *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 798655714Skris to be added. Now both 'req' and 'ca' can use new objects defined in the 798755714Skris config file. 798855714Skris [Steve Henson] 798955714Skris 799055714Skris *) Add cool BIO that does syslog (or event log on NT). 799155714Skris [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 799255714Skris 799355714Skris *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 799455714Skris TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 799555714Skris TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 799655714Skris Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 799755714Skris [Ben Laurie] 799855714Skris 799955714Skris *) Add preliminary config info for new extension code. 800055714Skris [Steve Henson] 800155714Skris 800255714Skris *) Make RSA_NO_PADDING really use no padding. 800355714Skris [Ulf Moeller <ulf@fitug.de>] 800455714Skris 800555714Skris *) Generate errors when private/public key check is done. 800655714Skris [Ben Laurie] 800755714Skris 800855714Skris *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 800955714Skris for some CRL extensions and new objects added. 801055714Skris [Steve Henson] 801155714Skris 801255714Skris *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 801355714Skris key usage extension and fuller support for authority key id. 801455714Skris [Steve Henson] 801555714Skris 801655714Skris *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 801755714Skris padding method for RSA, which is recommended for new applications in PKCS 801855714Skris #1 v2.0 (RFC 2437, October 1998). 801955714Skris OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 802055714Skris foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 802155714Skris against Bleichbacher's attack on RSA. 802255714Skris [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 802355714Skris Ben Laurie] 802455714Skris 802555714Skris *) Updates to the new SSL compression code 802655714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 802755714Skris 802855714Skris *) Fix so that the version number in the master secret, when passed 802955714Skris via RSA, checks that if TLS was proposed, but we roll back to SSLv3 803055714Skris (because the server will not accept higher), that the version number 803155714Skris is 0x03,0x01, not 0x03,0x00 803255714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 803355714Skris 803455714Skris *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 803555714Skris leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 803655714Skris in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 803755714Skris [Steve Henson] 803855714Skris 803955714Skris *) Support for RAW extensions where an arbitrary extension can be 804055714Skris created by including its DER encoding. See apps/openssl.cnf for 804155714Skris an example. 804255714Skris [Steve Henson] 804355714Skris 804455714Skris *) Make sure latest Perl versions don't interpret some generated C array 804555714Skris code as Perl array code in the crypto/err/err_genc.pl script. 804655714Skris [Lars Weber <3weber@informatik.uni-hamburg.de>] 804755714Skris 804855714Skris *) Modify ms/do_ms.bat to not generate assembly language makefiles since 804955714Skris not many people have the assembler. Various Win32 compilation fixes and 805055714Skris update to the INSTALL.W32 file with (hopefully) more accurate Win32 805155714Skris build instructions. 805255714Skris [Steve Henson] 805355714Skris 805455714Skris *) Modify configure script 'Configure' to automatically create crypto/date.h 805555714Skris file under Win32 and also build pem.h from pem.org. New script 805655714Skris util/mkfiles.pl to create the MINFO file on environments that can't do a 805755714Skris 'make files': perl util/mkfiles.pl >MINFO should work. 805855714Skris [Steve Henson] 805955714Skris 806055714Skris *) Major rework of DES function declarations, in the pursuit of correctness 806155714Skris and purity. As a result, many evil casts evaporated, and some weirdness, 806255714Skris too. You may find this causes warnings in your code. Zapping your evil 806355714Skris casts will probably fix them. Mostly. 806455714Skris [Ben Laurie] 806555714Skris 806655714Skris *) Fix for a typo in asn1.h. Bug fix to object creation script 806755714Skris obj_dat.pl. It considered a zero in an object definition to mean 806855714Skris "end of object": none of the objects in objects.h have any zeros 806955714Skris so it wasn't spotted. 807055714Skris [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 807155714Skris 807255714Skris *) Add support for Triple DES Cipher Block Chaining with Output Feedback 807355714Skris Masking (CBCM). In the absence of test vectors, the best I have been able 807455714Skris to do is check that the decrypt undoes the encrypt, so far. Send me test 807555714Skris vectors if you have them. 807655714Skris [Ben Laurie] 807755714Skris 807855714Skris *) Correct calculation of key length for export ciphers (too much space was 807955714Skris allocated for null ciphers). This has not been tested! 808055714Skris [Ben Laurie] 808155714Skris 808255714Skris *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 808355714Skris message is now correct (it understands "crypto" and "ssl" on its 808455714Skris command line). There is also now an "update" option. This will update 808555714Skris the util/ssleay.num and util/libeay.num files with any new functions. 808655714Skris If you do a: 808755714Skris perl util/mkdef.pl crypto ssl update 808855714Skris it will update them. 808955714Skris [Steve Henson] 809055714Skris 809155714Skris *) Overhauled the Perl interface (perl/*): 809255714Skris - ported BN stuff to OpenSSL's different BN library 809355714Skris - made the perl/ source tree CVS-aware 809455714Skris - renamed the package from SSLeay to OpenSSL (the files still contain 809555714Skris their history because I've copied them in the repository) 809655714Skris - removed obsolete files (the test scripts will be replaced 809755714Skris by better Test::Harness variants in the future) 809855714Skris [Ralf S. Engelschall] 809955714Skris 810055714Skris *) First cut for a very conservative source tree cleanup: 810155714Skris 1. merge various obsolete readme texts into doc/ssleay.txt 810255714Skris where we collect the old documents and readme texts. 810355714Skris 2. remove the first part of files where I'm already sure that we no 810455714Skris longer need them because of three reasons: either they are just temporary 810555714Skris files which were left by Eric or they are preserved original files where 810655714Skris I've verified that the diff is also available in the CVS via "cvs diff 810755714Skris -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 810855714Skris the crypto/md/ stuff). 810955714Skris [Ralf S. Engelschall] 811055714Skris 811155714Skris *) More extension code. Incomplete support for subject and issuer alt 811255714Skris name, issuer and authority key id. Change the i2v function parameters 811355714Skris and add an extra 'crl' parameter in the X509V3_CTX structure: guess 811455714Skris what that's for :-) Fix to ASN1 macro which messed up 811555714Skris IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 811655714Skris [Steve Henson] 811755714Skris 811855714Skris *) Preliminary support for ENUMERATED type. This is largely copied from the 811955714Skris INTEGER code. 812055714Skris [Steve Henson] 812155714Skris 812255714Skris *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 812355714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 812455714Skris 812555714Skris *) Make sure `make rehash' target really finds the `openssl' program. 812655714Skris [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 812755714Skris 812855714Skris *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 812955714Skris like to hear about it if this slows down other processors. 813055714Skris [Ben Laurie] 813155714Skris 813255714Skris *) Add CygWin32 platform information to Configure script. 813355714Skris [Alan Batie <batie@aahz.jf.intel.com>] 813455714Skris 813555714Skris *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 813655714Skris [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 813755714Skris 813855714Skris *) New program nseq to manipulate netscape certificate sequences 813955714Skris [Steve Henson] 814055714Skris 814155714Skris *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 814255714Skris few typos. 814355714Skris [Steve Henson] 814455714Skris 814555714Skris *) Fixes to BN code. Previously the default was to define BN_RECURSION 814655714Skris but the BN code had some problems that would cause failures when 814755714Skris doing certificate verification and some other functions. 814855714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 814955714Skris 815055714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 815155714Skris [Steve Henson] 815255714Skris 815355714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 815455714Skris [Steve Henson] 815555714Skris 815655714Skris *) Add several PKIX and private extended key usage OIDs. 815755714Skris [Steve Henson] 815855714Skris 815955714Skris *) Modify the 'ca' program to handle the new extension code. Modify 816055714Skris openssl.cnf for new extension format, add comments. 816155714Skris [Steve Henson] 816255714Skris 816355714Skris *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 816455714Skris and add a sample to openssl.cnf so req -x509 now adds appropriate 816555714Skris CA extensions. 816655714Skris [Steve Henson] 816755714Skris 816855714Skris *) Continued X509 V3 changes. Add to other makefiles, integrate with the 816955714Skris error code, add initial support to X509_print() and x509 application. 817055714Skris [Steve Henson] 817155714Skris 817255714Skris *) Takes a deep breath and start addding X509 V3 extension support code. Add 817355714Skris files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 817455714Skris stuff is currently isolated and isn't even compiled yet. 817555714Skris [Steve Henson] 817655714Skris 817755714Skris *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 817855714Skris ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 817955714Skris Removed the versions check from X509 routines when loading extensions: 818055714Skris this allows certain broken certificates that don't set the version 818155714Skris properly to be processed. 818255714Skris [Steve Henson] 818355714Skris 818455714Skris *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 818555714Skris Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 818655714Skris can still be regenerated with "make depend". 818755714Skris [Ben Laurie] 818855714Skris 818955714Skris *) Spelling mistake in C version of CAST-128. 819055714Skris [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 819155714Skris 819255714Skris *) Changes to the error generation code. The perl script err-code.pl 819355714Skris now reads in the old error codes and retains the old numbers, only 819455714Skris adding new ones if necessary. It also only changes the .err files if new 819555714Skris codes are added. The makefiles have been modified to only insert errors 819655714Skris when needed (to avoid needlessly modifying header files). This is done 819755714Skris by only inserting errors if the .err file is newer than the auto generated 819855714Skris C file. To rebuild all the error codes from scratch (the old behaviour) 819955714Skris either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 820055714Skris or delete all the .err files. 820155714Skris [Steve Henson] 820255714Skris 820355714Skris *) CAST-128 was incorrectly implemented for short keys. The C version has 820455714Skris been fixed, but is untested. The assembler versions are also fixed, but 820555714Skris new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 820655714Skris to regenerate it if needed. 820755714Skris [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 820855714Skris Hagino <itojun@kame.net>] 820955714Skris 821055714Skris *) File was opened incorrectly in randfile.c. 821155714Skris [Ulf M�ller <ulf@fitug.de>] 821255714Skris 821355714Skris *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 821455714Skris functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 821555714Skris GeneralizedTime. ASN1_TIME is the proper type used in certificates et 821655714Skris al: it's just almost always a UTCTime. Note this patch adds new error 821755714Skris codes so do a "make errors" if there are problems. 821855714Skris [Steve Henson] 821955714Skris 822055714Skris *) Correct Linux 1 recognition in config. 822155714Skris [Ulf M�ller <ulf@fitug.de>] 822255714Skris 822355714Skris *) Remove pointless MD5 hash when using DSA keys in ca. 822455714Skris [Anonymous <nobody@replay.com>] 822555714Skris 822655714Skris *) Generate an error if given an empty string as a cert directory. Also 822755714Skris generate an error if handed NULL (previously returned 0 to indicate an 822855714Skris error, but didn't set one). 822955714Skris [Ben Laurie, reported by Anonymous <nobody@replay.com>] 823055714Skris 823155714Skris *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 823255714Skris [Ben Laurie] 823355714Skris 823455714Skris *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 823555714Skris parameters. This was causing a warning which killed off the Win32 compile. 823655714Skris [Steve Henson] 823755714Skris 823855714Skris *) Remove C++ style comments from crypto/bn/bn_local.h. 823955714Skris [Neil Costigan <neil.costigan@celocom.com>] 824055714Skris 824155714Skris *) The function OBJ_txt2nid was broken. It was supposed to return a nid 824255714Skris based on a text string, looking up short and long names and finally 824355714Skris "dot" format. The "dot" format stuff didn't work. Added new function 824455714Skris OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 824555714Skris OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 824655714Skris OID is not part of the table. 824755714Skris [Steve Henson] 824855714Skris 824955714Skris *) Add prototypes to X509 lookup/verify methods, fixing a bug in 825055714Skris X509_LOOKUP_by_alias(). 825155714Skris [Ben Laurie] 825255714Skris 825355714Skris *) Sort openssl functions by name. 825455714Skris [Ben Laurie] 825555714Skris 825655714Skris *) Get the gendsa program working (hopefully) and add it to app list. Remove 825755714Skris encryption from sample DSA keys (in case anyone is interested the password 825855714Skris was "1234"). 825955714Skris [Steve Henson] 826055714Skris 826155714Skris *) Make _all_ *_free functions accept a NULL pointer. 826255714Skris [Frans Heymans <fheymans@isaserver.be>] 826355714Skris 826455714Skris *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 826555714Skris NULL pointers. 826655714Skris [Anonymous <nobody@replay.com>] 826755714Skris 826855714Skris *) s_server should send the CAfile as acceptable CAs, not its own cert. 826955714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 827055714Skris 827155714Skris *) Don't blow it for numeric -newkey arguments to apps/req. 827255714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 827355714Skris 827455714Skris *) Temp key "for export" tests were wrong in s3_srvr.c. 827555714Skris [Anonymous <nobody@replay.com>] 827655714Skris 827755714Skris *) Add prototype for temp key callback functions 827855714Skris SSL_CTX_set_tmp_{rsa,dh}_callback(). 827955714Skris [Ben Laurie] 828055714Skris 828155714Skris *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 828255714Skris DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 828355714Skris [Steve Henson] 828455714Skris 828555714Skris *) X509_name_add_entry() freed the wrong thing after an error. 828655714Skris [Arne Ansper <arne@ats.cyber.ee>] 828755714Skris 828855714Skris *) rsa_eay.c would attempt to free a NULL context. 828955714Skris [Arne Ansper <arne@ats.cyber.ee>] 829055714Skris 829155714Skris *) BIO_s_socket() had a broken should_retry() on Windoze. 829255714Skris [Arne Ansper <arne@ats.cyber.ee>] 829355714Skris 829455714Skris *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 829555714Skris [Arne Ansper <arne@ats.cyber.ee>] 829655714Skris 829755714Skris *) Make sure the already existing X509_STORE->depth variable is initialized 829855714Skris in X509_STORE_new(), but document the fact that this variable is still 829955714Skris unused in the certificate verification process. 830055714Skris [Ralf S. Engelschall] 830155714Skris 830255714Skris *) Fix the various library and apps files to free up pkeys obtained from 830355714Skris X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 830455714Skris [Steve Henson] 830555714Skris 830655714Skris *) Fix reference counting in X509_PUBKEY_get(). This makes 830755714Skris demos/maurice/example2.c work, amongst others, probably. 830855714Skris [Steve Henson and Ben Laurie] 830955714Skris 831055714Skris *) First cut of a cleanup for apps/. First the `ssleay' program is now named 831155714Skris `openssl' and second, the shortcut symlinks for the `openssl <command>' 831255714Skris are no longer created. This way we have a single and consistent command 831355714Skris line interface `openssl <command>', similar to `cvs <command>'. 831455714Skris [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 831555714Skris 831655714Skris *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 831755714Skris BIT STRING wrapper always have zero unused bits. 831855714Skris [Steve Henson] 831955714Skris 832055714Skris *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 832155714Skris [Steve Henson] 832255714Skris 832355714Skris *) Make the top-level INSTALL documentation easier to understand. 832455714Skris [Paul Sutton] 832555714Skris 832655714Skris *) Makefiles updated to exit if an error occurs in a sub-directory 832755714Skris make (including if user presses ^C) [Paul Sutton] 832855714Skris 832955714Skris *) Make Montgomery context stuff explicit in RSA data structure. 833055714Skris [Ben Laurie] 833155714Skris 833255714Skris *) Fix build order of pem and err to allow for generated pem.h. 833355714Skris [Ben Laurie] 833455714Skris 833555714Skris *) Fix renumbering bug in X509_NAME_delete_entry(). 833655714Skris [Ben Laurie] 833755714Skris 833855714Skris *) Enhanced the err-ins.pl script so it makes the error library number 833955714Skris global and can add a library name. This is needed for external ASN1 and 834055714Skris other error libraries. 834155714Skris [Steve Henson] 834255714Skris 834355714Skris *) Fixed sk_insert which never worked properly. 834455714Skris [Steve Henson] 834555714Skris 834655714Skris *) Fix ASN1 macros so they can handle indefinite length construted 834755714Skris EXPLICIT tags. Some non standard certificates use these: they can now 834855714Skris be read in. 834955714Skris [Steve Henson] 835055714Skris 835155714Skris *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 835255714Skris into a single doc/ssleay.txt bundle. This way the information is still 835355714Skris preserved but no longer messes up this directory. Now it's new room for 835455714Skris the new set of documenation files. 835555714Skris [Ralf S. Engelschall] 835655714Skris 835755714Skris *) SETs were incorrectly DER encoded. This was a major pain, because they 835855714Skris shared code with SEQUENCEs, which aren't coded the same. This means that 835955714Skris almost everything to do with SETs or SEQUENCEs has either changed name or 836055714Skris number of arguments. 836155714Skris [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 836255714Skris 836355714Skris *) Fix test data to work with the above. 836455714Skris [Ben Laurie] 836555714Skris 836655714Skris *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 836755714Skris was already fixed by Eric for 0.9.1 it seems. 836855714Skris [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 836955714Skris 837055714Skris *) Autodetect FreeBSD3. 837155714Skris [Ben Laurie] 837255714Skris 837355714Skris *) Fix various bugs in Configure. This affects the following platforms: 837455714Skris nextstep 837555714Skris ncr-scde 837655714Skris unixware-2.0 837755714Skris unixware-2.0-pentium 837855714Skris sco5-cc. 837955714Skris [Ben Laurie] 838055714Skris 838155714Skris *) Eliminate generated files from CVS. Reorder tests to regenerate files 838255714Skris before they are needed. 838355714Skris [Ben Laurie] 838455714Skris 838555714Skris *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 838655714Skris [Ben Laurie] 838755714Skris 838855714Skris 838955714Skris Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 839055714Skris 839155714Skris *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 839255714Skris changed SSLeay to OpenSSL in version strings. 839355714Skris [Ralf S. Engelschall] 839455714Skris 839555714Skris *) Some fixups to the top-level documents. 839655714Skris [Paul Sutton] 839755714Skris 839855714Skris *) Fixed the nasty bug where rsaref.h was not found under compile-time 839955714Skris because the symlink to include/ was missing. 840055714Skris [Ralf S. Engelschall] 840155714Skris 840255714Skris *) Incorporated the popular no-RSA/DSA-only patches 840355714Skris which allow to compile a RSA-free SSLeay. 840455714Skris [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 840555714Skris 840655714Skris *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 840755714Skris when "ssleay" is still not found. 840855714Skris [Ralf S. Engelschall] 840955714Skris 841055714Skris *) Added more platforms to Configure: Cray T3E, HPUX 11, 841155714Skris [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 841255714Skris 841355714Skris *) Updated the README file. 841455714Skris [Ralf S. Engelschall] 841555714Skris 841655714Skris *) Added various .cvsignore files in the CVS repository subdirs 841755714Skris to make a "cvs update" really silent. 841855714Skris [Ralf S. Engelschall] 841955714Skris 842055714Skris *) Recompiled the error-definition header files and added 842155714Skris missing symbols to the Win32 linker tables. 842255714Skris [Ralf S. Engelschall] 842355714Skris 842455714Skris *) Cleaned up the top-level documents; 842555714Skris o new files: CHANGES and LICENSE 842655714Skris o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 842755714Skris o merged COPYRIGHT into LICENSE 842855714Skris o removed obsolete TODO file 842955714Skris o renamed MICROSOFT to INSTALL.W32 843055714Skris [Ralf S. Engelschall] 843155714Skris 843255714Skris *) Removed dummy files from the 0.9.1b source tree: 843355714Skris crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 843455714Skris crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 843555714Skris crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 843655714Skris crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 843755714Skris util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 843855714Skris [Ralf S. Engelschall] 843955714Skris 844055714Skris *) Added various platform portability fixes. 844155714Skris [Mark J. Cox] 844255714Skris 844355714Skris *) The Genesis of the OpenSSL rpject: 844455714Skris We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 844555714Skris Young and Tim J. Hudson created while they were working for C2Net until 844655714Skris summer 1998. 844755714Skris [The OpenSSL Project] 844855714Skris 844955714Skris 845055714Skris Changes between 0.9.0b and 0.9.1b [not released] 845155714Skris 845255714Skris *) Updated a few CA certificates under certs/ 845355714Skris [Eric A. Young] 845455714Skris 845555714Skris *) Changed some BIGNUM api stuff. 845655714Skris [Eric A. Young] 845755714Skris 845855714Skris *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 845955714Skris DGUX x86, Linux Alpha, etc. 846055714Skris [Eric A. Young] 846155714Skris 846255714Skris *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 846355714Skris RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 846455714Skris available). 846555714Skris [Eric A. Young] 846655714Skris 846755714Skris *) Add -strparse option to asn1pars program which parses nested 846855714Skris binary structures 846955714Skris [Dr Stephen Henson <shenson@bigfoot.com>] 847055714Skris 847155714Skris *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 847255714Skris [Eric A. Young] 847355714Skris 847455714Skris *) DSA fix for "ca" program. 847555714Skris [Eric A. Young] 847655714Skris 847755714Skris *) Added "-genkey" option to "dsaparam" program. 847855714Skris [Eric A. Young] 847955714Skris 848055714Skris *) Added RIPE MD160 (rmd160) message digest. 848155714Skris [Eric A. Young] 848255714Skris 848355714Skris *) Added -a (all) option to "ssleay version" command. 848455714Skris [Eric A. Young] 848555714Skris 848655714Skris *) Added PLATFORM define which is the id given to Configure. 848755714Skris [Eric A. Young] 848855714Skris 848955714Skris *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 849055714Skris [Eric A. Young] 849155714Skris 849255714Skris *) Extended the ASN.1 parser routines. 849355714Skris [Eric A. Young] 849455714Skris 849555714Skris *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 849655714Skris [Eric A. Young] 849755714Skris 849855714Skris *) Added a BN_CTX to the BN library. 849955714Skris [Eric A. Young] 850055714Skris 850155714Skris *) Fixed the weak key values in DES library 850255714Skris [Eric A. Young] 850355714Skris 850455714Skris *) Changed API in EVP library for cipher aliases. 850555714Skris [Eric A. Young] 850655714Skris 850755714Skris *) Added support for RC2/64bit cipher. 850855714Skris [Eric A. Young] 850955714Skris 851055714Skris *) Converted the lhash library to the crypto/mem.c functions. 851155714Skris [Eric A. Young] 851255714Skris 851355714Skris *) Added more recognized ASN.1 object ids. 851455714Skris [Eric A. Young] 851555714Skris 851655714Skris *) Added more RSA padding checks for SSL/TLS. 851755714Skris [Eric A. Young] 851855714Skris 851955714Skris *) Added BIO proxy/filter functionality. 852055714Skris [Eric A. Young] 852155714Skris 852255714Skris *) Added extra_certs to SSL_CTX which can be used 852355714Skris send extra CA certificates to the client in the CA cert chain sending 852455714Skris process. It can be configured with SSL_CTX_add_extra_chain_cert(). 852555714Skris [Eric A. Young] 852655714Skris 852755714Skris *) Now Fortezza is denied in the authentication phase because 852855714Skris this is key exchange mechanism is not supported by SSLeay at all. 852955714Skris [Eric A. Young] 853055714Skris 853155714Skris *) Additional PKCS1 checks. 853255714Skris [Eric A. Young] 853355714Skris 853455714Skris *) Support the string "TLSv1" for all TLS v1 ciphers. 853555714Skris [Eric A. Young] 853655714Skris 853755714Skris *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 853855714Skris ex_data index of the SSL context in the X509_STORE_CTX ex_data. 853955714Skris [Eric A. Young] 854055714Skris 854155714Skris *) Fixed a few memory leaks. 854255714Skris [Eric A. Young] 854355714Skris 854455714Skris *) Fixed various code and comment typos. 854555714Skris [Eric A. Young] 854655714Skris 854755714Skris *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 854855714Skris bytes sent in the client random. 854955714Skris [Edward Bishop <ebishop@spyglass.com>] 855055714Skris 8551