CHANGES revision 109998
155714Skris 255714Skris OpenSSL CHANGES 355714Skris _______________ 455714Skris 5109998Smarkm Changes between 0.9.6h and 0.9.7 [31 Dec 2002] 6109998Smarkm 7109998Smarkm *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED 8109998Smarkm code (06) was taken as the first octet of the session ID and the last 9109998Smarkm octet was ignored consequently. As a result SSLv2 client side session 10109998Smarkm caching could not have worked due to the session ID mismatch between 11109998Smarkm client and server. 12109998Smarkm Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 13109998Smarkm PR #377. 14109998Smarkm [Lutz Jaenicke] 15109998Smarkm 16109998Smarkm *) Change the declaration of needed Kerberos libraries to use EX_LIBS 17109998Smarkm instead of the special (and badly supported) LIBKRB5. LIBKRB5 is 18109998Smarkm removed entirely. 19109998Smarkm [Richard Levitte] 20109998Smarkm 21109998Smarkm *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it 22109998Smarkm seems that in spite of existing for more than a year, many application 23109998Smarkm author have done nothing to provide the necessary callbacks, which 24109998Smarkm means that this particular engine will not work properly anywhere. 25109998Smarkm This is a very unfortunate situation which forces us, in the name 26109998Smarkm of usability, to give the hw_ncipher.c a static lock, which is part 27109998Smarkm of libcrypto. 28109998Smarkm NOTE: This is for the 0.9.7 series ONLY. This hack will never 29109998Smarkm appear in 0.9.8 or later. We EXPECT application authors to have 30109998Smarkm dealt properly with this when 0.9.8 is released (unless we actually 31109998Smarkm make such changes in the libcrypto locking code that changes will 32109998Smarkm have to be made anyway). 33109998Smarkm [Richard Levitte] 34109998Smarkm 35109998Smarkm *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content 36109998Smarkm octets have been read, EOF or an error occurs. Without this change 37109998Smarkm some truncated ASN1 structures will not produce an error. 38109998Smarkm [Steve Henson] 39109998Smarkm 40109998Smarkm *) Disable Heimdal support, since it hasn't been fully implemented. 41109998Smarkm Still give the possibility to force the use of Heimdal, but with 42109998Smarkm warnings and a request that patches get sent to openssl-dev. 43109998Smarkm [Richard Levitte] 44109998Smarkm 45109998Smarkm *) Add the VC-CE target, introduce the WINCE sysname, and add 46109998Smarkm INSTALL.WCE and appropriate conditionals to make it build. 47109998Smarkm [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 48109998Smarkm 49109998Smarkm *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and 50109998Smarkm cygssl-x.y.z.dll, where x, y and z are the major, minor and 51109998Smarkm edit numbers of the version. 52109998Smarkm [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 53109998Smarkm 54109998Smarkm *) Introduce safe string copy and catenation functions 55109998Smarkm (BUF_strlcpy() and BUF_strlcat()). 56109998Smarkm [Ben Laurie (CHATS) and Richard Levitte] 57109998Smarkm 58109998Smarkm *) Avoid using fixed-size buffers for one-line DNs. 59109998Smarkm [Ben Laurie (CHATS)] 60109998Smarkm 61109998Smarkm *) Add BUF_MEM_grow_clean() to avoid information leakage when 62109998Smarkm resizing buffers containing secrets, and use where appropriate. 63109998Smarkm [Ben Laurie (CHATS)] 64109998Smarkm 65109998Smarkm *) Avoid using fixed size buffers for configuration file location. 66109998Smarkm [Ben Laurie (CHATS)] 67109998Smarkm 68109998Smarkm *) Avoid filename truncation for various CA files. 69109998Smarkm [Ben Laurie (CHATS)] 70109998Smarkm 71109998Smarkm *) Use sizeof in preference to magic numbers. 72109998Smarkm [Ben Laurie (CHATS)] 73109998Smarkm 74109998Smarkm *) Avoid filename truncation in cert requests. 75109998Smarkm [Ben Laurie (CHATS)] 76109998Smarkm 77109998Smarkm *) Add assertions to check for (supposedly impossible) buffer 78109998Smarkm overflows. 79109998Smarkm [Ben Laurie (CHATS)] 80109998Smarkm 81109998Smarkm *) Don't cache truncated DNS entries in the local cache (this could 82109998Smarkm potentially lead to a spoofing attack). 83109998Smarkm [Ben Laurie (CHATS)] 84109998Smarkm 85109998Smarkm *) Fix various buffers to be large enough for hex/decimal 86109998Smarkm representations in a platform independent manner. 87109998Smarkm [Ben Laurie (CHATS)] 88109998Smarkm 89109998Smarkm *) Add CRYPTO_realloc_clean() to avoid information leakage when 90109998Smarkm resizing buffers containing secrets, and use where appropriate. 91109998Smarkm [Ben Laurie (CHATS)] 92109998Smarkm 93109998Smarkm *) Add BIO_indent() to avoid much slightly worrying code to do 94109998Smarkm indents. 95109998Smarkm [Ben Laurie (CHATS)] 96109998Smarkm 97109998Smarkm *) Convert sprintf()/BIO_puts() to BIO_printf(). 98109998Smarkm [Ben Laurie (CHATS)] 99109998Smarkm 100109998Smarkm *) buffer_gets() could terminate with the buffer only half 101109998Smarkm full. Fixed. 102109998Smarkm [Ben Laurie (CHATS)] 103109998Smarkm 104109998Smarkm *) Add assertions to prevent user-supplied crypto functions from 105109998Smarkm overflowing internal buffers by having large block sizes, etc. 106109998Smarkm [Ben Laurie (CHATS)] 107109998Smarkm 108109998Smarkm *) New OPENSSL_assert() macro (similar to assert(), but enabled 109109998Smarkm unconditionally). 110109998Smarkm [Ben Laurie (CHATS)] 111109998Smarkm 112109998Smarkm *) Eliminate unused copy of key in RC4. 113109998Smarkm [Ben Laurie (CHATS)] 114109998Smarkm 115109998Smarkm *) Eliminate unused and incorrectly sized buffers for IV in pem.h. 116109998Smarkm [Ben Laurie (CHATS)] 117109998Smarkm 118109998Smarkm *) Fix off-by-one error in EGD path. 119109998Smarkm [Ben Laurie (CHATS)] 120109998Smarkm 121109998Smarkm *) If RANDFILE path is too long, ignore instead of truncating. 122109998Smarkm [Ben Laurie (CHATS)] 123109998Smarkm 124109998Smarkm *) Eliminate unused and incorrectly sized X.509 structure 125109998Smarkm CBCParameter. 126109998Smarkm [Ben Laurie (CHATS)] 127109998Smarkm 128109998Smarkm *) Eliminate unused and dangerous function knumber(). 129109998Smarkm [Ben Laurie (CHATS)] 130109998Smarkm 131109998Smarkm *) Eliminate unused and dangerous structure, KSSL_ERR. 132109998Smarkm [Ben Laurie (CHATS)] 133109998Smarkm 134109998Smarkm *) Protect against overlong session ID context length in an encoded 135109998Smarkm session object. Since these are local, this does not appear to be 136109998Smarkm exploitable. 137109998Smarkm [Ben Laurie (CHATS)] 138109998Smarkm 139109998Smarkm *) Change from security patch (see 0.9.6e below) that did not affect 140109998Smarkm the 0.9.6 release series: 141109998Smarkm 142109998Smarkm Remote buffer overflow in SSL3 protocol - an attacker could 143109998Smarkm supply an oversized master key in Kerberos-enabled versions. 144109998Smarkm (CAN-2002-0657) 145109998Smarkm [Ben Laurie (CHATS)] 146109998Smarkm 147109998Smarkm *) Change the SSL kerb5 codes to match RFC 2712. 148109998Smarkm [Richard Levitte] 149109998Smarkm 150109998Smarkm *) Make -nameopt work fully for req and add -reqopt switch. 151109998Smarkm [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson] 152109998Smarkm 153109998Smarkm *) The "block size" for block ciphers in CFB and OFB mode should be 1. 154109998Smarkm [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>] 155109998Smarkm 156109998Smarkm *) Make sure tests can be performed even if the corresponding algorithms 157109998Smarkm have been removed entirely. This was also the last step to make 158109998Smarkm OpenSSL compilable with DJGPP under all reasonable conditions. 159109998Smarkm [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>] 160109998Smarkm 161109998Smarkm *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT 162109998Smarkm to allow version independent disabling of normally unselected ciphers, 163109998Smarkm which may be activated as a side-effect of selecting a single cipher. 164109998Smarkm 165109998Smarkm (E.g., cipher list string "RSA" enables ciphersuites that are left 166109998Smarkm out of "ALL" because they do not provide symmetric encryption. 167109998Smarkm "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) 168109998Smarkm [Lutz Jaenicke, Bodo Moeller] 169109998Smarkm 170109998Smarkm *) Add appropriate support for separate platform-dependent build 171109998Smarkm directories. The recommended way to make a platform-dependent 172109998Smarkm build directory is the following (tested on Linux), maybe with 173109998Smarkm some local tweaks: 174109998Smarkm 175109998Smarkm # Place yourself outside of the OpenSSL source tree. In 176109998Smarkm # this example, the environment variable OPENSSL_SOURCE 177109998Smarkm # is assumed to contain the absolute OpenSSL source directory. 178109998Smarkm mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" 179109998Smarkm cd objtree/"`uname -s`-`uname -r`-`uname -m`" 180109998Smarkm (cd $OPENSSL_SOURCE; find . -type f -o -type l) | while read F; do 181109998Smarkm mkdir -p `dirname $F` 182109998Smarkm ln -s $OPENSSL_SOURCE/$F $F 183109998Smarkm done 184109998Smarkm 185109998Smarkm To be absolutely sure not to disturb the source tree, a "make clean" 186109998Smarkm is a good thing. If it isn't successfull, don't worry about it, 187109998Smarkm it probably means the source directory is very clean. 188109998Smarkm [Richard Levitte] 189109998Smarkm 190109998Smarkm *) Make sure any ENGINE control commands make local copies of string 191109998Smarkm pointers passed to them whenever necessary. Otherwise it is possible 192109998Smarkm the caller may have overwritten (or deallocated) the original string 193109998Smarkm data when a later ENGINE operation tries to use the stored values. 194109998Smarkm [G�tz Babin-Ebell <babinebell@trustcenter.de>] 195109998Smarkm 196109998Smarkm *) Improve diagnostics in file reading and command-line digests. 197109998Smarkm [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>] 198109998Smarkm 199109998Smarkm *) Add AES modes CFB and OFB to the object database. Correct an 200109998Smarkm error in AES-CFB decryption. 201109998Smarkm [Richard Levitte] 202109998Smarkm 203109998Smarkm *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this 204109998Smarkm allows existing EVP_CIPHER_CTX structures to be reused after 205109998Smarkm calling EVP_*Final(). This behaviour is used by encryption 206109998Smarkm BIOs and some applications. This has the side effect that 207109998Smarkm applications must explicitly clean up cipher contexts with 208109998Smarkm EVP_CIPHER_CTX_cleanup() or they will leak memory. 209109998Smarkm [Steve Henson] 210109998Smarkm 211109998Smarkm *) Check the values of dna and dnb in bn_mul_recursive before calling 212109998Smarkm bn_mul_comba (a non zero value means the a or b arrays do not contain 213109998Smarkm n2 elements) and fallback to bn_mul_normal if either is not zero. 214109998Smarkm [Steve Henson] 215109998Smarkm 216109998Smarkm *) Fix escaping of non-ASCII characters when using the -subj option 217109998Smarkm of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) 218109998Smarkm [Lutz Jaenicke] 219109998Smarkm 220109998Smarkm *) Make object definitions compliant to LDAP (RFC2256): SN is the short 221109998Smarkm form for "surname", serialNumber has no short form. 222109998Smarkm Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; 223109998Smarkm therefore remove "mail" short name for "internet 7". 224109998Smarkm The OID for unique identifiers in X509 certificates is 225109998Smarkm x500UniqueIdentifier, not uniqueIdentifier. 226109998Smarkm Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) 227109998Smarkm [Lutz Jaenicke] 228109998Smarkm 229109998Smarkm *) Add an "init" command to the ENGINE config module and auto initialize 230109998Smarkm ENGINEs. Without any "init" command the ENGINE will be initialized 231109998Smarkm after all ctrl commands have been executed on it. If init=1 the 232109998Smarkm ENGINE is initailized at that point (ctrls before that point are run 233109998Smarkm on the uninitialized ENGINE and after on the initialized one). If 234109998Smarkm init=0 then the ENGINE will not be iniatialized at all. 235109998Smarkm [Steve Henson] 236109998Smarkm 237109998Smarkm *) Fix the 'app_verify_callback' interface so that the user-defined 238109998Smarkm argument is actually passed to the callback: In the 239109998Smarkm SSL_CTX_set_cert_verify_callback() prototype, the callback 240109998Smarkm declaration has been changed from 241109998Smarkm int (*cb)() 242109998Smarkm into 243109998Smarkm int (*cb)(X509_STORE_CTX *,void *); 244109998Smarkm in ssl_verify_cert_chain (ssl/ssl_cert.c), the call 245109998Smarkm i=s->ctx->app_verify_callback(&ctx) 246109998Smarkm has been changed into 247109998Smarkm i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). 248109998Smarkm 249109998Smarkm To update applications using SSL_CTX_set_cert_verify_callback(), 250109998Smarkm a dummy argument can be added to their callback functions. 251109998Smarkm [D. K. Smetters <smetters@parc.xerox.com>] 252109998Smarkm 253109998Smarkm *) Added the '4758cca' ENGINE to support IBM 4758 cards. 254109998Smarkm [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe] 255109998Smarkm 256109998Smarkm *) Add and OPENSSL_LOAD_CONF define which will cause 257109998Smarkm OpenSSL_add_all_algorithms() to load the openssl.cnf config file. 258109998Smarkm This allows older applications to transparently support certain 259109998Smarkm OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. 260109998Smarkm Two new functions OPENSSL_add_all_algorithms_noconf() which will never 261109998Smarkm load the config file and OPENSSL_add_all_algorithms_conf() which will 262109998Smarkm always load it have also been added. 263109998Smarkm [Steve Henson] 264109998Smarkm 265109998Smarkm *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. 266109998Smarkm Adjust NIDs and EVP layer. 267109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 268109998Smarkm 269109998Smarkm *) Config modules support in openssl utility. 270109998Smarkm 271109998Smarkm Most commands now load modules from the config file, 272109998Smarkm though in a few (such as version) this isn't done 273109998Smarkm because it couldn't be used for anything. 274109998Smarkm 275109998Smarkm In the case of ca and req the config file used is 276109998Smarkm the same as the utility itself: that is the -config 277109998Smarkm command line option can be used to specify an 278109998Smarkm alternative file. 279109998Smarkm [Steve Henson] 280109998Smarkm 281109998Smarkm *) Move default behaviour from OPENSSL_config(). If appname is NULL 282109998Smarkm use "openssl_conf" if filename is NULL use default openssl config file. 283109998Smarkm [Steve Henson] 284109998Smarkm 285109998Smarkm *) Add an argument to OPENSSL_config() to allow the use of an alternative 286109998Smarkm config section name. Add a new flag to tolerate a missing config file 287109998Smarkm and move code to CONF_modules_load_file(). 288109998Smarkm [Steve Henson] 289109998Smarkm 290109998Smarkm *) Support for crypto accelerator cards from Accelerated Encryption 291109998Smarkm Processing, www.aep.ie. (Use engine 'aep') 292109998Smarkm The support was copied from 0.9.6c [engine] and adapted/corrected 293109998Smarkm to work with the new engine framework. 294109998Smarkm [AEP Inc. and Richard Levitte] 295109998Smarkm 296109998Smarkm *) Support for SureWare crypto accelerator cards from Baltimore 297109998Smarkm Technologies. (Use engine 'sureware') 298109998Smarkm The support was copied from 0.9.6c [engine] and adapted 299109998Smarkm to work with the new engine framework. 300109998Smarkm [Richard Levitte] 301109998Smarkm 302109998Smarkm *) Have the CHIL engine fork-safe (as defined by nCipher) and actually 303109998Smarkm make the newer ENGINE framework commands for the CHIL engine work. 304109998Smarkm [Toomas Kiisk <vix@cyber.ee> and Richard Levitte] 305109998Smarkm 306109998Smarkm *) Make it possible to produce shared libraries on ReliantUNIX. 307109998Smarkm [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte] 308109998Smarkm 309109998Smarkm *) Add the configuration target debug-linux-ppro. 310109998Smarkm Make 'openssl rsa' use the general key loading routines 311109998Smarkm implemented in apps.c, and make those routines able to 312109998Smarkm handle the key format FORMAT_NETSCAPE and the variant 313109998Smarkm FORMAT_IISSGC. 314109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 315109998Smarkm 316109998Smarkm *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 317109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 318109998Smarkm 319109998Smarkm *) Add -keyform to rsautl, and document -engine. 320109998Smarkm [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>] 321109998Smarkm 322109998Smarkm *) Change BIO_new_file (crypto/bio/bss_file.c) to use new 323109998Smarkm BIO_R_NO_SUCH_FILE error code rather than the generic 324109998Smarkm ERR_R_SYS_LIB error code if fopen() fails with ENOENT. 325109998Smarkm [Ben Laurie] 326109998Smarkm 327109998Smarkm *) Add new functions 328109998Smarkm ERR_peek_last_error 329109998Smarkm ERR_peek_last_error_line 330109998Smarkm ERR_peek_last_error_line_data. 331109998Smarkm These are similar to 332109998Smarkm ERR_peek_error 333109998Smarkm ERR_peek_error_line 334109998Smarkm ERR_peek_error_line_data, 335109998Smarkm but report on the latest error recorded rather than the first one 336109998Smarkm still in the error queue. 337109998Smarkm [Ben Laurie, Bodo Moeller] 338109998Smarkm 339109998Smarkm *) default_algorithms option in ENGINE config module. This allows things 340109998Smarkm like: 341109998Smarkm default_algorithms = ALL 342109998Smarkm default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS 343109998Smarkm [Steve Henson] 344109998Smarkm 345109998Smarkm *) Prelminary ENGINE config module. 346109998Smarkm [Steve Henson] 347109998Smarkm 348109998Smarkm *) New experimental application configuration code. 349109998Smarkm [Steve Henson] 350109998Smarkm 351109998Smarkm *) Change the AES code to follow the same name structure as all other 352109998Smarkm symmetric ciphers, and behave the same way. Move everything to 353109998Smarkm the directory crypto/aes, thereby obsoleting crypto/rijndael. 354109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 355109998Smarkm 356109998Smarkm *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. 357109998Smarkm [Ben Laurie and Theo de Raadt] 358109998Smarkm 359109998Smarkm *) Add option to output public keys in req command. 360109998Smarkm [Massimiliano Pala madwolf@openca.org] 361109998Smarkm 362109998Smarkm *) Use wNAFs in EC_POINTs_mul() for improved efficiency 363109998Smarkm (up to about 10% better than before for P-192 and P-224). 364109998Smarkm [Bodo Moeller] 365109998Smarkm 366109998Smarkm *) New functions/macros 367109998Smarkm 368109998Smarkm SSL_CTX_set_msg_callback(ctx, cb) 369109998Smarkm SSL_CTX_set_msg_callback_arg(ctx, arg) 370109998Smarkm SSL_set_msg_callback(ssl, cb) 371109998Smarkm SSL_set_msg_callback_arg(ssl, arg) 372109998Smarkm 373109998Smarkm to request calling a callback function 374109998Smarkm 375109998Smarkm void cb(int write_p, int version, int content_type, 376109998Smarkm const void *buf, size_t len, SSL *ssl, void *arg) 377109998Smarkm 378109998Smarkm whenever a protocol message has been completely received 379109998Smarkm (write_p == 0) or sent (write_p == 1). Here 'version' is the 380109998Smarkm protocol version according to which the SSL library interprets 381109998Smarkm the current protocol message (SSL2_VERSION, SSL3_VERSION, or 382109998Smarkm TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or 383109998Smarkm the content type as defined in the SSL 3.0/TLS 1.0 protocol 384109998Smarkm specification (change_cipher_spec(20), alert(21), handshake(22)). 385109998Smarkm 'buf' and 'len' point to the actual message, 'ssl' to the 386109998Smarkm SSL object, and 'arg' is the application-defined value set by 387109998Smarkm SSL[_CTX]_set_msg_callback_arg(). 388109998Smarkm 389109998Smarkm 'openssl s_client' and 'openssl s_server' have new '-msg' options 390109998Smarkm to enable a callback that displays all protocol messages. 391109998Smarkm [Bodo Moeller] 392109998Smarkm 393109998Smarkm *) Change the shared library support so shared libraries are built as 394109998Smarkm soon as the corresponding static library is finished, and thereby get 395109998Smarkm openssl and the test programs linked against the shared library. 396109998Smarkm This still only happens when the keyword "shard" has been given to 397109998Smarkm the configuration scripts. 398109998Smarkm 399109998Smarkm NOTE: shared library support is still an experimental thing, and 400109998Smarkm backward binary compatibility is still not guaranteed. 401109998Smarkm ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte] 402109998Smarkm 403109998Smarkm *) Add support for Subject Information Access extension. 404109998Smarkm [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 405109998Smarkm 406109998Smarkm *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero 407109998Smarkm additional bytes when new memory had to be allocated, not just 408109998Smarkm when reusing an existing buffer. 409109998Smarkm [Bodo Moeller] 410109998Smarkm 411109998Smarkm *) New command line and configuration option 'utf8' for the req command. 412109998Smarkm This allows field values to be specified as UTF8 strings. 413109998Smarkm [Steve Henson] 414109998Smarkm 415109998Smarkm *) Add -multi and -mr options to "openssl speed" - giving multiple parallel 416109998Smarkm runs for the former and machine-readable output for the latter. 417109998Smarkm [Ben Laurie] 418109998Smarkm 419109998Smarkm *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion 420109998Smarkm of the e-mail address in the DN (i.e., it will go into a certificate 421109998Smarkm extension only). The new configuration file option 'email_in_dn = no' 422109998Smarkm has the same effect. 423109998Smarkm [Massimiliano Pala madwolf@openca.org] 424109998Smarkm 425109998Smarkm *) Change all functions with names starting with des_ to be starting 426109998Smarkm with DES_ instead. Add wrappers that are compatible with libdes, 427109998Smarkm but are named _ossl_old_des_*. Finally, add macros that map the 428109998Smarkm des_* symbols to the corresponding _ossl_old_des_* if libdes 429109998Smarkm compatibility is desired. If OpenSSL 0.9.6c compatibility is 430109998Smarkm desired, the des_* symbols will be mapped to DES_*, with one 431109998Smarkm exception. 432109998Smarkm 433109998Smarkm Since we provide two compatibility mappings, the user needs to 434109998Smarkm define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes 435109998Smarkm compatibility is desired. The default (i.e., when that macro 436109998Smarkm isn't defined) is OpenSSL 0.9.6c compatibility. 437109998Smarkm 438109998Smarkm There are also macros that enable and disable the support of old 439109998Smarkm des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT 440109998Smarkm and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those 441109998Smarkm are defined, the default will apply: to support the old des routines. 442109998Smarkm 443109998Smarkm In either case, one must include openssl/des.h to get the correct 444109998Smarkm definitions. Do not try to just include openssl/des_old.h, that 445109998Smarkm won't work. 446109998Smarkm 447109998Smarkm NOTE: This is a major break of an old API into a new one. Software 448109998Smarkm authors are encouraged to switch to the DES_ style functions. Some 449109998Smarkm time in the future, des_old.h and the libdes compatibility functions 450109998Smarkm will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the 451109998Smarkm default), and then completely removed. 452109998Smarkm [Richard Levitte] 453109998Smarkm 454109998Smarkm *) Test for certificates which contain unsupported critical extensions. 455109998Smarkm If such a certificate is found during a verify operation it is 456109998Smarkm rejected by default: this behaviour can be overridden by either 457109998Smarkm handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or 458109998Smarkm by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function 459109998Smarkm X509_supported_extension() has also been added which returns 1 if a 460109998Smarkm particular extension is supported. 461109998Smarkm [Steve Henson] 462109998Smarkm 463109998Smarkm *) Modify the behaviour of EVP cipher functions in similar way to digests 464109998Smarkm to retain compatibility with existing code. 465109998Smarkm [Steve Henson] 466109998Smarkm 467109998Smarkm *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain 468109998Smarkm compatibility with existing code. In particular the 'ctx' parameter does 469109998Smarkm not have to be to be initialized before the call to EVP_DigestInit() and 470109998Smarkm it is tidied up after a call to EVP_DigestFinal(). New function 471109998Smarkm EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function 472109998Smarkm EVP_MD_CTX_copy() changed to not require the destination to be 473109998Smarkm initialized valid and new function EVP_MD_CTX_copy_ex() added which 474109998Smarkm requires the destination to be valid. 475109998Smarkm 476109998Smarkm Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), 477109998Smarkm EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). 478109998Smarkm [Steve Henson] 479109998Smarkm 480109998Smarkm *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it 481109998Smarkm so that complete 'Handshake' protocol structures are kept in memory 482109998Smarkm instead of overwriting 'msg_type' and 'length' with 'body' data. 483109998Smarkm [Bodo Moeller] 484109998Smarkm 485109998Smarkm *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. 486109998Smarkm [Massimo Santin via Richard Levitte] 487109998Smarkm 488109998Smarkm *) Major restructuring to the underlying ENGINE code. This includes 489109998Smarkm reduction of linker bloat, separation of pure "ENGINE" manipulation 490109998Smarkm (initialisation, etc) from functionality dealing with implementations 491109998Smarkm of specific crypto iterfaces. This change also introduces integrated 492109998Smarkm support for symmetric ciphers and digest implementations - so ENGINEs 493109998Smarkm can now accelerate these by providing EVP_CIPHER and EVP_MD 494109998Smarkm implementations of their own. This is detailed in crypto/engine/README 495109998Smarkm as it couldn't be adequately described here. However, there are a few 496109998Smarkm API changes worth noting - some RSA, DSA, DH, and RAND functions that 497109998Smarkm were changed in the original introduction of ENGINE code have now 498109998Smarkm reverted back - the hooking from this code to ENGINE is now a good 499109998Smarkm deal more passive and at run-time, operations deal directly with 500109998Smarkm RSA_METHODs, DSA_METHODs (etc) as they did before, rather than 501109998Smarkm dereferencing through an ENGINE pointer any more. Also, the ENGINE 502109998Smarkm functions dealing with BN_MOD_EXP[_CRT] handlers have been removed - 503109998Smarkm they were not being used by the framework as there is no concept of a 504109998Smarkm BIGNUM_METHOD and they could not be generalised to the new 505109998Smarkm 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, 506109998Smarkm ENGINE_cpy() has been removed as it cannot be consistently defined in 507109998Smarkm the new code. 508109998Smarkm [Geoff Thorpe] 509109998Smarkm 510109998Smarkm *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. 511109998Smarkm [Steve Henson] 512109998Smarkm 513109998Smarkm *) Change mkdef.pl to sort symbols that get the same entry number, 514109998Smarkm and make sure the automatically generated functions ERR_load_* 515109998Smarkm become part of libeay.num as well. 516109998Smarkm [Richard Levitte] 517109998Smarkm 518109998Smarkm *) New function SSL_renegotiate_pending(). This returns true once 519109998Smarkm renegotiation has been requested (either SSL_renegotiate() call 520109998Smarkm or HelloRequest/ClientHello receveived from the peer) and becomes 521109998Smarkm false once a handshake has been completed. 522109998Smarkm (For servers, SSL_renegotiate() followed by SSL_do_handshake() 523109998Smarkm sends a HelloRequest, but does not ensure that a handshake takes 524109998Smarkm place. SSL_renegotiate_pending() is useful for checking if the 525109998Smarkm client has followed the request.) 526109998Smarkm [Bodo Moeller] 527109998Smarkm 528109998Smarkm *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. 529109998Smarkm By default, clients may request session resumption even during 530109998Smarkm renegotiation (if session ID contexts permit); with this option, 531109998Smarkm session resumption is possible only in the first handshake. 532109998Smarkm 533109998Smarkm SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes 534109998Smarkm more bits available for options that should not be part of 535109998Smarkm SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). 536109998Smarkm [Bodo Moeller] 537109998Smarkm 538109998Smarkm *) Add some demos for certificate and certificate request creation. 539109998Smarkm [Steve Henson] 540109998Smarkm 541109998Smarkm *) Make maximum certificate chain size accepted from the peer application 542109998Smarkm settable (SSL*_get/set_max_cert_list()), as proposed by 543109998Smarkm "Douglas E. Engert" <deengert@anl.gov>. 544109998Smarkm [Lutz Jaenicke] 545109998Smarkm 546109998Smarkm *) Add support for shared libraries for Unixware-7 547109998Smarkm (Boyd Lynn Gerber <gerberb@zenez.com>). 548109998Smarkm [Lutz Jaenicke] 549109998Smarkm 550109998Smarkm *) Add a "destroy" handler to ENGINEs that allows structural cleanup to 551109998Smarkm be done prior to destruction. Use this to unload error strings from 552109998Smarkm ENGINEs that load their own error strings. NB: This adds two new API 553109998Smarkm functions to "get" and "set" this destroy handler in an ENGINE. 554109998Smarkm [Geoff Thorpe] 555109998Smarkm 556109998Smarkm *) Alter all existing ENGINE implementations (except "openssl" and 557109998Smarkm "openbsd") to dynamically instantiate their own error strings. This 558109998Smarkm makes them more flexible to be built both as statically-linked ENGINEs 559109998Smarkm and self-contained shared-libraries loadable via the "dynamic" ENGINE. 560109998Smarkm Also, add stub code to each that makes building them as self-contained 561109998Smarkm shared-libraries easier (see README.ENGINE). 562109998Smarkm [Geoff Thorpe] 563109998Smarkm 564109998Smarkm *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE 565109998Smarkm implementations into applications that are completely implemented in 566109998Smarkm self-contained shared-libraries. The "dynamic" ENGINE exposes control 567109998Smarkm commands that can be used to configure what shared-library to load and 568109998Smarkm to control aspects of the way it is handled. Also, made an update to 569109998Smarkm the README.ENGINE file that brings its information up-to-date and 570109998Smarkm provides some information and instructions on the "dynamic" ENGINE 571109998Smarkm (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). 572109998Smarkm [Geoff Thorpe] 573109998Smarkm 574109998Smarkm *) Make it possible to unload ranges of ERR strings with a new 575109998Smarkm "ERR_unload_strings" function. 576109998Smarkm [Geoff Thorpe] 577109998Smarkm 578109998Smarkm *) Add a copy() function to EVP_MD. 579109998Smarkm [Ben Laurie] 580109998Smarkm 581109998Smarkm *) Make EVP_MD routines take a context pointer instead of just the 582109998Smarkm md_data void pointer. 583109998Smarkm [Ben Laurie] 584109998Smarkm 585109998Smarkm *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates 586109998Smarkm that the digest can only process a single chunk of data 587109998Smarkm (typically because it is provided by a piece of 588109998Smarkm hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application 589109998Smarkm is only going to provide a single chunk of data, and hence the 590109998Smarkm framework needn't accumulate the data for oneshot drivers. 591109998Smarkm [Ben Laurie] 592109998Smarkm 593109998Smarkm *) As with "ERR", make it possible to replace the underlying "ex_data" 594109998Smarkm functions. This change also alters the storage and management of global 595109998Smarkm ex_data state - it's now all inside ex_data.c and all "class" code (eg. 596109998Smarkm RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class 597109998Smarkm index counters. The API functions that use this state have been changed 598109998Smarkm to take a "class_index" rather than pointers to the class's local STACK 599109998Smarkm and counter, and there is now an API function to dynamically create new 600109998Smarkm classes. This centralisation allows us to (a) plug a lot of the 601109998Smarkm thread-safety problems that existed, and (b) makes it possible to clean 602109998Smarkm up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) 603109998Smarkm such data would previously have always leaked in application code and 604109998Smarkm workarounds were in place to make the memory debugging turn a blind eye 605109998Smarkm to it. Application code that doesn't use this new function will still 606109998Smarkm leak as before, but their memory debugging output will announce it now 607109998Smarkm rather than letting it slide. 608109998Smarkm 609109998Smarkm Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change 610109998Smarkm induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now 611109998Smarkm has a return value to indicate success or failure. 612109998Smarkm [Geoff Thorpe] 613109998Smarkm 614109998Smarkm *) Make it possible to replace the underlying "ERR" functions such that the 615109998Smarkm global state (2 LHASH tables and 2 locks) is only used by the "default" 616109998Smarkm implementation. This change also adds two functions to "get" and "set" 617109998Smarkm the implementation prior to it being automatically set the first time 618109998Smarkm any other ERR function takes place. Ie. an application can call "get", 619109998Smarkm pass the return value to a module it has just loaded, and that module 620109998Smarkm can call its own "set" function using that value. This means the 621109998Smarkm module's "ERR" operations will use (and modify) the error state in the 622109998Smarkm application and not in its own statically linked copy of OpenSSL code. 623109998Smarkm [Geoff Thorpe] 624109998Smarkm 625109998Smarkm *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment 626109998Smarkm reference counts. This performs normal REF_PRINT/REF_CHECK macros on 627109998Smarkm the operation, and provides a more encapsulated way for external code 628109998Smarkm (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code 629109998Smarkm to use these functions rather than manually incrementing the counts. 630109998Smarkm 631109998Smarkm Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". 632109998Smarkm [Geoff Thorpe] 633109998Smarkm 634109998Smarkm *) Add EVP test program. 635109998Smarkm [Ben Laurie] 636109998Smarkm 637109998Smarkm *) Add symmetric cipher support to ENGINE. Expect the API to change! 638109998Smarkm [Ben Laurie] 639109998Smarkm 640109998Smarkm *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() 641109998Smarkm X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), 642109998Smarkm X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). 643109998Smarkm These allow a CRL to be built without having to access X509_CRL fields 644109998Smarkm directly. Modify 'ca' application to use new functions. 645109998Smarkm [Steve Henson] 646109998Smarkm 647109998Smarkm *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended 648109998Smarkm bug workarounds. Rollback attack detection is a security feature. 649109998Smarkm The problem will only arise on OpenSSL servers when TLSv1 is not 650109998Smarkm available (sslv3_server_method() or SSL_OP_NO_TLSv1). 651109998Smarkm Software authors not wanting to support TLSv1 will have special reasons 652109998Smarkm for their choice and can explicitly enable this option. 653109998Smarkm [Bodo Moeller, Lutz Jaenicke] 654109998Smarkm 655109998Smarkm *) Rationalise EVP so it can be extended: don't include a union of 656109998Smarkm cipher/digest structures, add init/cleanup functions for EVP_MD_CTX 657109998Smarkm (similar to those existing for EVP_CIPHER_CTX). 658109998Smarkm Usage example: 659109998Smarkm 660109998Smarkm EVP_MD_CTX md; 661109998Smarkm 662109998Smarkm EVP_MD_CTX_init(&md); /* new function call */ 663109998Smarkm EVP_DigestInit(&md, EVP_sha1()); 664109998Smarkm EVP_DigestUpdate(&md, in, len); 665109998Smarkm EVP_DigestFinal(&md, out, NULL); 666109998Smarkm EVP_MD_CTX_cleanup(&md); /* new function call */ 667109998Smarkm 668109998Smarkm [Ben Laurie] 669109998Smarkm 670109998Smarkm *) Make DES key schedule conform to the usual scheme, as well as 671109998Smarkm correcting its structure. This means that calls to DES functions 672109998Smarkm now have to pass a pointer to a des_key_schedule instead of a 673109998Smarkm plain des_key_schedule (which was actually always a pointer 674109998Smarkm anyway): E.g., 675109998Smarkm 676109998Smarkm des_key_schedule ks; 677109998Smarkm 678109998Smarkm des_set_key_checked(..., &ks); 679109998Smarkm des_ncbc_encrypt(..., &ks, ...); 680109998Smarkm 681109998Smarkm (Note that a later change renames 'des_...' into 'DES_...'.) 682109998Smarkm [Ben Laurie] 683109998Smarkm 684109998Smarkm *) Initial reduction of linker bloat: the use of some functions, such as 685109998Smarkm PEM causes large amounts of unused functions to be linked in due to 686109998Smarkm poor organisation. For example pem_all.c contains every PEM function 687109998Smarkm which has a knock on effect of linking in large amounts of (unused) 688109998Smarkm ASN1 code. Grouping together similar functions and splitting unrelated 689109998Smarkm functions prevents this. 690109998Smarkm [Steve Henson] 691109998Smarkm 692109998Smarkm *) Cleanup of EVP macros. 693109998Smarkm [Ben Laurie] 694109998Smarkm 695109998Smarkm *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the 696109998Smarkm correct _ecb suffix. 697109998Smarkm [Ben Laurie] 698109998Smarkm 699109998Smarkm *) Add initial OCSP responder support to ocsp application. The 700109998Smarkm revocation information is handled using the text based index 701109998Smarkm use by the ca application. The responder can either handle 702109998Smarkm requests generated internally, supplied in files (for example 703109998Smarkm via a CGI script) or using an internal minimal server. 704109998Smarkm [Steve Henson] 705109998Smarkm 706109998Smarkm *) Add configuration choices to get zlib compression for TLS. 707109998Smarkm [Richard Levitte] 708109998Smarkm 709109998Smarkm *) Changes to Kerberos SSL for RFC 2712 compliance: 710109998Smarkm 1. Implemented real KerberosWrapper, instead of just using 711109998Smarkm KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] 712109998Smarkm 2. Implemented optional authenticator field of KerberosWrapper. 713109998Smarkm 714109998Smarkm Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, 715109998Smarkm and authenticator structs; see crypto/krb5/. 716109998Smarkm 717109998Smarkm Generalized Kerberos calls to support multiple Kerberos libraries. 718109998Smarkm [Vern Staats <staatsvr@asc.hpc.mil>, 719109998Smarkm Jeffrey Altman <jaltman@columbia.edu> 720109998Smarkm via Richard Levitte] 721109998Smarkm 722109998Smarkm *) Cause 'openssl speed' to use fully hard-coded DSA keys as it 723109998Smarkm already does with RSA. testdsa.h now has 'priv_key/pub_key' 724109998Smarkm values for each of the key sizes rather than having just 725109998Smarkm parameters (and 'speed' generating keys each time). 726109998Smarkm [Geoff Thorpe] 727109998Smarkm 728109998Smarkm *) Speed up EVP routines. 729109998Smarkm Before: 730109998Smarkmencrypt 731109998Smarkmtype 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 732109998Smarkmdes-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k 733109998Smarkmdes-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k 734109998Smarkmdes-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k 735109998Smarkmdecrypt 736109998Smarkmdes-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k 737109998Smarkmdes-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k 738109998Smarkmdes-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k 739109998Smarkm After: 740109998Smarkmencrypt 741109998Smarkmdes-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k 742109998Smarkmdecrypt 743109998Smarkmdes-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k 744109998Smarkm [Ben Laurie] 745109998Smarkm 746109998Smarkm *) Added the OS2-EMX target. 747109998Smarkm ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte] 748109998Smarkm 749109998Smarkm *) Rewrite apps to use NCONF routines instead of the old CONF. New functions 750109998Smarkm to support NCONF routines in extension code. New function CONF_set_nconf() 751109998Smarkm to allow functions which take an NCONF to also handle the old LHASH 752109998Smarkm structure: this means that the old CONF compatible routines can be 753109998Smarkm retained (in particular wrt extensions) without having to duplicate the 754109998Smarkm code. New function X509V3_add_ext_nconf_sk to add extensions to a stack. 755109998Smarkm [Steve Henson] 756109998Smarkm 757109998Smarkm *) Enhance the general user interface with mechanisms for inner control 758109998Smarkm and with possibilities to have yes/no kind of prompts. 759109998Smarkm [Richard Levitte] 760109998Smarkm 761109998Smarkm *) Change all calls to low level digest routines in the library and 762109998Smarkm applications to use EVP. Add missing calls to HMAC_cleanup() and 763109998Smarkm don't assume HMAC_CTX can be copied using memcpy(). 764109998Smarkm [Verdon Walker <VWalker@novell.com>, Steve Henson] 765109998Smarkm 766109998Smarkm *) Add the possibility to control engines through control names but with 767109998Smarkm arbitrary arguments instead of just a string. 768109998Smarkm Change the key loaders to take a UI_METHOD instead of a callback 769109998Smarkm function pointer. NOTE: this breaks binary compatibility with earlier 770109998Smarkm versions of OpenSSL [engine]. 771109998Smarkm Adapt the nCipher code for these new conditions and add a card insertion 772109998Smarkm callback. 773109998Smarkm [Richard Levitte] 774109998Smarkm 775109998Smarkm *) Enhance the general user interface with mechanisms to better support 776109998Smarkm dialog box interfaces, application-defined prompts, the possibility 777109998Smarkm to use defaults (for example default passwords from somewhere else) 778109998Smarkm and interrupts/cancellations. 779109998Smarkm [Richard Levitte] 780109998Smarkm 781109998Smarkm *) Tidy up PKCS#12 attribute handling. Add support for the CSP name 782109998Smarkm attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. 783109998Smarkm [Steve Henson] 784109998Smarkm 785109998Smarkm *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also 786109998Smarkm tidy up some unnecessarily weird code in 'sk_new()'). 787109998Smarkm [Geoff, reported by Diego Tartara <dtartara@novamens.com>] 788109998Smarkm 789109998Smarkm *) Change the key loading routines for ENGINEs to use the same kind 790109998Smarkm callback (pem_password_cb) as all other routines that need this 791109998Smarkm kind of callback. 792109998Smarkm [Richard Levitte] 793109998Smarkm 794109998Smarkm *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with 795109998Smarkm 256 bit (=32 byte) keys. Of course seeding with more entropy bytes 796109998Smarkm than this minimum value is recommended. 797109998Smarkm [Lutz Jaenicke] 798109998Smarkm 799109998Smarkm *) New random seeder for OpenVMS, using the system process statistics 800109998Smarkm that are easily reachable. 801109998Smarkm [Richard Levitte] 802109998Smarkm 803109998Smarkm *) Windows apparently can't transparently handle global 804109998Smarkm variables defined in DLLs. Initialisations such as: 805109998Smarkm 806109998Smarkm const ASN1_ITEM *it = &ASN1_INTEGER_it; 807109998Smarkm 808109998Smarkm wont compile. This is used by the any applications that need to 809109998Smarkm declare their own ASN1 modules. This was fixed by adding the option 810109998Smarkm EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly 811109998Smarkm needed for static libraries under Win32. 812109998Smarkm [Steve Henson] 813109998Smarkm 814109998Smarkm *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle 815109998Smarkm setting of purpose and trust fields. New X509_STORE trust and 816109998Smarkm purpose functions and tidy up setting in other SSL functions. 817109998Smarkm [Steve Henson] 818109998Smarkm 819109998Smarkm *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE 820109998Smarkm structure. These are inherited by X509_STORE_CTX when it is 821109998Smarkm initialised. This allows various defaults to be set in the 822109998Smarkm X509_STORE structure (such as flags for CRL checking and custom 823109998Smarkm purpose or trust settings) for functions which only use X509_STORE_CTX 824109998Smarkm internally such as S/MIME. 825109998Smarkm 826109998Smarkm Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and 827109998Smarkm trust settings if they are not set in X509_STORE. This allows X509_STORE 828109998Smarkm purposes and trust (in S/MIME for example) to override any set by default. 829109998Smarkm 830109998Smarkm Add command line options for CRL checking to smime, s_client and s_server 831109998Smarkm applications. 832109998Smarkm [Steve Henson] 833109998Smarkm 834109998Smarkm *) Initial CRL based revocation checking. If the CRL checking flag(s) 835109998Smarkm are set then the CRL is looked up in the X509_STORE structure and 836109998Smarkm its validity and signature checked, then if the certificate is found 837109998Smarkm in the CRL the verify fails with a revoked error. 838109998Smarkm 839109998Smarkm Various new CRL related callbacks added to X509_STORE_CTX structure. 840109998Smarkm 841109998Smarkm Command line options added to 'verify' application to support this. 842109998Smarkm 843109998Smarkm This needs some additional work, such as being able to handle multiple 844109998Smarkm CRLs with different times, extension based lookup (rather than just 845109998Smarkm by subject name) and ultimately more complete V2 CRL extension 846109998Smarkm handling. 847109998Smarkm [Steve Henson] 848109998Smarkm 849109998Smarkm *) Add a general user interface API (crypto/ui/). This is designed 850109998Smarkm to replace things like des_read_password and friends (backward 851109998Smarkm compatibility functions using this new API are provided). 852109998Smarkm The purpose is to remove prompting functions from the DES code 853109998Smarkm section as well as provide for prompting through dialog boxes in 854109998Smarkm a window system and the like. 855109998Smarkm [Richard Levitte] 856109998Smarkm 857109998Smarkm *) Add "ex_data" support to ENGINE so implementations can add state at a 858109998Smarkm per-structure level rather than having to store it globally. 859109998Smarkm [Geoff] 860109998Smarkm 861109998Smarkm *) Make it possible for ENGINE structures to be copied when retrieved by 862109998Smarkm ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. 863109998Smarkm This causes the "original" ENGINE structure to act like a template, 864109998Smarkm analogous to the RSA vs. RSA_METHOD type of separation. Because of this 865109998Smarkm operational state can be localised to each ENGINE structure, despite the 866109998Smarkm fact they all share the same "methods". New ENGINE structures returned in 867109998Smarkm this case have no functional references and the return value is the single 868109998Smarkm structural reference. This matches the single structural reference returned 869109998Smarkm by ENGINE_by_id() normally, when it is incremented on the pre-existing 870109998Smarkm ENGINE structure. 871109998Smarkm [Geoff] 872109998Smarkm 873109998Smarkm *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this 874109998Smarkm needs to match any other type at all we need to manually clear the 875109998Smarkm tag cache. 876109998Smarkm [Steve Henson] 877109998Smarkm 878109998Smarkm *) Changes to the "openssl engine" utility to include; 879109998Smarkm - verbosity levels ('-v', '-vv', and '-vvv') that provide information 880109998Smarkm about an ENGINE's available control commands. 881109998Smarkm - executing control commands from command line arguments using the 882109998Smarkm '-pre' and '-post' switches. '-post' is only used if '-t' is 883109998Smarkm specified and the ENGINE is successfully initialised. The syntax for 884109998Smarkm the individual commands are colon-separated, for example; 885109998Smarkm openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so 886109998Smarkm [Geoff] 887109998Smarkm 888109998Smarkm *) New dynamic control command support for ENGINEs. ENGINEs can now 889109998Smarkm declare their own commands (numbers), names (strings), descriptions, 890109998Smarkm and input types for run-time discovery by calling applications. A 891109998Smarkm subset of these commands are implicitly classed as "executable" 892109998Smarkm depending on their input type, and only these can be invoked through 893109998Smarkm the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this 894109998Smarkm can be based on user input, config files, etc). The distinction is 895109998Smarkm that "executable" commands cannot return anything other than a boolean 896109998Smarkm result and can only support numeric or string input, whereas some 897109998Smarkm discoverable commands may only be for direct use through 898109998Smarkm ENGINE_ctrl(), eg. supporting the exchange of binary data, function 899109998Smarkm pointers, or other custom uses. The "executable" commands are to 900109998Smarkm support parameterisations of ENGINE behaviour that can be 901109998Smarkm unambiguously defined by ENGINEs and used consistently across any 902109998Smarkm OpenSSL-based application. Commands have been added to all the 903109998Smarkm existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow 904109998Smarkm control over shared-library paths without source code alterations. 905109998Smarkm [Geoff] 906109998Smarkm 907109998Smarkm *) Changed all ENGINE implementations to dynamically allocate their 908109998Smarkm ENGINEs rather than declaring them statically. Apart from this being 909109998Smarkm necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, 910109998Smarkm this also allows the implementations to compile without using the 911109998Smarkm internal engine_int.h header. 912109998Smarkm [Geoff] 913109998Smarkm 914109998Smarkm *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a 915109998Smarkm 'const' value. Any code that should be able to modify a RAND_METHOD 916109998Smarkm should already have non-const pointers to it (ie. they should only 917109998Smarkm modify their own ones). 918109998Smarkm [Geoff] 919109998Smarkm 920109998Smarkm *) Made a variety of little tweaks to the ENGINE code. 921109998Smarkm - "atalla" and "ubsec" string definitions were moved from header files 922109998Smarkm to C code. "nuron" string definitions were placed in variables 923109998Smarkm rather than hard-coded - allowing parameterisation of these values 924109998Smarkm later on via ctrl() commands. 925109998Smarkm - Removed unused "#if 0"'d code. 926109998Smarkm - Fixed engine list iteration code so it uses ENGINE_free() to release 927109998Smarkm structural references. 928109998Smarkm - Constified the RAND_METHOD element of ENGINE structures. 929109998Smarkm - Constified various get/set functions as appropriate and added 930109998Smarkm missing functions (including a catch-all ENGINE_cpy that duplicates 931109998Smarkm all ENGINE values onto a new ENGINE except reference counts/state). 932109998Smarkm - Removed NULL parameter checks in get/set functions. Setting a method 933109998Smarkm or function to NULL is a way of cancelling out a previously set 934109998Smarkm value. Passing a NULL ENGINE parameter is just plain stupid anyway 935109998Smarkm and doesn't justify the extra error symbols and code. 936109998Smarkm - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for 937109998Smarkm flags from engine_int.h to engine.h. 938109998Smarkm - Changed prototypes for ENGINE handler functions (init(), finish(), 939109998Smarkm ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. 940109998Smarkm [Geoff] 941109998Smarkm 942109998Smarkm *) Implement binary inversion algorithm for BN_mod_inverse in addition 943109998Smarkm to the algorithm using long division. The binary algorithm can be 944109998Smarkm used only if the modulus is odd. On 32-bit systems, it is faster 945109998Smarkm only for relatively small moduli (roughly 20-30% for 128-bit moduli, 946109998Smarkm roughly 5-15% for 256-bit moduli), so we use it only for moduli 947109998Smarkm up to 450 bits. In 64-bit environments, the binary algorithm 948109998Smarkm appears to be advantageous for much longer moduli; here we use it 949109998Smarkm for moduli up to 2048 bits. 950109998Smarkm [Bodo Moeller] 951109998Smarkm 952109998Smarkm *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code 953109998Smarkm could not support the combine flag in choice fields. 954109998Smarkm [Steve Henson] 955109998Smarkm 956109998Smarkm *) Add a 'copy_extensions' option to the 'ca' utility. This copies 957109998Smarkm extensions from a certificate request to the certificate. 958109998Smarkm [Steve Henson] 959109998Smarkm 960109998Smarkm *) Allow multiple 'certopt' and 'nameopt' options to be separated 961109998Smarkm by commas. Add 'namopt' and 'certopt' options to the 'ca' config 962109998Smarkm file: this allows the display of the certificate about to be 963109998Smarkm signed to be customised, to allow certain fields to be included 964109998Smarkm or excluded and extension details. The old system didn't display 965109998Smarkm multicharacter strings properly, omitted fields not in the policy 966109998Smarkm and couldn't display additional details such as extensions. 967109998Smarkm [Steve Henson] 968109998Smarkm 969109998Smarkm *) Function EC_POINTs_mul for multiple scalar multiplication 970109998Smarkm of an arbitrary number of elliptic curve points 971109998Smarkm \sum scalars[i]*points[i], 972109998Smarkm optionally including the generator defined for the EC_GROUP: 973109998Smarkm scalar*generator + \sum scalars[i]*points[i]. 974109998Smarkm 975109998Smarkm EC_POINT_mul is a simple wrapper function for the typical case 976109998Smarkm that the point list has just one item (besides the optional 977109998Smarkm generator). 978109998Smarkm [Bodo Moeller] 979109998Smarkm 980109998Smarkm *) First EC_METHODs for curves over GF(p): 981109998Smarkm 982109998Smarkm EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr 983109998Smarkm operations and provides various method functions that can also 984109998Smarkm operate with faster implementations of modular arithmetic. 985109998Smarkm 986109998Smarkm EC_GFp_mont_method() reuses most functions that are part of 987109998Smarkm EC_GFp_simple_method, but uses Montgomery arithmetic. 988109998Smarkm 989109998Smarkm [Bodo Moeller; point addition and point doubling 990109998Smarkm implementation directly derived from source code provided by 991109998Smarkm Lenka Fibikova <fibikova@exp-math.uni-essen.de>] 992109998Smarkm 993109998Smarkm *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, 994109998Smarkm crypto/ec/ec_lib.c): 995109998Smarkm 996109998Smarkm Curves are EC_GROUP objects (with an optional group generator) 997109998Smarkm based on EC_METHODs that are built into the library. 998109998Smarkm 999109998Smarkm Points are EC_POINT objects based on EC_GROUP objects. 1000109998Smarkm 1001109998Smarkm Most of the framework would be able to handle curves over arbitrary 1002109998Smarkm finite fields, but as there are no obvious types for fields other 1003109998Smarkm than GF(p), some functions are limited to that for now. 1004109998Smarkm [Bodo Moeller] 1005109998Smarkm 1006109998Smarkm *) Add the -HTTP option to s_server. It is similar to -WWW, but requires 1007109998Smarkm that the file contains a complete HTTP response. 1008109998Smarkm [Richard Levitte] 1009109998Smarkm 1010109998Smarkm *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl 1011109998Smarkm change the def and num file printf format specifier from "%-40sXXX" 1012109998Smarkm to "%-39s XXX". The latter will always guarantee a space after the 1013109998Smarkm field while the former will cause them to run together if the field 1014109998Smarkm is 40 of more characters long. 1015109998Smarkm [Steve Henson] 1016109998Smarkm 1017109998Smarkm *) Constify the cipher and digest 'method' functions and structures 1018109998Smarkm and modify related functions to take constant EVP_MD and EVP_CIPHER 1019109998Smarkm pointers. 1020109998Smarkm [Steve Henson] 1021109998Smarkm 1022109998Smarkm *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them 1023109998Smarkm in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. 1024109998Smarkm [Bodo Moeller] 1025109998Smarkm 1026109998Smarkm *) Modify EVP_Digest*() routines so they now return values. Although the 1027109998Smarkm internal software routines can never fail additional hardware versions 1028109998Smarkm might. 1029109998Smarkm [Steve Henson] 1030109998Smarkm 1031109998Smarkm *) Clean up crypto/err/err.h and change some error codes to avoid conflicts: 1032109998Smarkm 1033109998Smarkm Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 1034109998Smarkm (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. 1035109998Smarkm 1036109998Smarkm ASN1 error codes 1037109998Smarkm ERR_R_NESTED_ASN1_ERROR 1038109998Smarkm ... 1039109998Smarkm ERR_R_MISSING_ASN1_EOS 1040109998Smarkm were 4 .. 9, conflicting with 1041109998Smarkm ERR_LIB_RSA (= ERR_R_RSA_LIB) 1042109998Smarkm ... 1043109998Smarkm ERR_LIB_PEM (= ERR_R_PEM_LIB). 1044109998Smarkm They are now 58 .. 63 (i.e., just below ERR_R_FATAL). 1045109998Smarkm 1046109998Smarkm Add new error code 'ERR_R_INTERNAL_ERROR'. 1047109998Smarkm [Bodo Moeller] 1048109998Smarkm 1049109998Smarkm *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock 1050109998Smarkm suffices. 1051109998Smarkm [Bodo Moeller] 1052109998Smarkm 1053109998Smarkm *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This 1054109998Smarkm sets the subject name for a new request or supersedes the 1055109998Smarkm subject name in a given request. Formats that can be parsed are 1056109998Smarkm 'CN=Some Name, OU=myOU, C=IT' 1057109998Smarkm and 1058109998Smarkm 'CN=Some Name/OU=myOU/C=IT'. 1059109998Smarkm 1060109998Smarkm Add options '-batch' and '-verbose' to 'openssl req'. 1061109998Smarkm [Massimiliano Pala <madwolf@hackmasters.net>] 1062109998Smarkm 1063109998Smarkm *) Introduce the possibility to access global variables through 1064109998Smarkm functions on platform were that's the best way to handle exporting 1065109998Smarkm global variables in shared libraries. To enable this functionality, 1066109998Smarkm one must configure with "EXPORT_VAR_AS_FN" or defined the C macro 1067109998Smarkm "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter 1068109998Smarkm is normally done by Configure or something similar). 1069109998Smarkm 1070109998Smarkm To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL 1071109998Smarkm in the source file (foo.c) like this: 1072109998Smarkm 1073109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; 1074109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(double,bar); 1075109998Smarkm 1076109998Smarkm To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL 1077109998Smarkm and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: 1078109998Smarkm 1079109998Smarkm OPENSSL_DECLARE_GLOBAL(int,foo); 1080109998Smarkm #define foo OPENSSL_GLOBAL_REF(foo) 1081109998Smarkm OPENSSL_DECLARE_GLOBAL(double,bar); 1082109998Smarkm #define bar OPENSSL_GLOBAL_REF(bar) 1083109998Smarkm 1084109998Smarkm The #defines are very important, and therefore so is including the 1085109998Smarkm header file everywhere where the defined globals are used. 1086109998Smarkm 1087109998Smarkm The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition 1088109998Smarkm of ASN.1 items, but that structure is a bit different. 1089109998Smarkm 1090109998Smarkm The largest change is in util/mkdef.pl which has been enhanced with 1091109998Smarkm better and easier to understand logic to choose which symbols should 1092109998Smarkm go into the Windows .def files as well as a number of fixes and code 1093109998Smarkm cleanup (among others, algorithm keywords are now sorted 1094109998Smarkm lexicographically to avoid constant rewrites). 1095109998Smarkm [Richard Levitte] 1096109998Smarkm 1097109998Smarkm *) In BN_div() keep a copy of the sign of 'num' before writing the 1098109998Smarkm result to 'rm' because if rm==num the value will be overwritten 1099109998Smarkm and produce the wrong result if 'num' is negative: this caused 1100109998Smarkm problems with BN_mod() and BN_nnmod(). 1101109998Smarkm [Steve Henson] 1102109998Smarkm 1103109998Smarkm *) Function OCSP_request_verify(). This checks the signature on an 1104109998Smarkm OCSP request and verifies the signer certificate. The signer 1105109998Smarkm certificate is just checked for a generic purpose and OCSP request 1106109998Smarkm trust settings. 1107109998Smarkm [Steve Henson] 1108109998Smarkm 1109109998Smarkm *) Add OCSP_check_validity() function to check the validity of OCSP 1110109998Smarkm responses. OCSP responses are prepared in real time and may only 1111109998Smarkm be a few seconds old. Simply checking that the current time lies 1112109998Smarkm between thisUpdate and nextUpdate max reject otherwise valid responses 1113109998Smarkm caused by either OCSP responder or client clock inaccuracy. Instead 1114109998Smarkm we allow thisUpdate and nextUpdate to fall within a certain period of 1115109998Smarkm the current time. The age of the response can also optionally be 1116109998Smarkm checked. Two new options -validity_period and -status_age added to 1117109998Smarkm ocsp utility. 1118109998Smarkm [Steve Henson] 1119109998Smarkm 1120109998Smarkm *) If signature or public key algorithm is unrecognized print out its 1121109998Smarkm OID rather that just UNKNOWN. 1122109998Smarkm [Steve Henson] 1123109998Smarkm 1124109998Smarkm *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and 1125109998Smarkm OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate 1126109998Smarkm ID to be generated from the issuer certificate alone which can then be 1127109998Smarkm passed to OCSP_id_issuer_cmp(). 1128109998Smarkm [Steve Henson] 1129109998Smarkm 1130109998Smarkm *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new 1131109998Smarkm ASN1 modules to export functions returning ASN1_ITEM pointers 1132109998Smarkm instead of the ASN1_ITEM structures themselves. This adds several 1133109998Smarkm new macros which allow the underlying ASN1 function/structure to 1134109998Smarkm be accessed transparently. As a result code should not use ASN1_ITEM 1135109998Smarkm references directly (such as &X509_it) but instead use the relevant 1136109998Smarkm macros (such as ASN1_ITEM_rptr(X509)). This option is to allow 1137109998Smarkm use of the new ASN1 code on platforms where exporting structures 1138109998Smarkm is problematical (for example in shared libraries) but exporting 1139109998Smarkm functions returning pointers to structures is not. 1140109998Smarkm [Steve Henson] 1141109998Smarkm 1142109998Smarkm *) Add support for overriding the generation of SSL/TLS session IDs. 1143109998Smarkm These callbacks can be registered either in an SSL_CTX or per SSL. 1144109998Smarkm The purpose of this is to allow applications to control, if they wish, 1145109998Smarkm the arbitrary values chosen for use as session IDs, particularly as it 1146109998Smarkm can be useful for session caching in multiple-server environments. A 1147109998Smarkm command-line switch for testing this (and any client code that wishes 1148109998Smarkm to use such a feature) has been added to "s_server". 1149109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 1150109998Smarkm 1151109998Smarkm *) Modify mkdef.pl to recognise and parse preprocessor conditionals 1152109998Smarkm of the form '#if defined(...) || defined(...) || ...' and 1153109998Smarkm '#if !defined(...) && !defined(...) && ...'. This also avoids 1154109998Smarkm the growing number of special cases it was previously handling. 1155109998Smarkm [Richard Levitte] 1156109998Smarkm 1157109998Smarkm *) Make all configuration macros available for application by making 1158109998Smarkm sure they are available in opensslconf.h, by giving them names starting 1159109998Smarkm with "OPENSSL_" to avoid conflicts with other packages and by making 1160109998Smarkm sure e_os2.h will cover all platform-specific cases together with 1161109998Smarkm opensslconf.h. 1162109998Smarkm Additionally, it is now possible to define configuration/platform- 1163109998Smarkm specific names (called "system identities"). In the C code, these 1164109998Smarkm are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another 1165109998Smarkm macro with the name beginning with "OPENSSL_SYS_", which is determined 1166109998Smarkm from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on 1167109998Smarkm what is available. 1168109998Smarkm [Richard Levitte] 1169109998Smarkm 1170109998Smarkm *) New option -set_serial to 'req' and 'x509' this allows the serial 1171109998Smarkm number to use to be specified on the command line. Previously self 1172109998Smarkm signed certificates were hard coded with serial number 0 and the 1173109998Smarkm CA options of 'x509' had to use a serial number in a file which was 1174109998Smarkm auto incremented. 1175109998Smarkm [Steve Henson] 1176109998Smarkm 1177109998Smarkm *) New options to 'ca' utility to support V2 CRL entry extensions. 1178109998Smarkm Currently CRL reason, invalidity date and hold instruction are 1179109998Smarkm supported. Add new CRL extensions to V3 code and some new objects. 1180109998Smarkm [Steve Henson] 1181109998Smarkm 1182109998Smarkm *) New function EVP_CIPHER_CTX_set_padding() this is used to 1183109998Smarkm disable standard block padding (aka PKCS#5 padding) in the EVP 1184109998Smarkm API, which was previously mandatory. This means that the data is 1185109998Smarkm not padded in any way and so the total length much be a multiple 1186109998Smarkm of the block size, otherwise an error occurs. 1187109998Smarkm [Steve Henson] 1188109998Smarkm 1189109998Smarkm *) Initial (incomplete) OCSP SSL support. 1190109998Smarkm [Steve Henson] 1191109998Smarkm 1192109998Smarkm *) New function OCSP_parse_url(). This splits up a URL into its host, 1193109998Smarkm port and path components: primarily to parse OCSP URLs. New -url 1194109998Smarkm option to ocsp utility. 1195109998Smarkm [Steve Henson] 1196109998Smarkm 1197109998Smarkm *) New nonce behavior. The return value of OCSP_check_nonce() now 1198109998Smarkm reflects the various checks performed. Applications can decide 1199109998Smarkm whether to tolerate certain situations such as an absent nonce 1200109998Smarkm in a response when one was present in a request: the ocsp application 1201109998Smarkm just prints out a warning. New function OCSP_add1_basic_nonce() 1202109998Smarkm this is to allow responders to include a nonce in a response even if 1203109998Smarkm the request is nonce-less. 1204109998Smarkm [Steve Henson] 1205109998Smarkm 1206109998Smarkm *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are 1207109998Smarkm skipped when using openssl x509 multiple times on a single input file, 1208109998Smarkm e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs". 1209109998Smarkm [Bodo Moeller] 1210109998Smarkm 1211109998Smarkm *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() 1212109998Smarkm set string type: to handle setting ASN1_TIME structures. Fix ca 1213109998Smarkm utility to correctly initialize revocation date of CRLs. 1214109998Smarkm [Steve Henson] 1215109998Smarkm 1216109998Smarkm *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override 1217109998Smarkm the clients preferred ciphersuites and rather use its own preferences. 1218109998Smarkm Should help to work around M$ SGC (Server Gated Cryptography) bug in 1219109998Smarkm Internet Explorer by ensuring unchanged hash method during stepup. 1220109998Smarkm (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) 1221109998Smarkm [Lutz Jaenicke] 1222109998Smarkm 1223109998Smarkm *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael 1224109998Smarkm to aes and add a new 'exist' option to print out symbols that don't 1225109998Smarkm appear to exist. 1226109998Smarkm [Steve Henson] 1227109998Smarkm 1228109998Smarkm *) Additional options to ocsp utility to allow flags to be set and 1229109998Smarkm additional certificates supplied. 1230109998Smarkm [Steve Henson] 1231109998Smarkm 1232109998Smarkm *) Add the option -VAfile to 'openssl ocsp', so the user can give the 1233109998Smarkm OCSP client a number of certificate to only verify the response 1234109998Smarkm signature against. 1235109998Smarkm [Richard Levitte] 1236109998Smarkm 1237109998Smarkm *) Update Rijndael code to version 3.0 and change EVP AES ciphers to 1238109998Smarkm handle the new API. Currently only ECB, CBC modes supported. Add new 1239109998Smarkm AES OIDs. 1240109998Smarkm 1241109998Smarkm Add TLS AES ciphersuites as described in RFC3268, "Advanced 1242109998Smarkm Encryption Standard (AES) Ciphersuites for Transport Layer 1243109998Smarkm Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were 1244109998Smarkm not enabled by default and were not part of the "ALL" ciphersuite 1245109998Smarkm alias because they were not yet official; they could be 1246109998Smarkm explicitly requested by specifying the "AESdraft" ciphersuite 1247109998Smarkm group alias. In the final release of OpenSSL 0.9.7, the group 1248109998Smarkm alias is called "AES" and is part of "ALL".) 1249109998Smarkm [Ben Laurie, Steve Henson, Bodo Moeller] 1250109998Smarkm 1251109998Smarkm *) New function OCSP_copy_nonce() to copy nonce value (if present) from 1252109998Smarkm request to response. 1253109998Smarkm [Steve Henson] 1254109998Smarkm 1255109998Smarkm *) Functions for OCSP responders. OCSP_request_onereq_count(), 1256109998Smarkm OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() 1257109998Smarkm extract information from a certificate request. OCSP_response_create() 1258109998Smarkm creates a response and optionally adds a basic response structure. 1259109998Smarkm OCSP_basic_add1_status() adds a complete single response to a basic 1260109998Smarkm response and returns the OCSP_SINGLERESP structure just added (to allow 1261109998Smarkm extensions to be included for example). OCSP_basic_add1_cert() adds a 1262109998Smarkm certificate to a basic response and OCSP_basic_sign() signs a basic 1263109998Smarkm response with various flags. New helper functions ASN1_TIME_check() 1264109998Smarkm (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() 1265109998Smarkm (converts ASN1_TIME to GeneralizedTime). 1266109998Smarkm [Steve Henson] 1267109998Smarkm 1268109998Smarkm *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() 1269109998Smarkm in a single operation. X509_get0_pubkey_bitstr() extracts the public_key 1270109998Smarkm structure from a certificate. X509_pubkey_digest() digests the public_key 1271109998Smarkm contents: this is used in various key identifiers. 1272109998Smarkm [Steve Henson] 1273109998Smarkm 1274109998Smarkm *) Make sk_sort() tolerate a NULL argument. 1275109998Smarkm [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>] 1276109998Smarkm 1277109998Smarkm *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates 1278109998Smarkm passed by the function are trusted implicitly. If any of them signed the 1279109998Smarkm response then it is assumed to be valid and is not verified. 1280109998Smarkm [Steve Henson] 1281109998Smarkm 1282109998Smarkm *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT 1283109998Smarkm to data. This was previously part of the PKCS7 ASN1 code. This 1284109998Smarkm was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. 1285109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 1286109998Smarkm <support@securenetterm.com>] 1287109998Smarkm 1288109998Smarkm *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 1289109998Smarkm routines: without these tracing memory leaks is very painful. 1290109998Smarkm Fix leaks in PKCS12 and PKCS7 routines. 1291109998Smarkm [Steve Henson] 1292109998Smarkm 1293109998Smarkm *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). 1294109998Smarkm Previously it initialised the 'type' argument to V_ASN1_UTCTIME which 1295109998Smarkm effectively meant GeneralizedTime would never be used. Now it 1296109998Smarkm is initialised to -1 but X509_time_adj() now has to check the value 1297109998Smarkm and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or 1298109998Smarkm V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. 1299109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 1300109998Smarkm <support@securenetterm.com>] 1301109998Smarkm 1302109998Smarkm *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously 1303109998Smarkm result in a zero length in the ASN1_INTEGER structure which was 1304109998Smarkm not consistent with the structure when d2i_ASN1_INTEGER() was used 1305109998Smarkm and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() 1306109998Smarkm to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() 1307109998Smarkm where it did not print out a minus for negative ASN1_INTEGER. 1308109998Smarkm [Steve Henson] 1309109998Smarkm 1310109998Smarkm *) Add summary printout to ocsp utility. The various functions which 1311109998Smarkm convert status values to strings have been renamed to: 1312109998Smarkm OCSP_response_status_str(), OCSP_cert_status_str() and 1313109998Smarkm OCSP_crl_reason_str() and are no longer static. New options 1314109998Smarkm to verify nonce values and to disable verification. OCSP response 1315109998Smarkm printout format cleaned up. 1316109998Smarkm [Steve Henson] 1317109998Smarkm 1318109998Smarkm *) Add additional OCSP certificate checks. These are those specified 1319109998Smarkm in RFC2560. This consists of two separate checks: the CA of the 1320109998Smarkm certificate being checked must either be the OCSP signer certificate 1321109998Smarkm or the issuer of the OCSP signer certificate. In the latter case the 1322109998Smarkm OCSP signer certificate must contain the OCSP signing extended key 1323109998Smarkm usage. This check is performed by attempting to match the OCSP 1324109998Smarkm signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash 1325109998Smarkm in the OCSP_CERTID structures of the response. 1326109998Smarkm [Steve Henson] 1327109998Smarkm 1328109998Smarkm *) Initial OCSP certificate verification added to OCSP_basic_verify() 1329109998Smarkm and related routines. This uses the standard OpenSSL certificate 1330109998Smarkm verify routines to perform initial checks (just CA validity) and 1331109998Smarkm to obtain the certificate chain. Then additional checks will be 1332109998Smarkm performed on the chain. Currently the root CA is checked to see 1333109998Smarkm if it is explicitly trusted for OCSP signing. This is used to set 1334109998Smarkm a root CA as a global signing root: that is any certificate that 1335109998Smarkm chains to that CA is an acceptable OCSP signing certificate. 1336109998Smarkm [Steve Henson] 1337109998Smarkm 1338109998Smarkm *) New '-extfile ...' option to 'openssl ca' for reading X.509v3 1339109998Smarkm extensions from a separate configuration file. 1340109998Smarkm As when reading extensions from the main configuration file, 1341109998Smarkm the '-extensions ...' option may be used for specifying the 1342109998Smarkm section to use. 1343109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 1344109998Smarkm 1345109998Smarkm *) New OCSP utility. Allows OCSP requests to be generated or 1346109998Smarkm read. The request can be sent to a responder and the output 1347109998Smarkm parsed, outputed or printed in text form. Not complete yet: 1348109998Smarkm still needs to check the OCSP response validity. 1349109998Smarkm [Steve Henson] 1350109998Smarkm 1351109998Smarkm *) New subcommands for 'openssl ca': 1352109998Smarkm 'openssl ca -status <serial>' prints the status of the cert with 1353109998Smarkm the given serial number (according to the index file). 1354109998Smarkm 'openssl ca -updatedb' updates the expiry status of certificates 1355109998Smarkm in the index file. 1356109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 1357109998Smarkm 1358109998Smarkm *) New '-newreq-nodes' command option to CA.pl. This is like 1359109998Smarkm '-newreq', but calls 'openssl req' with the '-nodes' option 1360109998Smarkm so that the resulting key is not encrypted. 1361109998Smarkm [Damien Miller <djm@mindrot.org>] 1362109998Smarkm 1363109998Smarkm *) New configuration for the GNU Hurd. 1364109998Smarkm [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte] 1365109998Smarkm 1366109998Smarkm *) Initial code to implement OCSP basic response verify. This 1367109998Smarkm is currently incomplete. Currently just finds the signer's 1368109998Smarkm certificate and verifies the signature on the response. 1369109998Smarkm [Steve Henson] 1370109998Smarkm 1371109998Smarkm *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in 1372109998Smarkm value of OPENSSLDIR. This is available via the new '-d' option 1373109998Smarkm to 'openssl version', and is also included in 'openssl version -a'. 1374109998Smarkm [Bodo Moeller] 1375109998Smarkm 1376109998Smarkm *) Allowing defining memory allocation callbacks that will be given 1377109998Smarkm file name and line number information in additional arguments 1378109998Smarkm (a const char* and an int). The basic functionality remains, as 1379109998Smarkm well as the original possibility to just replace malloc(), 1380109998Smarkm realloc() and free() by functions that do not know about these 1381109998Smarkm additional arguments. To register and find out the current 1382109998Smarkm settings for extended allocation functions, the following 1383109998Smarkm functions are provided: 1384109998Smarkm 1385109998Smarkm CRYPTO_set_mem_ex_functions 1386109998Smarkm CRYPTO_set_locked_mem_ex_functions 1387109998Smarkm CRYPTO_get_mem_ex_functions 1388109998Smarkm CRYPTO_get_locked_mem_ex_functions 1389109998Smarkm 1390109998Smarkm These work the same way as CRYPTO_set_mem_functions and friends. 1391109998Smarkm CRYPTO_get_[locked_]mem_functions now writes 0 where such an 1392109998Smarkm extended allocation function is enabled. 1393109998Smarkm Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where 1394109998Smarkm a conventional allocation function is enabled. 1395109998Smarkm [Richard Levitte, Bodo Moeller] 1396109998Smarkm 1397109998Smarkm *) Finish off removing the remaining LHASH function pointer casts. 1398109998Smarkm There should no longer be any prototype-casting required when using 1399109998Smarkm the LHASH abstraction, and any casts that remain are "bugs". See 1400109998Smarkm the callback types and macros at the head of lhash.h for details 1401109998Smarkm (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). 1402109998Smarkm [Geoff Thorpe] 1403109998Smarkm 1404109998Smarkm *) Add automatic query of EGD sockets in RAND_poll() for the unix variant. 1405109998Smarkm If /dev/[u]random devices are not available or do not return enough 1406109998Smarkm entropy, EGD style sockets (served by EGD or PRNGD) will automatically 1407109998Smarkm be queried. 1408109998Smarkm The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and 1409109998Smarkm /etc/entropy will be queried once each in this sequence, quering stops 1410109998Smarkm when enough entropy was collected without querying more sockets. 1411109998Smarkm [Lutz Jaenicke] 1412109998Smarkm 1413109998Smarkm *) Change the Unix RAND_poll() variant to be able to poll several 1414109998Smarkm random devices, as specified by DEVRANDOM, until a sufficient amount 1415109998Smarkm of data has been collected. We spend at most 10 ms on each file 1416109998Smarkm (select timeout) and read in non-blocking mode. DEVRANDOM now 1417109998Smarkm defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" 1418109998Smarkm (previously it was just the string "/dev/urandom"), so on typical 1419109998Smarkm platforms the 10 ms delay will never occur. 1420109998Smarkm Also separate out the Unix variant to its own file, rand_unix.c. 1421109998Smarkm For VMS, there's a currently-empty rand_vms.c. 1422109998Smarkm [Richard Levitte] 1423109998Smarkm 1424109998Smarkm *) Move OCSP client related routines to ocsp_cl.c. These 1425109998Smarkm provide utility functions which an application needing 1426109998Smarkm to issue a request to an OCSP responder and analyse the 1427109998Smarkm response will typically need: as opposed to those which an 1428109998Smarkm OCSP responder itself would need which will be added later. 1429109998Smarkm 1430109998Smarkm OCSP_request_sign() signs an OCSP request with an API similar 1431109998Smarkm to PKCS7_sign(). OCSP_response_status() returns status of OCSP 1432109998Smarkm response. OCSP_response_get1_basic() extracts basic response 1433109998Smarkm from response. OCSP_resp_find_status(): finds and extracts status 1434109998Smarkm information from an OCSP_CERTID structure (which will be created 1435109998Smarkm when the request structure is built). These are built from lower 1436109998Smarkm level functions which work on OCSP_SINGLERESP structures but 1437109998Smarkm wont normally be used unless the application wishes to examine 1438109998Smarkm extensions in the OCSP response for example. 1439109998Smarkm 1440109998Smarkm Replace nonce routines with a pair of functions. 1441109998Smarkm OCSP_request_add1_nonce() adds a nonce value and optionally 1442109998Smarkm generates a random value. OCSP_check_nonce() checks the 1443109998Smarkm validity of the nonce in an OCSP response. 1444109998Smarkm [Steve Henson] 1445109998Smarkm 1446109998Smarkm *) Change function OCSP_request_add() to OCSP_request_add0_id(). 1447109998Smarkm This doesn't copy the supplied OCSP_CERTID and avoids the 1448109998Smarkm need to free up the newly created id. Change return type 1449109998Smarkm to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. 1450109998Smarkm This can then be used to add extensions to the request. 1451109998Smarkm Deleted OCSP_request_new(), since most of its functionality 1452109998Smarkm is now in OCSP_REQUEST_new() (and the case insensitive name 1453109998Smarkm clash) apart from the ability to set the request name which 1454109998Smarkm will be added elsewhere. 1455109998Smarkm [Steve Henson] 1456109998Smarkm 1457109998Smarkm *) Update OCSP API. Remove obsolete extensions argument from 1458109998Smarkm various functions. Extensions are now handled using the new 1459109998Smarkm OCSP extension code. New simple OCSP HTTP function which 1460109998Smarkm can be used to send requests and parse the response. 1461109998Smarkm [Steve Henson] 1462109998Smarkm 1463109998Smarkm *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new 1464109998Smarkm ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN 1465109998Smarkm uses the special reorder version of SET OF to sort the attributes 1466109998Smarkm and reorder them to match the encoded order. This resolves a long 1467109998Smarkm standing problem: a verify on a PKCS7 structure just after signing 1468109998Smarkm it used to fail because the attribute order did not match the 1469109998Smarkm encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: 1470109998Smarkm it uses the received order. This is necessary to tolerate some broken 1471109998Smarkm software that does not order SET OF. This is handled by encoding 1472109998Smarkm as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) 1473109998Smarkm to produce the required SET OF. 1474109998Smarkm [Steve Henson] 1475109998Smarkm 1476109998Smarkm *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and 1477109998Smarkm OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header 1478109998Smarkm files to get correct declarations of the ASN.1 item variables. 1479109998Smarkm [Richard Levitte] 1480109998Smarkm 1481109998Smarkm *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many 1482109998Smarkm PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: 1483109998Smarkm asn1_check_tlen() would sometimes attempt to use 'ctx' when it was 1484109998Smarkm NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). 1485109998Smarkm New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant 1486109998Smarkm ASN1_ITEM and no wrapper functions. 1487109998Smarkm [Steve Henson] 1488109998Smarkm 1489109998Smarkm *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These 1490109998Smarkm replace the old function pointer based I/O routines. Change most of 1491109998Smarkm the *_d2i_bio() and *_d2i_fp() functions to use these. 1492109998Smarkm [Steve Henson] 1493109998Smarkm 1494109998Smarkm *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor 1495109998Smarkm lines, recognice more "algorithms" that can be deselected, and make 1496109998Smarkm it complain about algorithm deselection that isn't recognised. 1497109998Smarkm [Richard Levitte] 1498109998Smarkm 1499109998Smarkm *) New ASN1 functions to handle dup, sign, verify, digest, pack and 1500109998Smarkm unpack operations in terms of ASN1_ITEM. Modify existing wrappers 1501109998Smarkm to use new functions. Add NO_ASN1_OLD which can be set to remove 1502109998Smarkm some old style ASN1 functions: this can be used to determine if old 1503109998Smarkm code will still work when these eventually go away. 1504109998Smarkm [Steve Henson] 1505109998Smarkm 1506109998Smarkm *) New extension functions for OCSP structures, these follow the 1507109998Smarkm same conventions as certificates and CRLs. 1508109998Smarkm [Steve Henson] 1509109998Smarkm 1510109998Smarkm *) New function X509V3_add1_i2d(). This automatically encodes and 1511109998Smarkm adds an extension. Its behaviour can be customised with various 1512109998Smarkm flags to append, replace or delete. Various wrappers added for 1513109998Smarkm certifcates and CRLs. 1514109998Smarkm [Steve Henson] 1515109998Smarkm 1516109998Smarkm *) Fix to avoid calling the underlying ASN1 print routine when 1517109998Smarkm an extension cannot be parsed. Correct a typo in the 1518109998Smarkm OCSP_SERVICELOC extension. Tidy up print OCSP format. 1519109998Smarkm [Steve Henson] 1520109998Smarkm 1521109998Smarkm *) Make mkdef.pl parse some of the ASN1 macros and add apropriate 1522109998Smarkm entries for variables. 1523109998Smarkm [Steve Henson] 1524109998Smarkm 1525109998Smarkm *) Add functionality to apps/openssl.c for detecting locking 1526109998Smarkm problems: As the program is single-threaded, all we have 1527109998Smarkm to do is register a locking callback using an array for 1528109998Smarkm storing which locks are currently held by the program. 1529109998Smarkm [Bodo Moeller] 1530109998Smarkm 1531109998Smarkm *) Use a lock around the call to CRYPTO_get_ex_new_index() in 1532109998Smarkm SSL_get_ex_data_X509_STORE_idx(), which is used in 1533109998Smarkm ssl_verify_cert_chain() and thus can be called at any time 1534109998Smarkm during TLS/SSL handshakes so that thread-safety is essential. 1535109998Smarkm Unfortunately, the ex_data design is not at all suited 1536109998Smarkm for multi-threaded use, so it probably should be abolished. 1537109998Smarkm [Bodo Moeller] 1538109998Smarkm 1539109998Smarkm *) Added Broadcom "ubsec" ENGINE to OpenSSL. 1540109998Smarkm [Broadcom, tweaked and integrated by Geoff Thorpe] 1541109998Smarkm 1542109998Smarkm *) Move common extension printing code to new function 1543109998Smarkm X509V3_print_extensions(). Reorganise OCSP print routines and 1544109998Smarkm implement some needed OCSP ASN1 functions. Add OCSP extensions. 1545109998Smarkm [Steve Henson] 1546109998Smarkm 1547109998Smarkm *) New function X509_signature_print() to remove duplication in some 1548109998Smarkm print routines. 1549109998Smarkm [Steve Henson] 1550109998Smarkm 1551109998Smarkm *) Add a special meaning when SET OF and SEQUENCE OF flags are both 1552109998Smarkm set (this was treated exactly the same as SET OF previously). This 1553109998Smarkm is used to reorder the STACK representing the structure to match the 1554109998Smarkm encoding. This will be used to get round a problem where a PKCS7 1555109998Smarkm structure which was signed could not be verified because the STACK 1556109998Smarkm order did not reflect the encoded order. 1557109998Smarkm [Steve Henson] 1558109998Smarkm 1559109998Smarkm *) Reimplement the OCSP ASN1 module using the new code. 1560109998Smarkm [Steve Henson] 1561109998Smarkm 1562109998Smarkm *) Update the X509V3 code to permit the use of an ASN1_ITEM structure 1563109998Smarkm for its ASN1 operations. The old style function pointers still exist 1564109998Smarkm for now but they will eventually go away. 1565109998Smarkm [Steve Henson] 1566109998Smarkm 1567109998Smarkm *) Merge in replacement ASN1 code from the ASN1 branch. This almost 1568109998Smarkm completely replaces the old ASN1 functionality with a table driven 1569109998Smarkm encoder and decoder which interprets an ASN1_ITEM structure describing 1570109998Smarkm the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is 1571109998Smarkm largely maintained. Almost all of the old asn1_mac.h macro based ASN1 1572109998Smarkm has also been converted to the new form. 1573109998Smarkm [Steve Henson] 1574109998Smarkm 1575109998Smarkm *) Change BN_mod_exp_recp so that negative moduli are tolerated 1576109998Smarkm (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set 1577109998Smarkm so that BN_mod_exp_mont and BN_mod_exp_mont_word work 1578109998Smarkm for negative moduli. 1579109998Smarkm [Bodo Moeller] 1580109998Smarkm 1581109998Smarkm *) Fix BN_uadd and BN_usub: Always return non-negative results instead 1582109998Smarkm of not touching the result's sign bit. 1583109998Smarkm [Bodo Moeller] 1584109998Smarkm 1585109998Smarkm *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be 1586109998Smarkm set. 1587109998Smarkm [Bodo Moeller] 1588109998Smarkm 1589109998Smarkm *) Changed the LHASH code to use prototypes for callbacks, and created 1590109998Smarkm macros to declare and implement thin (optionally static) functions 1591109998Smarkm that provide type-safety and avoid function pointer casting for the 1592109998Smarkm type-specific callbacks. 1593109998Smarkm [Geoff Thorpe] 1594109998Smarkm 1595109998Smarkm *) Added Kerberos Cipher Suites to be used with TLS, as written in 1596109998Smarkm RFC 2712. 1597109998Smarkm [Veers Staats <staatsvr@asc.hpc.mil>, 1598109998Smarkm Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte] 1599109998Smarkm 1600109998Smarkm *) Reformat the FAQ so the different questions and answers can be divided 1601109998Smarkm in sections depending on the subject. 1602109998Smarkm [Richard Levitte] 1603109998Smarkm 1604109998Smarkm *) Have the zlib compression code load ZLIB.DLL dynamically under 1605109998Smarkm Windows. 1606109998Smarkm [Richard Levitte] 1607109998Smarkm 1608109998Smarkm *) New function BN_mod_sqrt for computing square roots modulo a prime 1609109998Smarkm (using the probabilistic Tonelli-Shanks algorithm unless 1610109998Smarkm p == 3 (mod 4) or p == 5 (mod 8), which are cases that can 1611109998Smarkm be handled deterministically). 1612109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 1613109998Smarkm 1614109998Smarkm *) Make BN_mod_inverse faster by explicitly handling small quotients 1615109998Smarkm in the Euclid loop. (Speed gain about 20% for small moduli [256 or 1616109998Smarkm 512 bits], about 30% for larger ones [1024 or 2048 bits].) 1617109998Smarkm [Bodo Moeller] 1618109998Smarkm 1619109998Smarkm *) New function BN_kronecker. 1620109998Smarkm [Bodo Moeller] 1621109998Smarkm 1622109998Smarkm *) Fix BN_gcd so that it works on negative inputs; the result is 1623109998Smarkm positive unless both parameters are zero. 1624109998Smarkm Previously something reasonably close to an infinite loop was 1625109998Smarkm possible because numbers could be growing instead of shrinking 1626109998Smarkm in the implementation of Euclid's algorithm. 1627109998Smarkm [Bodo Moeller] 1628109998Smarkm 1629109998Smarkm *) Fix BN_is_word() and BN_is_one() macros to take into account the 1630109998Smarkm sign of the number in question. 1631109998Smarkm 1632109998Smarkm Fix BN_is_word(a,w) to work correctly for w == 0. 1633109998Smarkm 1634109998Smarkm The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) 1635109998Smarkm because its test if the absolute value of 'a' equals 'w'. 1636109998Smarkm Note that BN_abs_is_word does *not* handle w == 0 reliably; 1637109998Smarkm it exists mostly for use in the implementations of BN_is_zero(), 1638109998Smarkm BN_is_one(), and BN_is_word(). 1639109998Smarkm [Bodo Moeller] 1640109998Smarkm 1641109998Smarkm *) New function BN_swap. 1642109998Smarkm [Bodo Moeller] 1643109998Smarkm 1644109998Smarkm *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that 1645109998Smarkm the exponentiation functions are more likely to produce reasonable 1646109998Smarkm results on negative inputs. 1647109998Smarkm [Bodo Moeller] 1648109998Smarkm 1649109998Smarkm *) Change BN_mod_mul so that the result is always non-negative. 1650109998Smarkm Previously, it could be negative if one of the factors was negative; 1651109998Smarkm I don't think anyone really wanted that behaviour. 1652109998Smarkm [Bodo Moeller] 1653109998Smarkm 1654109998Smarkm *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c 1655109998Smarkm (except for exponentiation, which stays in crypto/bn/bn_exp.c, 1656109998Smarkm and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c) 1657109998Smarkm and add new functions: 1658109998Smarkm 1659109998Smarkm BN_nnmod 1660109998Smarkm BN_mod_sqr 1661109998Smarkm BN_mod_add 1662109998Smarkm BN_mod_add_quick 1663109998Smarkm BN_mod_sub 1664109998Smarkm BN_mod_sub_quick 1665109998Smarkm BN_mod_lshift1 1666109998Smarkm BN_mod_lshift1_quick 1667109998Smarkm BN_mod_lshift 1668109998Smarkm BN_mod_lshift_quick 1669109998Smarkm 1670109998Smarkm These functions always generate non-negative results. 1671109998Smarkm 1672109998Smarkm BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r 1673109998Smarkm such that |m| < r < 0, BN_nnmod will output rem + |m| instead). 1674109998Smarkm 1675109998Smarkm BN_mod_XXX_quick(r, a, [b,] m) generates the same result as 1676109998Smarkm BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b] 1677109998Smarkm be reduced modulo m. 1678109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 1679109998Smarkm 1680109998Smarkm *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 1681109998Smarkm was actually never needed) and in BN_mul(). The removal in BN_mul() 1682109998Smarkm required a small change in bn_mul_part_recursive() and the addition 1683109998Smarkm of the functions bn_cmp_part_words(), bn_sub_part_words() and 1684109998Smarkm bn_add_part_words(), which do the same thing as bn_cmp_words(), 1685109998Smarkm bn_sub_words() and bn_add_words() except they take arrays with 1686109998Smarkm differing sizes. 1687109998Smarkm [Richard Levitte] 1688109998Smarkm 1689109998Smarkm *) In 'openssl passwd', verify passwords read from the terminal 1690109998Smarkm unless the '-salt' option is used (which usually means that 1691109998Smarkm verification would just waste user's time since the resulting 1692109998Smarkm hash is going to be compared with some given password hash) 1693109998Smarkm or the new '-noverify' option is used. 1694109998Smarkm 1695109998Smarkm This is an incompatible change, but it does not affect 1696109998Smarkm non-interactive use of 'openssl passwd' (passwords on the command 1697109998Smarkm line, '-stdin' option, '-in ...' option) and thus should not 1698109998Smarkm cause any problems. 1699109998Smarkm [Bodo Moeller] 1700109998Smarkm 1701109998Smarkm *) Remove all references to RSAref, since there's no more need for it. 1702109998Smarkm [Richard Levitte] 1703109998Smarkm 1704109998Smarkm *) Make DSO load along a path given through an environment variable 1705109998Smarkm (SHLIB_PATH) with shl_load(). 1706109998Smarkm [Richard Levitte] 1707109998Smarkm 1708109998Smarkm *) Constify the ENGINE code as a result of BIGNUM constification. 1709109998Smarkm Also constify the RSA code and most things related to it. In a 1710109998Smarkm few places, most notable in the depth of the ASN.1 code, ugly 1711109998Smarkm casts back to non-const were required (to be solved at a later 1712109998Smarkm time) 1713109998Smarkm [Richard Levitte] 1714109998Smarkm 1715109998Smarkm *) Make it so the openssl application has all engines loaded by default. 1716109998Smarkm [Richard Levitte] 1717109998Smarkm 1718109998Smarkm *) Constify the BIGNUM routines a little more. 1719109998Smarkm [Richard Levitte] 1720109998Smarkm 1721109998Smarkm *) Add the following functions: 1722109998Smarkm 1723109998Smarkm ENGINE_load_cswift() 1724109998Smarkm ENGINE_load_chil() 1725109998Smarkm ENGINE_load_atalla() 1726109998Smarkm ENGINE_load_nuron() 1727109998Smarkm ENGINE_load_builtin_engines() 1728109998Smarkm 1729109998Smarkm That way, an application can itself choose if external engines that 1730109998Smarkm are built-in in OpenSSL shall ever be used or not. The benefit is 1731109998Smarkm that applications won't have to be linked with libdl or other dso 1732109998Smarkm libraries unless it's really needed. 1733109998Smarkm 1734109998Smarkm Changed 'openssl engine' to load all engines on demand. 1735109998Smarkm Changed the engine header files to avoid the duplication of some 1736109998Smarkm declarations (they differed!). 1737109998Smarkm [Richard Levitte] 1738109998Smarkm 1739109998Smarkm *) 'openssl engine' can now list capabilities. 1740109998Smarkm [Richard Levitte] 1741109998Smarkm 1742109998Smarkm *) Better error reporting in 'openssl engine'. 1743109998Smarkm [Richard Levitte] 1744109998Smarkm 1745109998Smarkm *) Never call load_dh_param(NULL) in s_server. 1746109998Smarkm [Bodo Moeller] 1747109998Smarkm 1748109998Smarkm *) Add engine application. It can currently list engines by name and 1749109998Smarkm identity, and test if they are actually available. 1750109998Smarkm [Richard Levitte] 1751109998Smarkm 1752109998Smarkm *) Improve RPM specification file by forcing symbolic linking and making 1753109998Smarkm sure the installed documentation is also owned by root.root. 1754109998Smarkm [Damien Miller <djm@mindrot.org>] 1755109998Smarkm 1756109998Smarkm *) Give the OpenSSL applications more possibilities to make use of 1757109998Smarkm keys (public as well as private) handled by engines. 1758109998Smarkm [Richard Levitte] 1759109998Smarkm 1760109998Smarkm *) Add OCSP code that comes from CertCo. 1761109998Smarkm [Richard Levitte] 1762109998Smarkm 1763109998Smarkm *) Add VMS support for the Rijndael code. 1764109998Smarkm [Richard Levitte] 1765109998Smarkm 1766109998Smarkm *) Added untested support for Nuron crypto accelerator. 1767109998Smarkm [Ben Laurie] 1768109998Smarkm 1769109998Smarkm *) Add support for external cryptographic devices. This code was 1770109998Smarkm previously distributed separately as the "engine" branch. 1771109998Smarkm [Geoff Thorpe, Richard Levitte] 1772109998Smarkm 1773109998Smarkm *) Rework the filename-translation in the DSO code. It is now possible to 1774109998Smarkm have far greater control over how a "name" is turned into a filename 1775109998Smarkm depending on the operating environment and any oddities about the 1776109998Smarkm different shared library filenames on each system. 1777109998Smarkm [Geoff Thorpe] 1778109998Smarkm 1779109998Smarkm *) Support threads on FreeBSD-elf in Configure. 1780109998Smarkm [Richard Levitte] 1781109998Smarkm 1782109998Smarkm *) Fix for SHA1 assembly problem with MASM: it produces 1783109998Smarkm warnings about corrupt line number information when assembling 1784109998Smarkm with debugging information. This is caused by the overlapping 1785109998Smarkm of two sections. 1786109998Smarkm [Bernd Matthes <mainbug@celocom.de>, Steve Henson] 1787109998Smarkm 1788109998Smarkm *) NCONF changes. 1789109998Smarkm NCONF_get_number() has no error checking at all. As a replacement, 1790109998Smarkm NCONF_get_number_e() is defined (_e for "error checking") and is 1791109998Smarkm promoted strongly. The old NCONF_get_number is kept around for 1792109998Smarkm binary backward compatibility. 1793109998Smarkm Make it possible for methods to load from something other than a BIO, 1794109998Smarkm by providing a function pointer that is given a name instead of a BIO. 1795109998Smarkm For example, this could be used to load configuration data from an 1796109998Smarkm LDAP server. 1797109998Smarkm [Richard Levitte] 1798109998Smarkm 1799109998Smarkm *) Fix for non blocking accept BIOs. Added new I/O special reason 1800109998Smarkm BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs 1801109998Smarkm with non blocking I/O was not possible because no retry code was 1802109998Smarkm implemented. Also added new SSL code SSL_WANT_ACCEPT to cover 1803109998Smarkm this case. 1804109998Smarkm [Steve Henson] 1805109998Smarkm 1806109998Smarkm *) Added the beginnings of Rijndael support. 1807109998Smarkm [Ben Laurie] 1808109998Smarkm 1809109998Smarkm *) Fix for bug in DirectoryString mask setting. Add support for 1810109998Smarkm X509_NAME_print_ex() in 'req' and X509_print_ex() function 1811109998Smarkm to allow certificate printing to more controllable, additional 1812109998Smarkm 'certopt' option to 'x509' to allow new printing options to be 1813109998Smarkm set. 1814109998Smarkm [Steve Henson] 1815109998Smarkm 1816109998Smarkm *) Clean old EAY MD5 hack from e_os.h. 1817109998Smarkm [Richard Levitte] 1818109998Smarkm 1819109998Smarkm Changes between 0.9.6g and 0.9.6h [5 Dec 2002] 1820109998Smarkm 1821109998Smarkm *) New function OPENSSL_cleanse(), which is used to cleanse a section of 1822109998Smarkm memory from it's contents. This is done with a counter that will 1823109998Smarkm place alternating values in each byte. This can be used to solve 1824109998Smarkm two issues: 1) the removal of calls to memset() by highly optimizing 1825109998Smarkm compilers, and 2) cleansing with other values than 0, since those can 1826109998Smarkm be read through on certain media, for example a swap space on disk. 1827109998Smarkm [Geoff Thorpe] 1828109998Smarkm 1829109998Smarkm *) Bugfix: client side session caching did not work with external caching, 1830109998Smarkm because the session->cipher setting was not restored when reloading 1831109998Smarkm from the external cache. This problem was masked, when 1832109998Smarkm SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. 1833109998Smarkm (Found by Steve Haslam <steve@araqnid.ddts.net>.) 1834109998Smarkm [Lutz Jaenicke] 1835109998Smarkm 1836109998Smarkm *) Fix client_certificate (ssl/s2_clnt.c): The permissible total 1837109998Smarkm length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. 1838109998Smarkm [Zeev Lieber <zeev-l@yahoo.com>] 1839109998Smarkm 1840109998Smarkm *) Undo an undocumented change introduced in 0.9.6e which caused 1841109998Smarkm repeated calls to OpenSSL_add_all_ciphers() and 1842109998Smarkm OpenSSL_add_all_digests() to be ignored, even after calling 1843109998Smarkm EVP_cleanup(). 1844109998Smarkm [Richard Levitte] 1845109998Smarkm 1846109998Smarkm *) Change the default configuration reader to deal with last line not 1847109998Smarkm being properly terminated. 1848109998Smarkm [Richard Levitte] 1849109998Smarkm 1850109998Smarkm *) Change X509_NAME_cmp() so it applies the special rules on handling 1851109998Smarkm DN values that are of type PrintableString, as well as RDNs of type 1852109998Smarkm emailAddress where the value has the type ia5String. 1853109998Smarkm [stefank@valicert.com via Richard Levitte] 1854109998Smarkm 1855109998Smarkm *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half 1856109998Smarkm the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently 1857109998Smarkm doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be 1858109998Smarkm the bitwise-OR of the two for use by the majority of applications 1859109998Smarkm wanting this behaviour, and update the docs. The documented 1860109998Smarkm behaviour and actual behaviour were inconsistent and had been 1861109998Smarkm changing anyway, so this is more a bug-fix than a behavioural 1862109998Smarkm change. 1863109998Smarkm [Geoff Thorpe, diagnosed by Nadav Har'El] 1864109998Smarkm 1865109998Smarkm *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c 1866109998Smarkm (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). 1867109998Smarkm [Bodo Moeller] 1868109998Smarkm 1869109998Smarkm *) Fix initialization code race conditions in 1870109998Smarkm SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), 1871109998Smarkm SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), 1872109998Smarkm SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), 1873109998Smarkm TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), 1874109998Smarkm ssl2_get_cipher_by_char(), 1875109998Smarkm ssl3_get_cipher_by_char(). 1876109998Smarkm [Patrick McCormick <patrick@tellme.com>, Bodo Moeller] 1877109998Smarkm 1878109998Smarkm *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after 1879109998Smarkm the cached sessions are flushed, as the remove_cb() might use ex_data 1880109998Smarkm contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> 1881109998Smarkm (see [openssl.org #212]). 1882109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 1883109998Smarkm 1884109998Smarkm *) Fix typo in OBJ_txt2obj which incorrectly passed the content 1885109998Smarkm length, instead of the encoding length to d2i_ASN1_OBJECT. 1886109998Smarkm [Steve Henson] 1887109998Smarkm 1888101618Snectar Changes between 0.9.6f and 0.9.6g [9 Aug 2002] 1889101618Snectar 1890101618Snectar *) [In 0.9.6g-engine release:] 1891101618Snectar Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall'). 1892101618Snectar [Lynn Gazis <lgazis@rainbow.com>] 1893101618Snectar 1894101613Snectar Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 1895101613Snectar 1896101613Snectar *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 1897101613Snectar and get fix the header length calculation. 1898101613Snectar [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 1899101613Snectar Alon Kantor <alonk@checkpoint.com> (and others), 1900101613Snectar Steve Henson] 1901101613Snectar 1902101613Snectar *) Use proper error handling instead of 'assertions' in buffer 1903101613Snectar overflow checks added in 0.9.6e. This prevents DoS (the 1904101613Snectar assertions could call abort()). 1905101613Snectar [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] 1906101613Snectar 1907100936Snectar Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 1908100936Snectar 1909109998Smarkm *) Add various sanity checks to asn1_get_length() to reject 1910109998Smarkm the ASN1 length bytes if they exceed sizeof(long), will appear 1911109998Smarkm negative or the content length exceeds the length of the 1912109998Smarkm supplied buffer. 1913109998Smarkm [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 1914109998Smarkm 1915100936Snectar *) Fix cipher selection routines: ciphers without encryption had no flags 1916100936Snectar for the cipher strength set and where therefore not handled correctly 1917100936Snectar by the selection routines (PR #130). 1918100936Snectar [Lutz Jaenicke] 1919100936Snectar 1920100936Snectar *) Fix EVP_dsa_sha macro. 1921100936Snectar [Nils Larsch] 1922100936Snectar 1923100936Snectar *) New option 1924100936Snectar SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 1925100936Snectar for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 1926100936Snectar that was added in OpenSSL 0.9.6d. 1927100936Snectar 1928100936Snectar As the countermeasure turned out to be incompatible with some 1929100936Snectar broken SSL implementations, the new option is part of SSL_OP_ALL. 1930100936Snectar SSL_OP_ALL is usually employed when compatibility with weird SSL 1931100936Snectar implementations is desired (e.g. '-bugs' option to 's_client' and 1932100936Snectar 's_server'), so the new option is automatically set in many 1933100936Snectar applications. 1934100936Snectar [Bodo Moeller] 1935100936Snectar 1936100936Snectar *) Changes in security patch: 1937100936Snectar 1938100936Snectar Changes marked "(CHATS)" were sponsored by the Defense Advanced 1939100936Snectar Research Projects Agency (DARPA) and Air Force Research Laboratory, 1940100936Snectar Air Force Materiel Command, USAF, under agreement number 1941100936Snectar F30602-01-2-0537. 1942100936Snectar 1943100936Snectar *) Add various sanity checks to asn1_get_length() to reject 1944100936Snectar the ASN1 length bytes if they exceed sizeof(long), will appear 1945100936Snectar negative or the content length exceeds the length of the 1946109998Smarkm supplied buffer. (CAN-2002-0659) 1947100936Snectar [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 1948100936Snectar 1949100936Snectar *) Assertions for various potential buffer overflows, not known to 1950100936Snectar happen in practice. 1951100936Snectar [Ben Laurie (CHATS)] 1952100936Snectar 1953100936Snectar *) Various temporary buffers to hold ASCII versions of integers were 1954100936Snectar too small for 64 bit platforms. (CAN-2002-0655) 1955100936Snectar [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 1956100936Snectar 1957100936Snectar *) Remote buffer overflow in SSL3 protocol - an attacker could 1958100936Snectar supply an oversized session ID to a client. (CAN-2002-0656) 1959100936Snectar [Ben Laurie (CHATS)] 1960100936Snectar 1961100936Snectar *) Remote buffer overflow in SSL2 protocol - an attacker could 1962100936Snectar supply an oversized client master key. (CAN-2002-0656) 1963100936Snectar [Ben Laurie (CHATS)] 1964100936Snectar 1965100928Snectar Changes between 0.9.6c and 0.9.6d [9 May 2002] 1966100928Snectar 1967100928Snectar *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 1968100928Snectar encoded as NULL) with id-dsa-with-sha1. 1969100928Snectar [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 1970100928Snectar 1971100928Snectar *) Check various X509_...() return values in apps/req.c. 1972100928Snectar [Nils Larsch <nla@trustcenter.de>] 1973100928Snectar 1974100928Snectar *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 1975100928Snectar an end-of-file condition would erronously be flagged, when the CRLF 1976100928Snectar was just at the end of a processed block. The bug was discovered when 1977100928Snectar processing data through a buffering memory BIO handing the data to a 1978100928Snectar BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 1979100928Snectar <ptsekov@syntrex.com> and Nedelcho Stanev. 1980100928Snectar [Lutz Jaenicke] 1981100928Snectar 1982100928Snectar *) Implement a countermeasure against a vulnerability recently found 1983100928Snectar in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 1984100928Snectar before application data chunks to avoid the use of known IVs 1985100928Snectar with data potentially chosen by the attacker. 1986100928Snectar [Bodo Moeller] 1987100928Snectar 1988100928Snectar *) Fix length checks in ssl3_get_client_hello(). 1989100928Snectar [Bodo Moeller] 1990100928Snectar 1991100928Snectar *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 1992100928Snectar to prevent ssl3_read_internal() from incorrectly assuming that 1993100928Snectar ssl3_read_bytes() found application data while handshake 1994100928Snectar processing was enabled when in fact s->s3->in_read_app_data was 1995100928Snectar merely automatically cleared during the initial handshake. 1996100928Snectar [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 1997100928Snectar 1998100928Snectar *) Fix object definitions for Private and Enterprise: they were not 1999100928Snectar recognized in their shortname (=lowercase) representation. Extend 2000100928Snectar obj_dat.pl to issue an error when using undefined keywords instead 2001100928Snectar of silently ignoring the problem (Svenning Sorensen 2002100928Snectar <sss@sss.dnsalias.net>). 2003100928Snectar [Lutz Jaenicke] 2004100928Snectar 2005100928Snectar *) Fix DH_generate_parameters() so that it works for 'non-standard' 2006100928Snectar generators, i.e. generators other than 2 and 5. (Previously, the 2007100928Snectar code did not properly initialise the 'add' and 'rem' values to 2008100928Snectar BN_generate_prime().) 2009100928Snectar 2010100928Snectar In the new general case, we do not insist that 'generator' is 2011100928Snectar actually a primitive root: This requirement is rather pointless; 2012100928Snectar a generator of the order-q subgroup is just as good, if not 2013100928Snectar better. 2014100928Snectar [Bodo Moeller] 2015100928Snectar 2016100928Snectar *) Map new X509 verification errors to alerts. Discovered and submitted by 2017100928Snectar Tom Wu <tom@arcot.com>. 2018100928Snectar [Lutz Jaenicke] 2019100928Snectar 2020100928Snectar *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 2021100928Snectar returning non-zero before the data has been completely received 2022100928Snectar when using non-blocking I/O. 2023100928Snectar [Bodo Moeller; problem pointed out by John Hughes] 2024100928Snectar 2025100928Snectar *) Some of the ciphers missed the strength entry (SSL_LOW etc). 2026100928Snectar [Ben Laurie, Lutz Jaenicke] 2027100928Snectar 2028100928Snectar *) Fix bug in SSL_clear(): bad sessions were not removed (found by 2029100928Snectar Yoram Zahavi <YoramZ@gilian.com>). 2030100928Snectar [Lutz Jaenicke] 2031100928Snectar 2032100928Snectar *) Add information about CygWin 1.3 and on, and preserve proper 2033100928Snectar configuration for the versions before that. 2034100928Snectar [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 2035100928Snectar 2036100928Snectar *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 2037100928Snectar check whether we deal with a copy of a session and do not delete from 2038100928Snectar the cache in this case. Problem reported by "Izhar Shoshani Levi" 2039100928Snectar <izhar@checkpoint.com>. 2040100928Snectar [Lutz Jaenicke] 2041100928Snectar 2042100928Snectar *) Do not store session data into the internal session cache, if it 2043100928Snectar is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 2044100928Snectar flag is set). Proposed by Aslam <aslam@funk.com>. 2045100928Snectar [Lutz Jaenicke] 2046100928Snectar 2047100928Snectar *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 2048100928Snectar value is 0. 2049100928Snectar [Richard Levitte] 2050100928Snectar 2051109998Smarkm *) [In 0.9.6d-engine release:] 2052109998Smarkm Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 2053100928Snectar [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 2054100928Snectar 2055100928Snectar *) Add the configuration target linux-s390x. 2056100928Snectar [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 2057100928Snectar 2058100928Snectar *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 2059100928Snectar ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 2060100928Snectar variable as an indication that a ClientHello message has been 2061100928Snectar received. As the flag value will be lost between multiple 2062100928Snectar invocations of ssl3_accept when using non-blocking I/O, the 2063100928Snectar function may not be aware that a handshake has actually taken 2064100928Snectar place, thus preventing a new session from being added to the 2065100928Snectar session cache. 2066100928Snectar 2067100928Snectar To avoid this problem, we now set s->new_session to 2 instead of 2068100928Snectar using a local variable. 2069100928Snectar [Lutz Jaenicke, Bodo Moeller] 2070100928Snectar 2071100928Snectar *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 2072100928Snectar if the SSL_R_LENGTH_MISMATCH error is detected. 2073100928Snectar [Geoff Thorpe, Bodo Moeller] 2074100928Snectar 2075100928Snectar *) New 'shared_ldflag' column in Configure platform table. 2076100928Snectar [Richard Levitte] 2077100928Snectar 2078100928Snectar *) Fix EVP_CIPHER_mode macro. 2079100928Snectar ["Dan S. Camper" <dan@bti.net>] 2080100928Snectar 2081100928Snectar *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 2082100928Snectar type, we must throw them away by setting rr->length to 0. 2083100928Snectar [D P Chang <dpc@qualys.com>] 2084100928Snectar 208589837Skris Changes between 0.9.6b and 0.9.6c [21 dec 2001] 208689837Skris 208789837Skris *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 208889837Skris <Dominikus.Scherkl@biodata.com>. (The previous implementation 208989837Skris worked incorrectly for those cases where range = 10..._2 and 209089837Skris 3*range is two bits longer than range.) 209189837Skris [Bodo Moeller] 209289837Skris 209389837Skris *) Only add signing time to PKCS7 structures if it is not already 209489837Skris present. 209589837Skris [Steve Henson] 209689837Skris 209789837Skris *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 209889837Skris OBJ_ld_ce should be OBJ_id_ce. 209989837Skris Also some ip-pda OIDs in crypto/objects/objects.txt were 210089837Skris incorrect (cf. RFC 3039). 210189837Skris [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 210289837Skris 210389837Skris *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 210489837Skris returns early because it has nothing to do. 210589837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 210689837Skris 210789837Skris *) [In 0.9.6c-engine release:] 210889837Skris Fix mutex callback return values in crypto/engine/hw_ncipher.c. 210989837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 211089837Skris 211189837Skris *) [In 0.9.6c-engine release:] 211289837Skris Add support for Cryptographic Appliance's keyserver technology. 211389837Skris (Use engine 'keyclient') 211489837Skris [Cryptographic Appliances and Geoff Thorpe] 211589837Skris 211689837Skris *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 211789837Skris is called via tools/c89.sh because arguments have to be 211889837Skris rearranged (all '-L' options must appear before the first object 211989837Skris modules). 212089837Skris [Richard Shapiro <rshapiro@abinitio.com>] 212189837Skris 212289837Skris *) [In 0.9.6c-engine release:] 212389837Skris Add support for Broadcom crypto accelerator cards, backported 212489837Skris from 0.9.7. 212589837Skris [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 212689837Skris 212789837Skris *) [In 0.9.6c-engine release:] 212889837Skris Add support for SureWare crypto accelerator cards from 212989837Skris Baltimore Technologies. (Use engine 'sureware') 213089837Skris [Baltimore Technologies and Mark Cox] 213189837Skris 213289837Skris *) [In 0.9.6c-engine release:] 213389837Skris Add support for crypto accelerator cards from Accelerated 213489837Skris Encryption Processing, www.aep.ie. (Use engine 'aep') 213589837Skris [AEP Inc. and Mark Cox] 213689837Skris 213789837Skris *) Add a configuration entry for gcc on UnixWare. 213889837Skris [Gary Benson <gbenson@redhat.com>] 213989837Skris 214089837Skris *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 214189837Skris messages are stored in a single piece (fixed-length part and 214289837Skris variable-length part combined) and fix various bugs found on the way. 214389837Skris [Bodo Moeller] 214489837Skris 214589837Skris *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 214689837Skris instead. BIO_gethostbyname() does not know what timeouts are 214789837Skris appropriate, so entries would stay in cache even when they have 214889837Skris become invalid. 214989837Skris [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 215089837Skris 215189837Skris *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 215289837Skris faced with a pathologically small ClientHello fragment that does 215389837Skris not contain client_version: Instead of aborting with an error, 215489837Skris simply choose the highest available protocol version (i.e., 215589837Skris TLS 1.0 unless it is disabled). In practice, ClientHello 215689837Skris messages are never sent like this, but this change gives us 215789837Skris strictly correct behaviour at least for TLS. 215889837Skris [Bodo Moeller] 215989837Skris 216089837Skris *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 216189837Skris never resets s->method to s->ctx->method when called from within 216289837Skris one of the SSL handshake functions. 216389837Skris [Bodo Moeller; problem pointed out by Niko Baric] 216489837Skris 216589837Skris *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 216689837Skris (sent using the client's version number) if client_version is 216789837Skris smaller than the protocol version in use. Also change 216889837Skris ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 216989837Skris the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 217089837Skris the client will at least see that alert. 217189837Skris [Bodo Moeller] 217289837Skris 217389837Skris *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 217489837Skris correctly. 217589837Skris [Bodo Moeller] 217689837Skris 217789837Skris *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 217889837Skris client receives HelloRequest while in a handshake. 217989837Skris [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 218089837Skris 218189837Skris *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 218289837Skris should end in 'break', not 'goto end' which circuments various 218389837Skris cleanups done in state SSL_ST_OK. But session related stuff 218489837Skris must be disabled for SSL_ST_OK in the case that we just sent a 218589837Skris HelloRequest. 218689837Skris 218789837Skris Also avoid some overhead by not calling ssl_init_wbio_buffer() 218889837Skris before just sending a HelloRequest. 218989837Skris [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 219089837Skris 219189837Skris *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 219289837Skris reveal whether illegal block cipher padding was found or a MAC 219389837Skris verification error occured. (Neither SSLerr() codes nor alerts 219489837Skris are directly visible to potential attackers, but the information 219589837Skris may leak via logfiles.) 219689837Skris 219789837Skris Similar changes are not required for the SSL 2.0 implementation 219889837Skris because the number of padding bytes is sent in clear for SSL 2.0, 219989837Skris and the extra bytes are just ignored. However ssl/s2_pkt.c 220089837Skris failed to verify that the purported number of padding bytes is in 220189837Skris the legal range. 220289837Skris [Bodo Moeller] 220389837Skris 220489837Skris *) Add OpenUNIX-8 support including shared libraries 220589837Skris (Boyd Lynn Gerber <gerberb@zenez.com>). 220689837Skris [Lutz Jaenicke] 220789837Skris 220889837Skris *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 220989837Skris 'wristwatch attack' using huge encoding parameters (cf. 221089837Skris James H. Manger's CRYPTO 2001 paper). Note that the 221189837Skris RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 221289837Skris encoding parameters and hence was not vulnerable. 221389837Skris [Bodo Moeller] 221489837Skris 221589837Skris *) BN_sqr() bug fix. 221689837Skris [Ulf M�ller, reported by Jim Ellis <jim.ellis@cavium.com>] 221789837Skris 221889837Skris *) Rabin-Miller test analyses assume uniformly distributed witnesses, 221989837Skris so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 222089837Skris followed by modular reduction. 222189837Skris [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 222289837Skris 222389837Skris *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 222489837Skris equivalent based on BN_pseudo_rand() instead of BN_rand(). 222589837Skris [Bodo Moeller] 222689837Skris 222789837Skris *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 222889837Skris This function was broken, as the check for a new client hello message 222989837Skris to handle SGC did not allow these large messages. 223089837Skris (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 223189837Skris [Lutz Jaenicke] 223289837Skris 223389837Skris *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 223489837Skris [Lutz Jaenicke] 223589837Skris 223689837Skris *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 223789837Skris for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 223889837Skris [Lutz Jaenicke] 223989837Skris 224089837Skris *) Rework the configuration and shared library support for Tru64 Unix. 224189837Skris The configuration part makes use of modern compiler features and 224289837Skris still retains old compiler behavior for those that run older versions 224389837Skris of the OS. The shared library support part includes a variant that 224489837Skris uses the RPATH feature, and is available through the special 224589837Skris configuration target "alpha-cc-rpath", which will never be selected 224689837Skris automatically. 224789837Skris [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 224889837Skris 224989837Skris *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 225089837Skris with the same message size as in ssl3_get_certificate_request(). 225189837Skris Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 225289837Skris messages might inadvertently be reject as too long. 225389837Skris [Petr Lampa <lampa@fee.vutbr.cz>] 225489837Skris 225589837Skris *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 225689837Skris [Andy Polyakov] 225789837Skris 225889837Skris *) Modified SSL library such that the verify_callback that has been set 225989837Skris specificly for an SSL object with SSL_set_verify() is actually being 226089837Skris used. Before the change, a verify_callback set with this function was 226189837Skris ignored and the verify_callback() set in the SSL_CTX at the time of 226289837Skris the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 226389837Skris to allow the necessary settings. 226489837Skris [Lutz Jaenicke] 226589837Skris 226689837Skris *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 226789837Skris explicitly to NULL, as at least on Solaris 8 this seems not always to be 226889837Skris done automatically (in contradiction to the requirements of the C 226989837Skris standard). This made problems when used from OpenSSH. 227089837Skris [Lutz Jaenicke] 227189837Skris 227289837Skris *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 227389837Skris dh->length and always used 227489837Skris 227589837Skris BN_rand_range(priv_key, dh->p). 227689837Skris 227789837Skris BN_rand_range() is not necessary for Diffie-Hellman, and this 227889837Skris specific range makes Diffie-Hellman unnecessarily inefficient if 227989837Skris dh->length (recommended exponent length) is much smaller than the 228089837Skris length of dh->p. We could use BN_rand_range() if the order of 228189837Skris the subgroup was stored in the DH structure, but we only have 228289837Skris dh->length. 228389837Skris 228489837Skris So switch back to 228589837Skris 228689837Skris BN_rand(priv_key, l, ...) 228789837Skris 228889837Skris where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 228989837Skris otherwise. 229089837Skris [Bodo Moeller] 229189837Skris 229289837Skris *) In 229389837Skris 229489837Skris RSA_eay_public_encrypt 229589837Skris RSA_eay_private_decrypt 229689837Skris RSA_eay_private_encrypt (signing) 229789837Skris RSA_eay_public_decrypt (signature verification) 229889837Skris 229989837Skris (default implementations for RSA_public_encrypt, 230089837Skris RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 230189837Skris always reject numbers >= n. 230289837Skris [Bodo Moeller] 230389837Skris 230489837Skris *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 230589837Skris to synchronize access to 'locking_thread'. This is necessary on 230689837Skris systems where access to 'locking_thread' (an 'unsigned long' 230789837Skris variable) is not atomic. 230889837Skris [Bodo Moeller] 230989837Skris 231089837Skris *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 231189837Skris *before* setting the 'crypto_lock_rand' flag. The previous code had 231289837Skris a race condition if 0 is a valid thread ID. 231389837Skris [Travis Vitek <vitek@roguewave.com>] 231489837Skris 231589837Skris *) Add support for shared libraries under Irix. 231689837Skris [Albert Chin-A-Young <china@thewrittenword.com>] 231789837Skris 231889837Skris *) Add configuration option to build on Linux on both big-endian and 231989837Skris little-endian MIPS. 232089837Skris [Ralf Baechle <ralf@uni-koblenz.de>] 232189837Skris 232289837Skris *) Add the possibility to create shared libraries on HP-UX. 232389837Skris [Richard Levitte] 232489837Skris 232579998Skris Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 232679998Skris 232779998Skris *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 232879998Skris to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 232979998Skris Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 233079998Skris PRNG state recovery was possible based on the output of 233179998Skris one PRNG request appropriately sized to gain knowledge on 233279998Skris 'md' followed by enough consecutive 1-byte PRNG requests 233379998Skris to traverse all of 'state'. 233479998Skris 233579998Skris 1. When updating 'md_local' (the current thread's copy of 'md') 233679998Skris during PRNG output generation, hash all of the previous 233779998Skris 'md_local' value, not just the half used for PRNG output. 233879998Skris 233979998Skris 2. Make the number of bytes from 'state' included into the hash 234079998Skris independent from the number of PRNG bytes requested. 234179998Skris 234279998Skris The first measure alone would be sufficient to avoid 234379998Skris Markku-Juhani's attack. (Actually it had never occurred 234479998Skris to me that the half of 'md_local' used for chaining was the 234579998Skris half from which PRNG output bytes were taken -- I had always 234679998Skris assumed that the secret half would be used.) The second 234779998Skris measure makes sure that additional data from 'state' is never 234879998Skris mixed into 'md_local' in small portions; this heuristically 234979998Skris further strengthens the PRNG. 235079998Skris [Bodo Moeller] 235179998Skris 235279998Skris *) Fix crypto/bn/asm/mips3.s. 235379998Skris [Andy Polyakov] 235479998Skris 235579998Skris *) When only the key is given to "enc", the IV is undefined. Print out 235679998Skris an error message in this case. 235779998Skris [Lutz Jaenicke] 235879998Skris 235979998Skris *) Handle special case when X509_NAME is empty in X509 printing routines. 236079998Skris [Steve Henson] 236179998Skris 236279998Skris *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 236379998Skris positive and less than q. 236479998Skris [Bodo Moeller] 236579998Skris 236679998Skris *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 236779998Skris used: it isn't thread safe and the add_lock_callback should handle 236879998Skris that itself. 236979998Skris [Paul Rose <Paul.Rose@bridge.com>] 237079998Skris 237179998Skris *) Verify that incoming data obeys the block size in 237279998Skris ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 237379998Skris [Bodo Moeller] 237479998Skris 237579998Skris *) Fix OAEP check. 237679998Skris [Ulf M�ller, Bodo M�ller] 237779998Skris 237879998Skris *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 237989837Skris RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 238079998Skris when fixing the server behaviour for backwards-compatible 'client 238179998Skris hello' messages. (Note that the attack is impractical against 238279998Skris SSL 3.0 and TLS 1.0 anyway because length and version checking 238379998Skris means that the probability of guessing a valid ciphertext is 238479998Skris around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 238579998Skris paper.) 238679998Skris 238779998Skris Before 0.9.5, the countermeasure (hide the error by generating a 238879998Skris random 'decryption result') did not work properly because 238979998Skris ERR_clear_error() was missing, meaning that SSL_get_error() would 239079998Skris detect the supposedly ignored error. 239179998Skris 239279998Skris Both problems are now fixed. 239379998Skris [Bodo Moeller] 239479998Skris 239579998Skris *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 239679998Skris (previously it was 1024). 239779998Skris [Bodo Moeller] 239879998Skris 239979998Skris *) Fix for compatibility mode trust settings: ignore trust settings 240079998Skris unless some valid trust or reject settings are present. 240179998Skris [Steve Henson] 240279998Skris 240379998Skris *) Fix for blowfish EVP: its a variable length cipher. 240479998Skris [Steve Henson] 240579998Skris 240679998Skris *) Fix various bugs related to DSA S/MIME verification. Handle missing 240779998Skris parameters in DSA public key structures and return an error in the 240879998Skris DSA routines if parameters are absent. 240979998Skris [Steve Henson] 241079998Skris 241179998Skris *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 241279998Skris in the current directory if neither $RANDFILE nor $HOME was set. 241379998Skris RAND_file_name() in 0.9.6a returned NULL in this case. This has 241479998Skris caused some confusion to Windows users who haven't defined $HOME. 241579998Skris Thus RAND_file_name() is changed again: e_os.h can define a 241679998Skris DEFAULT_HOME, which will be used if $HOME is not set. 241779998Skris For Windows, we use "C:"; on other platforms, we still require 241879998Skris environment variables. 241979998Skris 242079998Skris *) Move 'if (!initialized) RAND_poll()' into regions protected by 242179998Skris CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 242279998Skris having multiple threads call RAND_poll() concurrently. 242379998Skris [Bodo Moeller] 242479998Skris 242579998Skris *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 242679998Skris combination of a flag and a thread ID variable. 242779998Skris Otherwise while one thread is in ssleay_rand_bytes (which sets the 242879998Skris flag), *other* threads can enter ssleay_add_bytes without obeying 242989837Skris the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 243079998Skris that they do not hold after the first thread unsets add_do_not_lock). 243179998Skris [Bodo Moeller] 243279998Skris 243379998Skris *) Change bctest again: '-x' expressions are not available in all 243479998Skris versions of 'test'. 243579998Skris [Bodo Moeller] 243679998Skris 243776866Skris Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 243872613Skris 243976866Skris *) Fix a couple of memory leaks in PKCS7_dataDecode() 244076866Skris [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 244176866Skris 244276866Skris *) Change Configure and Makefiles to provide EXE_EXT, which will contain 244376866Skris the default extension for executables, if any. Also, make the perl 244476866Skris scripts that use symlink() to test if it really exists and use "cp" 244576866Skris if it doesn't. All this made OpenSSL compilable and installable in 244676866Skris CygWin. 244776866Skris [Richard Levitte] 244876866Skris 244976866Skris *) Fix for asn1_GetSequence() for indefinite length constructed data. 245076866Skris If SEQUENCE is length is indefinite just set c->slen to the total 245176866Skris amount of data available. 245276866Skris [Steve Henson, reported by shige@FreeBSD.org] 245376866Skris [This change does not apply to 0.9.7.] 245476866Skris 245576866Skris *) Change bctest to avoid here-documents inside command substitution 245676866Skris (workaround for FreeBSD /bin/sh bug). 245776866Skris For compatibility with Ultrix, avoid shell functions (introduced 245876866Skris in the bctest version that searches along $PATH). 245976866Skris [Bodo Moeller] 246076866Skris 246176866Skris *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 246276866Skris with des_encrypt() defined on some operating systems, like Solaris 246376866Skris and UnixWare. 246476866Skris [Richard Levitte] 246576866Skris 246676866Skris *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 246776866Skris On the Importance of Eliminating Errors in Cryptographic 246876866Skris Computations, J. Cryptology 14 (2001) 2, 101-119, 246976866Skris http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 247076866Skris [Ulf Moeller] 247176866Skris 247276866Skris *) MIPS assembler BIGNUM division bug fix. 247376866Skris [Andy Polyakov] 247476866Skris 247576866Skris *) Disabled incorrect Alpha assembler code. 247676866Skris [Richard Levitte] 247776866Skris 247876866Skris *) Fix PKCS#7 decode routines so they correctly update the length 247976866Skris after reading an EOC for the EXPLICIT tag. 248076866Skris [Steve Henson] 248176866Skris [This change does not apply to 0.9.7.] 248276866Skris 248376866Skris *) Fix bug in PKCS#12 key generation routines. This was triggered 248476866Skris if a 3DES key was generated with a 0 initial byte. Include 248576866Skris PKCS12_BROKEN_KEYGEN compilation option to retain the old 248676866Skris (but broken) behaviour. 248776866Skris [Steve Henson] 248876866Skris 248976866Skris *) Enhance bctest to search for a working bc along $PATH and print 249076866Skris it when found. 249176866Skris [Tim Rice <tim@multitalents.net> via Richard Levitte] 249276866Skris 249376866Skris *) Fix memory leaks in err.c: free err_data string if necessary; 249476866Skris don't write to the wrong index in ERR_set_error_data. 249576866Skris [Bodo Moeller] 249676866Skris 249776866Skris *) Implement ssl23_peek (analogous to ssl23_read), which previously 249876866Skris did not exist. 249976866Skris [Bodo Moeller] 250076866Skris 250176866Skris *) Replace rdtsc with _emit statements for VC++ version 5. 250276866Skris [Jeremy Cooper <jeremy@baymoo.org>] 250376866Skris 250476866Skris *) Make it possible to reuse SSLv2 sessions. 250576866Skris [Richard Levitte] 250676866Skris 250776866Skris *) In copy_email() check for >= 0 as a return value for 250876866Skris X509_NAME_get_index_by_NID() since 0 is a valid index. 250976866Skris [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 251076866Skris 251176866Skris *) Avoid coredump with unsupported or invalid public keys by checking if 251276866Skris X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 251376866Skris PKCS7_verify() fails with non detached data. 251476866Skris [Steve Henson] 251576866Skris 251676866Skris *) Don't use getenv in library functions when run as setuid/setgid. 251776866Skris New function OPENSSL_issetugid(). 251876866Skris [Ulf Moeller] 251976866Skris 252076866Skris *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 252176866Skris due to incorrect handling of multi-threading: 252276866Skris 252376866Skris 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 252476866Skris 252576866Skris 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 252676866Skris 252776866Skris 3. Count how many times MemCheck_off() has been called so that 252876866Skris nested use can be treated correctly. This also avoids 252976866Skris inband-signalling in the previous code (which relied on the 253076866Skris assumption that thread ID 0 is impossible). 253176866Skris [Bodo Moeller] 253276866Skris 253376866Skris *) Add "-rand" option also to s_client and s_server. 253476866Skris [Lutz Jaenicke] 253576866Skris 253676866Skris *) Fix CPU detection on Irix 6.x. 253776866Skris [Kurt Hockenbury <khockenb@stevens-tech.edu> and 253876866Skris "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 253976866Skris 254076866Skris *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 254176866Skris was empty. 254276866Skris [Steve Henson] 254376866Skris [This change does not apply to 0.9.7.] 254476866Skris 254576866Skris *) Use the cached encoding of an X509_NAME structure rather than 254676866Skris copying it. This is apparently the reason for the libsafe "errors" 254776866Skris but the code is actually correct. 254876866Skris [Steve Henson] 254976866Skris 255072613Skris *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 255172613Skris Bleichenbacher's DSA attack. 255276866Skris Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 255376866Skris to be set and top=0 forces the highest bit to be set; top=-1 is new 255476866Skris and leaves the highest bit random. 255576866Skris [Ulf Moeller, Bodo Moeller] 255672613Skris 255772613Skris *) In the NCONF_...-based implementations for CONF_... queries 255872613Skris (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 255972613Skris a temporary CONF structure with the data component set to NULL 256072613Skris (which gives segmentation faults in lh_retrieve). 256172613Skris Instead, use NULL for the CONF pointer in CONF_get_string and 256272613Skris CONF_get_number (which may use environment variables) and directly 256372613Skris return NULL from CONF_get_section. 256472613Skris [Bodo Moeller] 256572613Skris 256672613Skris *) Fix potential buffer overrun for EBCDIC. 256772613Skris [Ulf Moeller] 256872613Skris 256972613Skris *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 257072613Skris keyUsage if basicConstraints absent for a CA. 257172613Skris [Steve Henson] 257272613Skris 257372613Skris *) Make SMIME_write_PKCS7() write mail header values with a format that 257472613Skris is more generally accepted (no spaces before the semicolon), since 257572613Skris some programs can't parse those values properly otherwise. Also make 257672613Skris sure BIO's that break lines after each write do not create invalid 257772613Skris headers. 257872613Skris [Richard Levitte] 257972613Skris 258072613Skris *) Make the CRL encoding routines work with empty SEQUENCE OF. The 258172613Skris macros previously used would not encode an empty SEQUENCE OF 258272613Skris and break the signature. 258372613Skris [Steve Henson] 258476866Skris [This change does not apply to 0.9.7.] 258572613Skris 258672613Skris *) Zero the premaster secret after deriving the master secret in 258772613Skris DH ciphersuites. 258872613Skris [Steve Henson] 258972613Skris 259072613Skris *) Add some EVP_add_digest_alias registrations (as found in 259172613Skris OpenSSL_add_all_digests()) to SSL_library_init() 259272613Skris aka OpenSSL_add_ssl_algorithms(). This provides improved 259372613Skris compatibility with peers using X.509 certificates 259472613Skris with unconventional AlgorithmIdentifier OIDs. 259572613Skris [Bodo Moeller] 259672613Skris 259772613Skris *) Fix for Irix with NO_ASM. 259872613Skris ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 259972613Skris 260072613Skris *) ./config script fixes. 260172613Skris [Ulf Moeller, Richard Levitte] 260272613Skris 260372613Skris *) Fix 'openssl passwd -1'. 260472613Skris [Bodo Moeller] 260572613Skris 260672613Skris *) Change PKCS12_key_gen_asc() so it can cope with non null 260772613Skris terminated strings whose length is passed in the passlen 260872613Skris parameter, for example from PEM callbacks. This was done 260972613Skris by adding an extra length parameter to asc2uni(). 261072613Skris [Steve Henson, reported by <oddissey@samsung.co.kr>] 261172613Skris 261272613Skris *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 261372613Skris call failed, free the DSA structure. 261472613Skris [Bodo Moeller] 261572613Skris 261672613Skris *) Fix to uni2asc() to cope with zero length Unicode strings. 261772613Skris These are present in some PKCS#12 files. 261872613Skris [Steve Henson] 261972613Skris 262072613Skris *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 262172613Skris Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 262272613Skris when writing a 32767 byte record. 262372613Skris [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 262472613Skris 262572613Skris *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 262672613Skris obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 262772613Skris 262872613Skris (RSA objects have a reference count access to which is protected 262972613Skris by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 263072613Skris so they are meant to be shared between threads.) 263172613Skris [Bodo Moeller, Geoff Thorpe; original patch submitted by 263272613Skris "Reddie, Steven" <Steven.Reddie@ca.com>] 263372613Skris 263472613Skris *) Fix a deadlock in CRYPTO_mem_leaks(). 263572613Skris [Bodo Moeller] 263672613Skris 263776866Skris *) Use better test patterns in bntest. 263876866Skris [Ulf M�ller] 263976866Skris 264072613Skris *) rand_win.c fix for Borland C. 264172613Skris [Ulf M�ller] 264272613Skris 264372613Skris *) BN_rshift bugfix for n == 0. 264472613Skris [Bodo Moeller] 264572613Skris 264676866Skris *) Add a 'bctest' script that checks for some known 'bc' bugs 264776866Skris so that 'make test' does not abort just because 'bc' is broken. 264876866Skris [Bodo Moeller] 264976866Skris 265072613Skris *) Store verify_result within SSL_SESSION also for client side to 265172613Skris avoid potential security hole. (Re-used sessions on the client side 265272613Skris always resulted in verify_result==X509_V_OK, not using the original 265372613Skris result of the server certificate verification.) 265472613Skris [Lutz Jaenicke] 265572613Skris 265672613Skris *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 265772613Skris SSL3_RT_APPLICATION_DATA, return 0. 265872613Skris Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 265972613Skris [Bodo Moeller] 266072613Skris 266172613Skris *) Fix SSL_peek: 266272613Skris Both ssl2_peek and ssl3_peek, which were totally broken in earlier 266372613Skris releases, have been re-implemented by renaming the previous 266472613Skris implementations of ssl2_read and ssl3_read to ssl2_read_internal 266572613Skris and ssl3_read_internal, respectively, and adding 'peek' parameters 266672613Skris to them. The new ssl[23]_{read,peek} functions are calls to 266772613Skris ssl[23]_read_internal with the 'peek' flag set appropriately. 266872613Skris A 'peek' parameter has also been added to ssl3_read_bytes, which 266972613Skris does the actual work for ssl3_read_internal. 267072613Skris [Bodo Moeller] 267172613Skris 267276866Skris *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 267376866Skris the method-specific "init()" handler. Also clean up ex_data after 267476866Skris calling the method-specific "finish()" handler. Previously, this was 267576866Skris happening the other way round. 267676866Skris [Geoff Thorpe] 267776866Skris 267872613Skris *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 267972613Skris The previous value, 12, was not always sufficient for BN_mod_exp(). 268072613Skris [Bodo Moeller] 268172613Skris 268276866Skris *) Make sure that shared libraries get the internal name engine with 268376866Skris the full version number and not just 0. This should mark the 268476866Skris shared libraries as not backward compatible. Of course, this should 268576866Skris be changed again when we can guarantee backward binary compatibility. 268676866Skris [Richard Levitte] 268776866Skris 268872613Skris *) Fix typo in get_cert_by_subject() in by_dir.c 268972613Skris [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 269072613Skris 269176866Skris *) Rework the system to generate shared libraries: 269276866Skris 269376866Skris - Make note of the expected extension for the shared libraries and 269476866Skris if there is a need for symbolic links from for example libcrypto.so.0 269576866Skris to libcrypto.so.0.9.7. There is extended info in Configure for 269676866Skris that. 269776866Skris 269876866Skris - Make as few rebuilds of the shared libraries as possible. 269976866Skris 270076866Skris - Still avoid linking the OpenSSL programs with the shared libraries. 270176866Skris 270276866Skris - When installing, install the shared libraries separately from the 270376866Skris static ones. 270476866Skris [Richard Levitte] 270576866Skris 270672613Skris *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 270772613Skris 270872613Skris Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 270972613Skris and not in SSL_clear because the latter is also used by the 271072613Skris accept/connect functions; previously, the settings made by 271172613Skris SSL_set_read_ahead would be lost during the handshake. 271272613Skris [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 271372613Skris 271472613Skris *) Correct util/mkdef.pl to be selective about disabled algorithms. 271572613Skris Previously, it would create entries for disableed algorithms no 271672613Skris matter what. 271772613Skris [Richard Levitte] 271872613Skris 271972613Skris *) Added several new manual pages for SSL_* function. 272072613Skris [Lutz Jaenicke] 272172613Skris 272268651Skris Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 272368651Skris 272468651Skris *) In ssl23_get_client_hello, generate an error message when faced 272568651Skris with an initial SSL 3.0/TLS record that is too small to contain the 272668651Skris first two bytes of the ClientHello message, i.e. client_version. 272768651Skris (Note that this is a pathologic case that probably has never happened 272868651Skris in real life.) The previous approach was to use the version number 272968651Skris from the record header as a substitute; but our protocol choice 273068651Skris should not depend on that one because it is not authenticated 273168651Skris by the Finished messages. 273268651Skris [Bodo Moeller] 273368651Skris 273468651Skris *) More robust randomness gathering functions for Windows. 273568651Skris [Jeffrey Altman <jaltman@columbia.edu>] 273668651Skris 273768651Skris *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 273868651Skris not set then we don't setup the error code for issuer check errors 273968651Skris to avoid possibly overwriting other errors which the callback does 274068651Skris handle. If an application does set the flag then we assume it knows 274168651Skris what it is doing and can handle the new informational codes 274268651Skris appropriately. 274368651Skris [Steve Henson] 274468651Skris 274568651Skris *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 274668651Skris a general "ANY" type, as such it should be able to decode anything 274768651Skris including tagged types. However it didn't check the class so it would 274868651Skris wrongly interpret tagged types in the same way as their universal 274968651Skris counterpart and unknown types were just rejected. Changed so that the 275068651Skris tagged and unknown types are handled in the same way as a SEQUENCE: 275168651Skris that is the encoding is stored intact. There is also a new type 275268651Skris "V_ASN1_OTHER" which is used when the class is not universal, in this 275368651Skris case we have no idea what the actual type is so we just lump them all 275468651Skris together. 275568651Skris [Steve Henson] 275668651Skris 275768651Skris *) On VMS, stdout may very well lead to a file that is written to 275868651Skris in a record-oriented fashion. That means that every write() will 275968651Skris write a separate record, which will be read separately by the 276068651Skris programs trying to read from it. This can be very confusing. 276168651Skris 276268651Skris The solution is to put a BIO filter in the way that will buffer 276368651Skris text until a linefeed is reached, and then write everything a 276468651Skris line at a time, so every record written will be an actual line, 276568651Skris not chunks of lines and not (usually doesn't happen, but I've 276668651Skris seen it once) several lines in one record. BIO_f_linebuffer() is 276768651Skris the answer. 276868651Skris 276968651Skris Currently, it's a VMS-only method, because that's where it has 277068651Skris been tested well enough. 277168651Skris [Richard Levitte] 277268651Skris 277368651Skris *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 277468651Skris it can return incorrect results. 277568651Skris (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 277668651Skris but it was in 0.9.6-beta[12].) 277768651Skris [Bodo Moeller] 277868651Skris 277968651Skris *) Disable the check for content being present when verifying detached 278068651Skris signatures in pk7_smime.c. Some versions of Netscape (wrongly) 278168651Skris include zero length content when signing messages. 278268651Skris [Steve Henson] 278368651Skris 278468651Skris *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 278568651Skris BIO_ctrl (for BIO pairs). 278668651Skris [Bodo M�ller] 278768651Skris 278868651Skris *) Add DSO method for VMS. 278968651Skris [Richard Levitte] 279068651Skris 279168651Skris *) Bug fix: Montgomery multiplication could produce results with the 279268651Skris wrong sign. 279368651Skris [Ulf M�ller] 279468651Skris 279568651Skris *) Add RPM specification openssl.spec and modify it to build three 279668651Skris packages. The default package contains applications, application 279768651Skris documentation and run-time libraries. The devel package contains 279868651Skris include files, static libraries and function documentation. The 279968651Skris doc package contains the contents of the doc directory. The original 280068651Skris openssl.spec was provided by Damien Miller <djm@mindrot.org>. 280168651Skris [Richard Levitte] 280268651Skris 280368651Skris *) Add a large number of documentation files for many SSL routines. 280468651Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 280568651Skris 280668651Skris *) Add a configuration entry for Sony News 4. 280768651Skris [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 280868651Skris 280968651Skris *) Don't set the two most significant bits to one when generating a 281068651Skris random number < q in the DSA library. 281168651Skris [Ulf M�ller] 281268651Skris 281368651Skris *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 281468651Skris behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 281568651Skris the underlying transport is blocking) if a handshake took place. 281668651Skris (The default behaviour is needed by applications such as s_client 281768651Skris and s_server that use select() to determine when to use SSL_read; 281868651Skris but for applications that know in advance when to expect data, it 281968651Skris just makes things more complicated.) 282068651Skris [Bodo Moeller] 282168651Skris 282268651Skris *) Add RAND_egd_bytes(), which gives control over the number of bytes read 282368651Skris from EGD. 282468651Skris [Ben Laurie] 282568651Skris 282668651Skris *) Add a few more EBCDIC conditionals that make `req' and `x509' 282768651Skris work better on such systems. 282868651Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 282968651Skris 283068651Skris *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 283168651Skris Update PKCS12_parse() so it copies the friendlyName and the 283268651Skris keyid to the certificates aux info. 283368651Skris [Steve Henson] 283468651Skris 283568651Skris *) Fix bug in PKCS7_verify() which caused an infinite loop 283668651Skris if there was more than one signature. 283768651Skris [Sven Uszpelkat <su@celocom.de>] 283868651Skris 283968651Skris *) Major change in util/mkdef.pl to include extra information 284068651Skris about each symbol, as well as presentig variables as well 284168651Skris as functions. This change means that there's n more need 284268651Skris to rebuild the .num files when some algorithms are excluded. 284368651Skris [Richard Levitte] 284468651Skris 284568651Skris *) Allow the verify time to be set by an application, 284668651Skris rather than always using the current time. 284768651Skris [Steve Henson] 284868651Skris 284968651Skris *) Phase 2 verify code reorganisation. The certificate 285068651Skris verify code now looks up an issuer certificate by a 285168651Skris number of criteria: subject name, authority key id 285268651Skris and key usage. It also verifies self signed certificates 285368651Skris by the same criteria. The main comparison function is 285468651Skris X509_check_issued() which performs these checks. 285568651Skris 285668651Skris Lot of changes were necessary in order to support this 285768651Skris without completely rewriting the lookup code. 285868651Skris 285968651Skris Authority and subject key identifier are now cached. 286068651Skris 286168651Skris The LHASH 'certs' is X509_STORE has now been replaced 286268651Skris by a STACK_OF(X509_OBJECT). This is mainly because an 286368651Skris LHASH can't store or retrieve multiple objects with 286468651Skris the same hash value. 286568651Skris 286668651Skris As a result various functions (which were all internal 286768651Skris use only) have changed to handle the new X509_STORE 286868651Skris structure. This will break anything that messed round 286968651Skris with X509_STORE internally. 287068651Skris 287168651Skris The functions X509_STORE_add_cert() now checks for an 287268651Skris exact match, rather than just subject name. 287368651Skris 287468651Skris The X509_STORE API doesn't directly support the retrieval 287568651Skris of multiple certificates matching a given criteria, however 287668651Skris this can be worked round by performing a lookup first 287768651Skris (which will fill the cache with candidate certificates) 287868651Skris and then examining the cache for matches. This is probably 287968651Skris the best we can do without throwing out X509_LOOKUP 288068651Skris entirely (maybe later...). 288168651Skris 288268651Skris The X509_VERIFY_CTX structure has been enhanced considerably. 288368651Skris 288468651Skris All certificate lookup operations now go via a get_issuer() 288568651Skris callback. Although this currently uses an X509_STORE it 288668651Skris can be replaced by custom lookups. This is a simple way 288768651Skris to bypass the X509_STORE hackery necessary to make this 288868651Skris work and makes it possible to use more efficient techniques 288968651Skris in future. A very simple version which uses a simple 289068651Skris STACK for its trusted certificate store is also provided 289168651Skris using X509_STORE_CTX_trusted_stack(). 289268651Skris 289368651Skris The verify_cb() and verify() callbacks now have equivalents 289468651Skris in the X509_STORE_CTX structure. 289568651Skris 289668651Skris X509_STORE_CTX also has a 'flags' field which can be used 289768651Skris to customise the verify behaviour. 289868651Skris [Steve Henson] 289968651Skris 290068651Skris *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 290168651Skris excludes S/MIME capabilities. 290268651Skris [Steve Henson] 290368651Skris 290468651Skris *) When a certificate request is read in keep a copy of the 290568651Skris original encoding of the signed data and use it when outputing 290668651Skris again. Signatures then use the original encoding rather than 290768651Skris a decoded, encoded version which may cause problems if the 290868651Skris request is improperly encoded. 290968651Skris [Steve Henson] 291068651Skris 291168651Skris *) For consistency with other BIO_puts implementations, call 291268651Skris buffer_write(b, ...) directly in buffer_puts instead of calling 291368651Skris BIO_write(b, ...). 291468651Skris 291568651Skris In BIO_puts, increment b->num_write as in BIO_write. 291668651Skris [Peter.Sylvester@EdelWeb.fr] 291768651Skris 291868651Skris *) Fix BN_mul_word for the case where the word is 0. (We have to use 291968651Skris BN_zero, we may not return a BIGNUM with an array consisting of 292068651Skris words set to zero.) 292168651Skris [Bodo Moeller] 292268651Skris 292368651Skris *) Avoid calling abort() from within the library when problems are 292468651Skris detected, except if preprocessor symbols have been defined 292568651Skris (such as REF_CHECK, BN_DEBUG etc.). 292668651Skris [Bodo Moeller] 292768651Skris 292868651Skris *) New openssl application 'rsautl'. This utility can be 292968651Skris used for low level RSA operations. DER public key 293068651Skris BIO/fp routines also added. 293168651Skris [Steve Henson] 293268651Skris 293368651Skris *) New Configure entry and patches for compiling on QNX 4. 293468651Skris [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 293568651Skris 293668651Skris *) A demo state-machine implementation was sponsored by 293768651Skris Nuron (http://www.nuron.com/) and is now available in 293868651Skris demos/state_machine. 293968651Skris [Ben Laurie] 294068651Skris 294168651Skris *) New options added to the 'dgst' utility for signature 294268651Skris generation and verification. 294368651Skris [Steve Henson] 294468651Skris 294568651Skris *) Unrecognized PKCS#7 content types are now handled via a 294668651Skris catch all ASN1_TYPE structure. This allows unsupported 294768651Skris types to be stored as a "blob" and an application can 294868651Skris encode and decode it manually. 294968651Skris [Steve Henson] 295068651Skris 295168651Skris *) Fix various signed/unsigned issues to make a_strex.c 295268651Skris compile under VC++. 295368651Skris [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 295468651Skris 295568651Skris *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 295668651Skris length if passed a buffer. ASN1_INTEGER_to_BN failed 295768651Skris if passed a NULL BN and its argument was negative. 295868651Skris [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 295968651Skris 296068651Skris *) Modification to PKCS#7 encoding routines to output definite 296168651Skris length encoding. Since currently the whole structures are in 296268651Skris memory there's not real point in using indefinite length 296368651Skris constructed encoding. However if OpenSSL is compiled with 296468651Skris the flag PKCS7_INDEFINITE_ENCODING the old form is used. 296568651Skris [Steve Henson] 296668651Skris 296768651Skris *) Added BIO_vprintf() and BIO_vsnprintf(). 296868651Skris [Richard Levitte] 296968651Skris 297068651Skris *) Added more prefixes to parse for in the the strings written 297168651Skris through a logging bio, to cover all the levels that are available 297268651Skris through syslog. The prefixes are now: 297368651Skris 297468651Skris PANIC, EMERG, EMR => LOG_EMERG 297568651Skris ALERT, ALR => LOG_ALERT 297668651Skris CRIT, CRI => LOG_CRIT 297768651Skris ERROR, ERR => LOG_ERR 297868651Skris WARNING, WARN, WAR => LOG_WARNING 297968651Skris NOTICE, NOTE, NOT => LOG_NOTICE 298068651Skris INFO, INF => LOG_INFO 298168651Skris DEBUG, DBG => LOG_DEBUG 298268651Skris 298368651Skris and as before, if none of those prefixes are present at the 298468651Skris beginning of the string, LOG_ERR is chosen. 298568651Skris 298668651Skris On Win32, the LOG_* levels are mapped according to this: 298768651Skris 298868651Skris LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 298968651Skris LOG_WARNING => EVENTLOG_WARNING_TYPE 299068651Skris LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 299168651Skris 299268651Skris [Richard Levitte] 299368651Skris 299468651Skris *) Made it possible to reconfigure with just the configuration 299568651Skris argument "reconf" or "reconfigure". The command line arguments 299668651Skris are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 299768651Skris and are retrieved from there when reconfiguring. 299868651Skris [Richard Levitte] 299968651Skris 300068651Skris *) MD4 implemented. 300168651Skris [Assar Westerlund <assar@sics.se>, Richard Levitte] 300268651Skris 300368651Skris *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 300468651Skris [Richard Levitte] 300568651Skris 300668651Skris *) The obj_dat.pl script was messing up the sorting of object 300768651Skris names. The reason was that it compared the quoted version 300868651Skris of strings as a result "OCSP" > "OCSP Signing" because 300968651Skris " > SPACE. Changed script to store unquoted versions of 301068651Skris names and add quotes on output. It was also omitting some 301168651Skris names from the lookup table if they were given a default 301268651Skris value (that is if SN is missing it is given the same 301368651Skris value as LN and vice versa), these are now added on the 301468651Skris grounds that if an object has a name we should be able to 301568651Skris look it up. Finally added warning output when duplicate 301668651Skris short or long names are found. 301768651Skris [Steve Henson] 301868651Skris 301968651Skris *) Changes needed for Tandem NSK. 302068651Skris [Scott Uroff <scott@xypro.com>] 302168651Skris 302268651Skris *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 302368651Skris RSA_padding_check_SSLv23(), special padding was never detected 302468651Skris and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 302568651Skris version rollback attacks was not effective. 302668651Skris 302768651Skris In s23_clnt.c, don't use special rollback-attack detection padding 302868651Skris (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 302968651Skris client; similarly, in s23_srvr.c, don't do the rollback check if 303068651Skris SSL 2.0 is the only protocol enabled in the server. 303168651Skris [Bodo Moeller] 303268651Skris 303368651Skris *) Make it possible to get hexdumps of unprintable data with 'openssl 303468651Skris asn1parse'. By implication, the functions ASN1_parse_dump() and 303568651Skris BIO_dump_indent() are added. 303668651Skris [Richard Levitte] 303768651Skris 303868651Skris *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 303968651Skris these print out strings and name structures based on various 304068651Skris flags including RFC2253 support and proper handling of 304168651Skris multibyte characters. Added options to the 'x509' utility 304268651Skris to allow the various flags to be set. 304368651Skris [Steve Henson] 304468651Skris 304568651Skris *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 304668651Skris Also change the functions X509_cmp_current_time() and 304768651Skris X509_gmtime_adj() work with an ASN1_TIME structure, 304868651Skris this will enable certificates using GeneralizedTime in validity 304968651Skris dates to be checked. 305068651Skris [Steve Henson] 305168651Skris 305268651Skris *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 305368651Skris negative public key encodings) on by default, 305468651Skris NO_NEG_PUBKEY_BUG can be set to disable it. 305568651Skris [Steve Henson] 305668651Skris 305768651Skris *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 305868651Skris content octets. An i2c_ASN1_OBJECT is unnecessary because 305968651Skris the encoding can be trivially obtained from the structure. 306068651Skris [Steve Henson] 306168651Skris 306268651Skris *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 306368651Skris not read locks (CRYPTO_r_[un]lock). 306468651Skris [Bodo Moeller] 306568651Skris 306668651Skris *) A first attempt at creating official support for shared 306768651Skris libraries through configuration. I've kept it so the 306868651Skris default is static libraries only, and the OpenSSL programs 306968651Skris are always statically linked for now, but there are 307068651Skris preparations for dynamic linking in place. 307189837Skris This has been tested on Linux and Tru64. 307268651Skris [Richard Levitte] 307368651Skris 307468651Skris *) Randomness polling function for Win9x, as described in: 307568651Skris Peter Gutmann, Software Generation of Practically Strong 307668651Skris Random Numbers. 307768651Skris [Ulf M�ller] 307868651Skris 307968651Skris *) Fix so PRNG is seeded in req if using an already existing 308068651Skris DSA key. 308168651Skris [Steve Henson] 308268651Skris 308368651Skris *) New options to smime application. -inform and -outform 308468651Skris allow alternative formats for the S/MIME message including 308568651Skris PEM and DER. The -content option allows the content to be 308668651Skris specified separately. This should allow things like Netscape 308768651Skris form signing output easier to verify. 308868651Skris [Steve Henson] 308968651Skris 309068651Skris *) Fix the ASN1 encoding of tags using the 'long form'. 309168651Skris [Steve Henson] 309268651Skris 309368651Skris *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 309468651Skris STRING types. These convert content octets to and from the 309568651Skris underlying type. The actual tag and length octets are 309668651Skris already assumed to have been read in and checked. These 309768651Skris are needed because all other string types have virtually 309868651Skris identical handling apart from the tag. By having versions 309968651Skris of the ASN1 functions that just operate on content octets 310068651Skris IMPLICIT tagging can be handled properly. It also allows 310168651Skris the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 310268651Skris and ASN1_INTEGER are identical apart from the tag. 310368651Skris [Steve Henson] 310468651Skris 310568651Skris *) Change the handling of OID objects as follows: 310668651Skris 310768651Skris - New object identifiers are inserted in objects.txt, following 310868651Skris the syntax given in objects.README. 310968651Skris - objects.pl is used to process obj_mac.num and create a new 311068651Skris obj_mac.h. 311168651Skris - obj_dat.pl is used to create a new obj_dat.h, using the data in 311268651Skris obj_mac.h. 311368651Skris 311468651Skris This is currently kind of a hack, and the perl code in objects.pl 311568651Skris isn't very elegant, but it works as I intended. The simplest way 311668651Skris to check that it worked correctly is to look in obj_dat.h and 311768651Skris check the array nid_objs and make sure the objects haven't moved 311868651Skris around (this is important!). Additions are OK, as well as 311968651Skris consistent name changes. 312068651Skris [Richard Levitte] 312168651Skris 312268651Skris *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 312368651Skris [Bodo Moeller] 312468651Skris 312568651Skris *) Addition of the command line parameter '-rand file' to 'openssl req'. 312668651Skris The given file adds to whatever has already been seeded into the 312768651Skris random pool through the RANDFILE configuration file option or 312868651Skris environment variable, or the default random state file. 312968651Skris [Richard Levitte] 313068651Skris 313168651Skris *) mkstack.pl now sorts each macro group into lexical order. 313268651Skris Previously the output order depended on the order the files 313368651Skris appeared in the directory, resulting in needless rewriting 313468651Skris of safestack.h . 313568651Skris [Steve Henson] 313668651Skris 313768651Skris *) Patches to make OpenSSL compile under Win32 again. Mostly 313868651Skris work arounds for the VC++ problem that it treats func() as 313968651Skris func(void). Also stripped out the parts of mkdef.pl that 314068651Skris added extra typesafe functions: these no longer exist. 314168651Skris [Steve Henson] 314268651Skris 314368651Skris *) Reorganisation of the stack code. The macros are now all 314468651Skris collected in safestack.h . Each macro is defined in terms of 314568651Skris a "stack macro" of the form SKM_<name>(type, a, b). The 314668651Skris DEBUG_SAFESTACK is now handled in terms of function casts, 314768651Skris this has the advantage of retaining type safety without the 314868651Skris use of additional functions. If DEBUG_SAFESTACK is not defined 314968651Skris then the non typesafe macros are used instead. Also modified the 315068651Skris mkstack.pl script to handle the new form. Needs testing to see 315168651Skris if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 315268651Skris the default if no major problems. Similar behaviour for ASN1_SET_OF 315368651Skris and PKCS12_STACK_OF. 315468651Skris [Steve Henson] 315568651Skris 315668651Skris *) When some versions of IIS use the 'NET' form of private key the 315768651Skris key derivation algorithm is different. Normally MD5(password) is 315868651Skris used as a 128 bit RC4 key. In the modified case 315968651Skris MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 316068651Skris new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 316168651Skris as the old Netscape_RSA functions except they have an additional 316268651Skris 'sgckey' parameter which uses the modified algorithm. Also added 316368651Skris an -sgckey command line option to the rsa utility. Thanks to 316468651Skris Adrian Peck <bertie@ncipher.com> for posting details of the modified 316568651Skris algorithm to openssl-dev. 316668651Skris [Steve Henson] 316768651Skris 316868651Skris *) The evp_local.h macros were using 'c.##kname' which resulted in 316968651Skris invalid expansion on some systems (SCO 5.0.5 for example). 317068651Skris Corrected to 'c.kname'. 317168651Skris [Phillip Porch <root@theporch.com>] 317268651Skris 317368651Skris *) New X509_get1_email() and X509_REQ_get1_email() functions that return 317468651Skris a STACK of email addresses from a certificate or request, these look 317568651Skris in the subject name and the subject alternative name extensions and 317668651Skris omit any duplicate addresses. 317768651Skris [Steve Henson] 317868651Skris 317968651Skris *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 318068651Skris This makes DSA verification about 2 % faster. 318168651Skris [Bodo Moeller] 318268651Skris 318368651Skris *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 318468651Skris (meaning that now 2^5 values will be precomputed, which is only 4 KB 318568651Skris plus overhead for 1024 bit moduli). 318668651Skris This makes exponentiations about 0.5 % faster for 1024 bit 318768651Skris exponents (as measured by "openssl speed rsa2048"). 318868651Skris [Bodo Moeller] 318968651Skris 319068651Skris *) Rename memory handling macros to avoid conflicts with other 319168651Skris software: 319268651Skris Malloc => OPENSSL_malloc 319368651Skris Malloc_locked => OPENSSL_malloc_locked 319468651Skris Realloc => OPENSSL_realloc 319568651Skris Free => OPENSSL_free 319668651Skris [Richard Levitte] 319768651Skris 319868651Skris *) New function BN_mod_exp_mont_word for small bases (roughly 15% 319968651Skris faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 320068651Skris [Bodo Moeller] 320168651Skris 320268651Skris *) CygWin32 support. 320368651Skris [John Jarvie <jjarvie@newsguy.com>] 320468651Skris 320568651Skris *) The type-safe stack code has been rejigged. It is now only compiled 320668651Skris in when OpenSSL is configured with the DEBUG_SAFESTACK option and 320768651Skris by default all type-specific stack functions are "#define"d back to 320868651Skris standard stack functions. This results in more streamlined output 320968651Skris but retains the type-safety checking possibilities of the original 321068651Skris approach. 321168651Skris [Geoff Thorpe] 321268651Skris 321368651Skris *) The STACK code has been cleaned up, and certain type declarations 321468651Skris that didn't make a lot of sense have been brought in line. This has 321568651Skris also involved a cleanup of sorts in safestack.h to more correctly 321668651Skris map type-safe stack functions onto their plain stack counterparts. 321768651Skris This work has also resulted in a variety of "const"ifications of 321868651Skris lots of the code, especially "_cmp" operations which should normally 321968651Skris be prototyped with "const" parameters anyway. 322068651Skris [Geoff Thorpe] 322168651Skris 322268651Skris *) When generating bytes for the first time in md_rand.c, 'stir the pool' 322368651Skris by seeding with STATE_SIZE dummy bytes (with zero entropy count). 322468651Skris (The PRNG state consists of two parts, the large pool 'state' and 'md', 322568651Skris where all of 'md' is used each time the PRNG is used, but 'state' 322668651Skris is used only indexed by a cyclic counter. As entropy may not be 322768651Skris well distributed from the beginning, 'md' is important as a 322868651Skris chaining variable. However, the output function chains only half 322968651Skris of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 323068651Skris all of 'md', and seeding with STATE_SIZE dummy bytes will result 323168651Skris in all of 'state' being rewritten, with the new values depending 323268651Skris on virtually all of 'md'. This overcomes the 80 bit limitation.) 323368651Skris [Bodo Moeller] 323468651Skris 323568651Skris *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 323668651Skris the handshake is continued after ssl_verify_cert_chain(); 323768651Skris otherwise, if SSL_VERIFY_NONE is set, remaining error codes 323868651Skris can lead to 'unexplainable' connection aborts later. 323968651Skris [Bodo Moeller; problem tracked down by Lutz Jaenicke] 324068651Skris 324168651Skris *) Major EVP API cipher revision. 324268651Skris Add hooks for extra EVP features. This allows various cipher 324368651Skris parameters to be set in the EVP interface. Support added for variable 324468651Skris key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 324568651Skris setting of RC2 and RC5 parameters. 324668651Skris 324768651Skris Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 324868651Skris ciphers. 324968651Skris 325068651Skris Remove lots of duplicated code from the EVP library. For example *every* 325168651Skris cipher init() function handles the 'iv' in the same way according to the 325268651Skris cipher mode. They also all do nothing if the 'key' parameter is NULL and 325368651Skris for CFB and OFB modes they zero ctx->num. 325468651Skris 325568651Skris New functionality allows removal of S/MIME code RC2 hack. 325668651Skris 325768651Skris Most of the routines have the same form and so can be declared in terms 325868651Skris of macros. 325968651Skris 326068651Skris By shifting this to the top level EVP_CipherInit() it can be removed from 326168651Skris all individual ciphers. If the cipher wants to handle IVs or keys 326268651Skris differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 326368651Skris flags. 326468651Skris 326568651Skris Change lots of functions like EVP_EncryptUpdate() to now return a 326668651Skris value: although software versions of the algorithms cannot fail 326768651Skris any installed hardware versions can. 326868651Skris [Steve Henson] 326968651Skris 327068651Skris *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 327168651Skris this option is set, tolerate broken clients that send the negotiated 327268651Skris protocol version number instead of the requested protocol version 327368651Skris number. 327468651Skris [Bodo Moeller] 327568651Skris 327668651Skris *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 327768651Skris i.e. non-zero for export ciphersuites, zero otherwise. 327868651Skris Previous versions had this flag inverted, inconsistent with 327968651Skris rsa_tmp_cb (..._TMP_RSA_CB). 328068651Skris [Bodo Moeller; problem reported by Amit Chopra] 328168651Skris 328268651Skris *) Add missing DSA library text string. Work around for some IIS 328368651Skris key files with invalid SEQUENCE encoding. 328468651Skris [Steve Henson] 328568651Skris 328668651Skris *) Add a document (doc/standards.txt) that list all kinds of standards 328768651Skris and so on that are implemented in OpenSSL. 328868651Skris [Richard Levitte] 328968651Skris 329068651Skris *) Enhance c_rehash script. Old version would mishandle certificates 329168651Skris with the same subject name hash and wouldn't handle CRLs at all. 329268651Skris Added -fingerprint option to crl utility, to support new c_rehash 329368651Skris features. 329468651Skris [Steve Henson] 329568651Skris 329668651Skris *) Eliminate non-ANSI declarations in crypto.h and stack.h. 329768651Skris [Ulf M�ller] 329868651Skris 329968651Skris *) Fix for SSL server purpose checking. Server checking was 330068651Skris rejecting certificates which had extended key usage present 330168651Skris but no ssl client purpose. 330268651Skris [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 330368651Skris 330468651Skris *) Make PKCS#12 code work with no password. The PKCS#12 spec 330568651Skris is a little unclear about how a blank password is handled. 330668651Skris Since the password in encoded as a BMPString with terminating 330768651Skris double NULL a zero length password would end up as just the 330868651Skris double NULL. However no password at all is different and is 330968651Skris handled differently in the PKCS#12 key generation code. NS 331068651Skris treats a blank password as zero length. MSIE treats it as no 331168651Skris password on export: but it will try both on import. We now do 331268651Skris the same: PKCS12_parse() tries zero length and no password if 331368651Skris the password is set to "" or NULL (NULL is now a valid password: 331468651Skris it wasn't before) as does the pkcs12 application. 331568651Skris [Steve Henson] 331668651Skris 331768651Skris *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 331868651Skris perror when PEM_read_bio_X509_REQ fails, the error message must 331968651Skris be obtained from the error queue. 332068651Skris [Bodo Moeller] 332168651Skris 332268651Skris *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 332368651Skris it in ERR_remove_state if appropriate, and change ERR_get_state 332468651Skris accordingly to avoid race conditions (this is necessary because 332568651Skris thread_hash is no longer constant once set). 332668651Skris [Bodo Moeller] 332768651Skris 332868651Skris *) Bugfix for linux-elf makefile.one. 332968651Skris [Ulf M�ller] 333068651Skris 333168651Skris *) RSA_get_default_method() will now cause a default 333268651Skris RSA_METHOD to be chosen if one doesn't exist already. 333368651Skris Previously this was only set during a call to RSA_new() 333468651Skris or RSA_new_method(NULL) meaning it was possible for 333568651Skris RSA_get_default_method() to return NULL. 333668651Skris [Geoff Thorpe] 333768651Skris 333868651Skris *) Added native name translation to the existing DSO code 333968651Skris that will convert (if the flag to do so is set) filenames 334068651Skris that are sufficiently small and have no path information 334168651Skris into a canonical native form. Eg. "blah" converted to 334268651Skris "libblah.so" or "blah.dll" etc. 334368651Skris [Geoff Thorpe] 334468651Skris 334568651Skris *) New function ERR_error_string_n(e, buf, len) which is like 334668651Skris ERR_error_string(e, buf), but writes at most 'len' bytes 334768651Skris including the 0 terminator. For ERR_error_string_n, 'buf' 334868651Skris may not be NULL. 334968651Skris [Damien Miller <djm@mindrot.org>, Bodo Moeller] 335068651Skris 335168651Skris *) CONF library reworked to become more general. A new CONF 335268651Skris configuration file reader "class" is implemented as well as a 335368651Skris new functions (NCONF_*, for "New CONF") to handle it. The now 335468651Skris old CONF_* functions are still there, but are reimplemented to 335568651Skris work in terms of the new functions. Also, a set of functions 335668651Skris to handle the internal storage of the configuration data is 335768651Skris provided to make it easier to write new configuration file 335868651Skris reader "classes" (I can definitely see something reading a 335968651Skris configuration file in XML format, for example), called _CONF_*, 336068651Skris or "the configuration storage API"... 336168651Skris 336268651Skris The new configuration file reading functions are: 336368651Skris 336468651Skris NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 336568651Skris NCONF_get_section, NCONF_get_string, NCONF_get_numbre 336668651Skris 336768651Skris NCONF_default, NCONF_WIN32 336868651Skris 336968651Skris NCONF_dump_fp, NCONF_dump_bio 337068651Skris 337168651Skris NCONF_default and NCONF_WIN32 are method (or "class") choosers, 337268651Skris NCONF_new creates a new CONF object. This works in the same way 337368651Skris as other interfaces in OpenSSL, like the BIO interface. 337468651Skris NCONF_dump_* dump the internal storage of the configuration file, 337568651Skris which is useful for debugging. All other functions take the same 337668651Skris arguments as the old CONF_* functions wth the exception of the 337768651Skris first that must be a `CONF *' instead of a `LHASH *'. 337868651Skris 337968651Skris To make it easer to use the new classes with the old CONF_* functions, 338068651Skris the function CONF_set_default_method is provided. 338168651Skris [Richard Levitte] 338268651Skris 338368651Skris *) Add '-tls1' option to 'openssl ciphers', which was already 338468651Skris mentioned in the documentation but had not been implemented. 338568651Skris (This option is not yet really useful because even the additional 338668651Skris experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 338768651Skris [Bodo Moeller] 338868651Skris 338968651Skris *) Initial DSO code added into libcrypto for letting OpenSSL (and 339068651Skris OpenSSL-based applications) load shared libraries and bind to 339168651Skris them in a portable way. 339268651Skris [Geoff Thorpe, with contributions from Richard Levitte] 339368651Skris 339459191Skris Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 339559191Skris 339659191Skris *) Make sure _lrotl and _lrotr are only used with MSVC. 339759191Skris 339859191Skris *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 339959191Skris (the default implementation of RAND_status). 340059191Skris 340159191Skris *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 340259191Skris to '-clrext' (= clear extensions), as intended and documented. 340359191Skris [Bodo Moeller; inconsistency pointed out by Michael Attili 340459191Skris <attili@amaxo.com>] 340559191Skris 340659191Skris *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 340759191Skris was larger than the MD block size. 340859191Skris [Steve Henson, pointed out by Yost William <YostW@tce.com>] 340959191Skris 341059191Skris *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 341159191Skris fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 341259191Skris using the passed key: if the passed key was a private key the result 341359191Skris of X509_print(), for example, would be to print out all the private key 341459191Skris components. 341559191Skris [Steve Henson] 341659191Skris 341759191Skris *) des_quad_cksum() byte order bug fix. 341859191Skris [Ulf M�ller, using the problem description in krb4-0.9.7, where 341959191Skris the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 342059191Skris 342159191Skris *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 342259191Skris discouraged. 342359191Skris [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 342459191Skris 342559191Skris *) For easily testing in shell scripts whether some command 342659191Skris 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 342759191Skris returns with exit code 0 iff no command of the given name is available. 342859191Skris 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 342959191Skris the output goes to stdout and nothing is printed to stderr. 343059191Skris Additional arguments are always ignored. 343159191Skris 343259191Skris Since for each cipher there is a command of the same name, 343359191Skris the 'no-cipher' compilation switches can be tested this way. 343459191Skris 343559191Skris ('openssl no-XXX' is not able to detect pseudo-commands such 343659191Skris as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 343759191Skris [Bodo Moeller] 343859191Skris 343959191Skris *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 344059191Skris [Bodo Moeller] 344159191Skris 344259191Skris *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 344359191Skris is set; it will be thrown away anyway because each handshake creates 344459191Skris its own key. 344559191Skris ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 344659191Skris to parameters -- in previous versions (since OpenSSL 0.9.3) the 344759191Skris 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 344859191Skris you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 344959191Skris [Bodo Moeller] 345059191Skris 345159191Skris *) New s_client option -ign_eof: EOF at stdin is ignored, and 345259191Skris 'Q' and 'R' lose their special meanings (quit/renegotiate). 345359191Skris This is part of what -quiet does; unlike -quiet, -ign_eof 345459191Skris does not suppress any output. 345559191Skris [Richard Levitte] 345659191Skris 345759191Skris *) Add compatibility options to the purpose and trust code. The 345859191Skris purpose X509_PURPOSE_ANY is "any purpose" which automatically 345959191Skris accepts a certificate or CA, this was the previous behaviour, 346059191Skris with all the associated security issues. 346159191Skris 346259191Skris X509_TRUST_COMPAT is the old trust behaviour: only and 346359191Skris automatically trust self signed roots in certificate store. A 346459191Skris new trust setting X509_TRUST_DEFAULT is used to specify that 346559191Skris a purpose has no associated trust setting and it should instead 346659191Skris use the value in the default purpose. 346759191Skris [Steve Henson] 346859191Skris 346959191Skris *) Fix the PKCS#8 DSA private key code so it decodes keys again 347059191Skris and fix a memory leak. 347159191Skris [Steve Henson] 347259191Skris 347359191Skris *) In util/mkerr.pl (which implements 'make errors'), preserve 347459191Skris reason strings from the previous version of the .c file, as 347559191Skris the default to have only downcase letters (and digits) in 347659191Skris automatically generated reasons codes is not always appropriate. 347759191Skris [Bodo Moeller] 347859191Skris 347959191Skris *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 348059191Skris using strerror. Previously, ERR_reason_error_string() returned 348159191Skris library names as reason strings for SYSerr; but SYSerr is a special 348259191Skris case where small numbers are errno values, not library numbers. 348359191Skris [Bodo Moeller] 348459191Skris 348559191Skris *) Add '-dsaparam' option to 'openssl dhparam' application. This 348659191Skris converts DSA parameters into DH parameters. (When creating parameters, 348759191Skris DSA_generate_parameters is used.) 348859191Skris [Bodo Moeller] 348959191Skris 349059191Skris *) Include 'length' (recommended exponent length) in C code generated 349159191Skris by 'openssl dhparam -C'. 349259191Skris [Bodo Moeller] 349359191Skris 349459191Skris *) The second argument to set_label in perlasm was already being used 349559191Skris so couldn't be used as a "file scope" flag. Moved to third argument 349659191Skris which was free. 349759191Skris [Steve Henson] 349859191Skris 349959191Skris *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 350059191Skris instead of RAND_bytes for encryption IVs and salts. 350159191Skris [Bodo Moeller] 350259191Skris 350359191Skris *) Include RAND_status() into RAND_METHOD instead of implementing 350459191Skris it only for md_rand.c Otherwise replacing the PRNG by calling 350559191Skris RAND_set_rand_method would be impossible. 350659191Skris [Bodo Moeller] 350759191Skris 350859191Skris *) Don't let DSA_generate_key() enter an infinite loop if the random 350959191Skris number generation fails. 351059191Skris [Bodo Moeller] 351159191Skris 351259191Skris *) New 'rand' application for creating pseudo-random output. 351359191Skris [Bodo Moeller] 351459191Skris 351559191Skris *) Added configuration support for Linux/IA64 351659191Skris [Rolf Haberrecker <rolf@suse.de>] 351759191Skris 351859191Skris *) Assembler module support for Mingw32. 351959191Skris [Ulf M�ller] 352059191Skris 352159191Skris *) Shared library support for HPUX (in shlib/). 352259191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 352359191Skris 352459191Skris *) Shared library support for Solaris gcc. 352559191Skris [Lutz Behnke <behnke@trustcenter.de>] 352659191Skris 352759191Skris Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 352859191Skris 352959191Skris *) PKCS7_encrypt() was adding text MIME headers twice because they 353059191Skris were added manually and by SMIME_crlf_copy(). 353159191Skris [Steve Henson] 353259191Skris 353359191Skris *) In bntest.c don't call BN_rand with zero bits argument. 353459191Skris [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 353559191Skris 353659191Skris *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 353759191Skris case was implemented. This caused BN_div_recp() to fail occasionally. 353859191Skris [Ulf M�ller] 353959191Skris 354059191Skris *) Add an optional second argument to the set_label() in the perl 354159191Skris assembly language builder. If this argument exists and is set 354259191Skris to 1 it signals that the assembler should use a symbol whose 354359191Skris scope is the entire file, not just the current function. This 354459191Skris is needed with MASM which uses the format label:: for this scope. 354559191Skris [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 354659191Skris 354759191Skris *) Change the ASN1 types so they are typedefs by default. Before 354859191Skris almost all types were #define'd to ASN1_STRING which was causing 354959191Skris STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 355059191Skris for example. 355159191Skris [Steve Henson] 355259191Skris 355359191Skris *) Change names of new functions to the new get1/get0 naming 355459191Skris convention: After 'get1', the caller owns a reference count 355559191Skris and has to call ..._free; 'get0' returns a pointer to some 355659191Skris data structure without incrementing reference counters. 355759191Skris (Some of the existing 'get' functions increment a reference 355859191Skris counter, some don't.) 355959191Skris Similarly, 'set1' and 'add1' functions increase reference 356059191Skris counters or duplicate objects. 356159191Skris [Steve Henson] 356259191Skris 356359191Skris *) Allow for the possibility of temp RSA key generation failure: 356459191Skris the code used to assume it always worked and crashed on failure. 356559191Skris [Steve Henson] 356659191Skris 356759191Skris *) Fix potential buffer overrun problem in BIO_printf(). 356859191Skris [Ulf M�ller, using public domain code by Patrick Powell; problem 356959191Skris pointed out by David Sacerdote <das33@cornell.edu>] 357059191Skris 357159191Skris *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 357259191Skris RAND_egd() and RAND_status(). In the command line application, 357359191Skris the EGD socket can be specified like a seed file using RANDFILE 357459191Skris or -rand. 357559191Skris [Ulf M�ller] 357659191Skris 357759191Skris *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 357859191Skris Some CAs (e.g. Verisign) distribute certificates in this form. 357959191Skris [Steve Henson] 358059191Skris 358159191Skris *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 358259191Skris list to exclude them. This means that no special compilation option 358359191Skris is needed to use anonymous DH: it just needs to be included in the 358459191Skris cipher list. 358559191Skris [Steve Henson] 358659191Skris 358759191Skris *) Change the EVP_MD_CTX_type macro so its meaning consistent with 358859191Skris EVP_MD_type. The old functionality is available in a new macro called 358959191Skris EVP_MD_md(). Change code that uses it and update docs. 359059191Skris [Steve Henson] 359159191Skris 359259191Skris *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 359359191Skris where the 'void *' argument is replaced by a function pointer argument. 359459191Skris Previously 'void *' was abused to point to functions, which works on 359559191Skris many platforms, but is not correct. As these functions are usually 359659191Skris called by macros defined in OpenSSL header files, most source code 359759191Skris should work without changes. 359859191Skris [Richard Levitte] 359959191Skris 360059191Skris *) <openssl/opensslconf.h> (which is created by Configure) now contains 360159191Skris sections with information on -D... compiler switches used for 360259191Skris compiling the library so that applications can see them. To enable 360359191Skris one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 360459191Skris must be defined. E.g., 360559191Skris #define OPENSSL_ALGORITHM_DEFINES 360659191Skris #include <openssl/opensslconf.h> 360759191Skris defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 360859191Skris [Richard Levitte, Ulf and Bodo M�ller] 360959191Skris 361059191Skris *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 361159191Skris record layer. 361259191Skris [Bodo Moeller] 361359191Skris 361459191Skris *) Change the 'other' type in certificate aux info to a STACK_OF 361559191Skris X509_ALGOR. Although not an AlgorithmIdentifier as such it has 361659191Skris the required ASN1 format: arbitrary types determined by an OID. 361759191Skris [Steve Henson] 361859191Skris 361959191Skris *) Add some PEM_write_X509_REQ_NEW() functions and a command line 362059191Skris argument to 'req'. This is not because the function is newer or 362159191Skris better than others it just uses the work 'NEW' in the certificate 362259191Skris request header lines. Some software needs this. 362359191Skris [Steve Henson] 362459191Skris 362559191Skris *) Reorganise password command line arguments: now passwords can be 362659191Skris obtained from various sources. Delete the PEM_cb function and make 362759191Skris it the default behaviour: i.e. if the callback is NULL and the 362859191Skris usrdata argument is not NULL interpret it as a null terminated pass 362959191Skris phrase. If usrdata and the callback are NULL then the pass phrase 363059191Skris is prompted for as usual. 363159191Skris [Steve Henson] 363259191Skris 363359191Skris *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 363459191Skris the support is automatically enabled. The resulting binaries will 363559191Skris autodetect the card and use it if present. 363659191Skris [Ben Laurie and Compaq Inc.] 363759191Skris 363859191Skris *) Work around for Netscape hang bug. This sends certificate request 363959191Skris and server done in one record. Since this is perfectly legal in the 364059191Skris SSL/TLS protocol it isn't a "bug" option and is on by default. See 364159191Skris the bugs/SSLv3 entry for more info. 364259191Skris [Steve Henson] 364359191Skris 364459191Skris *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 364559191Skris [Andy Polyakov] 364659191Skris 364759191Skris *) Add -rand argument to smime and pkcs12 applications and read/write 364859191Skris of seed file. 364959191Skris [Steve Henson] 365059191Skris 365159191Skris *) New 'passwd' tool for crypt(3) and apr1 password hashes. 365259191Skris [Bodo Moeller] 365359191Skris 365459191Skris *) Add command line password options to the remaining applications. 365559191Skris [Steve Henson] 365659191Skris 365759191Skris *) Bug fix for BN_div_recp() for numerators with an even number of 365859191Skris bits. 365959191Skris [Ulf M�ller] 366059191Skris 366159191Skris *) More tests in bntest.c, and changed test_bn output. 366259191Skris [Ulf M�ller] 366359191Skris 366459191Skris *) ./config recognizes MacOS X now. 366559191Skris [Andy Polyakov] 366659191Skris 366759191Skris *) Bug fix for BN_div() when the first words of num and divsor are 366859191Skris equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 366959191Skris [Ulf M�ller] 367059191Skris 367159191Skris *) Add support for various broken PKCS#8 formats, and command line 367259191Skris options to produce them. 367359191Skris [Steve Henson] 367459191Skris 367559191Skris *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 367659191Skris get temporary BIGNUMs from a BN_CTX. 367759191Skris [Ulf M�ller] 367859191Skris 367959191Skris *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 368059191Skris for p == 0. 368159191Skris [Ulf M�ller] 368259191Skris 368359191Skris *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 368459191Skris include a #define from the old name to the new. The original intent 368559191Skris was that statically linked binaries could for example just call 368659191Skris SSLeay_add_all_ciphers() to just add ciphers to the table and not 368759191Skris link with digests. This never worked becayse SSLeay_add_all_digests() 368859191Skris and SSLeay_add_all_ciphers() were in the same source file so calling 368959191Skris one would link with the other. They are now in separate source files. 369059191Skris [Steve Henson] 369159191Skris 369259191Skris *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 369359191Skris [Steve Henson] 369459191Skris 369559191Skris *) Use a less unusual form of the Miller-Rabin primality test (it used 369659191Skris a binary algorithm for exponentiation integrated into the Miller-Rabin 369759191Skris loop, our standard modexp algorithms are faster). 369859191Skris [Bodo Moeller] 369959191Skris 370059191Skris *) Support for the EBCDIC character set completed. 370159191Skris [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 370259191Skris 370359191Skris *) Source code cleanups: use const where appropriate, eliminate casts, 370459191Skris use void * instead of char * in lhash. 370559191Skris [Ulf M�ller] 370659191Skris 370759191Skris *) Bugfix: ssl3_send_server_key_exchange was not restartable 370859191Skris (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 370959191Skris this the server could overwrite ephemeral keys that the client 371059191Skris has already seen). 371159191Skris [Bodo Moeller] 371259191Skris 371359191Skris *) Turn DSA_is_prime into a macro that calls BN_is_prime, 371459191Skris using 50 iterations of the Rabin-Miller test. 371559191Skris 371659191Skris DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 371759191Skris iterations of the Rabin-Miller test as required by the appendix 371859191Skris to FIPS PUB 186[-1]) instead of DSA_is_prime. 371959191Skris As BN_is_prime_fasttest includes trial division, DSA parameter 372059191Skris generation becomes much faster. 372159191Skris 372259191Skris This implies a change for the callback functions in DSA_is_prime 372359191Skris and DSA_generate_parameters: The callback function is called once 372459191Skris for each positive witness in the Rabin-Miller test, not just 372559191Skris occasionally in the inner loop; and the parameters to the 372659191Skris callback function now provide an iteration count for the outer 372759191Skris loop rather than for the current invocation of the inner loop. 372859191Skris DSA_generate_parameters additionally can call the callback 372959191Skris function with an 'iteration count' of -1, meaning that a 373059191Skris candidate has passed the trial division test (when q is generated 373159191Skris from an application-provided seed, trial division is skipped). 373259191Skris [Bodo Moeller] 373359191Skris 373459191Skris *) New function BN_is_prime_fasttest that optionally does trial 373559191Skris division before starting the Rabin-Miller test and has 373659191Skris an additional BN_CTX * argument (whereas BN_is_prime always 373759191Skris has to allocate at least one BN_CTX). 373859191Skris 'callback(1, -1, cb_arg)' is called when a number has passed the 373959191Skris trial division stage. 374059191Skris [Bodo Moeller] 374159191Skris 374259191Skris *) Fix for bug in CRL encoding. The validity dates weren't being handled 374359191Skris as ASN1_TIME. 374459191Skris [Steve Henson] 374559191Skris 374659191Skris *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 374759191Skris [Steve Henson] 374859191Skris 374959191Skris *) New function BN_pseudo_rand(). 375059191Skris [Ulf M�ller] 375159191Skris 375259191Skris *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 375359191Skris bignum version of BN_from_montgomery() with the working code from 375459191Skris SSLeay 0.9.0 (the word based version is faster anyway), and clean up 375559191Skris the comments. 375659191Skris [Ulf M�ller] 375759191Skris 375859191Skris *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 375959191Skris made it impossible to use the same SSL_SESSION data structure in 376059191Skris SSL2 clients in multiple threads. 376159191Skris [Bodo Moeller] 376259191Skris 376359191Skris *) The return value of RAND_load_file() no longer counts bytes obtained 376459191Skris by stat(). RAND_load_file(..., -1) is new and uses the complete file 376559191Skris to seed the PRNG (previously an explicit byte count was required). 376659191Skris [Ulf M�ller, Bodo M�ller] 376759191Skris 376859191Skris *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 376959191Skris used (char *) instead of (void *) and had casts all over the place. 377059191Skris [Steve Henson] 377159191Skris 377259191Skris *) Make BN_generate_prime() return NULL on error if ret!=NULL. 377359191Skris [Ulf M�ller] 377459191Skris 377559191Skris *) Retain source code compatibility for BN_prime_checks macro: 377659191Skris BN_is_prime(..., BN_prime_checks, ...) now uses 377759191Skris BN_prime_checks_for_size to determine the appropriate number of 377859191Skris Rabin-Miller iterations. 377959191Skris [Ulf M�ller] 378059191Skris 378159191Skris *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 378259191Skris DH_CHECK_P_NOT_SAFE_PRIME. 378359191Skris (Check if this is true? OpenPGP calls them "strong".) 378459191Skris [Ulf M�ller] 378559191Skris 378659191Skris *) Merge the functionality of "dh" and "gendh" programs into a new program 378759191Skris "dhparam". The old programs are retained for now but will handle DH keys 378859191Skris (instead of parameters) in future. 378959191Skris [Steve Henson] 379059191Skris 379159191Skris *) Make the ciphers, s_server and s_client programs check the return values 379259191Skris when a new cipher list is set. 379359191Skris [Steve Henson] 379459191Skris 379559191Skris *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 379659191Skris ciphers. Before when the 56bit ciphers were enabled the sorting was 379759191Skris wrong. 379859191Skris 379959191Skris The syntax for the cipher sorting has been extended to support sorting by 380059191Skris cipher-strength (using the strength_bits hard coded in the tables). 380159191Skris The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 380259191Skris 380359191Skris Fix a bug in the cipher-command parser: when supplying a cipher command 380459191Skris string with an "undefined" symbol (neither command nor alphanumeric 380559191Skris [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 380659191Skris an error is flagged. 380759191Skris 380859191Skris Due to the strength-sorting extension, the code of the 380959191Skris ssl_create_cipher_list() function was completely rearranged. I hope that 381059191Skris the readability was also increased :-) 381159191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 381259191Skris 381359191Skris *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 381459191Skris for the first serial number and places 2 in the serial number file. This 381559191Skris avoids problems when the root CA is created with serial number zero and 381659191Skris the first user certificate has the same issuer name and serial number 381759191Skris as the root CA. 381859191Skris [Steve Henson] 381959191Skris 382059191Skris *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 382159191Skris the new code. Add documentation for this stuff. 382259191Skris [Steve Henson] 382359191Skris 382459191Skris *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 382559191Skris X509_*() to X509at_*() on the grounds that they don't handle X509 382659191Skris structures and behave in an analagous way to the X509v3 functions: 382759191Skris they shouldn't be called directly but wrapper functions should be used 382859191Skris instead. 382959191Skris 383059191Skris So we also now have some wrapper functions that call the X509at functions 383159191Skris when passed certificate requests. (TO DO: similar things can be done with 383259191Skris PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 383359191Skris things. Some of these need some d2i or i2d and print functionality 383459191Skris because they handle more complex structures.) 383559191Skris [Steve Henson] 383659191Skris 383759191Skris *) Add missing #ifndefs that caused missing symbols when building libssl 383859191Skris as a shared library without RSA. Use #ifndef NO_SSL2 instead of 383959191Skris NO_RSA in ssl/s2*.c. 384059191Skris [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 384159191Skris 384259191Skris *) Precautions against using the PRNG uninitialized: RAND_bytes() now 384359191Skris has a return value which indicates the quality of the random data 384459191Skris (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 384559191Skris error queue. New function RAND_pseudo_bytes() generates output that is 384659191Skris guaranteed to be unique but not unpredictable. RAND_add is like 384759191Skris RAND_seed, but takes an extra argument for an entropy estimate 384859191Skris (RAND_seed always assumes full entropy). 384959191Skris [Ulf M�ller] 385059191Skris 385159191Skris *) Do more iterations of Rabin-Miller probable prime test (specifically, 385259191Skris 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 385359191Skris instead of only 2 for all lengths; see BN_prime_checks_for_size definition 385459191Skris in crypto/bn/bn_prime.c for the complete table). This guarantees a 385559191Skris false-positive rate of at most 2^-80 for random input. 385659191Skris [Bodo Moeller] 385759191Skris 385859191Skris *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 385959191Skris [Bodo Moeller] 386059191Skris 386159191Skris *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 386259191Skris in the 0.9.5 release), this returns the chain 386359191Skris from an X509_CTX structure with a dup of the stack and all 386459191Skris the X509 reference counts upped: so the stack will exist 386559191Skris after X509_CTX_cleanup() has been called. Modify pkcs12.c 386659191Skris to use this. 386759191Skris 386859191Skris Also make SSL_SESSION_print() print out the verify return 386959191Skris code. 387059191Skris [Steve Henson] 387159191Skris 387259191Skris *) Add manpage for the pkcs12 command. Also change the default 387359191Skris behaviour so MAC iteration counts are used unless the new 387459191Skris -nomaciter option is used. This improves file security and 387559191Skris only older versions of MSIE (4.0 for example) need it. 387659191Skris [Steve Henson] 387759191Skris 387859191Skris *) Honor the no-xxx Configure options when creating .DEF files. 387959191Skris [Ulf M�ller] 388059191Skris 388159191Skris *) Add PKCS#10 attributes to field table: challengePassword, 388259191Skris unstructuredName and unstructuredAddress. These are taken from 388359191Skris draft PKCS#9 v2.0 but are compatible with v1.2 provided no 388459191Skris international characters are used. 388559191Skris 388659191Skris More changes to X509_ATTRIBUTE code: allow the setting of types 388759191Skris based on strings. Remove the 'loc' parameter when adding 388859191Skris attributes because these will be a SET OF encoding which is sorted 388959191Skris in ASN1 order. 389059191Skris [Steve Henson] 389159191Skris 389259191Skris *) Initial changes to the 'req' utility to allow request generation 389359191Skris automation. This will allow an application to just generate a template 389459191Skris file containing all the field values and have req construct the 389559191Skris request. 389659191Skris 389759191Skris Initial support for X509_ATTRIBUTE handling. Stacks of these are 389859191Skris used all over the place including certificate requests and PKCS#7 389959191Skris structures. They are currently handled manually where necessary with 390059191Skris some primitive wrappers for PKCS#7. The new functions behave in a 390159191Skris manner analogous to the X509 extension functions: they allow 390259191Skris attributes to be looked up by NID and added. 390359191Skris 390459191Skris Later something similar to the X509V3 code would be desirable to 390559191Skris automatically handle the encoding, decoding and printing of the 390659191Skris more complex types. The string types like challengePassword can 390759191Skris be handled by the string table functions. 390859191Skris 390959191Skris Also modified the multi byte string table handling. Now there is 391059191Skris a 'global mask' which masks out certain types. The table itself 391159191Skris can use the flag STABLE_NO_MASK to ignore the mask setting: this 391259191Skris is useful when for example there is only one permissible type 391359191Skris (as in countryName) and using the mask might result in no valid 391459191Skris types at all. 391559191Skris [Steve Henson] 391659191Skris 391759191Skris *) Clean up 'Finished' handling, and add functions SSL_get_finished and 391859191Skris SSL_get_peer_finished to allow applications to obtain the latest 391959191Skris Finished messages sent to the peer or expected from the peer, 392059191Skris respectively. (SSL_get_peer_finished is usually the Finished message 392159191Skris actually received from the peer, otherwise the protocol will be aborted.) 392259191Skris 392359191Skris As the Finished message are message digests of the complete handshake 392459191Skris (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 392559191Skris be used for external authentication procedures when the authentication 392659191Skris provided by SSL/TLS is not desired or is not enough. 392759191Skris [Bodo Moeller] 392859191Skris 392959191Skris *) Enhanced support for Alpha Linux is added. Now ./config checks if 393059191Skris the host supports BWX extension and if Compaq C is present on the 393159191Skris $PATH. Just exploiting of the BWX extension results in 20-30% 393259191Skris performance kick for some algorithms, e.g. DES and RC4 to mention 393359191Skris a couple. Compaq C in turn generates ~20% faster code for MD5 and 393459191Skris SHA1. 393559191Skris [Andy Polyakov] 393659191Skris 393759191Skris *) Add support for MS "fast SGC". This is arguably a violation of the 393859191Skris SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 393959191Skris weak crypto and after checking the certificate is SGC a second one 394059191Skris with strong crypto. MS SGC stops the first handshake after receiving 394159191Skris the server certificate message and sends a second client hello. Since 394259191Skris a server will typically do all the time consuming operations before 394359191Skris expecting any further messages from the client (server key exchange 394459191Skris is the most expensive) there is little difference between the two. 394559191Skris 394659191Skris To get OpenSSL to support MS SGC we have to permit a second client 394759191Skris hello message after we have sent server done. In addition we have to 394859191Skris reset the MAC if we do get this second client hello. 394959191Skris [Steve Henson] 395059191Skris 395159191Skris *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 395259191Skris if a DER encoded private key is RSA or DSA traditional format. Changed 395359191Skris d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 395459191Skris format DER encoded private key. Newer code should use PKCS#8 format which 395559191Skris has the key type encoded in the ASN1 structure. Added DER private key 395659191Skris support to pkcs8 application. 395759191Skris [Steve Henson] 395859191Skris 395959191Skris *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 396059191Skris ciphersuites has been selected (as required by the SSL 3/TLS 1 396159191Skris specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 396259191Skris is set, we interpret this as a request to violate the specification 396359191Skris (the worst that can happen is a handshake failure, and 'correct' 396459191Skris behaviour would result in a handshake failure anyway). 396559191Skris [Bodo Moeller] 396659191Skris 396759191Skris *) In SSL_CTX_add_session, take into account that there might be multiple 396859191Skris SSL_SESSION structures with the same session ID (e.g. when two threads 396959191Skris concurrently obtain them from an external cache). 397059191Skris The internal cache can handle only one SSL_SESSION with a given ID, 397159191Skris so if there's a conflict, we now throw out the old one to achieve 397259191Skris consistency. 397359191Skris [Bodo Moeller] 397459191Skris 397559191Skris *) Add OIDs for idea and blowfish in CBC mode. This will allow both 397659191Skris to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 397759191Skris some routines that use cipher OIDs: some ciphers do not have OIDs 397859191Skris defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 397959191Skris example. 398059191Skris [Steve Henson] 398159191Skris 398259191Skris *) Simplify the trust setting structure and code. Now we just have 398359191Skris two sequences of OIDs for trusted and rejected settings. These will 398459191Skris typically have values the same as the extended key usage extension 398559191Skris and any application specific purposes. 398659191Skris 398759191Skris The trust checking code now has a default behaviour: it will just 398859191Skris check for an object with the same NID as the passed id. Functions can 398959191Skris be provided to override either the default behaviour or the behaviour 399059191Skris for a given id. SSL client, server and email already have functions 399159191Skris in place for compatibility: they check the NID and also return "trusted" 399259191Skris if the certificate is self signed. 399359191Skris [Steve Henson] 399459191Skris 399559191Skris *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 399659191Skris traditional format into an EVP_PKEY structure. 399759191Skris [Steve Henson] 399859191Skris 399959191Skris *) Add a password callback function PEM_cb() which either prompts for 400059191Skris a password if usr_data is NULL or otherwise assumes it is a null 400159191Skris terminated password. Allow passwords to be passed on command line 400259191Skris environment or config files in a few more utilities. 400359191Skris [Steve Henson] 400459191Skris 400559191Skris *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 400659191Skris keys. Add some short names for PKCS#8 PBE algorithms and allow them 400759191Skris to be specified on the command line for the pkcs8 and pkcs12 utilities. 400859191Skris Update documentation. 400959191Skris [Steve Henson] 401059191Skris 401159191Skris *) Support for ASN1 "NULL" type. This could be handled before by using 401259191Skris ASN1_TYPE but there wasn't any function that would try to read a NULL 401359191Skris and produce an error if it couldn't. For compatibility we also have 401459191Skris ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 401559191Skris don't allocate anything because they don't need to. 401659191Skris [Steve Henson] 401759191Skris 401859191Skris *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 401959191Skris for details. 402059191Skris [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 402159191Skris 402259191Skris *) Rebuild of the memory allocation routines used by OpenSSL code and 402359191Skris possibly others as well. The purpose is to make an interface that 402459191Skris provide hooks so anyone can build a separate set of allocation and 402559191Skris deallocation routines to be used by OpenSSL, for example memory 402659191Skris pool implementations, or something else, which was previously hard 402759191Skris since Malloc(), Realloc() and Free() were defined as macros having 402859191Skris the values malloc, realloc and free, respectively (except for Win32 402959191Skris compilations). The same is provided for memory debugging code. 403059191Skris OpenSSL already comes with functionality to find memory leaks, but 403159191Skris this gives people a chance to debug other memory problems. 403259191Skris 403359191Skris With these changes, a new set of functions and macros have appeared: 403459191Skris 403568651Skris CRYPTO_set_mem_debug_functions() [F] 403668651Skris CRYPTO_get_mem_debug_functions() [F] 403768651Skris CRYPTO_dbg_set_options() [F] 403868651Skris CRYPTO_dbg_get_options() [F] 403968651Skris CRYPTO_malloc_debug_init() [M] 404059191Skris 404159191Skris The memory debug functions are NULL by default, unless the library 404259191Skris is compiled with CRYPTO_MDEBUG or friends is defined. If someone 404359191Skris wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 404459191Skris gives the standard debugging functions that come with OpenSSL) or 404559191Skris CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 404659191Skris provided by the library user) must be used. When the standard 404759191Skris debugging functions are used, CRYPTO_dbg_set_options can be used to 404859191Skris request additional information: 404959191Skris CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 405059191Skris the CRYPTO_MDEBUG_xxx macro when compiling the library. 405159191Skris 405259191Skris Also, things like CRYPTO_set_mem_functions will always give the 405359191Skris expected result (the new set of functions is used for allocation 405459191Skris and deallocation) at all times, regardless of platform and compiler 405559191Skris options. 405659191Skris 405759191Skris To finish it up, some functions that were never use in any other 405859191Skris way than through macros have a new API and new semantic: 405959191Skris 406059191Skris CRYPTO_dbg_malloc() 406159191Skris CRYPTO_dbg_realloc() 406259191Skris CRYPTO_dbg_free() 406359191Skris 406459191Skris All macros of value have retained their old syntax. 406559191Skris [Richard Levitte and Bodo Moeller] 406659191Skris 406759191Skris *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 406859191Skris ordering of SMIMECapabilities wasn't in "strength order" and there 406959191Skris was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 407059191Skris algorithm. 407159191Skris [Steve Henson] 407259191Skris 407359191Skris *) Some ASN1 types with illegal zero length encoding (INTEGER, 407459191Skris ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 407559191Skris [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 407659191Skris 407759191Skris *) Merge in my S/MIME library for OpenSSL. This provides a simple 407859191Skris S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 407959191Skris functionality to handle multipart/signed properly) and a utility 408059191Skris called 'smime' to call all this stuff. This is based on code I 408159191Skris originally wrote for Celo who have kindly allowed it to be 408259191Skris included in OpenSSL. 408359191Skris [Steve Henson] 408459191Skris 408559191Skris *) Add variants des_set_key_checked and des_set_key_unchecked of 408659191Skris des_set_key (aka des_key_sched). Global variable des_check_key 408759191Skris decides which of these is called by des_set_key; this way 408859191Skris des_check_key behaves as it always did, but applications and 408959191Skris the library itself, which was buggy for des_check_key == 1, 409059191Skris have a cleaner way to pick the version they need. 409159191Skris [Bodo Moeller] 409259191Skris 409359191Skris *) New function PKCS12_newpass() which changes the password of a 409459191Skris PKCS12 structure. 409559191Skris [Steve Henson] 409659191Skris 409759191Skris *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 409859191Skris dynamic mix. In both cases the ids can be used as an index into the 409959191Skris table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 410059191Skris functions so they accept a list of the field values and the 410159191Skris application doesn't need to directly manipulate the X509_TRUST 410259191Skris structure. 410359191Skris [Steve Henson] 410459191Skris 410559191Skris *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 410659191Skris need initialising. 410759191Skris [Steve Henson] 410859191Skris 410959191Skris *) Modify the way the V3 extension code looks up extensions. This now 411059191Skris works in a similar way to the object code: we have some "standard" 411159191Skris extensions in a static table which is searched with OBJ_bsearch() 411259191Skris and the application can add dynamic ones if needed. The file 411359191Skris crypto/x509v3/ext_dat.h now has the info: this file needs to be 411459191Skris updated whenever a new extension is added to the core code and kept 411559191Skris in ext_nid order. There is a simple program 'tabtest.c' which checks 411659191Skris this. New extensions are not added too often so this file can readily 411759191Skris be maintained manually. 411859191Skris 411959191Skris There are two big advantages in doing things this way. The extensions 412059191Skris can be looked up immediately and no longer need to be "added" using 412159191Skris X509V3_add_standard_extensions(): this function now does nothing. 412259191Skris [Side note: I get *lots* of email saying the extension code doesn't 412359191Skris work because people forget to call this function] 412459191Skris Also no dynamic allocation is done unless new extensions are added: 412559191Skris so if we don't add custom extensions there is no need to call 412659191Skris X509V3_EXT_cleanup(). 412759191Skris [Steve Henson] 412859191Skris 412959191Skris *) Modify enc utility's salting as follows: make salting the default. Add a 413059191Skris magic header, so unsalted files fail gracefully instead of just decrypting 413159191Skris to garbage. This is because not salting is a big security hole, so people 413259191Skris should be discouraged from doing it. 413359191Skris [Ben Laurie] 413459191Skris 413559191Skris *) Fixes and enhancements to the 'x509' utility. It allowed a message 413659191Skris digest to be passed on the command line but it only used this 413759191Skris parameter when signing a certificate. Modified so all relevant 413859191Skris operations are affected by the digest parameter including the 413959191Skris -fingerprint and -x509toreq options. Also -x509toreq choked if a 414059191Skris DSA key was used because it didn't fix the digest. 414159191Skris [Steve Henson] 414259191Skris 414359191Skris *) Initial certificate chain verify code. Currently tests the untrusted 414459191Skris certificates for consistency with the verify purpose (which is set 414559191Skris when the X509_STORE_CTX structure is set up) and checks the pathlength. 414659191Skris 414759191Skris There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 414859191Skris this is because it will reject chains with invalid extensions whereas 414959191Skris every previous version of OpenSSL and SSLeay made no checks at all. 415059191Skris 415159191Skris Trust code: checks the root CA for the relevant trust settings. Trust 415259191Skris settings have an initial value consistent with the verify purpose: e.g. 415359191Skris if the verify purpose is for SSL client use it expects the CA to be 415459191Skris trusted for SSL client use. However the default value can be changed to 415559191Skris permit custom trust settings: one example of this would be to only trust 415659191Skris certificates from a specific "secure" set of CAs. 415759191Skris 415859191Skris Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 415959191Skris which should be used for version portability: especially since the 416059191Skris verify structure is likely to change more often now. 416159191Skris 416259191Skris SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 416359191Skris to set them. If not set then assume SSL clients will verify SSL servers 416459191Skris and vice versa. 416559191Skris 416659191Skris Two new options to the verify program: -untrusted allows a set of 416759191Skris untrusted certificates to be passed in and -purpose which sets the 416859191Skris intended purpose of the certificate. If a purpose is set then the 416959191Skris new chain verify code is used to check extension consistency. 417059191Skris [Steve Henson] 417159191Skris 417259191Skris *) Support for the authority information access extension. 417359191Skris [Steve Henson] 417459191Skris 417559191Skris *) Modify RSA and DSA PEM read routines to transparently handle 417659191Skris PKCS#8 format private keys. New *_PUBKEY_* functions that handle 417759191Skris public keys in a format compatible with certificate 417859191Skris SubjectPublicKeyInfo structures. Unfortunately there were already 417959191Skris functions called *_PublicKey_* which used various odd formats so 418059191Skris these are retained for compatibility: however the DSA variants were 418159191Skris never in a public release so they have been deleted. Changed dsa/rsa 418259191Skris utilities to handle the new format: note no releases ever handled public 418359191Skris keys so we should be OK. 418459191Skris 418559191Skris The primary motivation for this change is to avoid the same fiasco 418659191Skris that dogs private keys: there are several incompatible private key 418759191Skris formats some of which are standard and some OpenSSL specific and 418859191Skris require various evil hacks to allow partial transparent handling and 418959191Skris even then it doesn't work with DER formats. Given the option anything 419059191Skris other than PKCS#8 should be dumped: but the other formats have to 419159191Skris stay in the name of compatibility. 419259191Skris 419359191Skris With public keys and the benefit of hindsight one standard format 419459191Skris is used which works with EVP_PKEY, RSA or DSA structures: though 419559191Skris it clearly returns an error if you try to read the wrong kind of key. 419659191Skris 419759191Skris Added a -pubkey option to the 'x509' utility to output the public key. 419859191Skris Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 419959191Skris (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 420059191Skris EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 420159191Skris that do the same as the EVP_PKEY_assign_*() except they up the 420259191Skris reference count of the added key (they don't "swallow" the 420359191Skris supplied key). 420459191Skris [Steve Henson] 420559191Skris 420659191Skris *) Fixes to crypto/x509/by_file.c the code to read in certificates and 420759191Skris CRLs would fail if the file contained no certificates or no CRLs: 420859191Skris added a new function to read in both types and return the number 420959191Skris read: this means that if none are read it will be an error. The 421059191Skris DER versions of the certificate and CRL reader would always fail 421159191Skris because it isn't possible to mix certificates and CRLs in DER format 421259191Skris without choking one or the other routine. Changed this to just read 421359191Skris a certificate: this is the best we can do. Also modified the code 421459191Skris in apps/verify.c to take notice of return codes: it was previously 421559191Skris attempting to read in certificates from NULL pointers and ignoring 421659191Skris any errors: this is one reason why the cert and CRL reader seemed 421759191Skris to work. It doesn't check return codes from the default certificate 421859191Skris routines: these may well fail if the certificates aren't installed. 421959191Skris [Steve Henson] 422059191Skris 422159191Skris *) Code to support otherName option in GeneralName. 422259191Skris [Steve Henson] 422359191Skris 422459191Skris *) First update to verify code. Change the verify utility 422559191Skris so it warns if it is passed a self signed certificate: 422659191Skris for consistency with the normal behaviour. X509_verify 422759191Skris has been modified to it will now verify a self signed 422859191Skris certificate if *exactly* the same certificate appears 422959191Skris in the store: it was previously impossible to trust a 423059191Skris single self signed certificate. This means that: 423159191Skris openssl verify ss.pem 423259191Skris now gives a warning about a self signed certificate but 423359191Skris openssl verify -CAfile ss.pem ss.pem 423459191Skris is OK. 423559191Skris [Steve Henson] 423659191Skris 423759191Skris *) For servers, store verify_result in SSL_SESSION data structure 423859191Skris (and add it to external session representation). 423959191Skris This is needed when client certificate verifications fails, 424059191Skris but an application-provided verification callback (set by 424159191Skris SSL_CTX_set_cert_verify_callback) allows accepting the session 424259191Skris anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 424359191Skris but returns 1): When the session is reused, we have to set 424459191Skris ssl->verify_result to the appropriate error code to avoid 424559191Skris security holes. 424659191Skris [Bodo Moeller, problem pointed out by Lutz Jaenicke] 424759191Skris 424859191Skris *) Fix a bug in the new PKCS#7 code: it didn't consider the 424959191Skris case in PKCS7_dataInit() where the signed PKCS7 structure 425059191Skris didn't contain any existing data because it was being created. 425159191Skris [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 425259191Skris 425359191Skris *) Add a salt to the key derivation routines in enc.c. This 425459191Skris forms the first 8 bytes of the encrypted file. Also add a 425559191Skris -S option to allow a salt to be input on the command line. 425659191Skris [Steve Henson] 425759191Skris 425859191Skris *) New function X509_cmp(). Oddly enough there wasn't a function 425959191Skris to compare two certificates. We do this by working out the SHA1 426059191Skris hash and comparing that. X509_cmp() will be needed by the trust 426159191Skris code. 426259191Skris [Steve Henson] 426359191Skris 426459191Skris *) SSL_get1_session() is like SSL_get_session(), but increments 426559191Skris the reference count in the SSL_SESSION returned. 426659191Skris [Geoff Thorpe <geoff@eu.c2.net>] 426759191Skris 426859191Skris *) Fix for 'req': it was adding a null to request attributes. 426959191Skris Also change the X509_LOOKUP and X509_INFO code to handle 427059191Skris certificate auxiliary information. 427159191Skris [Steve Henson] 427259191Skris 427359191Skris *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 427459191Skris the 'enc' command. 427559191Skris [Steve Henson] 427659191Skris 427759191Skris *) Add the possibility to add extra information to the memory leak 427859191Skris detecting output, to form tracebacks, showing from where each 427959191Skris allocation was originated: CRYPTO_push_info("constant string") adds 428059191Skris the string plus current file name and line number to a per-thread 428159191Skris stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 428259191Skris is like calling CYRPTO_pop_info() until the stack is empty. 428359191Skris Also updated memory leak detection code to be multi-thread-safe. 428459191Skris [Richard Levitte] 428559191Skris 428659191Skris *) Add options -text and -noout to pkcs7 utility and delete the 428759191Skris encryption options which never did anything. Update docs. 428859191Skris [Steve Henson] 428959191Skris 429059191Skris *) Add options to some of the utilities to allow the pass phrase 429159191Skris to be included on either the command line (not recommended on 429259191Skris OSes like Unix) or read from the environment. Update the 429359191Skris manpages and fix a few bugs. 429459191Skris [Steve Henson] 429559191Skris 429659191Skris *) Add a few manpages for some of the openssl commands. 429759191Skris [Steve Henson] 429859191Skris 429959191Skris *) Fix the -revoke option in ca. It was freeing up memory twice, 430059191Skris leaking and not finding already revoked certificates. 430159191Skris [Steve Henson] 430259191Skris 430359191Skris *) Extensive changes to support certificate auxiliary information. 430459191Skris This involves the use of X509_CERT_AUX structure and X509_AUX 430559191Skris functions. An X509_AUX function such as PEM_read_X509_AUX() 430659191Skris can still read in a certificate file in the usual way but it 430759191Skris will also read in any additional "auxiliary information". By 430859191Skris doing things this way a fair degree of compatibility can be 430959191Skris retained: existing certificates can have this information added 431059191Skris using the new 'x509' options. 431159191Skris 431259191Skris Current auxiliary information includes an "alias" and some trust 431359191Skris settings. The trust settings will ultimately be used in enhanced 431459191Skris certificate chain verification routines: currently a certificate 431559191Skris can only be trusted if it is self signed and then it is trusted 431659191Skris for all purposes. 431759191Skris [Steve Henson] 431859191Skris 431959191Skris *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 432059191Skris The problem was that one of the replacement routines had not been working 432159191Skris since SSLeay releases. For now the offending routine has been replaced 432259191Skris with non-optimised assembler. Even so, this now gives around 95% 432359191Skris performance improvement for 1024 bit RSA signs. 432459191Skris [Mark Cox] 432559191Skris 432659191Skris *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 432759191Skris handling. Most clients have the effective key size in bits equal to 432859191Skris the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 432959191Skris A few however don't do this and instead use the size of the decrypted key 433059191Skris to determine the RC2 key length and the AlgorithmIdentifier to determine 433159191Skris the effective key length. In this case the effective key length can still 433259191Skris be 40 bits but the key length can be 168 bits for example. This is fixed 433359191Skris by manually forcing an RC2 key into the EVP_PKEY structure because the 433459191Skris EVP code can't currently handle unusual RC2 key sizes: it always assumes 433559191Skris the key length and effective key length are equal. 433659191Skris [Steve Henson] 433759191Skris 433859191Skris *) Add a bunch of functions that should simplify the creation of 433959191Skris X509_NAME structures. Now you should be able to do: 434059191Skris X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 434159191Skris and have it automatically work out the correct field type and fill in 434259191Skris the structures. The more adventurous can try: 434359191Skris X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 434459191Skris and it will (hopefully) work out the correct multibyte encoding. 434559191Skris [Steve Henson] 434659191Skris 434759191Skris *) Change the 'req' utility to use the new field handling and multibyte 434859191Skris copy routines. Before the DN field creation was handled in an ad hoc 434959191Skris way in req, ca, and x509 which was rather broken and didn't support 435059191Skris BMPStrings or UTF8Strings. Since some software doesn't implement 435159191Skris BMPStrings or UTF8Strings yet, they can be enabled using the config file 435259191Skris using the dirstring_type option. See the new comment in the default 435359191Skris openssl.cnf for more info. 435459191Skris [Steve Henson] 435559191Skris 435659191Skris *) Make crypto/rand/md_rand.c more robust: 435759191Skris - Assure unique random numbers after fork(). 435859191Skris - Make sure that concurrent threads access the global counter and 435959191Skris md serializably so that we never lose entropy in them 436059191Skris or use exactly the same state in multiple threads. 436159191Skris Access to the large state is not always serializable because 436259191Skris the additional locking could be a performance killer, and 436359191Skris md should be large enough anyway. 436459191Skris [Bodo Moeller] 436559191Skris 436659191Skris *) New file apps/app_rand.c with commonly needed functionality 436759191Skris for handling the random seed file. 436859191Skris 436959191Skris Use the random seed file in some applications that previously did not: 437059191Skris ca, 437159191Skris dsaparam -genkey (which also ignored its '-rand' option), 437259191Skris s_client, 437359191Skris s_server, 437459191Skris x509 (when signing). 437559191Skris Except on systems with /dev/urandom, it is crucial to have a random 437659191Skris seed file at least for key creation, DSA signing, and for DH exchanges; 437759191Skris for RSA signatures we could do without one. 437859191Skris 437959191Skris gendh and gendsa (unlike genrsa) used to read only the first byte 438059191Skris of each file listed in the '-rand' option. The function as previously 438159191Skris found in genrsa is now in app_rand.c and is used by all programs 438259191Skris that support '-rand'. 438359191Skris [Bodo Moeller] 438459191Skris 438559191Skris *) In RAND_write_file, use mode 0600 for creating files; 438659191Skris don't just chmod when it may be too late. 438759191Skris [Bodo Moeller] 438859191Skris 438959191Skris *) Report an error from X509_STORE_load_locations 439059191Skris when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 439159191Skris [Bill Perry] 439259191Skris 439359191Skris *) New function ASN1_mbstring_copy() this copies a string in either 439459191Skris ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 439559191Skris into an ASN1_STRING type. A mask of permissible types is passed 439659191Skris and it chooses the "minimal" type to use or an error if not type 439759191Skris is suitable. 439859191Skris [Steve Henson] 439959191Skris 440059191Skris *) Add function equivalents to the various macros in asn1.h. The old 440159191Skris macros are retained with an M_ prefix. Code inside the library can 440259191Skris use the M_ macros. External code (including the openssl utility) 440359191Skris should *NOT* in order to be "shared library friendly". 440459191Skris [Steve Henson] 440559191Skris 440659191Skris *) Add various functions that can check a certificate's extensions 440759191Skris to see if it usable for various purposes such as SSL client, 440859191Skris server or S/MIME and CAs of these types. This is currently 440959191Skris VERY EXPERIMENTAL but will ultimately be used for certificate chain 441059191Skris verification. Also added a -purpose flag to x509 utility to 441159191Skris print out all the purposes. 441259191Skris [Steve Henson] 441359191Skris 441459191Skris *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 441559191Skris functions. 441659191Skris [Steve Henson] 441759191Skris 441859191Skris *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 441959191Skris for, obtain and decode and extension and obtain its critical flag. 442059191Skris This allows all the necessary extension code to be handled in a 442159191Skris single function call. 442259191Skris [Steve Henson] 442359191Skris 442459191Skris *) RC4 tune-up featuring 30-40% performance improvement on most RISC 442559191Skris platforms. See crypto/rc4/rc4_enc.c for further details. 442659191Skris [Andy Polyakov] 442759191Skris 442859191Skris *) New -noout option to asn1parse. This causes no output to be produced 442959191Skris its main use is when combined with -strparse and -out to extract data 443059191Skris from a file (which may not be in ASN.1 format). 443159191Skris [Steve Henson] 443259191Skris 443359191Skris *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 443459191Skris when producing the local key id. 443559191Skris [Richard Levitte <levitte@stacken.kth.se>] 443659191Skris 443759191Skris *) New option -dhparam in s_server. This allows a DH parameter file to be 443859191Skris stated explicitly. If it is not stated then it tries the first server 443959191Skris certificate file. The previous behaviour hard coded the filename 444059191Skris "server.pem". 444159191Skris [Steve Henson] 444259191Skris 444359191Skris *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 444459191Skris a public key to be input or output. For example: 444559191Skris openssl rsa -in key.pem -pubout -out pubkey.pem 444659191Skris Also added necessary DSA public key functions to handle this. 444759191Skris [Steve Henson] 444859191Skris 444959191Skris *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 445059191Skris in the message. This was handled by allowing 445159191Skris X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 445259191Skris [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 445359191Skris 445459191Skris *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 445559191Skris to the end of the strings whereas this didn't. This would cause problems 445659191Skris if strings read with d2i_ASN1_bytes() were later modified. 445759191Skris [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 445859191Skris 445959191Skris *) Fix for base64 decode bug. When a base64 bio reads only one line of 446059191Skris data and it contains EOF it will end up returning an error. This is 446159191Skris caused by input 46 bytes long. The cause is due to the way base64 446259191Skris BIOs find the start of base64 encoded data. They do this by trying a 446359191Skris trial decode on each line until they find one that works. When they 446459191Skris do a flag is set and it starts again knowing it can pass all the 446559191Skris data directly through the decoder. Unfortunately it doesn't reset 446659191Skris the context it uses. This means that if EOF is reached an attempt 446759191Skris is made to pass two EOFs through the context and this causes the 446859191Skris resulting error. This can also cause other problems as well. As is 446959191Skris usual with these problems it takes *ages* to find and the fix is 447059191Skris trivial: move one line. 447159191Skris [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 447259191Skris 447359191Skris *) Ugly workaround to get s_client and s_server working under Windows. The 447459191Skris old code wouldn't work because it needed to select() on sockets and the 447559191Skris tty (for keypresses and to see if data could be written). Win32 only 447659191Skris supports select() on sockets so we select() with a 1s timeout on the 447759191Skris sockets and then see if any characters are waiting to be read, if none 447859191Skris are present then we retry, we also assume we can always write data to 447959191Skris the tty. This isn't nice because the code then blocks until we've 448059191Skris received a complete line of data and it is effectively polling the 448159191Skris keyboard at 1s intervals: however it's quite a bit better than not 448259191Skris working at all :-) A dedicated Windows application might handle this 448359191Skris with an event loop for example. 448459191Skris [Steve Henson] 448559191Skris 448659191Skris *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 448759191Skris and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 448859191Skris will be called when RSA_sign() and RSA_verify() are used. This is useful 448959191Skris if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 449059191Skris For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 449159191Skris should *not* be used: RSA_sign() and RSA_verify() must be used instead. 449259191Skris This necessitated the support of an extra signature type NID_md5_sha1 449359191Skris for SSL signatures and modifications to the SSL library to use it instead 449459191Skris of calling RSA_public_decrypt() and RSA_private_encrypt(). 449559191Skris [Steve Henson] 449659191Skris 449759191Skris *) Add new -verify -CAfile and -CApath options to the crl program, these 449859191Skris will lookup a CRL issuers certificate and verify the signature in a 449959191Skris similar way to the verify program. Tidy up the crl program so it 450059191Skris no longer accesses structures directly. Make the ASN1 CRL parsing a bit 450159191Skris less strict. It will now permit CRL extensions even if it is not 450259191Skris a V2 CRL: this will allow it to tolerate some broken CRLs. 450359191Skris [Steve Henson] 450459191Skris 450559191Skris *) Initialize all non-automatic variables each time one of the openssl 450659191Skris sub-programs is started (this is necessary as they may be started 450759191Skris multiple times from the "OpenSSL>" prompt). 450859191Skris [Lennart Bang, Bodo Moeller] 450959191Skris 451059191Skris *) Preliminary compilation option RSA_NULL which disables RSA crypto without 451159191Skris removing all other RSA functionality (this is what NO_RSA does). This 451259191Skris is so (for example) those in the US can disable those operations covered 451359191Skris by the RSA patent while allowing storage and parsing of RSA keys and RSA 451459191Skris key generation. 451559191Skris [Steve Henson] 451659191Skris 451759191Skris *) Non-copying interface to BIO pairs. 451859191Skris (still largely untested) 451959191Skris [Bodo Moeller] 452059191Skris 452159191Skris *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 452259191Skris ASCII string. This was handled independently in various places before. 452359191Skris [Steve Henson] 452459191Skris 452559191Skris *) New functions UTF8_getc() and UTF8_putc() that parse and generate 452659191Skris UTF8 strings a character at a time. 452759191Skris [Steve Henson] 452859191Skris 452959191Skris *) Use client_version from client hello to select the protocol 453059191Skris (s23_srvr.c) and for RSA client key exchange verification 453159191Skris (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 453259191Skris [Bodo Moeller] 453359191Skris 453459191Skris *) Add various utility functions to handle SPKACs, these were previously 453559191Skris handled by poking round in the structure internals. Added new function 453659191Skris NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 453759191Skris print, verify and generate SPKACs. Based on an original idea from 453859191Skris Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 453959191Skris [Steve Henson] 454059191Skris 454159191Skris *) RIPEMD160 is operational on all platforms and is back in 'make test'. 454259191Skris [Andy Polyakov] 454359191Skris 454459191Skris *) Allow the config file extension section to be overwritten on the 454559191Skris command line. Based on an original idea from Massimiliano Pala 454659191Skris <madwolf@comune.modena.it>. The new option is called -extensions 454759191Skris and can be applied to ca, req and x509. Also -reqexts to override 454859191Skris the request extensions in req and -crlexts to override the crl extensions 454959191Skris in ca. 455059191Skris [Steve Henson] 455159191Skris 455259191Skris *) Add new feature to the SPKAC handling in ca. Now you can include 455359191Skris the same field multiple times by preceding it by "XXXX." for example: 455459191Skris 1.OU="Unit name 1" 455559191Skris 2.OU="Unit name 2" 455659191Skris this is the same syntax as used in the req config file. 455759191Skris [Steve Henson] 455859191Skris 455959191Skris *) Allow certificate extensions to be added to certificate requests. These 456059191Skris are specified in a 'req_extensions' option of the req section of the 456159191Skris config file. They can be printed out with the -text option to req but 456259191Skris are otherwise ignored at present. 456359191Skris [Steve Henson] 456459191Skris 456559191Skris *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 456659191Skris data read consists of only the final block it would not decrypted because 456759191Skris EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 456859191Skris A misplaced 'break' also meant the decrypted final block might not be 456959191Skris copied until the next read. 457059191Skris [Steve Henson] 457159191Skris 457259191Skris *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 457359191Skris a few extra parameters to the DH structure: these will be useful if 457459191Skris for example we want the value of 'q' or implement X9.42 DH. 457559191Skris [Steve Henson] 457659191Skris 457759191Skris *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 457859191Skris provides hooks that allow the default DSA functions or functions on a 457959191Skris "per key" basis to be replaced. This allows hardware acceleration and 458059191Skris hardware key storage to be handled without major modification to the 458159191Skris library. Also added low level modexp hooks and CRYPTO_EX structure and 458259191Skris associated functions. 458359191Skris [Steve Henson] 458459191Skris 458559191Skris *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 458659191Skris as "read only": it can't be written to and the buffer it points to will 458759191Skris not be freed. Reading from a read only BIO is much more efficient than 458859191Skris a normal memory BIO. This was added because there are several times when 458959191Skris an area of memory needs to be read from a BIO. The previous method was 459059191Skris to create a memory BIO and write the data to it, this results in two 459159191Skris copies of the data and an O(n^2) reading algorithm. There is a new 459259191Skris function BIO_new_mem_buf() which creates a read only memory BIO from 459359191Skris an area of memory. Also modified the PKCS#7 routines to use read only 459459191Skris memory BIOs. 459559191Skris [Steve Henson] 459659191Skris 459759191Skris *) Bugfix: ssl23_get_client_hello did not work properly when called in 459859191Skris state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 459959191Skris a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 460059191Skris but a retry condition occured while trying to read the rest. 460159191Skris [Bodo Moeller] 460259191Skris 460359191Skris *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 460459191Skris NID_pkcs7_encrypted by default: this was wrong since this should almost 460559191Skris always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 460659191Skris the encrypted data type: this is a more sensible place to put it and it 460759191Skris allows the PKCS#12 code to be tidied up that duplicated this 460859191Skris functionality. 460959191Skris [Steve Henson] 461059191Skris 461159191Skris *) Changed obj_dat.pl script so it takes its input and output files on 461259191Skris the command line. This should avoid shell escape redirection problems 461359191Skris under Win32. 461459191Skris [Steve Henson] 461559191Skris 461659191Skris *) Initial support for certificate extension requests, these are included 461759191Skris in things like Xenroll certificate requests. Included functions to allow 461859191Skris extensions to be obtained and added. 461959191Skris [Steve Henson] 462059191Skris 462159191Skris *) -crlf option to s_client and s_server for sending newlines as 462259191Skris CRLF (as required by many protocols). 462359191Skris [Bodo Moeller] 462459191Skris 462555714Skris Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 462655714Skris 462755714Skris *) Install libRSAglue.a when OpenSSL is built with RSAref. 462855714Skris [Ralf S. Engelschall] 462955714Skris 463055714Skris *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 463155714Skris [Andrija Antonijevic <TheAntony2@bigfoot.com>] 463255714Skris 463355714Skris *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 463455714Skris program. 463555714Skris [Steve Henson] 463655714Skris 463755714Skris *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 463855714Skris DH parameters/keys (q is lost during that conversion, but the resulting 463955714Skris DH parameters contain its length). 464055714Skris 464155714Skris For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 464255714Skris much faster than DH_generate_parameters (which creates parameters 464355714Skris where p = 2*q + 1), and also the smaller q makes DH computations 464455714Skris much more efficient (160-bit exponentiation instead of 1024-bit 464555714Skris exponentiation); so this provides a convenient way to support DHE 464655714Skris ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 464755714Skris utter importance to use 464855714Skris SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 464955714Skris or 465055714Skris SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 465155714Skris when such DH parameters are used, because otherwise small subgroup 465255714Skris attacks may become possible! 465355714Skris [Bodo Moeller] 465455714Skris 465555714Skris *) Avoid memory leak in i2d_DHparams. 465655714Skris [Bodo Moeller] 465755714Skris 465855714Skris *) Allow the -k option to be used more than once in the enc program: 465955714Skris this allows the same encrypted message to be read by multiple recipients. 466055714Skris [Steve Henson] 466155714Skris 466255714Skris *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 466355714Skris an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 466455714Skris it will always use the numerical form of the OID, even if it has a short 466555714Skris or long name. 466655714Skris [Steve Henson] 466755714Skris 466855714Skris *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 466955714Skris method only got called if p,q,dmp1,dmq1,iqmp components were present, 467055714Skris otherwise bn_mod_exp was called. In the case of hardware keys for example 467155714Skris no private key components need be present and it might store extra data 467259191Skris in the RSA structure, which cannot be accessed from bn_mod_exp. 467359191Skris By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 467459191Skris private key operations. 467555714Skris [Steve Henson] 467655714Skris 467755714Skris *) Added support for SPARC Linux. 467855714Skris [Andy Polyakov] 467955714Skris 468055714Skris *) pem_password_cb function type incompatibly changed from 468155714Skris typedef int pem_password_cb(char *buf, int size, int rwflag); 468255714Skris to 468355714Skris ....(char *buf, int size, int rwflag, void *userdata); 468455714Skris so that applications can pass data to their callbacks: 468555714Skris The PEM[_ASN1]_{read,write}... functions and macros now take an 468655714Skris additional void * argument, which is just handed through whenever 468755714Skris the password callback is called. 468859191Skris [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 468955714Skris 469055714Skris New function SSL_CTX_set_default_passwd_cb_userdata. 469155714Skris 469255714Skris Compatibility note: As many C implementations push function arguments 469355714Skris onto the stack in reverse order, the new library version is likely to 469455714Skris interoperate with programs that have been compiled with the old 469555714Skris pem_password_cb definition (PEM_whatever takes some data that 469655714Skris happens to be on the stack as its last argument, and the callback 469755714Skris just ignores this garbage); but there is no guarantee whatsoever that 469855714Skris this will work. 469955714Skris 470055714Skris *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 470155714Skris (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 470255714Skris problems not only on Windows, but also on some Unix platforms. 470355714Skris To avoid problematic command lines, these definitions are now in an 470455714Skris auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 470555714Skris for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 470655714Skris [Bodo Moeller] 470755714Skris 470855714Skris *) MIPS III/IV assembler module is reimplemented. 470955714Skris [Andy Polyakov] 471055714Skris 471155714Skris *) More DES library cleanups: remove references to srand/rand and 471255714Skris delete an unused file. 471355714Skris [Ulf M�ller] 471455714Skris 471555714Skris *) Add support for the the free Netwide assembler (NASM) under Win32, 471655714Skris since not many people have MASM (ml) and it can be hard to obtain. 471755714Skris This is currently experimental but it seems to work OK and pass all 471855714Skris the tests. Check out INSTALL.W32 for info. 471955714Skris [Steve Henson] 472055714Skris 472155714Skris *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 472255714Skris without temporary keys kept an extra copy of the server key, 472355714Skris and connections with temporary keys did not free everything in case 472455714Skris of an error. 472555714Skris [Bodo Moeller] 472655714Skris 472755714Skris *) New function RSA_check_key and new openssl rsa option -check 472855714Skris for verifying the consistency of RSA keys. 472955714Skris [Ulf Moeller, Bodo Moeller] 473055714Skris 473155714Skris *) Various changes to make Win32 compile work: 473255714Skris 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 473355714Skris 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 473455714Skris comparison" warnings. 473555714Skris 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 473655714Skris [Steve Henson] 473755714Skris 473855714Skris *) Add a debugging option to PKCS#5 v2 key generation function: when 473955714Skris you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 474055714Skris derived keys are printed to stderr. 474155714Skris [Steve Henson] 474255714Skris 474355714Skris *) Copy the flags in ASN1_STRING_dup(). 474455714Skris [Roman E. Pavlov <pre@mo.msk.ru>] 474555714Skris 474655714Skris *) The x509 application mishandled signing requests containing DSA 474755714Skris keys when the signing key was also DSA and the parameters didn't match. 474855714Skris 474955714Skris It was supposed to omit the parameters when they matched the signing key: 475055714Skris the verifying software was then supposed to automatically use the CA's 475155714Skris parameters if they were absent from the end user certificate. 475255714Skris 475355714Skris Omitting parameters is no longer recommended. The test was also 475455714Skris the wrong way round! This was probably due to unusual behaviour in 475555714Skris EVP_cmp_parameters() which returns 1 if the parameters match. 475655714Skris This meant that parameters were omitted when they *didn't* match and 475755714Skris the certificate was useless. Certificates signed with 'ca' didn't have 475855714Skris this bug. 475955714Skris [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 476055714Skris 476155714Skris *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 476255714Skris The interface is as follows: 476355714Skris Applications can use 476455714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 476555714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 476655714Skris "off" is now the default. 476755714Skris The library internally uses 476855714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 476955714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 477055714Skris to disable memory-checking temporarily. 477155714Skris 477255714Skris Some inconsistent states that previously were possible (and were 477355714Skris even the default) are now avoided. 477455714Skris 477555714Skris -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 477655714Skris with each memory chunk allocated; this is occasionally more helpful 477755714Skris than just having a counter. 477855714Skris 477955714Skris -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 478055714Skris 478155714Skris -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 478255714Skris extensions. 478355714Skris [Bodo Moeller] 478455714Skris 478555714Skris *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 478655714Skris which largely parallels "options", but is for changing API behaviour, 478755714Skris whereas "options" are about protocol behaviour. 478855714Skris Initial "mode" flags are: 478955714Skris 479055714Skris SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 479155714Skris a single record has been written. 479255714Skris SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 479355714Skris retries use the same buffer location. 479455714Skris (But all of the contents must be 479555714Skris copied!) 479655714Skris [Bodo Moeller] 479755714Skris 479879998Skris *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 479955714Skris worked. 480055714Skris 480155714Skris *) Fix problems with no-hmac etc. 480255714Skris [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 480355714Skris 480455714Skris *) New functions RSA_get_default_method(), RSA_set_method() and 480555714Skris RSA_get_method(). These allows replacement of RSA_METHODs without having 480655714Skris to mess around with the internals of an RSA structure. 480755714Skris [Steve Henson] 480855714Skris 480955714Skris *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 481055714Skris Also really enable memory leak checks in openssl.c and in some 481155714Skris test programs. 481255714Skris [Chad C. Mulligan, Bodo Moeller] 481355714Skris 481455714Skris *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 481555714Skris up the length of negative integers. This has now been simplified to just 481655714Skris store the length when it is first determined and use it later, rather 481755714Skris than trying to keep track of where data is copied and updating it to 481855714Skris point to the end. 481955714Skris [Steve Henson, reported by Brien Wheeler 482055714Skris <bwheeler@authentica-security.com>] 482155714Skris 482255714Skris *) Add a new function PKCS7_signatureVerify. This allows the verification 482355714Skris of a PKCS#7 signature but with the signing certificate passed to the 482455714Skris function itself. This contrasts with PKCS7_dataVerify which assumes the 482555714Skris certificate is present in the PKCS#7 structure. This isn't always the 482655714Skris case: certificates can be omitted from a PKCS#7 structure and be 482755714Skris distributed by "out of band" means (such as a certificate database). 482855714Skris [Steve Henson] 482955714Skris 483055714Skris *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 483155714Skris function prototypes in pem.h, also change util/mkdef.pl to add the 483255714Skris necessary function names. 483355714Skris [Steve Henson] 483455714Skris 483555714Skris *) mk1mf.pl (used by Windows builds) did not properly read the 483655714Skris options set by Configure in the top level Makefile, and Configure 483755714Skris was not even able to write more than one option correctly. 483855714Skris Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 483955714Skris [Bodo Moeller] 484055714Skris 484155714Skris *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 484255714Skris file to be loaded from a BIO or FILE pointer. The BIO version will 484355714Skris for example allow memory BIOs to contain config info. 484455714Skris [Steve Henson] 484555714Skris 484655714Skris *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 484755714Skris Whoever hopes to achieve shared-library compatibility across versions 484855714Skris must use this, not the compile-time macro. 484955714Skris (Exercise 0.9.4: Which is the minimum library version required by 485055714Skris such programs?) 485155714Skris Note: All this applies only to multi-threaded programs, others don't 485255714Skris need locks. 485355714Skris [Bodo Moeller] 485455714Skris 485555714Skris *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 485655714Skris through a BIO pair triggered the default case, i.e. 485755714Skris SSLerr(...,SSL_R_UNKNOWN_STATE). 485855714Skris [Bodo Moeller] 485955714Skris 486055714Skris *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 486155714Skris can use the SSL library even if none of the specific BIOs is 486255714Skris appropriate. 486355714Skris [Bodo Moeller] 486455714Skris 486555714Skris *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 486655714Skris for the encoded length. 486755714Skris [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 486855714Skris 486955714Skris *) Add initial documentation of the X509V3 functions. 487055714Skris [Steve Henson] 487155714Skris 487255714Skris *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 487355714Skris PEM_write_bio_PKCS8PrivateKey() that are equivalent to 487455714Skris PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 487555714Skris secure PKCS#8 private key format with a high iteration count. 487655714Skris [Steve Henson] 487755714Skris 487855714Skris *) Fix determination of Perl interpreter: A perl or perl5 487955714Skris _directory_ in $PATH was also accepted as the interpreter. 488055714Skris [Ralf S. Engelschall] 488155714Skris 488255714Skris *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 488355714Skris wrong with it but it was very old and did things like calling 488455714Skris PEM_ASN1_read() directly and used MD5 for the hash not to mention some 488555714Skris unusual formatting. 488655714Skris [Steve Henson] 488755714Skris 488855714Skris *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 488955714Skris to use the new extension code. 489055714Skris [Steve Henson] 489155714Skris 489255714Skris *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 489355714Skris with macros. This should make it easier to change their form, add extra 489455714Skris arguments etc. Fix a few PEM prototypes which didn't have cipher as a 489555714Skris constant. 489655714Skris [Steve Henson] 489755714Skris 489855714Skris *) Add to configuration table a new entry that can specify an alternative 489955714Skris name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 490055714Skris according to Mark Crispin <MRC@Panda.COM>. 490155714Skris [Bodo Moeller] 490255714Skris 490355714Skris#if 0 490455714Skris *) DES CBC did not update the IV. Weird. 490555714Skris [Ben Laurie] 490655714Skris#else 490755714Skris des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 490855714Skris Changing the behaviour of the former might break existing programs -- 490955714Skris where IV updating is needed, des_ncbc_encrypt can be used. 491055714Skris#endif 491155714Skris 491255714Skris *) When bntest is run from "make test" it drives bc to check its 491355714Skris calculations, as well as internally checking them. If an internal check 491455714Skris fails, it needs to cause bc to give a non-zero result or make test carries 491555714Skris on without noticing the failure. Fixed. 491655714Skris [Ben Laurie] 491755714Skris 491855714Skris *) DES library cleanups. 491955714Skris [Ulf M�ller] 492055714Skris 492155714Skris *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 492255714Skris used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 492355714Skris ciphers. NOTE: although the key derivation function has been verified 492455714Skris against some published test vectors it has not been extensively tested 492555714Skris yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 492655714Skris of v2.0. 492755714Skris [Steve Henson] 492855714Skris 492955714Skris *) Instead of "mkdir -p", which is not fully portable, use new 493055714Skris Perl script "util/mkdir-p.pl". 493155714Skris [Bodo Moeller] 493255714Skris 493355714Skris *) Rewrite the way password based encryption (PBE) is handled. It used to 493455714Skris assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 493555714Skris structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 493655714Skris but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 493755714Skris the 'parameter' field of the AlgorithmIdentifier is passed to the 493855714Skris underlying key generation function so it must do its own ASN1 parsing. 493955714Skris This has also changed the EVP_PBE_CipherInit() function which now has a 494055714Skris 'parameter' argument instead of literal salt and iteration count values 494155714Skris and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 494255714Skris [Steve Henson] 494355714Skris 494455714Skris *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 494555714Skris and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 494655714Skris Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 494755714Skris KEY" because this clashed with PKCS#8 unencrypted string. Since this 494855714Skris value was just used as a "magic string" and not used directly its 494955714Skris value doesn't matter. 495055714Skris [Steve Henson] 495155714Skris 495255714Skris *) Introduce some semblance of const correctness to BN. Shame C doesn't 495355714Skris support mutable. 495455714Skris [Ben Laurie] 495555714Skris 495655714Skris *) "linux-sparc64" configuration (ultrapenguin). 495755714Skris [Ray Miller <ray.miller@oucs.ox.ac.uk>] 495855714Skris "linux-sparc" configuration. 495955714Skris [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 496055714Skris 496155714Skris *) config now generates no-xxx options for missing ciphers. 496255714Skris [Ulf M�ller] 496355714Skris 496455714Skris *) Support the EBCDIC character set (work in progress). 496555714Skris File ebcdic.c not yet included because it has a different license. 496655714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 496755714Skris 496855714Skris *) Support BS2000/OSD-POSIX. 496955714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 497055714Skris 497155714Skris *) Make callbacks for key generation use void * instead of char *. 497255714Skris [Ben Laurie] 497355714Skris 497455714Skris *) Make S/MIME samples compile (not yet tested). 497555714Skris [Ben Laurie] 497655714Skris 497755714Skris *) Additional typesafe stacks. 497855714Skris [Ben Laurie] 497955714Skris 498055714Skris *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 498155714Skris [Bodo Moeller] 498255714Skris 498355714Skris 498455714Skris Changes between 0.9.3 and 0.9.3a [29 May 1999] 498555714Skris 498655714Skris *) New configuration variant "sco5-gcc". 498755714Skris 498855714Skris *) Updated some demos. 498955714Skris [Sean O Riordain, Wade Scholine] 499055714Skris 499155714Skris *) Add missing BIO_free at exit of pkcs12 application. 499255714Skris [Wu Zhigang] 499355714Skris 499455714Skris *) Fix memory leak in conf.c. 499555714Skris [Steve Henson] 499655714Skris 499755714Skris *) Updates for Win32 to assembler version of MD5. 499855714Skris [Steve Henson] 499955714Skris 500055714Skris *) Set #! path to perl in apps/der_chop to where we found it 500155714Skris instead of using a fixed path. 500255714Skris [Bodo Moeller] 500355714Skris 500455714Skris *) SHA library changes for irix64-mips4-cc. 500555714Skris [Andy Polyakov] 500655714Skris 500755714Skris *) Improvements for VMS support. 500855714Skris [Richard Levitte] 500955714Skris 501055714Skris 501155714Skris Changes between 0.9.2b and 0.9.3 [24 May 1999] 501255714Skris 501355714Skris *) Bignum library bug fix. IRIX 6 passes "make test" now! 501455714Skris This also avoids the problems with SC4.2 and unpatched SC5. 501555714Skris [Andy Polyakov <appro@fy.chalmers.se>] 501655714Skris 501755714Skris *) New functions sk_num, sk_value and sk_set to replace the previous macros. 501855714Skris These are required because of the typesafe stack would otherwise break 501955714Skris existing code. If old code used a structure member which used to be STACK 502055714Skris and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 502155714Skris sk_num or sk_value it would produce an error because the num, data members 502255714Skris are not present in STACK_OF. Now it just produces a warning. sk_set 502355714Skris replaces the old method of assigning a value to sk_value 502455714Skris (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 502555714Skris that does this will no longer work (and should use sk_set instead) but 502655714Skris this could be regarded as a "questionable" behaviour anyway. 502755714Skris [Steve Henson] 502855714Skris 502955714Skris *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 503055714Skris correctly handle encrypted S/MIME data. 503155714Skris [Steve Henson] 503255714Skris 503355714Skris *) Change type of various DES function arguments from des_cblock 503455714Skris (which means, in function argument declarations, pointer to char) 503555714Skris to des_cblock * (meaning pointer to array with 8 char elements), 503655714Skris which allows the compiler to do more typechecking; it was like 503755714Skris that back in SSLeay, but with lots of ugly casts. 503855714Skris 503955714Skris Introduce new type const_des_cblock. 504055714Skris [Bodo Moeller] 504155714Skris 504255714Skris *) Reorganise the PKCS#7 library and get rid of some of the more obvious 504355714Skris problems: find RecipientInfo structure that matches recipient certificate 504455714Skris and initialise the ASN1 structures properly based on passed cipher. 504555714Skris [Steve Henson] 504655714Skris 504755714Skris *) Belatedly make the BN tests actually check the results. 504855714Skris [Ben Laurie] 504955714Skris 505055714Skris *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 505155714Skris to and from BNs: it was completely broken. New compilation option 505255714Skris NEG_PUBKEY_BUG to allow for some broken certificates that encode public 505355714Skris key elements as negative integers. 505455714Skris [Steve Henson] 505555714Skris 505655714Skris *) Reorganize and speed up MD5. 505755714Skris [Andy Polyakov <appro@fy.chalmers.se>] 505855714Skris 505955714Skris *) VMS support. 506055714Skris [Richard Levitte <richard@levitte.org>] 506155714Skris 506255714Skris *) New option -out to asn1parse to allow the parsed structure to be 506355714Skris output to a file. This is most useful when combined with the -strparse 506455714Skris option to examine the output of things like OCTET STRINGS. 506555714Skris [Steve Henson] 506655714Skris 506755714Skris *) Make SSL library a little more fool-proof by not requiring any longer 506855714Skris that SSL_set_{accept,connect}_state be called before 506955714Skris SSL_{accept,connect} may be used (SSL_set_..._state is omitted 507055714Skris in many applications because usually everything *appeared* to work as 507155714Skris intended anyway -- now it really works as intended). 507255714Skris [Bodo Moeller] 507355714Skris 507455714Skris *) Move openssl.cnf out of lib/. 507555714Skris [Ulf M�ller] 507655714Skris 507755714Skris *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 507855714Skris -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 507955714Skris -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 508055714Skris [Ralf S. Engelschall] 508155714Skris 508255714Skris *) Various fixes to the EVP and PKCS#7 code. It may now be able to 508355714Skris handle PKCS#7 enveloped data properly. 508455714Skris [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 508555714Skris 508655714Skris *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 508755714Skris copying pointers. The cert_st handling is changed by this in 508855714Skris various ways (and thus what used to be known as ctx->default_cert 508955714Skris is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 509055714Skris any longer when s->cert does not give us what we need). 509155714Skris ssl_cert_instantiate becomes obsolete by this change. 509255714Skris As soon as we've got the new code right (possibly it already is?), 509355714Skris we have solved a couple of bugs of the earlier code where s->cert 509455714Skris was used as if it could not have been shared with other SSL structures. 509555714Skris 509655714Skris Note that using the SSL API in certain dirty ways now will result 509755714Skris in different behaviour than observed with earlier library versions: 509855714Skris Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 509955714Skris does not influence s as it used to. 510055714Skris 510155714Skris In order to clean up things more thoroughly, inside SSL_SESSION 510255714Skris we don't use CERT any longer, but a new structure SESS_CERT 510355714Skris that holds per-session data (if available); currently, this is 510455714Skris the peer's certificate chain and, for clients, the server's certificate 510555714Skris and temporary key. CERT holds only those values that can have 510655714Skris meaningful defaults in an SSL_CTX. 510755714Skris [Bodo Moeller] 510855714Skris 510955714Skris *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 511055714Skris from the internal representation. Various PKCS#7 fixes: remove some 511155714Skris evil casts and set the enc_dig_alg field properly based on the signing 511255714Skris key type. 511355714Skris [Steve Henson] 511455714Skris 511555714Skris *) Allow PKCS#12 password to be set from the command line or the 511655714Skris environment. Let 'ca' get its config file name from the environment 511755714Skris variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 511855714Skris and 'x509'). 511955714Skris [Steve Henson] 512055714Skris 512155714Skris *) Allow certificate policies extension to use an IA5STRING for the 512255714Skris organization field. This is contrary to the PKIX definition but 512355714Skris VeriSign uses it and IE5 only recognises this form. Document 'x509' 512455714Skris extension option. 512555714Skris [Steve Henson] 512655714Skris 512755714Skris *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 512855714Skris without disallowing inline assembler and the like for non-pedantic builds. 512955714Skris [Ben Laurie] 513055714Skris 513155714Skris *) Support Borland C++ builder. 513255714Skris [Janez Jere <jj@void.si>, modified by Ulf M�ller] 513355714Skris 513455714Skris *) Support Mingw32. 513555714Skris [Ulf M�ller] 513655714Skris 513755714Skris *) SHA-1 cleanups and performance enhancements. 513855714Skris [Andy Polyakov <appro@fy.chalmers.se>] 513955714Skris 514055714Skris *) Sparc v8plus assembler for the bignum library. 514155714Skris [Andy Polyakov <appro@fy.chalmers.se>] 514255714Skris 514355714Skris *) Accept any -xxx and +xxx compiler options in Configure. 514455714Skris [Ulf M�ller] 514555714Skris 514655714Skris *) Update HPUX configuration. 514755714Skris [Anonymous] 514855714Skris 514955714Skris *) Add missing sk_<type>_unshift() function to safestack.h 515055714Skris [Ralf S. Engelschall] 515155714Skris 515255714Skris *) New function SSL_CTX_use_certificate_chain_file that sets the 515355714Skris "extra_cert"s in addition to the certificate. (This makes sense 515455714Skris only for "PEM" format files, as chains as a whole are not 515555714Skris DER-encoded.) 515655714Skris [Bodo Moeller] 515755714Skris 515855714Skris *) Support verify_depth from the SSL API. 515955714Skris x509_vfy.c had what can be considered an off-by-one-error: 516055714Skris Its depth (which was not part of the external interface) 516155714Skris was actually counting the number of certificates in a chain; 516255714Skris now it really counts the depth. 516355714Skris [Bodo Moeller] 516455714Skris 516555714Skris *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 516655714Skris instead of X509err, which often resulted in confusing error 516755714Skris messages since the error codes are not globally unique 516855714Skris (e.g. an alleged error in ssl3_accept when a certificate 516955714Skris didn't match the private key). 517055714Skris 517155714Skris *) New function SSL_CTX_set_session_id_context that allows to set a default 517255714Skris value (so that you don't need SSL_set_session_id_context for each 517355714Skris connection using the SSL_CTX). 517455714Skris [Bodo Moeller] 517555714Skris 517655714Skris *) OAEP decoding bug fix. 517755714Skris [Ulf M�ller] 517855714Skris 517955714Skris *) Support INSTALL_PREFIX for package builders, as proposed by 518055714Skris David Harris. 518155714Skris [Bodo Moeller] 518255714Skris 518355714Skris *) New Configure options "threads" and "no-threads". For systems 518455714Skris where the proper compiler options are known (currently Solaris 518555714Skris and Linux), "threads" is the default. 518655714Skris [Bodo Moeller] 518755714Skris 518855714Skris *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 518955714Skris [Bodo Moeller] 519055714Skris 519155714Skris *) Install various scripts to $(OPENSSLDIR)/misc, not to 519255714Skris $(INSTALLTOP)/bin -- they shouldn't clutter directories 519355714Skris such as /usr/local/bin. 519455714Skris [Bodo Moeller] 519555714Skris 519655714Skris *) "make linux-shared" to build shared libraries. 519755714Skris [Niels Poppe <niels@netbox.org>] 519855714Skris 519955714Skris *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 520055714Skris [Ulf M�ller] 520155714Skris 520255714Skris *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 520355714Skris extension adding in x509 utility. 520455714Skris [Steve Henson] 520555714Skris 520655714Skris *) Remove NOPROTO sections and error code comments. 520755714Skris [Ulf M�ller] 520855714Skris 520955714Skris *) Partial rewrite of the DEF file generator to now parse the ANSI 521055714Skris prototypes. 521155714Skris [Steve Henson] 521255714Skris 521355714Skris *) New Configure options --prefix=DIR and --openssldir=DIR. 521455714Skris [Ulf M�ller] 521555714Skris 521655714Skris *) Complete rewrite of the error code script(s). It is all now handled 521755714Skris by one script at the top level which handles error code gathering, 521855714Skris header rewriting and C source file generation. It should be much better 521955714Skris than the old method: it now uses a modified version of Ulf's parser to 522055714Skris read the ANSI prototypes in all header files (thus the old K&R definitions 522155714Skris aren't needed for error creation any more) and do a better job of 522255714Skris translating function codes into names. The old 'ASN1 error code imbedded 522355714Skris in a comment' is no longer necessary and it doesn't use .err files which 522455714Skris have now been deleted. Also the error code call doesn't have to appear all 522555714Skris on one line (which resulted in some large lines...). 522655714Skris [Steve Henson] 522755714Skris 522855714Skris *) Change #include filenames from <foo.h> to <openssl/foo.h>. 522955714Skris [Bodo Moeller] 523055714Skris 523155714Skris *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 523255714Skris 0 (which usually indicates a closed connection), but continue reading. 523355714Skris [Bodo Moeller] 523455714Skris 523555714Skris *) Fix some race conditions. 523655714Skris [Bodo Moeller] 523755714Skris 523855714Skris *) Add support for CRL distribution points extension. Add Certificate 523955714Skris Policies and CRL distribution points documentation. 524055714Skris [Steve Henson] 524155714Skris 524255714Skris *) Move the autogenerated header file parts to crypto/opensslconf.h. 524355714Skris [Ulf M�ller] 524455714Skris 524555714Skris *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 524655714Skris 8 of keying material. Merlin has also confirmed interop with this fix 524755714Skris between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 524855714Skris [Merlin Hughes <merlin@baltimore.ie>] 524955714Skris 525055714Skris *) Fix lots of warnings. 525155714Skris [Richard Levitte <levitte@stacken.kth.se>] 525255714Skris 525355714Skris *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 525455714Skris the directory spec didn't end with a LIST_SEPARATOR_CHAR. 525555714Skris [Richard Levitte <levitte@stacken.kth.se>] 525655714Skris 525755714Skris *) Fix problems with sizeof(long) == 8. 525855714Skris [Andy Polyakov <appro@fy.chalmers.se>] 525955714Skris 526055714Skris *) Change functions to ANSI C. 526155714Skris [Ulf M�ller] 526255714Skris 526355714Skris *) Fix typos in error codes. 526455714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 526555714Skris 526655714Skris *) Remove defunct assembler files from Configure. 526755714Skris [Ulf M�ller] 526855714Skris 526955714Skris *) SPARC v8 assembler BIGNUM implementation. 527055714Skris [Andy Polyakov <appro@fy.chalmers.se>] 527155714Skris 527255714Skris *) Support for Certificate Policies extension: both print and set. 527355714Skris Various additions to support the r2i method this uses. 527455714Skris [Steve Henson] 527555714Skris 527655714Skris *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 527755714Skris return a const string when you are expecting an allocated buffer. 527855714Skris [Ben Laurie] 527955714Skris 528055714Skris *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 528155714Skris types DirectoryString and DisplayText. 528255714Skris [Steve Henson] 528355714Skris 528455714Skris *) Add code to allow r2i extensions to access the configuration database, 528555714Skris add an LHASH database driver and add several ctx helper functions. 528655714Skris [Steve Henson] 528755714Skris 528855714Skris *) Fix an evil bug in bn_expand2() which caused various BN functions to 528955714Skris fail when they extended the size of a BIGNUM. 529055714Skris [Steve Henson] 529155714Skris 529255714Skris *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 529355714Skris support typesafe stack. 529455714Skris [Steve Henson] 529555714Skris 529655714Skris *) Fix typo in SSL_[gs]et_options(). 529755714Skris [Nils Frostberg <nils@medcom.se>] 529855714Skris 529955714Skris *) Delete various functions and files that belonged to the (now obsolete) 530055714Skris old X509V3 handling code. 530155714Skris [Steve Henson] 530255714Skris 530355714Skris *) New Configure option "rsaref". 530455714Skris [Ulf M�ller] 530555714Skris 530655714Skris *) Don't auto-generate pem.h. 530755714Skris [Bodo Moeller] 530855714Skris 530955714Skris *) Introduce type-safe ASN.1 SETs. 531055714Skris [Ben Laurie] 531155714Skris 531255714Skris *) Convert various additional casted stacks to type-safe STACK_OF() variants. 531355714Skris [Ben Laurie, Ralf S. Engelschall, Steve Henson] 531455714Skris 531555714Skris *) Introduce type-safe STACKs. This will almost certainly break lots of code 531655714Skris that links with OpenSSL (well at least cause lots of warnings), but fear 531755714Skris not: the conversion is trivial, and it eliminates loads of evil casts. A 531855714Skris few STACKed things have been converted already. Feel free to convert more. 531955714Skris In the fullness of time, I'll do away with the STACK type altogether. 532055714Skris [Ben Laurie] 532155714Skris 532255714Skris *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 532355714Skris specified in <certfile> by updating the entry in the index.txt file. 532455714Skris This way one no longer has to edit the index.txt file manually for 532555714Skris revoking a certificate. The -revoke option does the gory details now. 532655714Skris [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 532755714Skris 532855714Skris *) Fix `openssl crl -noout -text' combination where `-noout' killed the 532955714Skris `-text' option at all and this way the `-noout -text' combination was 533055714Skris inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 533155714Skris [Ralf S. Engelschall] 533255714Skris 533355714Skris *) Make sure a corresponding plain text error message exists for the 533455714Skris X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 533555714Skris verify callback function determined that a certificate was revoked. 533655714Skris [Ralf S. Engelschall] 533755714Skris 533855714Skris *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 533955714Skris ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 534055714Skris all available cipers including rc5, which was forgotten until now. 534155714Skris In order to let the testing shell script know which algorithms 534255714Skris are available, a new (up to now undocumented) command 534355714Skris "openssl list-cipher-commands" is used. 534455714Skris [Bodo Moeller] 534555714Skris 534655714Skris *) Bugfix: s_client occasionally would sleep in select() when 534755714Skris it should have checked SSL_pending() first. 534855714Skris [Bodo Moeller] 534955714Skris 535055714Skris *) New functions DSA_do_sign and DSA_do_verify to provide access to 535155714Skris the raw DSA values prior to ASN.1 encoding. 535255714Skris [Ulf M�ller] 535355714Skris 535455714Skris *) Tweaks to Configure 535555714Skris [Niels Poppe <niels@netbox.org>] 535655714Skris 535755714Skris *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 535855714Skris yet... 535955714Skris [Steve Henson] 536055714Skris 536155714Skris *) New variables $(RANLIB) and $(PERL) in the Makefiles. 536255714Skris [Ulf M�ller] 536355714Skris 536455714Skris *) New config option to avoid instructions that are illegal on the 80386. 536555714Skris The default code is faster, but requires at least a 486. 536655714Skris [Ulf M�ller] 536755714Skris 536855714Skris *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 536955714Skris SSL2_SERVER_VERSION (not used at all) macros, which are now the 537055714Skris same as SSL2_VERSION anyway. 537155714Skris [Bodo Moeller] 537255714Skris 537355714Skris *) New "-showcerts" option for s_client. 537455714Skris [Bodo Moeller] 537555714Skris 537655714Skris *) Still more PKCS#12 integration. Add pkcs12 application to openssl 537755714Skris application. Various cleanups and fixes. 537855714Skris [Steve Henson] 537955714Skris 538055714Skris *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 538155714Skris modify error routines to work internally. Add error codes and PBE init 538255714Skris to library startup routines. 538355714Skris [Steve Henson] 538455714Skris 538555714Skris *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 538655714Skris packing functions to asn1 and evp. Changed function names and error 538755714Skris codes along the way. 538855714Skris [Steve Henson] 538955714Skris 539055714Skris *) PKCS12 integration: and so it begins... First of several patches to 539155714Skris slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 539255714Skris objects to objects.h 539355714Skris [Steve Henson] 539455714Skris 539555714Skris *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 539655714Skris and display support for Thawte strong extranet extension. 539755714Skris [Steve Henson] 539855714Skris 539955714Skris *) Add LinuxPPC support. 540055714Skris [Jeff Dubrule <igor@pobox.org>] 540155714Skris 540255714Skris *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 540355714Skris bn_div_words in alpha.s. 540455714Skris [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 540555714Skris 540655714Skris *) Make sure the RSA OAEP test is skipped under -DRSAref because 540755714Skris OAEP isn't supported when OpenSSL is built with RSAref. 540855714Skris [Ulf Moeller <ulf@fitug.de>] 540955714Skris 541055714Skris *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 541155714Skris so they no longer are missing under -DNOPROTO. 541255714Skris [Soren S. Jorvang <soren@t.dk>] 541355714Skris 541455714Skris 541555714Skris Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 541655714Skris 541755714Skris *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 541855714Skris doesn't work when the session is reused. Coming soon! 541955714Skris [Ben Laurie] 542055714Skris 542155714Skris *) Fix a security hole, that allows sessions to be reused in the wrong 542255714Skris context thus bypassing client cert protection! All software that uses 542355714Skris client certs and session caches in multiple contexts NEEDS PATCHING to 542455714Skris allow session reuse! A fuller solution is in the works. 542555714Skris [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 542655714Skris 542755714Skris *) Some more source tree cleanups (removed obsolete files 542855714Skris crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 542955714Skris permission on "config" script to be executable) and a fix for the INSTALL 543055714Skris document. 543155714Skris [Ulf Moeller <ulf@fitug.de>] 543255714Skris 543355714Skris *) Remove some legacy and erroneous uses of malloc, free instead of 543455714Skris Malloc, Free. 543555714Skris [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 543655714Skris 543755714Skris *) Make rsa_oaep_test return non-zero on error. 543855714Skris [Ulf Moeller <ulf@fitug.de>] 543955714Skris 544055714Skris *) Add support for native Solaris shared libraries. Configure 544155714Skris solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 544255714Skris if someone would make that last step automatic. 544355714Skris [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 544455714Skris 544555714Skris *) ctx_size was not built with the right compiler during "make links". Fixed. 544655714Skris [Ben Laurie] 544755714Skris 544855714Skris *) Change the meaning of 'ALL' in the cipher list. It now means "everything 544955714Skris except NULL ciphers". This means the default cipher list will no longer 545055714Skris enable NULL ciphers. They need to be specifically enabled e.g. with 545155714Skris the string "DEFAULT:eNULL". 545255714Skris [Steve Henson] 545355714Skris 545455714Skris *) Fix to RSA private encryption routines: if p < q then it would 545555714Skris occasionally produce an invalid result. This will only happen with 545655714Skris externally generated keys because OpenSSL (and SSLeay) ensure p > q. 545755714Skris [Steve Henson] 545855714Skris 545955714Skris *) Be less restrictive and allow also `perl util/perlpath.pl 546055714Skris /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 546155714Skris because this way one can also use an interpreter named `perl5' (which is 546255714Skris usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 546355714Skris installed as `perl'). 546455714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 546555714Skris 546655714Skris *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 546755714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 546855714Skris 546955714Skris *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 547055714Skris advapi32.lib to Win32 build and change the pem test comparision 547155714Skris to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 547255714Skris suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 547355714Skris and crypto/des/ede_cbcm_enc.c. 547455714Skris [Steve Henson] 547555714Skris 547655714Skris *) DES quad checksum was broken on big-endian architectures. Fixed. 547755714Skris [Ben Laurie] 547855714Skris 547955714Skris *) Comment out two functions in bio.h that aren't implemented. Fix up the 548055714Skris Win32 test batch file so it (might) work again. The Win32 test batch file 548155714Skris is horrible: I feel ill.... 548255714Skris [Steve Henson] 548355714Skris 548455714Skris *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 548555714Skris in e_os.h. Audit of header files to check ANSI and non ANSI 548655714Skris sections: 10 functions were absent from non ANSI section and not exported 548755714Skris from Windows DLLs. Fixed up libeay.num for new functions. 548855714Skris [Steve Henson] 548955714Skris 549055714Skris *) Make `openssl version' output lines consistent. 549155714Skris [Ralf S. Engelschall] 549255714Skris 549355714Skris *) Fix Win32 symbol export lists for BIO functions: Added 549455714Skris BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 549555714Skris to ms/libeay{16,32}.def. 549655714Skris [Ralf S. Engelschall] 549755714Skris 549855714Skris *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 549955714Skris fine under Unix and passes some trivial tests I've now added. But the 550055714Skris whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 550155714Skris added to make sure no one expects that this stuff really works in the 550255714Skris OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 550355714Skris up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 550455714Skris openssl_bio.xs. 550555714Skris [Ralf S. Engelschall] 550655714Skris 550755714Skris *) Fix the generation of two part addresses in perl. 550855714Skris [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 550955714Skris 551055714Skris *) Add config entry for Linux on MIPS. 551155714Skris [John Tobey <jtobey@channel1.com>] 551255714Skris 551355714Skris *) Make links whenever Configure is run, unless we are on Windoze. 551455714Skris [Ben Laurie] 551555714Skris 551655714Skris *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 551755714Skris Currently only issuerAltName and AuthorityKeyIdentifier make any sense 551855714Skris in CRLs. 551955714Skris [Steve Henson] 552055714Skris 552155714Skris *) Add a useful kludge to allow package maintainers to specify compiler and 552255714Skris other platforms details on the command line without having to patch the 552355714Skris Configure script everytime: One now can use ``perl Configure 552455714Skris <id>:<details>'', i.e. platform ids are allowed to have details appended 552555714Skris to them (seperated by colons). This is treated as there would be a static 552655714Skris pre-configured entry in Configure's %table under key <id> with value 552755714Skris <details> and ``perl Configure <id>'' is called. So, when you want to 552855714Skris perform a quick test-compile under FreeBSD 3.1 with pgcc and without 552955714Skris assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 553055714Skris now, which overrides the FreeBSD-elf entry on-the-fly. 553155714Skris [Ralf S. Engelschall] 553255714Skris 553355714Skris *) Disable new TLS1 ciphersuites by default: they aren't official yet. 553455714Skris [Ben Laurie] 553555714Skris 553655714Skris *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 553755714Skris on the `perl Configure ...' command line. This way one can compile 553855714Skris OpenSSL libraries with Position Independent Code (PIC) which is needed 553955714Skris for linking it into DSOs. 554055714Skris [Ralf S. Engelschall] 554155714Skris 554255714Skris *) Remarkably, export ciphers were totally broken and no-one had noticed! 554355714Skris Fixed. 554455714Skris [Ben Laurie] 554555714Skris 554655714Skris *) Cleaned up the LICENSE document: The official contact for any license 554755714Skris questions now is the OpenSSL core team under openssl-core@openssl.org. 554855714Skris And add a paragraph about the dual-license situation to make sure people 554955714Skris recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 555055714Skris to the OpenSSL toolkit. 555155714Skris [Ralf S. Engelschall] 555255714Skris 555355714Skris *) General source tree makefile cleanups: Made `making xxx in yyy...' 555455714Skris display consistent in the source tree and replaced `/bin/rm' by `rm'. 555555714Skris Additonally cleaned up the `make links' target: Remove unnecessary 555655714Skris semicolons, subsequent redundant removes, inline point.sh into mklink.sh 555755714Skris to speed processing and no longer clutter the display with confusing 555855714Skris stuff. Instead only the actually done links are displayed. 555955714Skris [Ralf S. Engelschall] 556055714Skris 556155714Skris *) Permit null encryption ciphersuites, used for authentication only. It used 556255714Skris to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 556355714Skris It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 556455714Skris encryption. 556555714Skris [Ben Laurie] 556655714Skris 556755714Skris *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 556855714Skris signed attributes when verifying signatures (this would break them), 556955714Skris the detached data encoding was wrong and public keys obtained using 557055714Skris X509_get_pubkey() weren't freed. 557155714Skris [Steve Henson] 557255714Skris 557355714Skris *) Add text documentation for the BUFFER functions. Also added a work around 557455714Skris to a Win95 console bug. This was triggered by the password read stuff: the 557555714Skris last character typed gets carried over to the next fread(). If you were 557655714Skris generating a new cert request using 'req' for example then the last 557755714Skris character of the passphrase would be CR which would then enter the first 557855714Skris field as blank. 557955714Skris [Steve Henson] 558055714Skris 558155714Skris *) Added the new `Includes OpenSSL Cryptography Software' button as 558255714Skris doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 558355714Skris button and can be used by applications based on OpenSSL to show the 558455714Skris relationship to the OpenSSL project. 558555714Skris [Ralf S. Engelschall] 558655714Skris 558755714Skris *) Remove confusing variables in function signatures in files 558855714Skris ssl/ssl_lib.c and ssl/ssl.h. 558955714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 559055714Skris 559155714Skris *) Don't install bss_file.c under PREFIX/include/ 559255714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 559355714Skris 559455714Skris *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 559555714Skris functions that return function pointers and has support for NT specific 559655714Skris stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 559755714Skris #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 559855714Skris unsigned to signed types: this was killing the Win32 compile. 559955714Skris [Steve Henson] 560055714Skris 560155714Skris *) Add new certificate file to stack functions, 560255714Skris SSL_add_dir_cert_subjects_to_stack() and 560355714Skris SSL_add_file_cert_subjects_to_stack(). These largely supplant 560455714Skris SSL_load_client_CA_file(), and can be used to add multiple certs easily 560555714Skris to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 560655714Skris This means that Apache-SSL and similar packages don't have to mess around 560755714Skris to add as many CAs as they want to the preferred list. 560855714Skris [Ben Laurie] 560955714Skris 561055714Skris *) Experiment with doxygen documentation. Currently only partially applied to 561155714Skris ssl/ssl_lib.c. 561255714Skris See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 561355714Skris openssl.doxy as the configuration file. 561455714Skris [Ben Laurie] 561555714Skris 561655714Skris *) Get rid of remaining C++-style comments which strict C compilers hate. 561755714Skris [Ralf S. Engelschall, pointed out by Carlos Amengual] 561855714Skris 561955714Skris *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 562055714Skris compiled in by default: it has problems with large keys. 562155714Skris [Steve Henson] 562255714Skris 562355714Skris *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 562455714Skris DH private keys and/or callback functions which directly correspond to 562555714Skris their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 562655714Skris is needed for applications which have to configure certificates on a 562755714Skris per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 562855714Skris (e.g. s_server). 562955714Skris For the RSA certificate situation is makes no difference, but 563055714Skris for the DSA certificate situation this fixes the "no shared cipher" 563155714Skris problem where the OpenSSL cipher selection procedure failed because the 563255714Skris temporary keys were not overtaken from the context and the API provided 563355714Skris no way to reconfigure them. 563455714Skris The new functions now let applications reconfigure the stuff and they 563555714Skris are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 563655714Skris SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 563755714Skris non-public-API function ssl_cert_instantiate() is used as a helper 563855714Skris function and also to reduce code redundancy inside ssl_rsa.c. 563955714Skris [Ralf S. Engelschall] 564055714Skris 564155714Skris *) Move s_server -dcert and -dkey options out of the undocumented feature 564255714Skris area because they are useful for the DSA situation and should be 564355714Skris recognized by the users. 564455714Skris [Ralf S. Engelschall] 564555714Skris 564655714Skris *) Fix the cipher decision scheme for export ciphers: the export bits are 564755714Skris *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 564855714Skris SSL_EXP_MASK. So, the original variable has to be used instead of the 564955714Skris already masked variable. 565055714Skris [Richard Levitte <levitte@stacken.kth.se>] 565155714Skris 565255714Skris *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 565355714Skris [Richard Levitte <levitte@stacken.kth.se>] 565455714Skris 565555714Skris *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 565655714Skris from `int' to `unsigned int' because it's a length and initialized by 565755714Skris EVP_DigestFinal() which expects an `unsigned int *'. 565855714Skris [Richard Levitte <levitte@stacken.kth.se>] 565955714Skris 566055714Skris *) Don't hard-code path to Perl interpreter on shebang line of Configure 566155714Skris script. Instead use the usual Shell->Perl transition trick. 566255714Skris [Ralf S. Engelschall] 566355714Skris 566455714Skris *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 566555714Skris (in addition to RSA certificates) to match the behaviour of `openssl dsa 566655714Skris -noout -modulus' as it's already the case for `openssl rsa -noout 566755714Skris -modulus'. For RSA the -modulus is the real "modulus" while for DSA 566855714Skris currently the public key is printed (a decision which was already done by 566955714Skris `openssl dsa -modulus' in the past) which serves a similar purpose. 567055714Skris Additionally the NO_RSA no longer completely removes the whole -modulus 567155714Skris option; it now only avoids using the RSA stuff. Same applies to NO_DSA 567255714Skris now, too. 567355714Skris [Ralf S. Engelschall] 567455714Skris 567555714Skris *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 567655714Skris BIO. See the source (crypto/evp/bio_ok.c) for more info. 567755714Skris [Arne Ansper <arne@ats.cyber.ee>] 567855714Skris 567955714Skris *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 568055714Skris to be added. Now both 'req' and 'ca' can use new objects defined in the 568155714Skris config file. 568255714Skris [Steve Henson] 568355714Skris 568455714Skris *) Add cool BIO that does syslog (or event log on NT). 568555714Skris [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 568655714Skris 568755714Skris *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 568855714Skris TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 568955714Skris TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 569055714Skris Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 569155714Skris [Ben Laurie] 569255714Skris 569355714Skris *) Add preliminary config info for new extension code. 569455714Skris [Steve Henson] 569555714Skris 569655714Skris *) Make RSA_NO_PADDING really use no padding. 569755714Skris [Ulf Moeller <ulf@fitug.de>] 569855714Skris 569955714Skris *) Generate errors when private/public key check is done. 570055714Skris [Ben Laurie] 570155714Skris 570255714Skris *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 570355714Skris for some CRL extensions and new objects added. 570455714Skris [Steve Henson] 570555714Skris 570655714Skris *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 570755714Skris key usage extension and fuller support for authority key id. 570855714Skris [Steve Henson] 570955714Skris 571055714Skris *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 571155714Skris padding method for RSA, which is recommended for new applications in PKCS 571255714Skris #1 v2.0 (RFC 2437, October 1998). 571355714Skris OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 571455714Skris foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 571555714Skris against Bleichbacher's attack on RSA. 571655714Skris [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 571755714Skris Ben Laurie] 571855714Skris 571955714Skris *) Updates to the new SSL compression code 572055714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 572155714Skris 572255714Skris *) Fix so that the version number in the master secret, when passed 572355714Skris via RSA, checks that if TLS was proposed, but we roll back to SSLv3 572455714Skris (because the server will not accept higher), that the version number 572555714Skris is 0x03,0x01, not 0x03,0x00 572655714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 572755714Skris 572855714Skris *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 572955714Skris leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 573055714Skris in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 573155714Skris [Steve Henson] 573255714Skris 573355714Skris *) Support for RAW extensions where an arbitrary extension can be 573455714Skris created by including its DER encoding. See apps/openssl.cnf for 573555714Skris an example. 573655714Skris [Steve Henson] 573755714Skris 573855714Skris *) Make sure latest Perl versions don't interpret some generated C array 573955714Skris code as Perl array code in the crypto/err/err_genc.pl script. 574055714Skris [Lars Weber <3weber@informatik.uni-hamburg.de>] 574155714Skris 574255714Skris *) Modify ms/do_ms.bat to not generate assembly language makefiles since 574355714Skris not many people have the assembler. Various Win32 compilation fixes and 574455714Skris update to the INSTALL.W32 file with (hopefully) more accurate Win32 574555714Skris build instructions. 574655714Skris [Steve Henson] 574755714Skris 574855714Skris *) Modify configure script 'Configure' to automatically create crypto/date.h 574955714Skris file under Win32 and also build pem.h from pem.org. New script 575055714Skris util/mkfiles.pl to create the MINFO file on environments that can't do a 575155714Skris 'make files': perl util/mkfiles.pl >MINFO should work. 575255714Skris [Steve Henson] 575355714Skris 575455714Skris *) Major rework of DES function declarations, in the pursuit of correctness 575555714Skris and purity. As a result, many evil casts evaporated, and some weirdness, 575655714Skris too. You may find this causes warnings in your code. Zapping your evil 575755714Skris casts will probably fix them. Mostly. 575855714Skris [Ben Laurie] 575955714Skris 576055714Skris *) Fix for a typo in asn1.h. Bug fix to object creation script 576155714Skris obj_dat.pl. It considered a zero in an object definition to mean 576255714Skris "end of object": none of the objects in objects.h have any zeros 576355714Skris so it wasn't spotted. 576455714Skris [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 576555714Skris 576655714Skris *) Add support for Triple DES Cipher Block Chaining with Output Feedback 576755714Skris Masking (CBCM). In the absence of test vectors, the best I have been able 576855714Skris to do is check that the decrypt undoes the encrypt, so far. Send me test 576955714Skris vectors if you have them. 577055714Skris [Ben Laurie] 577155714Skris 577255714Skris *) Correct calculation of key length for export ciphers (too much space was 577355714Skris allocated for null ciphers). This has not been tested! 577455714Skris [Ben Laurie] 577555714Skris 577655714Skris *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 577755714Skris message is now correct (it understands "crypto" and "ssl" on its 577855714Skris command line). There is also now an "update" option. This will update 577955714Skris the util/ssleay.num and util/libeay.num files with any new functions. 578055714Skris If you do a: 578155714Skris perl util/mkdef.pl crypto ssl update 578255714Skris it will update them. 578355714Skris [Steve Henson] 578455714Skris 578555714Skris *) Overhauled the Perl interface (perl/*): 578655714Skris - ported BN stuff to OpenSSL's different BN library 578755714Skris - made the perl/ source tree CVS-aware 578855714Skris - renamed the package from SSLeay to OpenSSL (the files still contain 578955714Skris their history because I've copied them in the repository) 579055714Skris - removed obsolete files (the test scripts will be replaced 579155714Skris by better Test::Harness variants in the future) 579255714Skris [Ralf S. Engelschall] 579355714Skris 579455714Skris *) First cut for a very conservative source tree cleanup: 579555714Skris 1. merge various obsolete readme texts into doc/ssleay.txt 579655714Skris where we collect the old documents and readme texts. 579755714Skris 2. remove the first part of files where I'm already sure that we no 579855714Skris longer need them because of three reasons: either they are just temporary 579955714Skris files which were left by Eric or they are preserved original files where 580055714Skris I've verified that the diff is also available in the CVS via "cvs diff 580155714Skris -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 580255714Skris the crypto/md/ stuff). 580355714Skris [Ralf S. Engelschall] 580455714Skris 580555714Skris *) More extension code. Incomplete support for subject and issuer alt 580655714Skris name, issuer and authority key id. Change the i2v function parameters 580755714Skris and add an extra 'crl' parameter in the X509V3_CTX structure: guess 580855714Skris what that's for :-) Fix to ASN1 macro which messed up 580955714Skris IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 581055714Skris [Steve Henson] 581155714Skris 581255714Skris *) Preliminary support for ENUMERATED type. This is largely copied from the 581355714Skris INTEGER code. 581455714Skris [Steve Henson] 581555714Skris 581655714Skris *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 581755714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 581855714Skris 581955714Skris *) Make sure `make rehash' target really finds the `openssl' program. 582055714Skris [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 582155714Skris 582255714Skris *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 582355714Skris like to hear about it if this slows down other processors. 582455714Skris [Ben Laurie] 582555714Skris 582655714Skris *) Add CygWin32 platform information to Configure script. 582755714Skris [Alan Batie <batie@aahz.jf.intel.com>] 582855714Skris 582955714Skris *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 583055714Skris [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 583155714Skris 583255714Skris *) New program nseq to manipulate netscape certificate sequences 583355714Skris [Steve Henson] 583455714Skris 583555714Skris *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 583655714Skris few typos. 583755714Skris [Steve Henson] 583855714Skris 583955714Skris *) Fixes to BN code. Previously the default was to define BN_RECURSION 584055714Skris but the BN code had some problems that would cause failures when 584155714Skris doing certificate verification and some other functions. 584255714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 584355714Skris 584455714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 584555714Skris [Steve Henson] 584655714Skris 584755714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 584855714Skris [Steve Henson] 584955714Skris 585055714Skris *) Add several PKIX and private extended key usage OIDs. 585155714Skris [Steve Henson] 585255714Skris 585355714Skris *) Modify the 'ca' program to handle the new extension code. Modify 585455714Skris openssl.cnf for new extension format, add comments. 585555714Skris [Steve Henson] 585655714Skris 585755714Skris *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 585855714Skris and add a sample to openssl.cnf so req -x509 now adds appropriate 585955714Skris CA extensions. 586055714Skris [Steve Henson] 586155714Skris 586255714Skris *) Continued X509 V3 changes. Add to other makefiles, integrate with the 586355714Skris error code, add initial support to X509_print() and x509 application. 586455714Skris [Steve Henson] 586555714Skris 586655714Skris *) Takes a deep breath and start addding X509 V3 extension support code. Add 586755714Skris files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 586855714Skris stuff is currently isolated and isn't even compiled yet. 586955714Skris [Steve Henson] 587055714Skris 587155714Skris *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 587255714Skris ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 587355714Skris Removed the versions check from X509 routines when loading extensions: 587455714Skris this allows certain broken certificates that don't set the version 587555714Skris properly to be processed. 587655714Skris [Steve Henson] 587755714Skris 587855714Skris *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 587955714Skris Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 588055714Skris can still be regenerated with "make depend". 588155714Skris [Ben Laurie] 588255714Skris 588355714Skris *) Spelling mistake in C version of CAST-128. 588455714Skris [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 588555714Skris 588655714Skris *) Changes to the error generation code. The perl script err-code.pl 588755714Skris now reads in the old error codes and retains the old numbers, only 588855714Skris adding new ones if necessary. It also only changes the .err files if new 588955714Skris codes are added. The makefiles have been modified to only insert errors 589055714Skris when needed (to avoid needlessly modifying header files). This is done 589155714Skris by only inserting errors if the .err file is newer than the auto generated 589255714Skris C file. To rebuild all the error codes from scratch (the old behaviour) 589355714Skris either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 589455714Skris or delete all the .err files. 589555714Skris [Steve Henson] 589655714Skris 589755714Skris *) CAST-128 was incorrectly implemented for short keys. The C version has 589855714Skris been fixed, but is untested. The assembler versions are also fixed, but 589955714Skris new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 590055714Skris to regenerate it if needed. 590155714Skris [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 590255714Skris Hagino <itojun@kame.net>] 590355714Skris 590455714Skris *) File was opened incorrectly in randfile.c. 590555714Skris [Ulf M�ller <ulf@fitug.de>] 590655714Skris 590755714Skris *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 590855714Skris functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 590955714Skris GeneralizedTime. ASN1_TIME is the proper type used in certificates et 591055714Skris al: it's just almost always a UTCTime. Note this patch adds new error 591155714Skris codes so do a "make errors" if there are problems. 591255714Skris [Steve Henson] 591355714Skris 591455714Skris *) Correct Linux 1 recognition in config. 591555714Skris [Ulf M�ller <ulf@fitug.de>] 591655714Skris 591755714Skris *) Remove pointless MD5 hash when using DSA keys in ca. 591855714Skris [Anonymous <nobody@replay.com>] 591955714Skris 592055714Skris *) Generate an error if given an empty string as a cert directory. Also 592155714Skris generate an error if handed NULL (previously returned 0 to indicate an 592255714Skris error, but didn't set one). 592355714Skris [Ben Laurie, reported by Anonymous <nobody@replay.com>] 592455714Skris 592555714Skris *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 592655714Skris [Ben Laurie] 592755714Skris 592855714Skris *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 592955714Skris parameters. This was causing a warning which killed off the Win32 compile. 593055714Skris [Steve Henson] 593155714Skris 593255714Skris *) Remove C++ style comments from crypto/bn/bn_local.h. 593355714Skris [Neil Costigan <neil.costigan@celocom.com>] 593455714Skris 593555714Skris *) The function OBJ_txt2nid was broken. It was supposed to return a nid 593655714Skris based on a text string, looking up short and long names and finally 593755714Skris "dot" format. The "dot" format stuff didn't work. Added new function 593855714Skris OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 593955714Skris OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 594055714Skris OID is not part of the table. 594155714Skris [Steve Henson] 594255714Skris 594355714Skris *) Add prototypes to X509 lookup/verify methods, fixing a bug in 594455714Skris X509_LOOKUP_by_alias(). 594555714Skris [Ben Laurie] 594655714Skris 594755714Skris *) Sort openssl functions by name. 594855714Skris [Ben Laurie] 594955714Skris 595055714Skris *) Get the gendsa program working (hopefully) and add it to app list. Remove 595155714Skris encryption from sample DSA keys (in case anyone is interested the password 595255714Skris was "1234"). 595355714Skris [Steve Henson] 595455714Skris 595555714Skris *) Make _all_ *_free functions accept a NULL pointer. 595655714Skris [Frans Heymans <fheymans@isaserver.be>] 595755714Skris 595855714Skris *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 595955714Skris NULL pointers. 596055714Skris [Anonymous <nobody@replay.com>] 596155714Skris 596255714Skris *) s_server should send the CAfile as acceptable CAs, not its own cert. 596355714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 596455714Skris 596555714Skris *) Don't blow it for numeric -newkey arguments to apps/req. 596655714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 596755714Skris 596855714Skris *) Temp key "for export" tests were wrong in s3_srvr.c. 596955714Skris [Anonymous <nobody@replay.com>] 597055714Skris 597155714Skris *) Add prototype for temp key callback functions 597255714Skris SSL_CTX_set_tmp_{rsa,dh}_callback(). 597355714Skris [Ben Laurie] 597455714Skris 597555714Skris *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 597655714Skris DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 597755714Skris [Steve Henson] 597855714Skris 597955714Skris *) X509_name_add_entry() freed the wrong thing after an error. 598055714Skris [Arne Ansper <arne@ats.cyber.ee>] 598155714Skris 598255714Skris *) rsa_eay.c would attempt to free a NULL context. 598355714Skris [Arne Ansper <arne@ats.cyber.ee>] 598455714Skris 598555714Skris *) BIO_s_socket() had a broken should_retry() on Windoze. 598655714Skris [Arne Ansper <arne@ats.cyber.ee>] 598755714Skris 598855714Skris *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 598955714Skris [Arne Ansper <arne@ats.cyber.ee>] 599055714Skris 599155714Skris *) Make sure the already existing X509_STORE->depth variable is initialized 599255714Skris in X509_STORE_new(), but document the fact that this variable is still 599355714Skris unused in the certificate verification process. 599455714Skris [Ralf S. Engelschall] 599555714Skris 599655714Skris *) Fix the various library and apps files to free up pkeys obtained from 599755714Skris X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 599855714Skris [Steve Henson] 599955714Skris 600055714Skris *) Fix reference counting in X509_PUBKEY_get(). This makes 600155714Skris demos/maurice/example2.c work, amongst others, probably. 600255714Skris [Steve Henson and Ben Laurie] 600355714Skris 600455714Skris *) First cut of a cleanup for apps/. First the `ssleay' program is now named 600555714Skris `openssl' and second, the shortcut symlinks for the `openssl <command>' 600655714Skris are no longer created. This way we have a single and consistent command 600755714Skris line interface `openssl <command>', similar to `cvs <command>'. 600855714Skris [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 600955714Skris 601055714Skris *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 601155714Skris BIT STRING wrapper always have zero unused bits. 601255714Skris [Steve Henson] 601355714Skris 601455714Skris *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 601555714Skris [Steve Henson] 601655714Skris 601755714Skris *) Make the top-level INSTALL documentation easier to understand. 601855714Skris [Paul Sutton] 601955714Skris 602055714Skris *) Makefiles updated to exit if an error occurs in a sub-directory 602155714Skris make (including if user presses ^C) [Paul Sutton] 602255714Skris 602355714Skris *) Make Montgomery context stuff explicit in RSA data structure. 602455714Skris [Ben Laurie] 602555714Skris 602655714Skris *) Fix build order of pem and err to allow for generated pem.h. 602755714Skris [Ben Laurie] 602855714Skris 602955714Skris *) Fix renumbering bug in X509_NAME_delete_entry(). 603055714Skris [Ben Laurie] 603155714Skris 603255714Skris *) Enhanced the err-ins.pl script so it makes the error library number 603355714Skris global and can add a library name. This is needed for external ASN1 and 603455714Skris other error libraries. 603555714Skris [Steve Henson] 603655714Skris 603755714Skris *) Fixed sk_insert which never worked properly. 603855714Skris [Steve Henson] 603955714Skris 604055714Skris *) Fix ASN1 macros so they can handle indefinite length construted 604155714Skris EXPLICIT tags. Some non standard certificates use these: they can now 604255714Skris be read in. 604355714Skris [Steve Henson] 604455714Skris 604555714Skris *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 604655714Skris into a single doc/ssleay.txt bundle. This way the information is still 604755714Skris preserved but no longer messes up this directory. Now it's new room for 604855714Skris the new set of documenation files. 604955714Skris [Ralf S. Engelschall] 605055714Skris 605155714Skris *) SETs were incorrectly DER encoded. This was a major pain, because they 605255714Skris shared code with SEQUENCEs, which aren't coded the same. This means that 605355714Skris almost everything to do with SETs or SEQUENCEs has either changed name or 605455714Skris number of arguments. 605555714Skris [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 605655714Skris 605755714Skris *) Fix test data to work with the above. 605855714Skris [Ben Laurie] 605955714Skris 606055714Skris *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 606155714Skris was already fixed by Eric for 0.9.1 it seems. 606255714Skris [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 606355714Skris 606455714Skris *) Autodetect FreeBSD3. 606555714Skris [Ben Laurie] 606655714Skris 606755714Skris *) Fix various bugs in Configure. This affects the following platforms: 606855714Skris nextstep 606955714Skris ncr-scde 607055714Skris unixware-2.0 607155714Skris unixware-2.0-pentium 607255714Skris sco5-cc. 607355714Skris [Ben Laurie] 607455714Skris 607555714Skris *) Eliminate generated files from CVS. Reorder tests to regenerate files 607655714Skris before they are needed. 607755714Skris [Ben Laurie] 607855714Skris 607955714Skris *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 608055714Skris [Ben Laurie] 608155714Skris 608255714Skris 608355714Skris Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 608455714Skris 608555714Skris *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 608655714Skris changed SSLeay to OpenSSL in version strings. 608755714Skris [Ralf S. Engelschall] 608855714Skris 608955714Skris *) Some fixups to the top-level documents. 609055714Skris [Paul Sutton] 609155714Skris 609255714Skris *) Fixed the nasty bug where rsaref.h was not found under compile-time 609355714Skris because the symlink to include/ was missing. 609455714Skris [Ralf S. Engelschall] 609555714Skris 609655714Skris *) Incorporated the popular no-RSA/DSA-only patches 609755714Skris which allow to compile a RSA-free SSLeay. 609855714Skris [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 609955714Skris 610055714Skris *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 610155714Skris when "ssleay" is still not found. 610255714Skris [Ralf S. Engelschall] 610355714Skris 610455714Skris *) Added more platforms to Configure: Cray T3E, HPUX 11, 610555714Skris [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 610655714Skris 610755714Skris *) Updated the README file. 610855714Skris [Ralf S. Engelschall] 610955714Skris 611055714Skris *) Added various .cvsignore files in the CVS repository subdirs 611155714Skris to make a "cvs update" really silent. 611255714Skris [Ralf S. Engelschall] 611355714Skris 611455714Skris *) Recompiled the error-definition header files and added 611555714Skris missing symbols to the Win32 linker tables. 611655714Skris [Ralf S. Engelschall] 611755714Skris 611855714Skris *) Cleaned up the top-level documents; 611955714Skris o new files: CHANGES and LICENSE 612055714Skris o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 612155714Skris o merged COPYRIGHT into LICENSE 612255714Skris o removed obsolete TODO file 612355714Skris o renamed MICROSOFT to INSTALL.W32 612455714Skris [Ralf S. Engelschall] 612555714Skris 612655714Skris *) Removed dummy files from the 0.9.1b source tree: 612755714Skris crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 612855714Skris crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 612955714Skris crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 613055714Skris crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 613155714Skris util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 613255714Skris [Ralf S. Engelschall] 613355714Skris 613455714Skris *) Added various platform portability fixes. 613555714Skris [Mark J. Cox] 613655714Skris 613755714Skris *) The Genesis of the OpenSSL rpject: 613855714Skris We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 613955714Skris Young and Tim J. Hudson created while they were working for C2Net until 614055714Skris summer 1998. 614155714Skris [The OpenSSL Project] 614255714Skris 614355714Skris 614455714Skris Changes between 0.9.0b and 0.9.1b [not released] 614555714Skris 614655714Skris *) Updated a few CA certificates under certs/ 614755714Skris [Eric A. Young] 614855714Skris 614955714Skris *) Changed some BIGNUM api stuff. 615055714Skris [Eric A. Young] 615155714Skris 615255714Skris *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 615355714Skris DGUX x86, Linux Alpha, etc. 615455714Skris [Eric A. Young] 615555714Skris 615655714Skris *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 615755714Skris RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 615855714Skris available). 615955714Skris [Eric A. Young] 616055714Skris 616155714Skris *) Add -strparse option to asn1pars program which parses nested 616255714Skris binary structures 616355714Skris [Dr Stephen Henson <shenson@bigfoot.com>] 616455714Skris 616555714Skris *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 616655714Skris [Eric A. Young] 616755714Skris 616855714Skris *) DSA fix for "ca" program. 616955714Skris [Eric A. Young] 617055714Skris 617155714Skris *) Added "-genkey" option to "dsaparam" program. 617255714Skris [Eric A. Young] 617355714Skris 617455714Skris *) Added RIPE MD160 (rmd160) message digest. 617555714Skris [Eric A. Young] 617655714Skris 617755714Skris *) Added -a (all) option to "ssleay version" command. 617855714Skris [Eric A. Young] 617955714Skris 618055714Skris *) Added PLATFORM define which is the id given to Configure. 618155714Skris [Eric A. Young] 618255714Skris 618355714Skris *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 618455714Skris [Eric A. Young] 618555714Skris 618655714Skris *) Extended the ASN.1 parser routines. 618755714Skris [Eric A. Young] 618855714Skris 618955714Skris *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 619055714Skris [Eric A. Young] 619155714Skris 619255714Skris *) Added a BN_CTX to the BN library. 619355714Skris [Eric A. Young] 619455714Skris 619555714Skris *) Fixed the weak key values in DES library 619655714Skris [Eric A. Young] 619755714Skris 619855714Skris *) Changed API in EVP library for cipher aliases. 619955714Skris [Eric A. Young] 620055714Skris 620155714Skris *) Added support for RC2/64bit cipher. 620255714Skris [Eric A. Young] 620355714Skris 620455714Skris *) Converted the lhash library to the crypto/mem.c functions. 620555714Skris [Eric A. Young] 620655714Skris 620755714Skris *) Added more recognized ASN.1 object ids. 620855714Skris [Eric A. Young] 620955714Skris 621055714Skris *) Added more RSA padding checks for SSL/TLS. 621155714Skris [Eric A. Young] 621255714Skris 621355714Skris *) Added BIO proxy/filter functionality. 621455714Skris [Eric A. Young] 621555714Skris 621655714Skris *) Added extra_certs to SSL_CTX which can be used 621755714Skris send extra CA certificates to the client in the CA cert chain sending 621855714Skris process. It can be configured with SSL_CTX_add_extra_chain_cert(). 621955714Skris [Eric A. Young] 622055714Skris 622155714Skris *) Now Fortezza is denied in the authentication phase because 622255714Skris this is key exchange mechanism is not supported by SSLeay at all. 622355714Skris [Eric A. Young] 622455714Skris 622555714Skris *) Additional PKCS1 checks. 622655714Skris [Eric A. Young] 622755714Skris 622855714Skris *) Support the string "TLSv1" for all TLS v1 ciphers. 622955714Skris [Eric A. Young] 623055714Skris 623155714Skris *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 623255714Skris ex_data index of the SSL context in the X509_STORE_CTX ex_data. 623355714Skris [Eric A. Young] 623455714Skris 623555714Skris *) Fixed a few memory leaks. 623655714Skris [Eric A. Young] 623755714Skris 623855714Skris *) Fixed various code and comment typos. 623955714Skris [Eric A. Young] 624055714Skris 624155714Skris *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 624255714Skris bytes sent in the client random. 624355714Skris [Edward Bishop <ebishop@spyglass.com>] 624455714Skris 6245