CHANGES revision 111147
155714Skris 255714Skris OpenSSL CHANGES 355714Skris _______________ 455714Skris 5111147Snectar Changes between 0.9.7 and 0.9.7a [19 Feb 2003] 6111147Snectar 7111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 8111147Snectar via timing by performing a MAC computation even if incorrrect 9111147Snectar block cipher padding has been found. This is a countermeasure 10111147Snectar against active attacks where the attacker has to distinguish 11111147Snectar between bad padding and a MAC verification error. (CAN-2003-0078) 12111147Snectar 13111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 14111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 15111147Snectar Martin Vuagnoux (EPFL, Ilion)] 16111147Snectar 17111147Snectar *) Make the no-err option work as intended. The intention with no-err 18111147Snectar is not to have the whole error stack handling routines removed from 19111147Snectar libcrypto, it's only intended to remove all the function name and 20111147Snectar reason texts, thereby removing some of the footprint that may not 21111147Snectar be interesting if those errors aren't displayed anyway. 22111147Snectar 23111147Snectar NOTE: it's still possible for any application or module to have it's 24111147Snectar own set of error texts inserted. The routines are there, just not 25111147Snectar used by default when no-err is given. 26111147Snectar [Richard Levitte] 27111147Snectar 28111147Snectar *) Add support for FreeBSD on IA64. 29111147Snectar [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454] 30111147Snectar 31111147Snectar *) Adjust DES_cbc_cksum() so it returns the same value as the MIT 32111147Snectar Kerberos function mit_des_cbc_cksum(). Before this change, 33111147Snectar the value returned by DES_cbc_cksum() was like the one from 34111147Snectar mit_des_cbc_cksum(), except the bytes were swapped. 35111147Snectar [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte] 36111147Snectar 37111147Snectar *) Allow an application to disable the automatic SSL chain building. 38111147Snectar Before this a rather primitive chain build was always performed in 39111147Snectar ssl3_output_cert_chain(): an application had no way to send the 40111147Snectar correct chain if the automatic operation produced an incorrect result. 41111147Snectar 42111147Snectar Now the chain builder is disabled if either: 43111147Snectar 44111147Snectar 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert(). 45111147Snectar 46111147Snectar 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set. 47111147Snectar 48111147Snectar The reasoning behind this is that an application would not want the 49111147Snectar auto chain building to take place if extra chain certificates are 50111147Snectar present and it might also want a means of sending no additional 51111147Snectar certificates (for example the chain has two certificates and the 52111147Snectar root is omitted). 53111147Snectar [Steve Henson] 54111147Snectar 55111147Snectar *) Add the possibility to build without the ENGINE framework. 56111147Snectar [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 57111147Snectar 58111147Snectar *) Under Win32 gmtime() can return NULL: check return value in 59111147Snectar OPENSSL_gmtime(). Add error code for case where gmtime() fails. 60111147Snectar [Steve Henson] 61111147Snectar 62111147Snectar *) DSA routines: under certain error conditions uninitialized BN objects 63111147Snectar could be freed. Solution: make sure initialization is performed early 64111147Snectar enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>, 65111147Snectar Nils Larsch <nla@trustcenter.de> via PR#459) 66111147Snectar [Lutz Jaenicke] 67111147Snectar 68111147Snectar *) Another fix for SSLv2 session ID handling: the session ID was incorrectly 69111147Snectar checked on reconnect on the client side, therefore session resumption 70111147Snectar could still fail with a "ssl session id is different" error. This 71111147Snectar behaviour is masked when SSL_OP_ALL is used due to 72111147Snectar SSL_OP_MICROSOFT_SESS_ID_BUG being set. 73111147Snectar Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 74111147Snectar followup to PR #377. 75111147Snectar [Lutz Jaenicke] 76111147Snectar 77111147Snectar *) IA-32 assembler support enhancements: unified ELF targets, support 78111147Snectar for SCO/Caldera platforms, fix for Cygwin shared build. 79111147Snectar [Andy Polyakov] 80111147Snectar 81111147Snectar *) Add support for FreeBSD on sparc64. As a consequence, support for 82111147Snectar FreeBSD on non-x86 processors is separate from x86 processors on 83111147Snectar the config script, much like the NetBSD support. 84111147Snectar [Richard Levitte & Kris Kennaway <kris@obsecurity.org>] 85111147Snectar 86109998Smarkm Changes between 0.9.6h and 0.9.7 [31 Dec 2002] 87109998Smarkm 88109998Smarkm *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED 89109998Smarkm code (06) was taken as the first octet of the session ID and the last 90109998Smarkm octet was ignored consequently. As a result SSLv2 client side session 91109998Smarkm caching could not have worked due to the session ID mismatch between 92109998Smarkm client and server. 93109998Smarkm Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 94109998Smarkm PR #377. 95109998Smarkm [Lutz Jaenicke] 96109998Smarkm 97109998Smarkm *) Change the declaration of needed Kerberos libraries to use EX_LIBS 98109998Smarkm instead of the special (and badly supported) LIBKRB5. LIBKRB5 is 99109998Smarkm removed entirely. 100109998Smarkm [Richard Levitte] 101109998Smarkm 102109998Smarkm *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it 103109998Smarkm seems that in spite of existing for more than a year, many application 104109998Smarkm author have done nothing to provide the necessary callbacks, which 105109998Smarkm means that this particular engine will not work properly anywhere. 106109998Smarkm This is a very unfortunate situation which forces us, in the name 107109998Smarkm of usability, to give the hw_ncipher.c a static lock, which is part 108109998Smarkm of libcrypto. 109109998Smarkm NOTE: This is for the 0.9.7 series ONLY. This hack will never 110109998Smarkm appear in 0.9.8 or later. We EXPECT application authors to have 111109998Smarkm dealt properly with this when 0.9.8 is released (unless we actually 112109998Smarkm make such changes in the libcrypto locking code that changes will 113109998Smarkm have to be made anyway). 114109998Smarkm [Richard Levitte] 115109998Smarkm 116109998Smarkm *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content 117109998Smarkm octets have been read, EOF or an error occurs. Without this change 118109998Smarkm some truncated ASN1 structures will not produce an error. 119109998Smarkm [Steve Henson] 120109998Smarkm 121109998Smarkm *) Disable Heimdal support, since it hasn't been fully implemented. 122109998Smarkm Still give the possibility to force the use of Heimdal, but with 123109998Smarkm warnings and a request that patches get sent to openssl-dev. 124109998Smarkm [Richard Levitte] 125109998Smarkm 126109998Smarkm *) Add the VC-CE target, introduce the WINCE sysname, and add 127109998Smarkm INSTALL.WCE and appropriate conditionals to make it build. 128109998Smarkm [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 129109998Smarkm 130109998Smarkm *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and 131109998Smarkm cygssl-x.y.z.dll, where x, y and z are the major, minor and 132109998Smarkm edit numbers of the version. 133109998Smarkm [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 134109998Smarkm 135109998Smarkm *) Introduce safe string copy and catenation functions 136109998Smarkm (BUF_strlcpy() and BUF_strlcat()). 137109998Smarkm [Ben Laurie (CHATS) and Richard Levitte] 138109998Smarkm 139109998Smarkm *) Avoid using fixed-size buffers for one-line DNs. 140109998Smarkm [Ben Laurie (CHATS)] 141109998Smarkm 142109998Smarkm *) Add BUF_MEM_grow_clean() to avoid information leakage when 143109998Smarkm resizing buffers containing secrets, and use where appropriate. 144109998Smarkm [Ben Laurie (CHATS)] 145109998Smarkm 146109998Smarkm *) Avoid using fixed size buffers for configuration file location. 147109998Smarkm [Ben Laurie (CHATS)] 148109998Smarkm 149109998Smarkm *) Avoid filename truncation for various CA files. 150109998Smarkm [Ben Laurie (CHATS)] 151109998Smarkm 152109998Smarkm *) Use sizeof in preference to magic numbers. 153109998Smarkm [Ben Laurie (CHATS)] 154109998Smarkm 155109998Smarkm *) Avoid filename truncation in cert requests. 156109998Smarkm [Ben Laurie (CHATS)] 157109998Smarkm 158109998Smarkm *) Add assertions to check for (supposedly impossible) buffer 159109998Smarkm overflows. 160109998Smarkm [Ben Laurie (CHATS)] 161109998Smarkm 162109998Smarkm *) Don't cache truncated DNS entries in the local cache (this could 163109998Smarkm potentially lead to a spoofing attack). 164109998Smarkm [Ben Laurie (CHATS)] 165109998Smarkm 166109998Smarkm *) Fix various buffers to be large enough for hex/decimal 167109998Smarkm representations in a platform independent manner. 168109998Smarkm [Ben Laurie (CHATS)] 169109998Smarkm 170109998Smarkm *) Add CRYPTO_realloc_clean() to avoid information leakage when 171109998Smarkm resizing buffers containing secrets, and use where appropriate. 172109998Smarkm [Ben Laurie (CHATS)] 173109998Smarkm 174109998Smarkm *) Add BIO_indent() to avoid much slightly worrying code to do 175109998Smarkm indents. 176109998Smarkm [Ben Laurie (CHATS)] 177109998Smarkm 178109998Smarkm *) Convert sprintf()/BIO_puts() to BIO_printf(). 179109998Smarkm [Ben Laurie (CHATS)] 180109998Smarkm 181109998Smarkm *) buffer_gets() could terminate with the buffer only half 182109998Smarkm full. Fixed. 183109998Smarkm [Ben Laurie (CHATS)] 184109998Smarkm 185109998Smarkm *) Add assertions to prevent user-supplied crypto functions from 186109998Smarkm overflowing internal buffers by having large block sizes, etc. 187109998Smarkm [Ben Laurie (CHATS)] 188109998Smarkm 189109998Smarkm *) New OPENSSL_assert() macro (similar to assert(), but enabled 190109998Smarkm unconditionally). 191109998Smarkm [Ben Laurie (CHATS)] 192109998Smarkm 193109998Smarkm *) Eliminate unused copy of key in RC4. 194109998Smarkm [Ben Laurie (CHATS)] 195109998Smarkm 196109998Smarkm *) Eliminate unused and incorrectly sized buffers for IV in pem.h. 197109998Smarkm [Ben Laurie (CHATS)] 198109998Smarkm 199109998Smarkm *) Fix off-by-one error in EGD path. 200109998Smarkm [Ben Laurie (CHATS)] 201109998Smarkm 202109998Smarkm *) If RANDFILE path is too long, ignore instead of truncating. 203109998Smarkm [Ben Laurie (CHATS)] 204109998Smarkm 205109998Smarkm *) Eliminate unused and incorrectly sized X.509 structure 206109998Smarkm CBCParameter. 207109998Smarkm [Ben Laurie (CHATS)] 208109998Smarkm 209109998Smarkm *) Eliminate unused and dangerous function knumber(). 210109998Smarkm [Ben Laurie (CHATS)] 211109998Smarkm 212109998Smarkm *) Eliminate unused and dangerous structure, KSSL_ERR. 213109998Smarkm [Ben Laurie (CHATS)] 214109998Smarkm 215109998Smarkm *) Protect against overlong session ID context length in an encoded 216109998Smarkm session object. Since these are local, this does not appear to be 217109998Smarkm exploitable. 218109998Smarkm [Ben Laurie (CHATS)] 219109998Smarkm 220109998Smarkm *) Change from security patch (see 0.9.6e below) that did not affect 221109998Smarkm the 0.9.6 release series: 222109998Smarkm 223109998Smarkm Remote buffer overflow in SSL3 protocol - an attacker could 224109998Smarkm supply an oversized master key in Kerberos-enabled versions. 225109998Smarkm (CAN-2002-0657) 226109998Smarkm [Ben Laurie (CHATS)] 227109998Smarkm 228109998Smarkm *) Change the SSL kerb5 codes to match RFC 2712. 229109998Smarkm [Richard Levitte] 230109998Smarkm 231109998Smarkm *) Make -nameopt work fully for req and add -reqopt switch. 232109998Smarkm [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson] 233109998Smarkm 234109998Smarkm *) The "block size" for block ciphers in CFB and OFB mode should be 1. 235109998Smarkm [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>] 236109998Smarkm 237109998Smarkm *) Make sure tests can be performed even if the corresponding algorithms 238109998Smarkm have been removed entirely. This was also the last step to make 239109998Smarkm OpenSSL compilable with DJGPP under all reasonable conditions. 240109998Smarkm [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>] 241109998Smarkm 242109998Smarkm *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT 243109998Smarkm to allow version independent disabling of normally unselected ciphers, 244109998Smarkm which may be activated as a side-effect of selecting a single cipher. 245109998Smarkm 246109998Smarkm (E.g., cipher list string "RSA" enables ciphersuites that are left 247109998Smarkm out of "ALL" because they do not provide symmetric encryption. 248109998Smarkm "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) 249109998Smarkm [Lutz Jaenicke, Bodo Moeller] 250109998Smarkm 251109998Smarkm *) Add appropriate support for separate platform-dependent build 252109998Smarkm directories. The recommended way to make a platform-dependent 253109998Smarkm build directory is the following (tested on Linux), maybe with 254109998Smarkm some local tweaks: 255109998Smarkm 256109998Smarkm # Place yourself outside of the OpenSSL source tree. In 257109998Smarkm # this example, the environment variable OPENSSL_SOURCE 258109998Smarkm # is assumed to contain the absolute OpenSSL source directory. 259109998Smarkm mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" 260109998Smarkm cd objtree/"`uname -s`-`uname -r`-`uname -m`" 261111147Snectar (cd $OPENSSL_SOURCE; find . -type f) | while read F; do 262109998Smarkm mkdir -p `dirname $F` 263109998Smarkm ln -s $OPENSSL_SOURCE/$F $F 264109998Smarkm done 265109998Smarkm 266109998Smarkm To be absolutely sure not to disturb the source tree, a "make clean" 267109998Smarkm is a good thing. If it isn't successfull, don't worry about it, 268109998Smarkm it probably means the source directory is very clean. 269109998Smarkm [Richard Levitte] 270109998Smarkm 271109998Smarkm *) Make sure any ENGINE control commands make local copies of string 272109998Smarkm pointers passed to them whenever necessary. Otherwise it is possible 273109998Smarkm the caller may have overwritten (or deallocated) the original string 274109998Smarkm data when a later ENGINE operation tries to use the stored values. 275109998Smarkm [G�tz Babin-Ebell <babinebell@trustcenter.de>] 276109998Smarkm 277109998Smarkm *) Improve diagnostics in file reading and command-line digests. 278109998Smarkm [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>] 279109998Smarkm 280109998Smarkm *) Add AES modes CFB and OFB to the object database. Correct an 281109998Smarkm error in AES-CFB decryption. 282109998Smarkm [Richard Levitte] 283109998Smarkm 284109998Smarkm *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this 285109998Smarkm allows existing EVP_CIPHER_CTX structures to be reused after 286109998Smarkm calling EVP_*Final(). This behaviour is used by encryption 287109998Smarkm BIOs and some applications. This has the side effect that 288109998Smarkm applications must explicitly clean up cipher contexts with 289109998Smarkm EVP_CIPHER_CTX_cleanup() or they will leak memory. 290109998Smarkm [Steve Henson] 291109998Smarkm 292109998Smarkm *) Check the values of dna and dnb in bn_mul_recursive before calling 293109998Smarkm bn_mul_comba (a non zero value means the a or b arrays do not contain 294109998Smarkm n2 elements) and fallback to bn_mul_normal if either is not zero. 295109998Smarkm [Steve Henson] 296109998Smarkm 297109998Smarkm *) Fix escaping of non-ASCII characters when using the -subj option 298109998Smarkm of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) 299109998Smarkm [Lutz Jaenicke] 300109998Smarkm 301109998Smarkm *) Make object definitions compliant to LDAP (RFC2256): SN is the short 302109998Smarkm form for "surname", serialNumber has no short form. 303109998Smarkm Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; 304109998Smarkm therefore remove "mail" short name for "internet 7". 305109998Smarkm The OID for unique identifiers in X509 certificates is 306109998Smarkm x500UniqueIdentifier, not uniqueIdentifier. 307109998Smarkm Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) 308109998Smarkm [Lutz Jaenicke] 309109998Smarkm 310109998Smarkm *) Add an "init" command to the ENGINE config module and auto initialize 311109998Smarkm ENGINEs. Without any "init" command the ENGINE will be initialized 312109998Smarkm after all ctrl commands have been executed on it. If init=1 the 313109998Smarkm ENGINE is initailized at that point (ctrls before that point are run 314109998Smarkm on the uninitialized ENGINE and after on the initialized one). If 315109998Smarkm init=0 then the ENGINE will not be iniatialized at all. 316109998Smarkm [Steve Henson] 317109998Smarkm 318109998Smarkm *) Fix the 'app_verify_callback' interface so that the user-defined 319109998Smarkm argument is actually passed to the callback: In the 320109998Smarkm SSL_CTX_set_cert_verify_callback() prototype, the callback 321109998Smarkm declaration has been changed from 322109998Smarkm int (*cb)() 323109998Smarkm into 324109998Smarkm int (*cb)(X509_STORE_CTX *,void *); 325109998Smarkm in ssl_verify_cert_chain (ssl/ssl_cert.c), the call 326109998Smarkm i=s->ctx->app_verify_callback(&ctx) 327109998Smarkm has been changed into 328109998Smarkm i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). 329109998Smarkm 330109998Smarkm To update applications using SSL_CTX_set_cert_verify_callback(), 331109998Smarkm a dummy argument can be added to their callback functions. 332109998Smarkm [D. K. Smetters <smetters@parc.xerox.com>] 333109998Smarkm 334109998Smarkm *) Added the '4758cca' ENGINE to support IBM 4758 cards. 335109998Smarkm [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe] 336109998Smarkm 337109998Smarkm *) Add and OPENSSL_LOAD_CONF define which will cause 338109998Smarkm OpenSSL_add_all_algorithms() to load the openssl.cnf config file. 339109998Smarkm This allows older applications to transparently support certain 340109998Smarkm OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. 341109998Smarkm Two new functions OPENSSL_add_all_algorithms_noconf() which will never 342109998Smarkm load the config file and OPENSSL_add_all_algorithms_conf() which will 343109998Smarkm always load it have also been added. 344109998Smarkm [Steve Henson] 345109998Smarkm 346109998Smarkm *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. 347109998Smarkm Adjust NIDs and EVP layer. 348109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 349109998Smarkm 350109998Smarkm *) Config modules support in openssl utility. 351109998Smarkm 352109998Smarkm Most commands now load modules from the config file, 353109998Smarkm though in a few (such as version) this isn't done 354109998Smarkm because it couldn't be used for anything. 355109998Smarkm 356109998Smarkm In the case of ca and req the config file used is 357109998Smarkm the same as the utility itself: that is the -config 358109998Smarkm command line option can be used to specify an 359109998Smarkm alternative file. 360109998Smarkm [Steve Henson] 361109998Smarkm 362109998Smarkm *) Move default behaviour from OPENSSL_config(). If appname is NULL 363109998Smarkm use "openssl_conf" if filename is NULL use default openssl config file. 364109998Smarkm [Steve Henson] 365109998Smarkm 366109998Smarkm *) Add an argument to OPENSSL_config() to allow the use of an alternative 367109998Smarkm config section name. Add a new flag to tolerate a missing config file 368109998Smarkm and move code to CONF_modules_load_file(). 369109998Smarkm [Steve Henson] 370109998Smarkm 371109998Smarkm *) Support for crypto accelerator cards from Accelerated Encryption 372109998Smarkm Processing, www.aep.ie. (Use engine 'aep') 373109998Smarkm The support was copied from 0.9.6c [engine] and adapted/corrected 374109998Smarkm to work with the new engine framework. 375109998Smarkm [AEP Inc. and Richard Levitte] 376109998Smarkm 377109998Smarkm *) Support for SureWare crypto accelerator cards from Baltimore 378109998Smarkm Technologies. (Use engine 'sureware') 379109998Smarkm The support was copied from 0.9.6c [engine] and adapted 380109998Smarkm to work with the new engine framework. 381109998Smarkm [Richard Levitte] 382109998Smarkm 383109998Smarkm *) Have the CHIL engine fork-safe (as defined by nCipher) and actually 384109998Smarkm make the newer ENGINE framework commands for the CHIL engine work. 385109998Smarkm [Toomas Kiisk <vix@cyber.ee> and Richard Levitte] 386109998Smarkm 387109998Smarkm *) Make it possible to produce shared libraries on ReliantUNIX. 388109998Smarkm [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte] 389109998Smarkm 390109998Smarkm *) Add the configuration target debug-linux-ppro. 391109998Smarkm Make 'openssl rsa' use the general key loading routines 392109998Smarkm implemented in apps.c, and make those routines able to 393109998Smarkm handle the key format FORMAT_NETSCAPE and the variant 394109998Smarkm FORMAT_IISSGC. 395109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 396109998Smarkm 397109998Smarkm *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 398109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 399109998Smarkm 400109998Smarkm *) Add -keyform to rsautl, and document -engine. 401109998Smarkm [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>] 402109998Smarkm 403109998Smarkm *) Change BIO_new_file (crypto/bio/bss_file.c) to use new 404109998Smarkm BIO_R_NO_SUCH_FILE error code rather than the generic 405109998Smarkm ERR_R_SYS_LIB error code if fopen() fails with ENOENT. 406109998Smarkm [Ben Laurie] 407109998Smarkm 408109998Smarkm *) Add new functions 409109998Smarkm ERR_peek_last_error 410109998Smarkm ERR_peek_last_error_line 411109998Smarkm ERR_peek_last_error_line_data. 412109998Smarkm These are similar to 413109998Smarkm ERR_peek_error 414109998Smarkm ERR_peek_error_line 415109998Smarkm ERR_peek_error_line_data, 416109998Smarkm but report on the latest error recorded rather than the first one 417109998Smarkm still in the error queue. 418109998Smarkm [Ben Laurie, Bodo Moeller] 419109998Smarkm 420109998Smarkm *) default_algorithms option in ENGINE config module. This allows things 421109998Smarkm like: 422109998Smarkm default_algorithms = ALL 423109998Smarkm default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS 424109998Smarkm [Steve Henson] 425109998Smarkm 426109998Smarkm *) Prelminary ENGINE config module. 427109998Smarkm [Steve Henson] 428109998Smarkm 429109998Smarkm *) New experimental application configuration code. 430109998Smarkm [Steve Henson] 431109998Smarkm 432109998Smarkm *) Change the AES code to follow the same name structure as all other 433109998Smarkm symmetric ciphers, and behave the same way. Move everything to 434109998Smarkm the directory crypto/aes, thereby obsoleting crypto/rijndael. 435109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 436109998Smarkm 437109998Smarkm *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. 438109998Smarkm [Ben Laurie and Theo de Raadt] 439109998Smarkm 440109998Smarkm *) Add option to output public keys in req command. 441109998Smarkm [Massimiliano Pala madwolf@openca.org] 442109998Smarkm 443109998Smarkm *) Use wNAFs in EC_POINTs_mul() for improved efficiency 444109998Smarkm (up to about 10% better than before for P-192 and P-224). 445109998Smarkm [Bodo Moeller] 446109998Smarkm 447109998Smarkm *) New functions/macros 448109998Smarkm 449109998Smarkm SSL_CTX_set_msg_callback(ctx, cb) 450109998Smarkm SSL_CTX_set_msg_callback_arg(ctx, arg) 451109998Smarkm SSL_set_msg_callback(ssl, cb) 452109998Smarkm SSL_set_msg_callback_arg(ssl, arg) 453109998Smarkm 454109998Smarkm to request calling a callback function 455109998Smarkm 456109998Smarkm void cb(int write_p, int version, int content_type, 457109998Smarkm const void *buf, size_t len, SSL *ssl, void *arg) 458109998Smarkm 459109998Smarkm whenever a protocol message has been completely received 460109998Smarkm (write_p == 0) or sent (write_p == 1). Here 'version' is the 461109998Smarkm protocol version according to which the SSL library interprets 462109998Smarkm the current protocol message (SSL2_VERSION, SSL3_VERSION, or 463109998Smarkm TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or 464109998Smarkm the content type as defined in the SSL 3.0/TLS 1.0 protocol 465109998Smarkm specification (change_cipher_spec(20), alert(21), handshake(22)). 466109998Smarkm 'buf' and 'len' point to the actual message, 'ssl' to the 467109998Smarkm SSL object, and 'arg' is the application-defined value set by 468109998Smarkm SSL[_CTX]_set_msg_callback_arg(). 469109998Smarkm 470109998Smarkm 'openssl s_client' and 'openssl s_server' have new '-msg' options 471109998Smarkm to enable a callback that displays all protocol messages. 472109998Smarkm [Bodo Moeller] 473109998Smarkm 474109998Smarkm *) Change the shared library support so shared libraries are built as 475109998Smarkm soon as the corresponding static library is finished, and thereby get 476109998Smarkm openssl and the test programs linked against the shared library. 477109998Smarkm This still only happens when the keyword "shard" has been given to 478109998Smarkm the configuration scripts. 479109998Smarkm 480109998Smarkm NOTE: shared library support is still an experimental thing, and 481109998Smarkm backward binary compatibility is still not guaranteed. 482109998Smarkm ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte] 483109998Smarkm 484109998Smarkm *) Add support for Subject Information Access extension. 485109998Smarkm [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 486109998Smarkm 487109998Smarkm *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero 488109998Smarkm additional bytes when new memory had to be allocated, not just 489109998Smarkm when reusing an existing buffer. 490109998Smarkm [Bodo Moeller] 491109998Smarkm 492109998Smarkm *) New command line and configuration option 'utf8' for the req command. 493109998Smarkm This allows field values to be specified as UTF8 strings. 494109998Smarkm [Steve Henson] 495109998Smarkm 496109998Smarkm *) Add -multi and -mr options to "openssl speed" - giving multiple parallel 497109998Smarkm runs for the former and machine-readable output for the latter. 498109998Smarkm [Ben Laurie] 499109998Smarkm 500109998Smarkm *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion 501109998Smarkm of the e-mail address in the DN (i.e., it will go into a certificate 502109998Smarkm extension only). The new configuration file option 'email_in_dn = no' 503109998Smarkm has the same effect. 504109998Smarkm [Massimiliano Pala madwolf@openca.org] 505109998Smarkm 506109998Smarkm *) Change all functions with names starting with des_ to be starting 507109998Smarkm with DES_ instead. Add wrappers that are compatible with libdes, 508109998Smarkm but are named _ossl_old_des_*. Finally, add macros that map the 509109998Smarkm des_* symbols to the corresponding _ossl_old_des_* if libdes 510109998Smarkm compatibility is desired. If OpenSSL 0.9.6c compatibility is 511109998Smarkm desired, the des_* symbols will be mapped to DES_*, with one 512109998Smarkm exception. 513109998Smarkm 514109998Smarkm Since we provide two compatibility mappings, the user needs to 515109998Smarkm define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes 516109998Smarkm compatibility is desired. The default (i.e., when that macro 517109998Smarkm isn't defined) is OpenSSL 0.9.6c compatibility. 518109998Smarkm 519109998Smarkm There are also macros that enable and disable the support of old 520109998Smarkm des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT 521109998Smarkm and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those 522109998Smarkm are defined, the default will apply: to support the old des routines. 523109998Smarkm 524109998Smarkm In either case, one must include openssl/des.h to get the correct 525109998Smarkm definitions. Do not try to just include openssl/des_old.h, that 526109998Smarkm won't work. 527109998Smarkm 528109998Smarkm NOTE: This is a major break of an old API into a new one. Software 529109998Smarkm authors are encouraged to switch to the DES_ style functions. Some 530109998Smarkm time in the future, des_old.h and the libdes compatibility functions 531109998Smarkm will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the 532109998Smarkm default), and then completely removed. 533109998Smarkm [Richard Levitte] 534109998Smarkm 535109998Smarkm *) Test for certificates which contain unsupported critical extensions. 536109998Smarkm If such a certificate is found during a verify operation it is 537109998Smarkm rejected by default: this behaviour can be overridden by either 538109998Smarkm handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or 539109998Smarkm by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function 540109998Smarkm X509_supported_extension() has also been added which returns 1 if a 541109998Smarkm particular extension is supported. 542109998Smarkm [Steve Henson] 543109998Smarkm 544109998Smarkm *) Modify the behaviour of EVP cipher functions in similar way to digests 545109998Smarkm to retain compatibility with existing code. 546109998Smarkm [Steve Henson] 547109998Smarkm 548109998Smarkm *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain 549109998Smarkm compatibility with existing code. In particular the 'ctx' parameter does 550109998Smarkm not have to be to be initialized before the call to EVP_DigestInit() and 551109998Smarkm it is tidied up after a call to EVP_DigestFinal(). New function 552109998Smarkm EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function 553109998Smarkm EVP_MD_CTX_copy() changed to not require the destination to be 554109998Smarkm initialized valid and new function EVP_MD_CTX_copy_ex() added which 555109998Smarkm requires the destination to be valid. 556109998Smarkm 557109998Smarkm Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), 558109998Smarkm EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). 559109998Smarkm [Steve Henson] 560109998Smarkm 561109998Smarkm *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it 562109998Smarkm so that complete 'Handshake' protocol structures are kept in memory 563109998Smarkm instead of overwriting 'msg_type' and 'length' with 'body' data. 564109998Smarkm [Bodo Moeller] 565109998Smarkm 566109998Smarkm *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. 567109998Smarkm [Massimo Santin via Richard Levitte] 568109998Smarkm 569109998Smarkm *) Major restructuring to the underlying ENGINE code. This includes 570109998Smarkm reduction of linker bloat, separation of pure "ENGINE" manipulation 571109998Smarkm (initialisation, etc) from functionality dealing with implementations 572109998Smarkm of specific crypto iterfaces. This change also introduces integrated 573109998Smarkm support for symmetric ciphers and digest implementations - so ENGINEs 574109998Smarkm can now accelerate these by providing EVP_CIPHER and EVP_MD 575109998Smarkm implementations of their own. This is detailed in crypto/engine/README 576109998Smarkm as it couldn't be adequately described here. However, there are a few 577109998Smarkm API changes worth noting - some RSA, DSA, DH, and RAND functions that 578109998Smarkm were changed in the original introduction of ENGINE code have now 579109998Smarkm reverted back - the hooking from this code to ENGINE is now a good 580109998Smarkm deal more passive and at run-time, operations deal directly with 581109998Smarkm RSA_METHODs, DSA_METHODs (etc) as they did before, rather than 582109998Smarkm dereferencing through an ENGINE pointer any more. Also, the ENGINE 583109998Smarkm functions dealing with BN_MOD_EXP[_CRT] handlers have been removed - 584109998Smarkm they were not being used by the framework as there is no concept of a 585109998Smarkm BIGNUM_METHOD and they could not be generalised to the new 586109998Smarkm 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, 587109998Smarkm ENGINE_cpy() has been removed as it cannot be consistently defined in 588109998Smarkm the new code. 589109998Smarkm [Geoff Thorpe] 590109998Smarkm 591109998Smarkm *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. 592109998Smarkm [Steve Henson] 593109998Smarkm 594109998Smarkm *) Change mkdef.pl to sort symbols that get the same entry number, 595109998Smarkm and make sure the automatically generated functions ERR_load_* 596109998Smarkm become part of libeay.num as well. 597109998Smarkm [Richard Levitte] 598109998Smarkm 599109998Smarkm *) New function SSL_renegotiate_pending(). This returns true once 600109998Smarkm renegotiation has been requested (either SSL_renegotiate() call 601109998Smarkm or HelloRequest/ClientHello receveived from the peer) and becomes 602109998Smarkm false once a handshake has been completed. 603109998Smarkm (For servers, SSL_renegotiate() followed by SSL_do_handshake() 604109998Smarkm sends a HelloRequest, but does not ensure that a handshake takes 605109998Smarkm place. SSL_renegotiate_pending() is useful for checking if the 606109998Smarkm client has followed the request.) 607109998Smarkm [Bodo Moeller] 608109998Smarkm 609109998Smarkm *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. 610109998Smarkm By default, clients may request session resumption even during 611109998Smarkm renegotiation (if session ID contexts permit); with this option, 612109998Smarkm session resumption is possible only in the first handshake. 613109998Smarkm 614109998Smarkm SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes 615109998Smarkm more bits available for options that should not be part of 616109998Smarkm SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). 617109998Smarkm [Bodo Moeller] 618109998Smarkm 619109998Smarkm *) Add some demos for certificate and certificate request creation. 620109998Smarkm [Steve Henson] 621109998Smarkm 622109998Smarkm *) Make maximum certificate chain size accepted from the peer application 623109998Smarkm settable (SSL*_get/set_max_cert_list()), as proposed by 624109998Smarkm "Douglas E. Engert" <deengert@anl.gov>. 625109998Smarkm [Lutz Jaenicke] 626109998Smarkm 627109998Smarkm *) Add support for shared libraries for Unixware-7 628109998Smarkm (Boyd Lynn Gerber <gerberb@zenez.com>). 629109998Smarkm [Lutz Jaenicke] 630109998Smarkm 631109998Smarkm *) Add a "destroy" handler to ENGINEs that allows structural cleanup to 632109998Smarkm be done prior to destruction. Use this to unload error strings from 633109998Smarkm ENGINEs that load their own error strings. NB: This adds two new API 634109998Smarkm functions to "get" and "set" this destroy handler in an ENGINE. 635109998Smarkm [Geoff Thorpe] 636109998Smarkm 637109998Smarkm *) Alter all existing ENGINE implementations (except "openssl" and 638109998Smarkm "openbsd") to dynamically instantiate their own error strings. This 639109998Smarkm makes them more flexible to be built both as statically-linked ENGINEs 640109998Smarkm and self-contained shared-libraries loadable via the "dynamic" ENGINE. 641109998Smarkm Also, add stub code to each that makes building them as self-contained 642109998Smarkm shared-libraries easier (see README.ENGINE). 643109998Smarkm [Geoff Thorpe] 644109998Smarkm 645109998Smarkm *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE 646109998Smarkm implementations into applications that are completely implemented in 647109998Smarkm self-contained shared-libraries. The "dynamic" ENGINE exposes control 648109998Smarkm commands that can be used to configure what shared-library to load and 649109998Smarkm to control aspects of the way it is handled. Also, made an update to 650109998Smarkm the README.ENGINE file that brings its information up-to-date and 651109998Smarkm provides some information and instructions on the "dynamic" ENGINE 652109998Smarkm (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). 653109998Smarkm [Geoff Thorpe] 654109998Smarkm 655109998Smarkm *) Make it possible to unload ranges of ERR strings with a new 656109998Smarkm "ERR_unload_strings" function. 657109998Smarkm [Geoff Thorpe] 658109998Smarkm 659109998Smarkm *) Add a copy() function to EVP_MD. 660109998Smarkm [Ben Laurie] 661109998Smarkm 662109998Smarkm *) Make EVP_MD routines take a context pointer instead of just the 663109998Smarkm md_data void pointer. 664109998Smarkm [Ben Laurie] 665109998Smarkm 666109998Smarkm *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates 667109998Smarkm that the digest can only process a single chunk of data 668109998Smarkm (typically because it is provided by a piece of 669109998Smarkm hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application 670109998Smarkm is only going to provide a single chunk of data, and hence the 671109998Smarkm framework needn't accumulate the data for oneshot drivers. 672109998Smarkm [Ben Laurie] 673109998Smarkm 674109998Smarkm *) As with "ERR", make it possible to replace the underlying "ex_data" 675109998Smarkm functions. This change also alters the storage and management of global 676109998Smarkm ex_data state - it's now all inside ex_data.c and all "class" code (eg. 677109998Smarkm RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class 678109998Smarkm index counters. The API functions that use this state have been changed 679109998Smarkm to take a "class_index" rather than pointers to the class's local STACK 680109998Smarkm and counter, and there is now an API function to dynamically create new 681109998Smarkm classes. This centralisation allows us to (a) plug a lot of the 682109998Smarkm thread-safety problems that existed, and (b) makes it possible to clean 683109998Smarkm up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) 684109998Smarkm such data would previously have always leaked in application code and 685109998Smarkm workarounds were in place to make the memory debugging turn a blind eye 686109998Smarkm to it. Application code that doesn't use this new function will still 687109998Smarkm leak as before, but their memory debugging output will announce it now 688109998Smarkm rather than letting it slide. 689109998Smarkm 690109998Smarkm Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change 691109998Smarkm induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now 692109998Smarkm has a return value to indicate success or failure. 693109998Smarkm [Geoff Thorpe] 694109998Smarkm 695109998Smarkm *) Make it possible to replace the underlying "ERR" functions such that the 696109998Smarkm global state (2 LHASH tables and 2 locks) is only used by the "default" 697109998Smarkm implementation. This change also adds two functions to "get" and "set" 698109998Smarkm the implementation prior to it being automatically set the first time 699109998Smarkm any other ERR function takes place. Ie. an application can call "get", 700109998Smarkm pass the return value to a module it has just loaded, and that module 701109998Smarkm can call its own "set" function using that value. This means the 702109998Smarkm module's "ERR" operations will use (and modify) the error state in the 703109998Smarkm application and not in its own statically linked copy of OpenSSL code. 704109998Smarkm [Geoff Thorpe] 705109998Smarkm 706109998Smarkm *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment 707109998Smarkm reference counts. This performs normal REF_PRINT/REF_CHECK macros on 708109998Smarkm the operation, and provides a more encapsulated way for external code 709109998Smarkm (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code 710109998Smarkm to use these functions rather than manually incrementing the counts. 711109998Smarkm 712109998Smarkm Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". 713109998Smarkm [Geoff Thorpe] 714109998Smarkm 715109998Smarkm *) Add EVP test program. 716109998Smarkm [Ben Laurie] 717109998Smarkm 718109998Smarkm *) Add symmetric cipher support to ENGINE. Expect the API to change! 719109998Smarkm [Ben Laurie] 720109998Smarkm 721109998Smarkm *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() 722109998Smarkm X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), 723109998Smarkm X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). 724109998Smarkm These allow a CRL to be built without having to access X509_CRL fields 725109998Smarkm directly. Modify 'ca' application to use new functions. 726109998Smarkm [Steve Henson] 727109998Smarkm 728109998Smarkm *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended 729109998Smarkm bug workarounds. Rollback attack detection is a security feature. 730109998Smarkm The problem will only arise on OpenSSL servers when TLSv1 is not 731109998Smarkm available (sslv3_server_method() or SSL_OP_NO_TLSv1). 732109998Smarkm Software authors not wanting to support TLSv1 will have special reasons 733109998Smarkm for their choice and can explicitly enable this option. 734109998Smarkm [Bodo Moeller, Lutz Jaenicke] 735109998Smarkm 736109998Smarkm *) Rationalise EVP so it can be extended: don't include a union of 737109998Smarkm cipher/digest structures, add init/cleanup functions for EVP_MD_CTX 738109998Smarkm (similar to those existing for EVP_CIPHER_CTX). 739109998Smarkm Usage example: 740109998Smarkm 741109998Smarkm EVP_MD_CTX md; 742109998Smarkm 743109998Smarkm EVP_MD_CTX_init(&md); /* new function call */ 744109998Smarkm EVP_DigestInit(&md, EVP_sha1()); 745109998Smarkm EVP_DigestUpdate(&md, in, len); 746109998Smarkm EVP_DigestFinal(&md, out, NULL); 747109998Smarkm EVP_MD_CTX_cleanup(&md); /* new function call */ 748109998Smarkm 749109998Smarkm [Ben Laurie] 750109998Smarkm 751109998Smarkm *) Make DES key schedule conform to the usual scheme, as well as 752109998Smarkm correcting its structure. This means that calls to DES functions 753109998Smarkm now have to pass a pointer to a des_key_schedule instead of a 754109998Smarkm plain des_key_schedule (which was actually always a pointer 755109998Smarkm anyway): E.g., 756109998Smarkm 757109998Smarkm des_key_schedule ks; 758109998Smarkm 759109998Smarkm des_set_key_checked(..., &ks); 760109998Smarkm des_ncbc_encrypt(..., &ks, ...); 761109998Smarkm 762109998Smarkm (Note that a later change renames 'des_...' into 'DES_...'.) 763109998Smarkm [Ben Laurie] 764109998Smarkm 765109998Smarkm *) Initial reduction of linker bloat: the use of some functions, such as 766109998Smarkm PEM causes large amounts of unused functions to be linked in due to 767109998Smarkm poor organisation. For example pem_all.c contains every PEM function 768109998Smarkm which has a knock on effect of linking in large amounts of (unused) 769109998Smarkm ASN1 code. Grouping together similar functions and splitting unrelated 770109998Smarkm functions prevents this. 771109998Smarkm [Steve Henson] 772109998Smarkm 773109998Smarkm *) Cleanup of EVP macros. 774109998Smarkm [Ben Laurie] 775109998Smarkm 776109998Smarkm *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the 777109998Smarkm correct _ecb suffix. 778109998Smarkm [Ben Laurie] 779109998Smarkm 780109998Smarkm *) Add initial OCSP responder support to ocsp application. The 781109998Smarkm revocation information is handled using the text based index 782109998Smarkm use by the ca application. The responder can either handle 783109998Smarkm requests generated internally, supplied in files (for example 784109998Smarkm via a CGI script) or using an internal minimal server. 785109998Smarkm [Steve Henson] 786109998Smarkm 787109998Smarkm *) Add configuration choices to get zlib compression for TLS. 788109998Smarkm [Richard Levitte] 789109998Smarkm 790109998Smarkm *) Changes to Kerberos SSL for RFC 2712 compliance: 791109998Smarkm 1. Implemented real KerberosWrapper, instead of just using 792109998Smarkm KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] 793109998Smarkm 2. Implemented optional authenticator field of KerberosWrapper. 794109998Smarkm 795109998Smarkm Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, 796109998Smarkm and authenticator structs; see crypto/krb5/. 797109998Smarkm 798109998Smarkm Generalized Kerberos calls to support multiple Kerberos libraries. 799109998Smarkm [Vern Staats <staatsvr@asc.hpc.mil>, 800109998Smarkm Jeffrey Altman <jaltman@columbia.edu> 801109998Smarkm via Richard Levitte] 802109998Smarkm 803109998Smarkm *) Cause 'openssl speed' to use fully hard-coded DSA keys as it 804109998Smarkm already does with RSA. testdsa.h now has 'priv_key/pub_key' 805109998Smarkm values for each of the key sizes rather than having just 806109998Smarkm parameters (and 'speed' generating keys each time). 807109998Smarkm [Geoff Thorpe] 808109998Smarkm 809109998Smarkm *) Speed up EVP routines. 810109998Smarkm Before: 811109998Smarkmencrypt 812109998Smarkmtype 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 813109998Smarkmdes-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k 814109998Smarkmdes-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k 815109998Smarkmdes-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k 816109998Smarkmdecrypt 817109998Smarkmdes-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k 818109998Smarkmdes-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k 819109998Smarkmdes-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k 820109998Smarkm After: 821109998Smarkmencrypt 822109998Smarkmdes-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k 823109998Smarkmdecrypt 824109998Smarkmdes-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k 825109998Smarkm [Ben Laurie] 826109998Smarkm 827109998Smarkm *) Added the OS2-EMX target. 828109998Smarkm ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte] 829109998Smarkm 830109998Smarkm *) Rewrite apps to use NCONF routines instead of the old CONF. New functions 831109998Smarkm to support NCONF routines in extension code. New function CONF_set_nconf() 832109998Smarkm to allow functions which take an NCONF to also handle the old LHASH 833109998Smarkm structure: this means that the old CONF compatible routines can be 834109998Smarkm retained (in particular wrt extensions) without having to duplicate the 835109998Smarkm code. New function X509V3_add_ext_nconf_sk to add extensions to a stack. 836109998Smarkm [Steve Henson] 837109998Smarkm 838109998Smarkm *) Enhance the general user interface with mechanisms for inner control 839109998Smarkm and with possibilities to have yes/no kind of prompts. 840109998Smarkm [Richard Levitte] 841109998Smarkm 842109998Smarkm *) Change all calls to low level digest routines in the library and 843109998Smarkm applications to use EVP. Add missing calls to HMAC_cleanup() and 844109998Smarkm don't assume HMAC_CTX can be copied using memcpy(). 845109998Smarkm [Verdon Walker <VWalker@novell.com>, Steve Henson] 846109998Smarkm 847109998Smarkm *) Add the possibility to control engines through control names but with 848109998Smarkm arbitrary arguments instead of just a string. 849109998Smarkm Change the key loaders to take a UI_METHOD instead of a callback 850109998Smarkm function pointer. NOTE: this breaks binary compatibility with earlier 851109998Smarkm versions of OpenSSL [engine]. 852109998Smarkm Adapt the nCipher code for these new conditions and add a card insertion 853109998Smarkm callback. 854109998Smarkm [Richard Levitte] 855109998Smarkm 856109998Smarkm *) Enhance the general user interface with mechanisms to better support 857109998Smarkm dialog box interfaces, application-defined prompts, the possibility 858109998Smarkm to use defaults (for example default passwords from somewhere else) 859109998Smarkm and interrupts/cancellations. 860109998Smarkm [Richard Levitte] 861109998Smarkm 862109998Smarkm *) Tidy up PKCS#12 attribute handling. Add support for the CSP name 863109998Smarkm attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. 864109998Smarkm [Steve Henson] 865109998Smarkm 866109998Smarkm *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also 867109998Smarkm tidy up some unnecessarily weird code in 'sk_new()'). 868109998Smarkm [Geoff, reported by Diego Tartara <dtartara@novamens.com>] 869109998Smarkm 870109998Smarkm *) Change the key loading routines for ENGINEs to use the same kind 871109998Smarkm callback (pem_password_cb) as all other routines that need this 872109998Smarkm kind of callback. 873109998Smarkm [Richard Levitte] 874109998Smarkm 875109998Smarkm *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with 876109998Smarkm 256 bit (=32 byte) keys. Of course seeding with more entropy bytes 877109998Smarkm than this minimum value is recommended. 878109998Smarkm [Lutz Jaenicke] 879109998Smarkm 880109998Smarkm *) New random seeder for OpenVMS, using the system process statistics 881109998Smarkm that are easily reachable. 882109998Smarkm [Richard Levitte] 883109998Smarkm 884109998Smarkm *) Windows apparently can't transparently handle global 885109998Smarkm variables defined in DLLs. Initialisations such as: 886109998Smarkm 887109998Smarkm const ASN1_ITEM *it = &ASN1_INTEGER_it; 888109998Smarkm 889109998Smarkm wont compile. This is used by the any applications that need to 890109998Smarkm declare their own ASN1 modules. This was fixed by adding the option 891109998Smarkm EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly 892109998Smarkm needed for static libraries under Win32. 893109998Smarkm [Steve Henson] 894109998Smarkm 895109998Smarkm *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle 896109998Smarkm setting of purpose and trust fields. New X509_STORE trust and 897109998Smarkm purpose functions and tidy up setting in other SSL functions. 898109998Smarkm [Steve Henson] 899109998Smarkm 900109998Smarkm *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE 901109998Smarkm structure. These are inherited by X509_STORE_CTX when it is 902109998Smarkm initialised. This allows various defaults to be set in the 903109998Smarkm X509_STORE structure (such as flags for CRL checking and custom 904109998Smarkm purpose or trust settings) for functions which only use X509_STORE_CTX 905109998Smarkm internally such as S/MIME. 906109998Smarkm 907109998Smarkm Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and 908109998Smarkm trust settings if they are not set in X509_STORE. This allows X509_STORE 909109998Smarkm purposes and trust (in S/MIME for example) to override any set by default. 910109998Smarkm 911109998Smarkm Add command line options for CRL checking to smime, s_client and s_server 912109998Smarkm applications. 913109998Smarkm [Steve Henson] 914109998Smarkm 915109998Smarkm *) Initial CRL based revocation checking. If the CRL checking flag(s) 916109998Smarkm are set then the CRL is looked up in the X509_STORE structure and 917109998Smarkm its validity and signature checked, then if the certificate is found 918109998Smarkm in the CRL the verify fails with a revoked error. 919109998Smarkm 920109998Smarkm Various new CRL related callbacks added to X509_STORE_CTX structure. 921109998Smarkm 922109998Smarkm Command line options added to 'verify' application to support this. 923109998Smarkm 924109998Smarkm This needs some additional work, such as being able to handle multiple 925109998Smarkm CRLs with different times, extension based lookup (rather than just 926109998Smarkm by subject name) and ultimately more complete V2 CRL extension 927109998Smarkm handling. 928109998Smarkm [Steve Henson] 929109998Smarkm 930109998Smarkm *) Add a general user interface API (crypto/ui/). This is designed 931109998Smarkm to replace things like des_read_password and friends (backward 932109998Smarkm compatibility functions using this new API are provided). 933109998Smarkm The purpose is to remove prompting functions from the DES code 934109998Smarkm section as well as provide for prompting through dialog boxes in 935109998Smarkm a window system and the like. 936109998Smarkm [Richard Levitte] 937109998Smarkm 938109998Smarkm *) Add "ex_data" support to ENGINE so implementations can add state at a 939109998Smarkm per-structure level rather than having to store it globally. 940109998Smarkm [Geoff] 941109998Smarkm 942109998Smarkm *) Make it possible for ENGINE structures to be copied when retrieved by 943109998Smarkm ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. 944109998Smarkm This causes the "original" ENGINE structure to act like a template, 945109998Smarkm analogous to the RSA vs. RSA_METHOD type of separation. Because of this 946109998Smarkm operational state can be localised to each ENGINE structure, despite the 947109998Smarkm fact they all share the same "methods". New ENGINE structures returned in 948109998Smarkm this case have no functional references and the return value is the single 949109998Smarkm structural reference. This matches the single structural reference returned 950109998Smarkm by ENGINE_by_id() normally, when it is incremented on the pre-existing 951109998Smarkm ENGINE structure. 952109998Smarkm [Geoff] 953109998Smarkm 954109998Smarkm *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this 955109998Smarkm needs to match any other type at all we need to manually clear the 956109998Smarkm tag cache. 957109998Smarkm [Steve Henson] 958109998Smarkm 959109998Smarkm *) Changes to the "openssl engine" utility to include; 960109998Smarkm - verbosity levels ('-v', '-vv', and '-vvv') that provide information 961109998Smarkm about an ENGINE's available control commands. 962109998Smarkm - executing control commands from command line arguments using the 963109998Smarkm '-pre' and '-post' switches. '-post' is only used if '-t' is 964109998Smarkm specified and the ENGINE is successfully initialised. The syntax for 965109998Smarkm the individual commands are colon-separated, for example; 966109998Smarkm openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so 967109998Smarkm [Geoff] 968109998Smarkm 969109998Smarkm *) New dynamic control command support for ENGINEs. ENGINEs can now 970109998Smarkm declare their own commands (numbers), names (strings), descriptions, 971109998Smarkm and input types for run-time discovery by calling applications. A 972109998Smarkm subset of these commands are implicitly classed as "executable" 973109998Smarkm depending on their input type, and only these can be invoked through 974109998Smarkm the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this 975109998Smarkm can be based on user input, config files, etc). The distinction is 976109998Smarkm that "executable" commands cannot return anything other than a boolean 977109998Smarkm result and can only support numeric or string input, whereas some 978109998Smarkm discoverable commands may only be for direct use through 979109998Smarkm ENGINE_ctrl(), eg. supporting the exchange of binary data, function 980109998Smarkm pointers, or other custom uses. The "executable" commands are to 981109998Smarkm support parameterisations of ENGINE behaviour that can be 982109998Smarkm unambiguously defined by ENGINEs and used consistently across any 983109998Smarkm OpenSSL-based application. Commands have been added to all the 984109998Smarkm existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow 985109998Smarkm control over shared-library paths without source code alterations. 986109998Smarkm [Geoff] 987109998Smarkm 988109998Smarkm *) Changed all ENGINE implementations to dynamically allocate their 989109998Smarkm ENGINEs rather than declaring them statically. Apart from this being 990109998Smarkm necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, 991109998Smarkm this also allows the implementations to compile without using the 992109998Smarkm internal engine_int.h header. 993109998Smarkm [Geoff] 994109998Smarkm 995109998Smarkm *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a 996109998Smarkm 'const' value. Any code that should be able to modify a RAND_METHOD 997109998Smarkm should already have non-const pointers to it (ie. they should only 998109998Smarkm modify their own ones). 999109998Smarkm [Geoff] 1000109998Smarkm 1001109998Smarkm *) Made a variety of little tweaks to the ENGINE code. 1002109998Smarkm - "atalla" and "ubsec" string definitions were moved from header files 1003109998Smarkm to C code. "nuron" string definitions were placed in variables 1004109998Smarkm rather than hard-coded - allowing parameterisation of these values 1005109998Smarkm later on via ctrl() commands. 1006109998Smarkm - Removed unused "#if 0"'d code. 1007109998Smarkm - Fixed engine list iteration code so it uses ENGINE_free() to release 1008109998Smarkm structural references. 1009109998Smarkm - Constified the RAND_METHOD element of ENGINE structures. 1010109998Smarkm - Constified various get/set functions as appropriate and added 1011109998Smarkm missing functions (including a catch-all ENGINE_cpy that duplicates 1012109998Smarkm all ENGINE values onto a new ENGINE except reference counts/state). 1013109998Smarkm - Removed NULL parameter checks in get/set functions. Setting a method 1014109998Smarkm or function to NULL is a way of cancelling out a previously set 1015109998Smarkm value. Passing a NULL ENGINE parameter is just plain stupid anyway 1016109998Smarkm and doesn't justify the extra error symbols and code. 1017109998Smarkm - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for 1018109998Smarkm flags from engine_int.h to engine.h. 1019109998Smarkm - Changed prototypes for ENGINE handler functions (init(), finish(), 1020109998Smarkm ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. 1021109998Smarkm [Geoff] 1022109998Smarkm 1023109998Smarkm *) Implement binary inversion algorithm for BN_mod_inverse in addition 1024109998Smarkm to the algorithm using long division. The binary algorithm can be 1025109998Smarkm used only if the modulus is odd. On 32-bit systems, it is faster 1026109998Smarkm only for relatively small moduli (roughly 20-30% for 128-bit moduli, 1027109998Smarkm roughly 5-15% for 256-bit moduli), so we use it only for moduli 1028109998Smarkm up to 450 bits. In 64-bit environments, the binary algorithm 1029109998Smarkm appears to be advantageous for much longer moduli; here we use it 1030109998Smarkm for moduli up to 2048 bits. 1031109998Smarkm [Bodo Moeller] 1032109998Smarkm 1033109998Smarkm *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code 1034109998Smarkm could not support the combine flag in choice fields. 1035109998Smarkm [Steve Henson] 1036109998Smarkm 1037109998Smarkm *) Add a 'copy_extensions' option to the 'ca' utility. This copies 1038109998Smarkm extensions from a certificate request to the certificate. 1039109998Smarkm [Steve Henson] 1040109998Smarkm 1041109998Smarkm *) Allow multiple 'certopt' and 'nameopt' options to be separated 1042109998Smarkm by commas. Add 'namopt' and 'certopt' options to the 'ca' config 1043109998Smarkm file: this allows the display of the certificate about to be 1044109998Smarkm signed to be customised, to allow certain fields to be included 1045109998Smarkm or excluded and extension details. The old system didn't display 1046109998Smarkm multicharacter strings properly, omitted fields not in the policy 1047109998Smarkm and couldn't display additional details such as extensions. 1048109998Smarkm [Steve Henson] 1049109998Smarkm 1050109998Smarkm *) Function EC_POINTs_mul for multiple scalar multiplication 1051109998Smarkm of an arbitrary number of elliptic curve points 1052109998Smarkm \sum scalars[i]*points[i], 1053109998Smarkm optionally including the generator defined for the EC_GROUP: 1054109998Smarkm scalar*generator + \sum scalars[i]*points[i]. 1055109998Smarkm 1056109998Smarkm EC_POINT_mul is a simple wrapper function for the typical case 1057109998Smarkm that the point list has just one item (besides the optional 1058109998Smarkm generator). 1059109998Smarkm [Bodo Moeller] 1060109998Smarkm 1061109998Smarkm *) First EC_METHODs for curves over GF(p): 1062109998Smarkm 1063109998Smarkm EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr 1064109998Smarkm operations and provides various method functions that can also 1065109998Smarkm operate with faster implementations of modular arithmetic. 1066109998Smarkm 1067109998Smarkm EC_GFp_mont_method() reuses most functions that are part of 1068109998Smarkm EC_GFp_simple_method, but uses Montgomery arithmetic. 1069109998Smarkm 1070109998Smarkm [Bodo Moeller; point addition and point doubling 1071109998Smarkm implementation directly derived from source code provided by 1072109998Smarkm Lenka Fibikova <fibikova@exp-math.uni-essen.de>] 1073109998Smarkm 1074109998Smarkm *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, 1075109998Smarkm crypto/ec/ec_lib.c): 1076109998Smarkm 1077109998Smarkm Curves are EC_GROUP objects (with an optional group generator) 1078109998Smarkm based on EC_METHODs that are built into the library. 1079109998Smarkm 1080109998Smarkm Points are EC_POINT objects based on EC_GROUP objects. 1081109998Smarkm 1082109998Smarkm Most of the framework would be able to handle curves over arbitrary 1083109998Smarkm finite fields, but as there are no obvious types for fields other 1084109998Smarkm than GF(p), some functions are limited to that for now. 1085109998Smarkm [Bodo Moeller] 1086109998Smarkm 1087109998Smarkm *) Add the -HTTP option to s_server. It is similar to -WWW, but requires 1088109998Smarkm that the file contains a complete HTTP response. 1089109998Smarkm [Richard Levitte] 1090109998Smarkm 1091109998Smarkm *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl 1092109998Smarkm change the def and num file printf format specifier from "%-40sXXX" 1093109998Smarkm to "%-39s XXX". The latter will always guarantee a space after the 1094109998Smarkm field while the former will cause them to run together if the field 1095109998Smarkm is 40 of more characters long. 1096109998Smarkm [Steve Henson] 1097109998Smarkm 1098109998Smarkm *) Constify the cipher and digest 'method' functions and structures 1099109998Smarkm and modify related functions to take constant EVP_MD and EVP_CIPHER 1100109998Smarkm pointers. 1101109998Smarkm [Steve Henson] 1102109998Smarkm 1103109998Smarkm *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them 1104109998Smarkm in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. 1105109998Smarkm [Bodo Moeller] 1106109998Smarkm 1107109998Smarkm *) Modify EVP_Digest*() routines so they now return values. Although the 1108109998Smarkm internal software routines can never fail additional hardware versions 1109109998Smarkm might. 1110109998Smarkm [Steve Henson] 1111109998Smarkm 1112109998Smarkm *) Clean up crypto/err/err.h and change some error codes to avoid conflicts: 1113109998Smarkm 1114109998Smarkm Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 1115109998Smarkm (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. 1116109998Smarkm 1117109998Smarkm ASN1 error codes 1118109998Smarkm ERR_R_NESTED_ASN1_ERROR 1119109998Smarkm ... 1120109998Smarkm ERR_R_MISSING_ASN1_EOS 1121109998Smarkm were 4 .. 9, conflicting with 1122109998Smarkm ERR_LIB_RSA (= ERR_R_RSA_LIB) 1123109998Smarkm ... 1124109998Smarkm ERR_LIB_PEM (= ERR_R_PEM_LIB). 1125109998Smarkm They are now 58 .. 63 (i.e., just below ERR_R_FATAL). 1126109998Smarkm 1127109998Smarkm Add new error code 'ERR_R_INTERNAL_ERROR'. 1128109998Smarkm [Bodo Moeller] 1129109998Smarkm 1130109998Smarkm *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock 1131109998Smarkm suffices. 1132109998Smarkm [Bodo Moeller] 1133109998Smarkm 1134109998Smarkm *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This 1135109998Smarkm sets the subject name for a new request or supersedes the 1136109998Smarkm subject name in a given request. Formats that can be parsed are 1137109998Smarkm 'CN=Some Name, OU=myOU, C=IT' 1138109998Smarkm and 1139109998Smarkm 'CN=Some Name/OU=myOU/C=IT'. 1140109998Smarkm 1141109998Smarkm Add options '-batch' and '-verbose' to 'openssl req'. 1142109998Smarkm [Massimiliano Pala <madwolf@hackmasters.net>] 1143109998Smarkm 1144109998Smarkm *) Introduce the possibility to access global variables through 1145109998Smarkm functions on platform were that's the best way to handle exporting 1146109998Smarkm global variables in shared libraries. To enable this functionality, 1147109998Smarkm one must configure with "EXPORT_VAR_AS_FN" or defined the C macro 1148109998Smarkm "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter 1149109998Smarkm is normally done by Configure or something similar). 1150109998Smarkm 1151109998Smarkm To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL 1152109998Smarkm in the source file (foo.c) like this: 1153109998Smarkm 1154109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; 1155109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(double,bar); 1156109998Smarkm 1157109998Smarkm To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL 1158109998Smarkm and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: 1159109998Smarkm 1160109998Smarkm OPENSSL_DECLARE_GLOBAL(int,foo); 1161109998Smarkm #define foo OPENSSL_GLOBAL_REF(foo) 1162109998Smarkm OPENSSL_DECLARE_GLOBAL(double,bar); 1163109998Smarkm #define bar OPENSSL_GLOBAL_REF(bar) 1164109998Smarkm 1165109998Smarkm The #defines are very important, and therefore so is including the 1166109998Smarkm header file everywhere where the defined globals are used. 1167109998Smarkm 1168109998Smarkm The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition 1169109998Smarkm of ASN.1 items, but that structure is a bit different. 1170109998Smarkm 1171109998Smarkm The largest change is in util/mkdef.pl which has been enhanced with 1172109998Smarkm better and easier to understand logic to choose which symbols should 1173109998Smarkm go into the Windows .def files as well as a number of fixes and code 1174109998Smarkm cleanup (among others, algorithm keywords are now sorted 1175109998Smarkm lexicographically to avoid constant rewrites). 1176109998Smarkm [Richard Levitte] 1177109998Smarkm 1178109998Smarkm *) In BN_div() keep a copy of the sign of 'num' before writing the 1179109998Smarkm result to 'rm' because if rm==num the value will be overwritten 1180109998Smarkm and produce the wrong result if 'num' is negative: this caused 1181109998Smarkm problems with BN_mod() and BN_nnmod(). 1182109998Smarkm [Steve Henson] 1183109998Smarkm 1184109998Smarkm *) Function OCSP_request_verify(). This checks the signature on an 1185109998Smarkm OCSP request and verifies the signer certificate. The signer 1186109998Smarkm certificate is just checked for a generic purpose and OCSP request 1187109998Smarkm trust settings. 1188109998Smarkm [Steve Henson] 1189109998Smarkm 1190109998Smarkm *) Add OCSP_check_validity() function to check the validity of OCSP 1191109998Smarkm responses. OCSP responses are prepared in real time and may only 1192109998Smarkm be a few seconds old. Simply checking that the current time lies 1193109998Smarkm between thisUpdate and nextUpdate max reject otherwise valid responses 1194109998Smarkm caused by either OCSP responder or client clock inaccuracy. Instead 1195109998Smarkm we allow thisUpdate and nextUpdate to fall within a certain period of 1196109998Smarkm the current time. The age of the response can also optionally be 1197109998Smarkm checked. Two new options -validity_period and -status_age added to 1198109998Smarkm ocsp utility. 1199109998Smarkm [Steve Henson] 1200109998Smarkm 1201109998Smarkm *) If signature or public key algorithm is unrecognized print out its 1202109998Smarkm OID rather that just UNKNOWN. 1203109998Smarkm [Steve Henson] 1204109998Smarkm 1205109998Smarkm *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and 1206109998Smarkm OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate 1207109998Smarkm ID to be generated from the issuer certificate alone which can then be 1208109998Smarkm passed to OCSP_id_issuer_cmp(). 1209109998Smarkm [Steve Henson] 1210109998Smarkm 1211109998Smarkm *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new 1212109998Smarkm ASN1 modules to export functions returning ASN1_ITEM pointers 1213109998Smarkm instead of the ASN1_ITEM structures themselves. This adds several 1214109998Smarkm new macros which allow the underlying ASN1 function/structure to 1215109998Smarkm be accessed transparently. As a result code should not use ASN1_ITEM 1216109998Smarkm references directly (such as &X509_it) but instead use the relevant 1217109998Smarkm macros (such as ASN1_ITEM_rptr(X509)). This option is to allow 1218109998Smarkm use of the new ASN1 code on platforms where exporting structures 1219109998Smarkm is problematical (for example in shared libraries) but exporting 1220109998Smarkm functions returning pointers to structures is not. 1221109998Smarkm [Steve Henson] 1222109998Smarkm 1223109998Smarkm *) Add support for overriding the generation of SSL/TLS session IDs. 1224109998Smarkm These callbacks can be registered either in an SSL_CTX or per SSL. 1225109998Smarkm The purpose of this is to allow applications to control, if they wish, 1226109998Smarkm the arbitrary values chosen for use as session IDs, particularly as it 1227109998Smarkm can be useful for session caching in multiple-server environments. A 1228109998Smarkm command-line switch for testing this (and any client code that wishes 1229109998Smarkm to use such a feature) has been added to "s_server". 1230109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 1231109998Smarkm 1232109998Smarkm *) Modify mkdef.pl to recognise and parse preprocessor conditionals 1233109998Smarkm of the form '#if defined(...) || defined(...) || ...' and 1234109998Smarkm '#if !defined(...) && !defined(...) && ...'. This also avoids 1235109998Smarkm the growing number of special cases it was previously handling. 1236109998Smarkm [Richard Levitte] 1237109998Smarkm 1238109998Smarkm *) Make all configuration macros available for application by making 1239109998Smarkm sure they are available in opensslconf.h, by giving them names starting 1240109998Smarkm with "OPENSSL_" to avoid conflicts with other packages and by making 1241109998Smarkm sure e_os2.h will cover all platform-specific cases together with 1242109998Smarkm opensslconf.h. 1243109998Smarkm Additionally, it is now possible to define configuration/platform- 1244109998Smarkm specific names (called "system identities"). In the C code, these 1245109998Smarkm are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another 1246109998Smarkm macro with the name beginning with "OPENSSL_SYS_", which is determined 1247109998Smarkm from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on 1248109998Smarkm what is available. 1249109998Smarkm [Richard Levitte] 1250109998Smarkm 1251109998Smarkm *) New option -set_serial to 'req' and 'x509' this allows the serial 1252109998Smarkm number to use to be specified on the command line. Previously self 1253109998Smarkm signed certificates were hard coded with serial number 0 and the 1254109998Smarkm CA options of 'x509' had to use a serial number in a file which was 1255109998Smarkm auto incremented. 1256109998Smarkm [Steve Henson] 1257109998Smarkm 1258109998Smarkm *) New options to 'ca' utility to support V2 CRL entry extensions. 1259109998Smarkm Currently CRL reason, invalidity date and hold instruction are 1260109998Smarkm supported. Add new CRL extensions to V3 code and some new objects. 1261109998Smarkm [Steve Henson] 1262109998Smarkm 1263109998Smarkm *) New function EVP_CIPHER_CTX_set_padding() this is used to 1264109998Smarkm disable standard block padding (aka PKCS#5 padding) in the EVP 1265109998Smarkm API, which was previously mandatory. This means that the data is 1266109998Smarkm not padded in any way and so the total length much be a multiple 1267109998Smarkm of the block size, otherwise an error occurs. 1268109998Smarkm [Steve Henson] 1269109998Smarkm 1270109998Smarkm *) Initial (incomplete) OCSP SSL support. 1271109998Smarkm [Steve Henson] 1272109998Smarkm 1273109998Smarkm *) New function OCSP_parse_url(). This splits up a URL into its host, 1274109998Smarkm port and path components: primarily to parse OCSP URLs. New -url 1275109998Smarkm option to ocsp utility. 1276109998Smarkm [Steve Henson] 1277109998Smarkm 1278109998Smarkm *) New nonce behavior. The return value of OCSP_check_nonce() now 1279109998Smarkm reflects the various checks performed. Applications can decide 1280109998Smarkm whether to tolerate certain situations such as an absent nonce 1281109998Smarkm in a response when one was present in a request: the ocsp application 1282109998Smarkm just prints out a warning. New function OCSP_add1_basic_nonce() 1283109998Smarkm this is to allow responders to include a nonce in a response even if 1284109998Smarkm the request is nonce-less. 1285109998Smarkm [Steve Henson] 1286109998Smarkm 1287109998Smarkm *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are 1288109998Smarkm skipped when using openssl x509 multiple times on a single input file, 1289109998Smarkm e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs". 1290109998Smarkm [Bodo Moeller] 1291109998Smarkm 1292109998Smarkm *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() 1293109998Smarkm set string type: to handle setting ASN1_TIME structures. Fix ca 1294109998Smarkm utility to correctly initialize revocation date of CRLs. 1295109998Smarkm [Steve Henson] 1296109998Smarkm 1297109998Smarkm *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override 1298109998Smarkm the clients preferred ciphersuites and rather use its own preferences. 1299109998Smarkm Should help to work around M$ SGC (Server Gated Cryptography) bug in 1300109998Smarkm Internet Explorer by ensuring unchanged hash method during stepup. 1301109998Smarkm (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) 1302109998Smarkm [Lutz Jaenicke] 1303109998Smarkm 1304109998Smarkm *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael 1305109998Smarkm to aes and add a new 'exist' option to print out symbols that don't 1306109998Smarkm appear to exist. 1307109998Smarkm [Steve Henson] 1308109998Smarkm 1309109998Smarkm *) Additional options to ocsp utility to allow flags to be set and 1310109998Smarkm additional certificates supplied. 1311109998Smarkm [Steve Henson] 1312109998Smarkm 1313109998Smarkm *) Add the option -VAfile to 'openssl ocsp', so the user can give the 1314109998Smarkm OCSP client a number of certificate to only verify the response 1315109998Smarkm signature against. 1316109998Smarkm [Richard Levitte] 1317109998Smarkm 1318109998Smarkm *) Update Rijndael code to version 3.0 and change EVP AES ciphers to 1319109998Smarkm handle the new API. Currently only ECB, CBC modes supported. Add new 1320109998Smarkm AES OIDs. 1321109998Smarkm 1322109998Smarkm Add TLS AES ciphersuites as described in RFC3268, "Advanced 1323109998Smarkm Encryption Standard (AES) Ciphersuites for Transport Layer 1324109998Smarkm Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were 1325109998Smarkm not enabled by default and were not part of the "ALL" ciphersuite 1326109998Smarkm alias because they were not yet official; they could be 1327109998Smarkm explicitly requested by specifying the "AESdraft" ciphersuite 1328109998Smarkm group alias. In the final release of OpenSSL 0.9.7, the group 1329109998Smarkm alias is called "AES" and is part of "ALL".) 1330109998Smarkm [Ben Laurie, Steve Henson, Bodo Moeller] 1331109998Smarkm 1332109998Smarkm *) New function OCSP_copy_nonce() to copy nonce value (if present) from 1333109998Smarkm request to response. 1334109998Smarkm [Steve Henson] 1335109998Smarkm 1336109998Smarkm *) Functions for OCSP responders. OCSP_request_onereq_count(), 1337109998Smarkm OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() 1338109998Smarkm extract information from a certificate request. OCSP_response_create() 1339109998Smarkm creates a response and optionally adds a basic response structure. 1340109998Smarkm OCSP_basic_add1_status() adds a complete single response to a basic 1341109998Smarkm response and returns the OCSP_SINGLERESP structure just added (to allow 1342109998Smarkm extensions to be included for example). OCSP_basic_add1_cert() adds a 1343109998Smarkm certificate to a basic response and OCSP_basic_sign() signs a basic 1344109998Smarkm response with various flags. New helper functions ASN1_TIME_check() 1345109998Smarkm (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() 1346109998Smarkm (converts ASN1_TIME to GeneralizedTime). 1347109998Smarkm [Steve Henson] 1348109998Smarkm 1349109998Smarkm *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() 1350109998Smarkm in a single operation. X509_get0_pubkey_bitstr() extracts the public_key 1351109998Smarkm structure from a certificate. X509_pubkey_digest() digests the public_key 1352109998Smarkm contents: this is used in various key identifiers. 1353109998Smarkm [Steve Henson] 1354109998Smarkm 1355109998Smarkm *) Make sk_sort() tolerate a NULL argument. 1356109998Smarkm [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>] 1357109998Smarkm 1358109998Smarkm *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates 1359109998Smarkm passed by the function are trusted implicitly. If any of them signed the 1360109998Smarkm response then it is assumed to be valid and is not verified. 1361109998Smarkm [Steve Henson] 1362109998Smarkm 1363109998Smarkm *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT 1364109998Smarkm to data. This was previously part of the PKCS7 ASN1 code. This 1365109998Smarkm was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. 1366109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 1367109998Smarkm <support@securenetterm.com>] 1368109998Smarkm 1369109998Smarkm *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 1370109998Smarkm routines: without these tracing memory leaks is very painful. 1371109998Smarkm Fix leaks in PKCS12 and PKCS7 routines. 1372109998Smarkm [Steve Henson] 1373109998Smarkm 1374109998Smarkm *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). 1375109998Smarkm Previously it initialised the 'type' argument to V_ASN1_UTCTIME which 1376109998Smarkm effectively meant GeneralizedTime would never be used. Now it 1377109998Smarkm is initialised to -1 but X509_time_adj() now has to check the value 1378109998Smarkm and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or 1379109998Smarkm V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. 1380109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 1381109998Smarkm <support@securenetterm.com>] 1382109998Smarkm 1383109998Smarkm *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously 1384109998Smarkm result in a zero length in the ASN1_INTEGER structure which was 1385109998Smarkm not consistent with the structure when d2i_ASN1_INTEGER() was used 1386109998Smarkm and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() 1387109998Smarkm to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() 1388109998Smarkm where it did not print out a minus for negative ASN1_INTEGER. 1389109998Smarkm [Steve Henson] 1390109998Smarkm 1391109998Smarkm *) Add summary printout to ocsp utility. The various functions which 1392109998Smarkm convert status values to strings have been renamed to: 1393109998Smarkm OCSP_response_status_str(), OCSP_cert_status_str() and 1394109998Smarkm OCSP_crl_reason_str() and are no longer static. New options 1395109998Smarkm to verify nonce values and to disable verification. OCSP response 1396109998Smarkm printout format cleaned up. 1397109998Smarkm [Steve Henson] 1398109998Smarkm 1399109998Smarkm *) Add additional OCSP certificate checks. These are those specified 1400109998Smarkm in RFC2560. This consists of two separate checks: the CA of the 1401109998Smarkm certificate being checked must either be the OCSP signer certificate 1402109998Smarkm or the issuer of the OCSP signer certificate. In the latter case the 1403109998Smarkm OCSP signer certificate must contain the OCSP signing extended key 1404109998Smarkm usage. This check is performed by attempting to match the OCSP 1405109998Smarkm signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash 1406109998Smarkm in the OCSP_CERTID structures of the response. 1407109998Smarkm [Steve Henson] 1408109998Smarkm 1409109998Smarkm *) Initial OCSP certificate verification added to OCSP_basic_verify() 1410109998Smarkm and related routines. This uses the standard OpenSSL certificate 1411109998Smarkm verify routines to perform initial checks (just CA validity) and 1412109998Smarkm to obtain the certificate chain. Then additional checks will be 1413109998Smarkm performed on the chain. Currently the root CA is checked to see 1414109998Smarkm if it is explicitly trusted for OCSP signing. This is used to set 1415109998Smarkm a root CA as a global signing root: that is any certificate that 1416109998Smarkm chains to that CA is an acceptable OCSP signing certificate. 1417109998Smarkm [Steve Henson] 1418109998Smarkm 1419109998Smarkm *) New '-extfile ...' option to 'openssl ca' for reading X.509v3 1420109998Smarkm extensions from a separate configuration file. 1421109998Smarkm As when reading extensions from the main configuration file, 1422109998Smarkm the '-extensions ...' option may be used for specifying the 1423109998Smarkm section to use. 1424109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 1425109998Smarkm 1426109998Smarkm *) New OCSP utility. Allows OCSP requests to be generated or 1427109998Smarkm read. The request can be sent to a responder and the output 1428109998Smarkm parsed, outputed or printed in text form. Not complete yet: 1429109998Smarkm still needs to check the OCSP response validity. 1430109998Smarkm [Steve Henson] 1431109998Smarkm 1432109998Smarkm *) New subcommands for 'openssl ca': 1433109998Smarkm 'openssl ca -status <serial>' prints the status of the cert with 1434109998Smarkm the given serial number (according to the index file). 1435109998Smarkm 'openssl ca -updatedb' updates the expiry status of certificates 1436109998Smarkm in the index file. 1437109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 1438109998Smarkm 1439109998Smarkm *) New '-newreq-nodes' command option to CA.pl. This is like 1440109998Smarkm '-newreq', but calls 'openssl req' with the '-nodes' option 1441109998Smarkm so that the resulting key is not encrypted. 1442109998Smarkm [Damien Miller <djm@mindrot.org>] 1443109998Smarkm 1444109998Smarkm *) New configuration for the GNU Hurd. 1445109998Smarkm [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte] 1446109998Smarkm 1447109998Smarkm *) Initial code to implement OCSP basic response verify. This 1448109998Smarkm is currently incomplete. Currently just finds the signer's 1449109998Smarkm certificate and verifies the signature on the response. 1450109998Smarkm [Steve Henson] 1451109998Smarkm 1452109998Smarkm *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in 1453109998Smarkm value of OPENSSLDIR. This is available via the new '-d' option 1454109998Smarkm to 'openssl version', and is also included in 'openssl version -a'. 1455109998Smarkm [Bodo Moeller] 1456109998Smarkm 1457109998Smarkm *) Allowing defining memory allocation callbacks that will be given 1458109998Smarkm file name and line number information in additional arguments 1459109998Smarkm (a const char* and an int). The basic functionality remains, as 1460109998Smarkm well as the original possibility to just replace malloc(), 1461109998Smarkm realloc() and free() by functions that do not know about these 1462109998Smarkm additional arguments. To register and find out the current 1463109998Smarkm settings for extended allocation functions, the following 1464109998Smarkm functions are provided: 1465109998Smarkm 1466109998Smarkm CRYPTO_set_mem_ex_functions 1467109998Smarkm CRYPTO_set_locked_mem_ex_functions 1468109998Smarkm CRYPTO_get_mem_ex_functions 1469109998Smarkm CRYPTO_get_locked_mem_ex_functions 1470109998Smarkm 1471109998Smarkm These work the same way as CRYPTO_set_mem_functions and friends. 1472109998Smarkm CRYPTO_get_[locked_]mem_functions now writes 0 where such an 1473109998Smarkm extended allocation function is enabled. 1474109998Smarkm Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where 1475109998Smarkm a conventional allocation function is enabled. 1476109998Smarkm [Richard Levitte, Bodo Moeller] 1477109998Smarkm 1478109998Smarkm *) Finish off removing the remaining LHASH function pointer casts. 1479109998Smarkm There should no longer be any prototype-casting required when using 1480109998Smarkm the LHASH abstraction, and any casts that remain are "bugs". See 1481109998Smarkm the callback types and macros at the head of lhash.h for details 1482109998Smarkm (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). 1483109998Smarkm [Geoff Thorpe] 1484109998Smarkm 1485109998Smarkm *) Add automatic query of EGD sockets in RAND_poll() for the unix variant. 1486109998Smarkm If /dev/[u]random devices are not available or do not return enough 1487109998Smarkm entropy, EGD style sockets (served by EGD or PRNGD) will automatically 1488109998Smarkm be queried. 1489109998Smarkm The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and 1490109998Smarkm /etc/entropy will be queried once each in this sequence, quering stops 1491109998Smarkm when enough entropy was collected without querying more sockets. 1492109998Smarkm [Lutz Jaenicke] 1493109998Smarkm 1494109998Smarkm *) Change the Unix RAND_poll() variant to be able to poll several 1495109998Smarkm random devices, as specified by DEVRANDOM, until a sufficient amount 1496109998Smarkm of data has been collected. We spend at most 10 ms on each file 1497109998Smarkm (select timeout) and read in non-blocking mode. DEVRANDOM now 1498109998Smarkm defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" 1499109998Smarkm (previously it was just the string "/dev/urandom"), so on typical 1500109998Smarkm platforms the 10 ms delay will never occur. 1501109998Smarkm Also separate out the Unix variant to its own file, rand_unix.c. 1502109998Smarkm For VMS, there's a currently-empty rand_vms.c. 1503109998Smarkm [Richard Levitte] 1504109998Smarkm 1505109998Smarkm *) Move OCSP client related routines to ocsp_cl.c. These 1506109998Smarkm provide utility functions which an application needing 1507109998Smarkm to issue a request to an OCSP responder and analyse the 1508109998Smarkm response will typically need: as opposed to those which an 1509109998Smarkm OCSP responder itself would need which will be added later. 1510109998Smarkm 1511109998Smarkm OCSP_request_sign() signs an OCSP request with an API similar 1512109998Smarkm to PKCS7_sign(). OCSP_response_status() returns status of OCSP 1513109998Smarkm response. OCSP_response_get1_basic() extracts basic response 1514109998Smarkm from response. OCSP_resp_find_status(): finds and extracts status 1515109998Smarkm information from an OCSP_CERTID structure (which will be created 1516109998Smarkm when the request structure is built). These are built from lower 1517109998Smarkm level functions which work on OCSP_SINGLERESP structures but 1518109998Smarkm wont normally be used unless the application wishes to examine 1519109998Smarkm extensions in the OCSP response for example. 1520109998Smarkm 1521109998Smarkm Replace nonce routines with a pair of functions. 1522109998Smarkm OCSP_request_add1_nonce() adds a nonce value and optionally 1523109998Smarkm generates a random value. OCSP_check_nonce() checks the 1524109998Smarkm validity of the nonce in an OCSP response. 1525109998Smarkm [Steve Henson] 1526109998Smarkm 1527109998Smarkm *) Change function OCSP_request_add() to OCSP_request_add0_id(). 1528109998Smarkm This doesn't copy the supplied OCSP_CERTID and avoids the 1529109998Smarkm need to free up the newly created id. Change return type 1530109998Smarkm to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. 1531109998Smarkm This can then be used to add extensions to the request. 1532109998Smarkm Deleted OCSP_request_new(), since most of its functionality 1533109998Smarkm is now in OCSP_REQUEST_new() (and the case insensitive name 1534109998Smarkm clash) apart from the ability to set the request name which 1535109998Smarkm will be added elsewhere. 1536109998Smarkm [Steve Henson] 1537109998Smarkm 1538109998Smarkm *) Update OCSP API. Remove obsolete extensions argument from 1539109998Smarkm various functions. Extensions are now handled using the new 1540109998Smarkm OCSP extension code. New simple OCSP HTTP function which 1541109998Smarkm can be used to send requests and parse the response. 1542109998Smarkm [Steve Henson] 1543109998Smarkm 1544109998Smarkm *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new 1545109998Smarkm ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN 1546109998Smarkm uses the special reorder version of SET OF to sort the attributes 1547109998Smarkm and reorder them to match the encoded order. This resolves a long 1548109998Smarkm standing problem: a verify on a PKCS7 structure just after signing 1549109998Smarkm it used to fail because the attribute order did not match the 1550109998Smarkm encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: 1551109998Smarkm it uses the received order. This is necessary to tolerate some broken 1552109998Smarkm software that does not order SET OF. This is handled by encoding 1553109998Smarkm as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) 1554109998Smarkm to produce the required SET OF. 1555109998Smarkm [Steve Henson] 1556109998Smarkm 1557109998Smarkm *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and 1558109998Smarkm OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header 1559109998Smarkm files to get correct declarations of the ASN.1 item variables. 1560109998Smarkm [Richard Levitte] 1561109998Smarkm 1562109998Smarkm *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many 1563109998Smarkm PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: 1564109998Smarkm asn1_check_tlen() would sometimes attempt to use 'ctx' when it was 1565109998Smarkm NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). 1566109998Smarkm New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant 1567109998Smarkm ASN1_ITEM and no wrapper functions. 1568109998Smarkm [Steve Henson] 1569109998Smarkm 1570109998Smarkm *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These 1571109998Smarkm replace the old function pointer based I/O routines. Change most of 1572109998Smarkm the *_d2i_bio() and *_d2i_fp() functions to use these. 1573109998Smarkm [Steve Henson] 1574109998Smarkm 1575109998Smarkm *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor 1576109998Smarkm lines, recognice more "algorithms" that can be deselected, and make 1577109998Smarkm it complain about algorithm deselection that isn't recognised. 1578109998Smarkm [Richard Levitte] 1579109998Smarkm 1580109998Smarkm *) New ASN1 functions to handle dup, sign, verify, digest, pack and 1581109998Smarkm unpack operations in terms of ASN1_ITEM. Modify existing wrappers 1582109998Smarkm to use new functions. Add NO_ASN1_OLD which can be set to remove 1583109998Smarkm some old style ASN1 functions: this can be used to determine if old 1584109998Smarkm code will still work when these eventually go away. 1585109998Smarkm [Steve Henson] 1586109998Smarkm 1587109998Smarkm *) New extension functions for OCSP structures, these follow the 1588109998Smarkm same conventions as certificates and CRLs. 1589109998Smarkm [Steve Henson] 1590109998Smarkm 1591109998Smarkm *) New function X509V3_add1_i2d(). This automatically encodes and 1592109998Smarkm adds an extension. Its behaviour can be customised with various 1593109998Smarkm flags to append, replace or delete. Various wrappers added for 1594109998Smarkm certifcates and CRLs. 1595109998Smarkm [Steve Henson] 1596109998Smarkm 1597109998Smarkm *) Fix to avoid calling the underlying ASN1 print routine when 1598109998Smarkm an extension cannot be parsed. Correct a typo in the 1599109998Smarkm OCSP_SERVICELOC extension. Tidy up print OCSP format. 1600109998Smarkm [Steve Henson] 1601109998Smarkm 1602109998Smarkm *) Make mkdef.pl parse some of the ASN1 macros and add apropriate 1603109998Smarkm entries for variables. 1604109998Smarkm [Steve Henson] 1605109998Smarkm 1606109998Smarkm *) Add functionality to apps/openssl.c for detecting locking 1607109998Smarkm problems: As the program is single-threaded, all we have 1608109998Smarkm to do is register a locking callback using an array for 1609109998Smarkm storing which locks are currently held by the program. 1610109998Smarkm [Bodo Moeller] 1611109998Smarkm 1612109998Smarkm *) Use a lock around the call to CRYPTO_get_ex_new_index() in 1613109998Smarkm SSL_get_ex_data_X509_STORE_idx(), which is used in 1614109998Smarkm ssl_verify_cert_chain() and thus can be called at any time 1615109998Smarkm during TLS/SSL handshakes so that thread-safety is essential. 1616109998Smarkm Unfortunately, the ex_data design is not at all suited 1617109998Smarkm for multi-threaded use, so it probably should be abolished. 1618109998Smarkm [Bodo Moeller] 1619109998Smarkm 1620109998Smarkm *) Added Broadcom "ubsec" ENGINE to OpenSSL. 1621109998Smarkm [Broadcom, tweaked and integrated by Geoff Thorpe] 1622109998Smarkm 1623109998Smarkm *) Move common extension printing code to new function 1624109998Smarkm X509V3_print_extensions(). Reorganise OCSP print routines and 1625109998Smarkm implement some needed OCSP ASN1 functions. Add OCSP extensions. 1626109998Smarkm [Steve Henson] 1627109998Smarkm 1628109998Smarkm *) New function X509_signature_print() to remove duplication in some 1629109998Smarkm print routines. 1630109998Smarkm [Steve Henson] 1631109998Smarkm 1632109998Smarkm *) Add a special meaning when SET OF and SEQUENCE OF flags are both 1633109998Smarkm set (this was treated exactly the same as SET OF previously). This 1634109998Smarkm is used to reorder the STACK representing the structure to match the 1635109998Smarkm encoding. This will be used to get round a problem where a PKCS7 1636109998Smarkm structure which was signed could not be verified because the STACK 1637109998Smarkm order did not reflect the encoded order. 1638109998Smarkm [Steve Henson] 1639109998Smarkm 1640109998Smarkm *) Reimplement the OCSP ASN1 module using the new code. 1641109998Smarkm [Steve Henson] 1642109998Smarkm 1643109998Smarkm *) Update the X509V3 code to permit the use of an ASN1_ITEM structure 1644109998Smarkm for its ASN1 operations. The old style function pointers still exist 1645109998Smarkm for now but they will eventually go away. 1646109998Smarkm [Steve Henson] 1647109998Smarkm 1648109998Smarkm *) Merge in replacement ASN1 code from the ASN1 branch. This almost 1649109998Smarkm completely replaces the old ASN1 functionality with a table driven 1650109998Smarkm encoder and decoder which interprets an ASN1_ITEM structure describing 1651109998Smarkm the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is 1652109998Smarkm largely maintained. Almost all of the old asn1_mac.h macro based ASN1 1653109998Smarkm has also been converted to the new form. 1654109998Smarkm [Steve Henson] 1655109998Smarkm 1656109998Smarkm *) Change BN_mod_exp_recp so that negative moduli are tolerated 1657109998Smarkm (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set 1658109998Smarkm so that BN_mod_exp_mont and BN_mod_exp_mont_word work 1659109998Smarkm for negative moduli. 1660109998Smarkm [Bodo Moeller] 1661109998Smarkm 1662109998Smarkm *) Fix BN_uadd and BN_usub: Always return non-negative results instead 1663109998Smarkm of not touching the result's sign bit. 1664109998Smarkm [Bodo Moeller] 1665109998Smarkm 1666109998Smarkm *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be 1667109998Smarkm set. 1668109998Smarkm [Bodo Moeller] 1669109998Smarkm 1670109998Smarkm *) Changed the LHASH code to use prototypes for callbacks, and created 1671109998Smarkm macros to declare and implement thin (optionally static) functions 1672109998Smarkm that provide type-safety and avoid function pointer casting for the 1673109998Smarkm type-specific callbacks. 1674109998Smarkm [Geoff Thorpe] 1675109998Smarkm 1676109998Smarkm *) Added Kerberos Cipher Suites to be used with TLS, as written in 1677109998Smarkm RFC 2712. 1678109998Smarkm [Veers Staats <staatsvr@asc.hpc.mil>, 1679109998Smarkm Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte] 1680109998Smarkm 1681109998Smarkm *) Reformat the FAQ so the different questions and answers can be divided 1682109998Smarkm in sections depending on the subject. 1683109998Smarkm [Richard Levitte] 1684109998Smarkm 1685109998Smarkm *) Have the zlib compression code load ZLIB.DLL dynamically under 1686109998Smarkm Windows. 1687109998Smarkm [Richard Levitte] 1688109998Smarkm 1689109998Smarkm *) New function BN_mod_sqrt for computing square roots modulo a prime 1690109998Smarkm (using the probabilistic Tonelli-Shanks algorithm unless 1691109998Smarkm p == 3 (mod 4) or p == 5 (mod 8), which are cases that can 1692109998Smarkm be handled deterministically). 1693109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 1694109998Smarkm 1695109998Smarkm *) Make BN_mod_inverse faster by explicitly handling small quotients 1696109998Smarkm in the Euclid loop. (Speed gain about 20% for small moduli [256 or 1697109998Smarkm 512 bits], about 30% for larger ones [1024 or 2048 bits].) 1698109998Smarkm [Bodo Moeller] 1699109998Smarkm 1700109998Smarkm *) New function BN_kronecker. 1701109998Smarkm [Bodo Moeller] 1702109998Smarkm 1703109998Smarkm *) Fix BN_gcd so that it works on negative inputs; the result is 1704109998Smarkm positive unless both parameters are zero. 1705109998Smarkm Previously something reasonably close to an infinite loop was 1706109998Smarkm possible because numbers could be growing instead of shrinking 1707109998Smarkm in the implementation of Euclid's algorithm. 1708109998Smarkm [Bodo Moeller] 1709109998Smarkm 1710109998Smarkm *) Fix BN_is_word() and BN_is_one() macros to take into account the 1711109998Smarkm sign of the number in question. 1712109998Smarkm 1713109998Smarkm Fix BN_is_word(a,w) to work correctly for w == 0. 1714109998Smarkm 1715109998Smarkm The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) 1716109998Smarkm because its test if the absolute value of 'a' equals 'w'. 1717109998Smarkm Note that BN_abs_is_word does *not* handle w == 0 reliably; 1718109998Smarkm it exists mostly for use in the implementations of BN_is_zero(), 1719109998Smarkm BN_is_one(), and BN_is_word(). 1720109998Smarkm [Bodo Moeller] 1721109998Smarkm 1722109998Smarkm *) New function BN_swap. 1723109998Smarkm [Bodo Moeller] 1724109998Smarkm 1725109998Smarkm *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that 1726109998Smarkm the exponentiation functions are more likely to produce reasonable 1727109998Smarkm results on negative inputs. 1728109998Smarkm [Bodo Moeller] 1729109998Smarkm 1730109998Smarkm *) Change BN_mod_mul so that the result is always non-negative. 1731109998Smarkm Previously, it could be negative if one of the factors was negative; 1732109998Smarkm I don't think anyone really wanted that behaviour. 1733109998Smarkm [Bodo Moeller] 1734109998Smarkm 1735109998Smarkm *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c 1736109998Smarkm (except for exponentiation, which stays in crypto/bn/bn_exp.c, 1737109998Smarkm and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c) 1738109998Smarkm and add new functions: 1739109998Smarkm 1740109998Smarkm BN_nnmod 1741109998Smarkm BN_mod_sqr 1742109998Smarkm BN_mod_add 1743109998Smarkm BN_mod_add_quick 1744109998Smarkm BN_mod_sub 1745109998Smarkm BN_mod_sub_quick 1746109998Smarkm BN_mod_lshift1 1747109998Smarkm BN_mod_lshift1_quick 1748109998Smarkm BN_mod_lshift 1749109998Smarkm BN_mod_lshift_quick 1750109998Smarkm 1751109998Smarkm These functions always generate non-negative results. 1752109998Smarkm 1753109998Smarkm BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r 1754109998Smarkm such that |m| < r < 0, BN_nnmod will output rem + |m| instead). 1755109998Smarkm 1756109998Smarkm BN_mod_XXX_quick(r, a, [b,] m) generates the same result as 1757109998Smarkm BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b] 1758109998Smarkm be reduced modulo m. 1759109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 1760109998Smarkm 1761111147Snectar#if 0 1762111147Snectar The following entry accidentily appeared in the CHANGES file 1763111147Snectar distributed with OpenSSL 0.9.7. The modifications described in 1764111147Snectar it do *not* apply to OpenSSL 0.9.7. 1765111147Snectar 1766109998Smarkm *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 1767109998Smarkm was actually never needed) and in BN_mul(). The removal in BN_mul() 1768109998Smarkm required a small change in bn_mul_part_recursive() and the addition 1769109998Smarkm of the functions bn_cmp_part_words(), bn_sub_part_words() and 1770109998Smarkm bn_add_part_words(), which do the same thing as bn_cmp_words(), 1771109998Smarkm bn_sub_words() and bn_add_words() except they take arrays with 1772109998Smarkm differing sizes. 1773109998Smarkm [Richard Levitte] 1774111147Snectar#endif 1775109998Smarkm 1776109998Smarkm *) In 'openssl passwd', verify passwords read from the terminal 1777109998Smarkm unless the '-salt' option is used (which usually means that 1778109998Smarkm verification would just waste user's time since the resulting 1779109998Smarkm hash is going to be compared with some given password hash) 1780109998Smarkm or the new '-noverify' option is used. 1781109998Smarkm 1782109998Smarkm This is an incompatible change, but it does not affect 1783109998Smarkm non-interactive use of 'openssl passwd' (passwords on the command 1784109998Smarkm line, '-stdin' option, '-in ...' option) and thus should not 1785109998Smarkm cause any problems. 1786109998Smarkm [Bodo Moeller] 1787109998Smarkm 1788109998Smarkm *) Remove all references to RSAref, since there's no more need for it. 1789109998Smarkm [Richard Levitte] 1790109998Smarkm 1791109998Smarkm *) Make DSO load along a path given through an environment variable 1792109998Smarkm (SHLIB_PATH) with shl_load(). 1793109998Smarkm [Richard Levitte] 1794109998Smarkm 1795109998Smarkm *) Constify the ENGINE code as a result of BIGNUM constification. 1796109998Smarkm Also constify the RSA code and most things related to it. In a 1797109998Smarkm few places, most notable in the depth of the ASN.1 code, ugly 1798109998Smarkm casts back to non-const were required (to be solved at a later 1799109998Smarkm time) 1800109998Smarkm [Richard Levitte] 1801109998Smarkm 1802109998Smarkm *) Make it so the openssl application has all engines loaded by default. 1803109998Smarkm [Richard Levitte] 1804109998Smarkm 1805109998Smarkm *) Constify the BIGNUM routines a little more. 1806109998Smarkm [Richard Levitte] 1807109998Smarkm 1808109998Smarkm *) Add the following functions: 1809109998Smarkm 1810109998Smarkm ENGINE_load_cswift() 1811109998Smarkm ENGINE_load_chil() 1812109998Smarkm ENGINE_load_atalla() 1813109998Smarkm ENGINE_load_nuron() 1814109998Smarkm ENGINE_load_builtin_engines() 1815109998Smarkm 1816109998Smarkm That way, an application can itself choose if external engines that 1817109998Smarkm are built-in in OpenSSL shall ever be used or not. The benefit is 1818109998Smarkm that applications won't have to be linked with libdl or other dso 1819109998Smarkm libraries unless it's really needed. 1820109998Smarkm 1821109998Smarkm Changed 'openssl engine' to load all engines on demand. 1822109998Smarkm Changed the engine header files to avoid the duplication of some 1823109998Smarkm declarations (they differed!). 1824109998Smarkm [Richard Levitte] 1825109998Smarkm 1826109998Smarkm *) 'openssl engine' can now list capabilities. 1827109998Smarkm [Richard Levitte] 1828109998Smarkm 1829109998Smarkm *) Better error reporting in 'openssl engine'. 1830109998Smarkm [Richard Levitte] 1831109998Smarkm 1832109998Smarkm *) Never call load_dh_param(NULL) in s_server. 1833109998Smarkm [Bodo Moeller] 1834109998Smarkm 1835109998Smarkm *) Add engine application. It can currently list engines by name and 1836109998Smarkm identity, and test if they are actually available. 1837109998Smarkm [Richard Levitte] 1838109998Smarkm 1839109998Smarkm *) Improve RPM specification file by forcing symbolic linking and making 1840109998Smarkm sure the installed documentation is also owned by root.root. 1841109998Smarkm [Damien Miller <djm@mindrot.org>] 1842109998Smarkm 1843109998Smarkm *) Give the OpenSSL applications more possibilities to make use of 1844109998Smarkm keys (public as well as private) handled by engines. 1845109998Smarkm [Richard Levitte] 1846109998Smarkm 1847109998Smarkm *) Add OCSP code that comes from CertCo. 1848109998Smarkm [Richard Levitte] 1849109998Smarkm 1850109998Smarkm *) Add VMS support for the Rijndael code. 1851109998Smarkm [Richard Levitte] 1852109998Smarkm 1853109998Smarkm *) Added untested support for Nuron crypto accelerator. 1854109998Smarkm [Ben Laurie] 1855109998Smarkm 1856109998Smarkm *) Add support for external cryptographic devices. This code was 1857109998Smarkm previously distributed separately as the "engine" branch. 1858109998Smarkm [Geoff Thorpe, Richard Levitte] 1859109998Smarkm 1860109998Smarkm *) Rework the filename-translation in the DSO code. It is now possible to 1861109998Smarkm have far greater control over how a "name" is turned into a filename 1862109998Smarkm depending on the operating environment and any oddities about the 1863109998Smarkm different shared library filenames on each system. 1864109998Smarkm [Geoff Thorpe] 1865109998Smarkm 1866109998Smarkm *) Support threads on FreeBSD-elf in Configure. 1867109998Smarkm [Richard Levitte] 1868109998Smarkm 1869109998Smarkm *) Fix for SHA1 assembly problem with MASM: it produces 1870109998Smarkm warnings about corrupt line number information when assembling 1871109998Smarkm with debugging information. This is caused by the overlapping 1872109998Smarkm of two sections. 1873109998Smarkm [Bernd Matthes <mainbug@celocom.de>, Steve Henson] 1874109998Smarkm 1875109998Smarkm *) NCONF changes. 1876109998Smarkm NCONF_get_number() has no error checking at all. As a replacement, 1877109998Smarkm NCONF_get_number_e() is defined (_e for "error checking") and is 1878109998Smarkm promoted strongly. The old NCONF_get_number is kept around for 1879109998Smarkm binary backward compatibility. 1880109998Smarkm Make it possible for methods to load from something other than a BIO, 1881109998Smarkm by providing a function pointer that is given a name instead of a BIO. 1882109998Smarkm For example, this could be used to load configuration data from an 1883109998Smarkm LDAP server. 1884109998Smarkm [Richard Levitte] 1885109998Smarkm 1886109998Smarkm *) Fix for non blocking accept BIOs. Added new I/O special reason 1887109998Smarkm BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs 1888109998Smarkm with non blocking I/O was not possible because no retry code was 1889109998Smarkm implemented. Also added new SSL code SSL_WANT_ACCEPT to cover 1890109998Smarkm this case. 1891109998Smarkm [Steve Henson] 1892109998Smarkm 1893109998Smarkm *) Added the beginnings of Rijndael support. 1894109998Smarkm [Ben Laurie] 1895109998Smarkm 1896109998Smarkm *) Fix for bug in DirectoryString mask setting. Add support for 1897109998Smarkm X509_NAME_print_ex() in 'req' and X509_print_ex() function 1898109998Smarkm to allow certificate printing to more controllable, additional 1899109998Smarkm 'certopt' option to 'x509' to allow new printing options to be 1900109998Smarkm set. 1901109998Smarkm [Steve Henson] 1902109998Smarkm 1903109998Smarkm *) Clean old EAY MD5 hack from e_os.h. 1904109998Smarkm [Richard Levitte] 1905109998Smarkm 1906111147Snectar Changes between 0.9.6h and 0.9.6i [19 Feb 2003] 1907111147Snectar 1908111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 1909111147Snectar via timing by performing a MAC computation even if incorrrect 1910111147Snectar block cipher padding has been found. This is a countermeasure 1911111147Snectar against active attacks where the attacker has to distinguish 1912111147Snectar between bad padding and a MAC verification error. (CAN-2003-0078) 1913111147Snectar 1914111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 1915111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 1916111147Snectar Martin Vuagnoux (EPFL, Ilion)] 1917111147Snectar 1918109998Smarkm Changes between 0.9.6g and 0.9.6h [5 Dec 2002] 1919109998Smarkm 1920109998Smarkm *) New function OPENSSL_cleanse(), which is used to cleanse a section of 1921109998Smarkm memory from it's contents. This is done with a counter that will 1922109998Smarkm place alternating values in each byte. This can be used to solve 1923109998Smarkm two issues: 1) the removal of calls to memset() by highly optimizing 1924109998Smarkm compilers, and 2) cleansing with other values than 0, since those can 1925109998Smarkm be read through on certain media, for example a swap space on disk. 1926109998Smarkm [Geoff Thorpe] 1927109998Smarkm 1928109998Smarkm *) Bugfix: client side session caching did not work with external caching, 1929109998Smarkm because the session->cipher setting was not restored when reloading 1930109998Smarkm from the external cache. This problem was masked, when 1931109998Smarkm SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. 1932109998Smarkm (Found by Steve Haslam <steve@araqnid.ddts.net>.) 1933109998Smarkm [Lutz Jaenicke] 1934109998Smarkm 1935109998Smarkm *) Fix client_certificate (ssl/s2_clnt.c): The permissible total 1936109998Smarkm length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. 1937109998Smarkm [Zeev Lieber <zeev-l@yahoo.com>] 1938109998Smarkm 1939109998Smarkm *) Undo an undocumented change introduced in 0.9.6e which caused 1940109998Smarkm repeated calls to OpenSSL_add_all_ciphers() and 1941109998Smarkm OpenSSL_add_all_digests() to be ignored, even after calling 1942109998Smarkm EVP_cleanup(). 1943109998Smarkm [Richard Levitte] 1944109998Smarkm 1945109998Smarkm *) Change the default configuration reader to deal with last line not 1946109998Smarkm being properly terminated. 1947109998Smarkm [Richard Levitte] 1948109998Smarkm 1949109998Smarkm *) Change X509_NAME_cmp() so it applies the special rules on handling 1950109998Smarkm DN values that are of type PrintableString, as well as RDNs of type 1951109998Smarkm emailAddress where the value has the type ia5String. 1952109998Smarkm [stefank@valicert.com via Richard Levitte] 1953109998Smarkm 1954109998Smarkm *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half 1955109998Smarkm the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently 1956109998Smarkm doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be 1957109998Smarkm the bitwise-OR of the two for use by the majority of applications 1958109998Smarkm wanting this behaviour, and update the docs. The documented 1959109998Smarkm behaviour and actual behaviour were inconsistent and had been 1960109998Smarkm changing anyway, so this is more a bug-fix than a behavioural 1961109998Smarkm change. 1962109998Smarkm [Geoff Thorpe, diagnosed by Nadav Har'El] 1963109998Smarkm 1964109998Smarkm *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c 1965109998Smarkm (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). 1966109998Smarkm [Bodo Moeller] 1967109998Smarkm 1968109998Smarkm *) Fix initialization code race conditions in 1969109998Smarkm SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), 1970109998Smarkm SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), 1971109998Smarkm SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), 1972109998Smarkm TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), 1973109998Smarkm ssl2_get_cipher_by_char(), 1974109998Smarkm ssl3_get_cipher_by_char(). 1975109998Smarkm [Patrick McCormick <patrick@tellme.com>, Bodo Moeller] 1976109998Smarkm 1977109998Smarkm *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after 1978109998Smarkm the cached sessions are flushed, as the remove_cb() might use ex_data 1979109998Smarkm contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> 1980109998Smarkm (see [openssl.org #212]). 1981109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 1982109998Smarkm 1983109998Smarkm *) Fix typo in OBJ_txt2obj which incorrectly passed the content 1984109998Smarkm length, instead of the encoding length to d2i_ASN1_OBJECT. 1985109998Smarkm [Steve Henson] 1986109998Smarkm 1987101618Snectar Changes between 0.9.6f and 0.9.6g [9 Aug 2002] 1988101618Snectar 1989101618Snectar *) [In 0.9.6g-engine release:] 1990101618Snectar Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall'). 1991101618Snectar [Lynn Gazis <lgazis@rainbow.com>] 1992101618Snectar 1993101613Snectar Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 1994101613Snectar 1995101613Snectar *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 1996101613Snectar and get fix the header length calculation. 1997101613Snectar [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 1998101613Snectar Alon Kantor <alonk@checkpoint.com> (and others), 1999101613Snectar Steve Henson] 2000101613Snectar 2001101613Snectar *) Use proper error handling instead of 'assertions' in buffer 2002101613Snectar overflow checks added in 0.9.6e. This prevents DoS (the 2003101613Snectar assertions could call abort()). 2004101613Snectar [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] 2005101613Snectar 2006100936Snectar Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 2007100936Snectar 2008109998Smarkm *) Add various sanity checks to asn1_get_length() to reject 2009109998Smarkm the ASN1 length bytes if they exceed sizeof(long), will appear 2010109998Smarkm negative or the content length exceeds the length of the 2011109998Smarkm supplied buffer. 2012109998Smarkm [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 2013109998Smarkm 2014100936Snectar *) Fix cipher selection routines: ciphers without encryption had no flags 2015100936Snectar for the cipher strength set and where therefore not handled correctly 2016100936Snectar by the selection routines (PR #130). 2017100936Snectar [Lutz Jaenicke] 2018100936Snectar 2019100936Snectar *) Fix EVP_dsa_sha macro. 2020100936Snectar [Nils Larsch] 2021100936Snectar 2022100936Snectar *) New option 2023100936Snectar SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 2024100936Snectar for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 2025100936Snectar that was added in OpenSSL 0.9.6d. 2026100936Snectar 2027100936Snectar As the countermeasure turned out to be incompatible with some 2028100936Snectar broken SSL implementations, the new option is part of SSL_OP_ALL. 2029100936Snectar SSL_OP_ALL is usually employed when compatibility with weird SSL 2030100936Snectar implementations is desired (e.g. '-bugs' option to 's_client' and 2031100936Snectar 's_server'), so the new option is automatically set in many 2032100936Snectar applications. 2033100936Snectar [Bodo Moeller] 2034100936Snectar 2035100936Snectar *) Changes in security patch: 2036100936Snectar 2037100936Snectar Changes marked "(CHATS)" were sponsored by the Defense Advanced 2038100936Snectar Research Projects Agency (DARPA) and Air Force Research Laboratory, 2039100936Snectar Air Force Materiel Command, USAF, under agreement number 2040100936Snectar F30602-01-2-0537. 2041100936Snectar 2042100936Snectar *) Add various sanity checks to asn1_get_length() to reject 2043100936Snectar the ASN1 length bytes if they exceed sizeof(long), will appear 2044100936Snectar negative or the content length exceeds the length of the 2045109998Smarkm supplied buffer. (CAN-2002-0659) 2046100936Snectar [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 2047100936Snectar 2048100936Snectar *) Assertions for various potential buffer overflows, not known to 2049100936Snectar happen in practice. 2050100936Snectar [Ben Laurie (CHATS)] 2051100936Snectar 2052100936Snectar *) Various temporary buffers to hold ASCII versions of integers were 2053100936Snectar too small for 64 bit platforms. (CAN-2002-0655) 2054100936Snectar [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 2055100936Snectar 2056100936Snectar *) Remote buffer overflow in SSL3 protocol - an attacker could 2057100936Snectar supply an oversized session ID to a client. (CAN-2002-0656) 2058100936Snectar [Ben Laurie (CHATS)] 2059100936Snectar 2060100936Snectar *) Remote buffer overflow in SSL2 protocol - an attacker could 2061100936Snectar supply an oversized client master key. (CAN-2002-0656) 2062100936Snectar [Ben Laurie (CHATS)] 2063100936Snectar 2064100928Snectar Changes between 0.9.6c and 0.9.6d [9 May 2002] 2065100928Snectar 2066100928Snectar *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 2067100928Snectar encoded as NULL) with id-dsa-with-sha1. 2068100928Snectar [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 2069100928Snectar 2070100928Snectar *) Check various X509_...() return values in apps/req.c. 2071100928Snectar [Nils Larsch <nla@trustcenter.de>] 2072100928Snectar 2073100928Snectar *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 2074100928Snectar an end-of-file condition would erronously be flagged, when the CRLF 2075100928Snectar was just at the end of a processed block. The bug was discovered when 2076100928Snectar processing data through a buffering memory BIO handing the data to a 2077100928Snectar BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 2078100928Snectar <ptsekov@syntrex.com> and Nedelcho Stanev. 2079100928Snectar [Lutz Jaenicke] 2080100928Snectar 2081100928Snectar *) Implement a countermeasure against a vulnerability recently found 2082100928Snectar in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 2083100928Snectar before application data chunks to avoid the use of known IVs 2084100928Snectar with data potentially chosen by the attacker. 2085100928Snectar [Bodo Moeller] 2086100928Snectar 2087100928Snectar *) Fix length checks in ssl3_get_client_hello(). 2088100928Snectar [Bodo Moeller] 2089100928Snectar 2090100928Snectar *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 2091100928Snectar to prevent ssl3_read_internal() from incorrectly assuming that 2092100928Snectar ssl3_read_bytes() found application data while handshake 2093100928Snectar processing was enabled when in fact s->s3->in_read_app_data was 2094100928Snectar merely automatically cleared during the initial handshake. 2095100928Snectar [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 2096100928Snectar 2097100928Snectar *) Fix object definitions for Private and Enterprise: they were not 2098100928Snectar recognized in their shortname (=lowercase) representation. Extend 2099100928Snectar obj_dat.pl to issue an error when using undefined keywords instead 2100100928Snectar of silently ignoring the problem (Svenning Sorensen 2101100928Snectar <sss@sss.dnsalias.net>). 2102100928Snectar [Lutz Jaenicke] 2103100928Snectar 2104100928Snectar *) Fix DH_generate_parameters() so that it works for 'non-standard' 2105100928Snectar generators, i.e. generators other than 2 and 5. (Previously, the 2106100928Snectar code did not properly initialise the 'add' and 'rem' values to 2107100928Snectar BN_generate_prime().) 2108100928Snectar 2109100928Snectar In the new general case, we do not insist that 'generator' is 2110100928Snectar actually a primitive root: This requirement is rather pointless; 2111100928Snectar a generator of the order-q subgroup is just as good, if not 2112100928Snectar better. 2113100928Snectar [Bodo Moeller] 2114100928Snectar 2115100928Snectar *) Map new X509 verification errors to alerts. Discovered and submitted by 2116100928Snectar Tom Wu <tom@arcot.com>. 2117100928Snectar [Lutz Jaenicke] 2118100928Snectar 2119100928Snectar *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 2120100928Snectar returning non-zero before the data has been completely received 2121100928Snectar when using non-blocking I/O. 2122100928Snectar [Bodo Moeller; problem pointed out by John Hughes] 2123100928Snectar 2124100928Snectar *) Some of the ciphers missed the strength entry (SSL_LOW etc). 2125100928Snectar [Ben Laurie, Lutz Jaenicke] 2126100928Snectar 2127100928Snectar *) Fix bug in SSL_clear(): bad sessions were not removed (found by 2128100928Snectar Yoram Zahavi <YoramZ@gilian.com>). 2129100928Snectar [Lutz Jaenicke] 2130100928Snectar 2131100928Snectar *) Add information about CygWin 1.3 and on, and preserve proper 2132100928Snectar configuration for the versions before that. 2133100928Snectar [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 2134100928Snectar 2135100928Snectar *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 2136100928Snectar check whether we deal with a copy of a session and do not delete from 2137100928Snectar the cache in this case. Problem reported by "Izhar Shoshani Levi" 2138100928Snectar <izhar@checkpoint.com>. 2139100928Snectar [Lutz Jaenicke] 2140100928Snectar 2141100928Snectar *) Do not store session data into the internal session cache, if it 2142100928Snectar is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 2143100928Snectar flag is set). Proposed by Aslam <aslam@funk.com>. 2144100928Snectar [Lutz Jaenicke] 2145100928Snectar 2146100928Snectar *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 2147100928Snectar value is 0. 2148100928Snectar [Richard Levitte] 2149100928Snectar 2150109998Smarkm *) [In 0.9.6d-engine release:] 2151109998Smarkm Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 2152100928Snectar [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 2153100928Snectar 2154100928Snectar *) Add the configuration target linux-s390x. 2155100928Snectar [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 2156100928Snectar 2157100928Snectar *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 2158100928Snectar ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 2159100928Snectar variable as an indication that a ClientHello message has been 2160100928Snectar received. As the flag value will be lost between multiple 2161100928Snectar invocations of ssl3_accept when using non-blocking I/O, the 2162100928Snectar function may not be aware that a handshake has actually taken 2163100928Snectar place, thus preventing a new session from being added to the 2164100928Snectar session cache. 2165100928Snectar 2166100928Snectar To avoid this problem, we now set s->new_session to 2 instead of 2167100928Snectar using a local variable. 2168100928Snectar [Lutz Jaenicke, Bodo Moeller] 2169100928Snectar 2170100928Snectar *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 2171100928Snectar if the SSL_R_LENGTH_MISMATCH error is detected. 2172100928Snectar [Geoff Thorpe, Bodo Moeller] 2173100928Snectar 2174100928Snectar *) New 'shared_ldflag' column in Configure platform table. 2175100928Snectar [Richard Levitte] 2176100928Snectar 2177100928Snectar *) Fix EVP_CIPHER_mode macro. 2178100928Snectar ["Dan S. Camper" <dan@bti.net>] 2179100928Snectar 2180100928Snectar *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 2181100928Snectar type, we must throw them away by setting rr->length to 0. 2182100928Snectar [D P Chang <dpc@qualys.com>] 2183100928Snectar 218489837Skris Changes between 0.9.6b and 0.9.6c [21 dec 2001] 218589837Skris 218689837Skris *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 218789837Skris <Dominikus.Scherkl@biodata.com>. (The previous implementation 218889837Skris worked incorrectly for those cases where range = 10..._2 and 218989837Skris 3*range is two bits longer than range.) 219089837Skris [Bodo Moeller] 219189837Skris 219289837Skris *) Only add signing time to PKCS7 structures if it is not already 219389837Skris present. 219489837Skris [Steve Henson] 219589837Skris 219689837Skris *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 219789837Skris OBJ_ld_ce should be OBJ_id_ce. 219889837Skris Also some ip-pda OIDs in crypto/objects/objects.txt were 219989837Skris incorrect (cf. RFC 3039). 220089837Skris [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 220189837Skris 220289837Skris *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 220389837Skris returns early because it has nothing to do. 220489837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 220589837Skris 220689837Skris *) [In 0.9.6c-engine release:] 220789837Skris Fix mutex callback return values in crypto/engine/hw_ncipher.c. 220889837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 220989837Skris 221089837Skris *) [In 0.9.6c-engine release:] 221189837Skris Add support for Cryptographic Appliance's keyserver technology. 221289837Skris (Use engine 'keyclient') 221389837Skris [Cryptographic Appliances and Geoff Thorpe] 221489837Skris 221589837Skris *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 221689837Skris is called via tools/c89.sh because arguments have to be 221789837Skris rearranged (all '-L' options must appear before the first object 221889837Skris modules). 221989837Skris [Richard Shapiro <rshapiro@abinitio.com>] 222089837Skris 222189837Skris *) [In 0.9.6c-engine release:] 222289837Skris Add support for Broadcom crypto accelerator cards, backported 222389837Skris from 0.9.7. 222489837Skris [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 222589837Skris 222689837Skris *) [In 0.9.6c-engine release:] 222789837Skris Add support for SureWare crypto accelerator cards from 222889837Skris Baltimore Technologies. (Use engine 'sureware') 222989837Skris [Baltimore Technologies and Mark Cox] 223089837Skris 223189837Skris *) [In 0.9.6c-engine release:] 223289837Skris Add support for crypto accelerator cards from Accelerated 223389837Skris Encryption Processing, www.aep.ie. (Use engine 'aep') 223489837Skris [AEP Inc. and Mark Cox] 223589837Skris 223689837Skris *) Add a configuration entry for gcc on UnixWare. 223789837Skris [Gary Benson <gbenson@redhat.com>] 223889837Skris 223989837Skris *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 224089837Skris messages are stored in a single piece (fixed-length part and 224189837Skris variable-length part combined) and fix various bugs found on the way. 224289837Skris [Bodo Moeller] 224389837Skris 224489837Skris *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 224589837Skris instead. BIO_gethostbyname() does not know what timeouts are 224689837Skris appropriate, so entries would stay in cache even when they have 224789837Skris become invalid. 224889837Skris [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 224989837Skris 225089837Skris *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 225189837Skris faced with a pathologically small ClientHello fragment that does 225289837Skris not contain client_version: Instead of aborting with an error, 225389837Skris simply choose the highest available protocol version (i.e., 225489837Skris TLS 1.0 unless it is disabled). In practice, ClientHello 225589837Skris messages are never sent like this, but this change gives us 225689837Skris strictly correct behaviour at least for TLS. 225789837Skris [Bodo Moeller] 225889837Skris 225989837Skris *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 226089837Skris never resets s->method to s->ctx->method when called from within 226189837Skris one of the SSL handshake functions. 226289837Skris [Bodo Moeller; problem pointed out by Niko Baric] 226389837Skris 226489837Skris *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 226589837Skris (sent using the client's version number) if client_version is 226689837Skris smaller than the protocol version in use. Also change 226789837Skris ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 226889837Skris the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 226989837Skris the client will at least see that alert. 227089837Skris [Bodo Moeller] 227189837Skris 227289837Skris *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 227389837Skris correctly. 227489837Skris [Bodo Moeller] 227589837Skris 227689837Skris *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 227789837Skris client receives HelloRequest while in a handshake. 227889837Skris [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 227989837Skris 228089837Skris *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 228189837Skris should end in 'break', not 'goto end' which circuments various 228289837Skris cleanups done in state SSL_ST_OK. But session related stuff 228389837Skris must be disabled for SSL_ST_OK in the case that we just sent a 228489837Skris HelloRequest. 228589837Skris 228689837Skris Also avoid some overhead by not calling ssl_init_wbio_buffer() 228789837Skris before just sending a HelloRequest. 228889837Skris [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 228989837Skris 229089837Skris *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 229189837Skris reveal whether illegal block cipher padding was found or a MAC 229289837Skris verification error occured. (Neither SSLerr() codes nor alerts 229389837Skris are directly visible to potential attackers, but the information 229489837Skris may leak via logfiles.) 229589837Skris 229689837Skris Similar changes are not required for the SSL 2.0 implementation 229789837Skris because the number of padding bytes is sent in clear for SSL 2.0, 229889837Skris and the extra bytes are just ignored. However ssl/s2_pkt.c 229989837Skris failed to verify that the purported number of padding bytes is in 230089837Skris the legal range. 230189837Skris [Bodo Moeller] 230289837Skris 230389837Skris *) Add OpenUNIX-8 support including shared libraries 230489837Skris (Boyd Lynn Gerber <gerberb@zenez.com>). 230589837Skris [Lutz Jaenicke] 230689837Skris 230789837Skris *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 230889837Skris 'wristwatch attack' using huge encoding parameters (cf. 230989837Skris James H. Manger's CRYPTO 2001 paper). Note that the 231089837Skris RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 231189837Skris encoding parameters and hence was not vulnerable. 231289837Skris [Bodo Moeller] 231389837Skris 231489837Skris *) BN_sqr() bug fix. 231589837Skris [Ulf M�ller, reported by Jim Ellis <jim.ellis@cavium.com>] 231689837Skris 231789837Skris *) Rabin-Miller test analyses assume uniformly distributed witnesses, 231889837Skris so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 231989837Skris followed by modular reduction. 232089837Skris [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 232189837Skris 232289837Skris *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 232389837Skris equivalent based on BN_pseudo_rand() instead of BN_rand(). 232489837Skris [Bodo Moeller] 232589837Skris 232689837Skris *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 232789837Skris This function was broken, as the check for a new client hello message 232889837Skris to handle SGC did not allow these large messages. 232989837Skris (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 233089837Skris [Lutz Jaenicke] 233189837Skris 233289837Skris *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 233389837Skris [Lutz Jaenicke] 233489837Skris 233589837Skris *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 233689837Skris for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 233789837Skris [Lutz Jaenicke] 233889837Skris 233989837Skris *) Rework the configuration and shared library support for Tru64 Unix. 234089837Skris The configuration part makes use of modern compiler features and 234189837Skris still retains old compiler behavior for those that run older versions 234289837Skris of the OS. The shared library support part includes a variant that 234389837Skris uses the RPATH feature, and is available through the special 234489837Skris configuration target "alpha-cc-rpath", which will never be selected 234589837Skris automatically. 234689837Skris [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 234789837Skris 234889837Skris *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 234989837Skris with the same message size as in ssl3_get_certificate_request(). 235089837Skris Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 235189837Skris messages might inadvertently be reject as too long. 235289837Skris [Petr Lampa <lampa@fee.vutbr.cz>] 235389837Skris 235489837Skris *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 235589837Skris [Andy Polyakov] 235689837Skris 235789837Skris *) Modified SSL library such that the verify_callback that has been set 235889837Skris specificly for an SSL object with SSL_set_verify() is actually being 235989837Skris used. Before the change, a verify_callback set with this function was 236089837Skris ignored and the verify_callback() set in the SSL_CTX at the time of 236189837Skris the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 236289837Skris to allow the necessary settings. 236389837Skris [Lutz Jaenicke] 236489837Skris 236589837Skris *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 236689837Skris explicitly to NULL, as at least on Solaris 8 this seems not always to be 236789837Skris done automatically (in contradiction to the requirements of the C 236889837Skris standard). This made problems when used from OpenSSH. 236989837Skris [Lutz Jaenicke] 237089837Skris 237189837Skris *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 237289837Skris dh->length and always used 237389837Skris 237489837Skris BN_rand_range(priv_key, dh->p). 237589837Skris 237689837Skris BN_rand_range() is not necessary for Diffie-Hellman, and this 237789837Skris specific range makes Diffie-Hellman unnecessarily inefficient if 237889837Skris dh->length (recommended exponent length) is much smaller than the 237989837Skris length of dh->p. We could use BN_rand_range() if the order of 238089837Skris the subgroup was stored in the DH structure, but we only have 238189837Skris dh->length. 238289837Skris 238389837Skris So switch back to 238489837Skris 238589837Skris BN_rand(priv_key, l, ...) 238689837Skris 238789837Skris where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 238889837Skris otherwise. 238989837Skris [Bodo Moeller] 239089837Skris 239189837Skris *) In 239289837Skris 239389837Skris RSA_eay_public_encrypt 239489837Skris RSA_eay_private_decrypt 239589837Skris RSA_eay_private_encrypt (signing) 239689837Skris RSA_eay_public_decrypt (signature verification) 239789837Skris 239889837Skris (default implementations for RSA_public_encrypt, 239989837Skris RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 240089837Skris always reject numbers >= n. 240189837Skris [Bodo Moeller] 240289837Skris 240389837Skris *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 240489837Skris to synchronize access to 'locking_thread'. This is necessary on 240589837Skris systems where access to 'locking_thread' (an 'unsigned long' 240689837Skris variable) is not atomic. 240789837Skris [Bodo Moeller] 240889837Skris 240989837Skris *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 241089837Skris *before* setting the 'crypto_lock_rand' flag. The previous code had 241189837Skris a race condition if 0 is a valid thread ID. 241289837Skris [Travis Vitek <vitek@roguewave.com>] 241389837Skris 241489837Skris *) Add support for shared libraries under Irix. 241589837Skris [Albert Chin-A-Young <china@thewrittenword.com>] 241689837Skris 241789837Skris *) Add configuration option to build on Linux on both big-endian and 241889837Skris little-endian MIPS. 241989837Skris [Ralf Baechle <ralf@uni-koblenz.de>] 242089837Skris 242189837Skris *) Add the possibility to create shared libraries on HP-UX. 242289837Skris [Richard Levitte] 242389837Skris 242479998Skris Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 242579998Skris 242679998Skris *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 242779998Skris to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 242879998Skris Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 242979998Skris PRNG state recovery was possible based on the output of 243079998Skris one PRNG request appropriately sized to gain knowledge on 243179998Skris 'md' followed by enough consecutive 1-byte PRNG requests 243279998Skris to traverse all of 'state'. 243379998Skris 243479998Skris 1. When updating 'md_local' (the current thread's copy of 'md') 243579998Skris during PRNG output generation, hash all of the previous 243679998Skris 'md_local' value, not just the half used for PRNG output. 243779998Skris 243879998Skris 2. Make the number of bytes from 'state' included into the hash 243979998Skris independent from the number of PRNG bytes requested. 244079998Skris 244179998Skris The first measure alone would be sufficient to avoid 244279998Skris Markku-Juhani's attack. (Actually it had never occurred 244379998Skris to me that the half of 'md_local' used for chaining was the 244479998Skris half from which PRNG output bytes were taken -- I had always 244579998Skris assumed that the secret half would be used.) The second 244679998Skris measure makes sure that additional data from 'state' is never 244779998Skris mixed into 'md_local' in small portions; this heuristically 244879998Skris further strengthens the PRNG. 244979998Skris [Bodo Moeller] 245079998Skris 245179998Skris *) Fix crypto/bn/asm/mips3.s. 245279998Skris [Andy Polyakov] 245379998Skris 245479998Skris *) When only the key is given to "enc", the IV is undefined. Print out 245579998Skris an error message in this case. 245679998Skris [Lutz Jaenicke] 245779998Skris 245879998Skris *) Handle special case when X509_NAME is empty in X509 printing routines. 245979998Skris [Steve Henson] 246079998Skris 246179998Skris *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 246279998Skris positive and less than q. 246379998Skris [Bodo Moeller] 246479998Skris 246579998Skris *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 246679998Skris used: it isn't thread safe and the add_lock_callback should handle 246779998Skris that itself. 246879998Skris [Paul Rose <Paul.Rose@bridge.com>] 246979998Skris 247079998Skris *) Verify that incoming data obeys the block size in 247179998Skris ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 247279998Skris [Bodo Moeller] 247379998Skris 247479998Skris *) Fix OAEP check. 247579998Skris [Ulf M�ller, Bodo M�ller] 247679998Skris 247779998Skris *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 247889837Skris RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 247979998Skris when fixing the server behaviour for backwards-compatible 'client 248079998Skris hello' messages. (Note that the attack is impractical against 248179998Skris SSL 3.0 and TLS 1.0 anyway because length and version checking 248279998Skris means that the probability of guessing a valid ciphertext is 248379998Skris around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 248479998Skris paper.) 248579998Skris 248679998Skris Before 0.9.5, the countermeasure (hide the error by generating a 248779998Skris random 'decryption result') did not work properly because 248879998Skris ERR_clear_error() was missing, meaning that SSL_get_error() would 248979998Skris detect the supposedly ignored error. 249079998Skris 249179998Skris Both problems are now fixed. 249279998Skris [Bodo Moeller] 249379998Skris 249479998Skris *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 249579998Skris (previously it was 1024). 249679998Skris [Bodo Moeller] 249779998Skris 249879998Skris *) Fix for compatibility mode trust settings: ignore trust settings 249979998Skris unless some valid trust or reject settings are present. 250079998Skris [Steve Henson] 250179998Skris 250279998Skris *) Fix for blowfish EVP: its a variable length cipher. 250379998Skris [Steve Henson] 250479998Skris 250579998Skris *) Fix various bugs related to DSA S/MIME verification. Handle missing 250679998Skris parameters in DSA public key structures and return an error in the 250779998Skris DSA routines if parameters are absent. 250879998Skris [Steve Henson] 250979998Skris 251079998Skris *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 251179998Skris in the current directory if neither $RANDFILE nor $HOME was set. 251279998Skris RAND_file_name() in 0.9.6a returned NULL in this case. This has 251379998Skris caused some confusion to Windows users who haven't defined $HOME. 251479998Skris Thus RAND_file_name() is changed again: e_os.h can define a 251579998Skris DEFAULT_HOME, which will be used if $HOME is not set. 251679998Skris For Windows, we use "C:"; on other platforms, we still require 251779998Skris environment variables. 251879998Skris 251979998Skris *) Move 'if (!initialized) RAND_poll()' into regions protected by 252079998Skris CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 252179998Skris having multiple threads call RAND_poll() concurrently. 252279998Skris [Bodo Moeller] 252379998Skris 252479998Skris *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 252579998Skris combination of a flag and a thread ID variable. 252679998Skris Otherwise while one thread is in ssleay_rand_bytes (which sets the 252779998Skris flag), *other* threads can enter ssleay_add_bytes without obeying 252889837Skris the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 252979998Skris that they do not hold after the first thread unsets add_do_not_lock). 253079998Skris [Bodo Moeller] 253179998Skris 253279998Skris *) Change bctest again: '-x' expressions are not available in all 253379998Skris versions of 'test'. 253479998Skris [Bodo Moeller] 253579998Skris 253676866Skris Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 253772613Skris 253876866Skris *) Fix a couple of memory leaks in PKCS7_dataDecode() 253976866Skris [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 254076866Skris 254176866Skris *) Change Configure and Makefiles to provide EXE_EXT, which will contain 254276866Skris the default extension for executables, if any. Also, make the perl 254376866Skris scripts that use symlink() to test if it really exists and use "cp" 254476866Skris if it doesn't. All this made OpenSSL compilable and installable in 254576866Skris CygWin. 254676866Skris [Richard Levitte] 254776866Skris 254876866Skris *) Fix for asn1_GetSequence() for indefinite length constructed data. 254976866Skris If SEQUENCE is length is indefinite just set c->slen to the total 255076866Skris amount of data available. 255176866Skris [Steve Henson, reported by shige@FreeBSD.org] 255276866Skris [This change does not apply to 0.9.7.] 255376866Skris 255476866Skris *) Change bctest to avoid here-documents inside command substitution 255576866Skris (workaround for FreeBSD /bin/sh bug). 255676866Skris For compatibility with Ultrix, avoid shell functions (introduced 255776866Skris in the bctest version that searches along $PATH). 255876866Skris [Bodo Moeller] 255976866Skris 256076866Skris *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 256176866Skris with des_encrypt() defined on some operating systems, like Solaris 256276866Skris and UnixWare. 256376866Skris [Richard Levitte] 256476866Skris 256576866Skris *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 256676866Skris On the Importance of Eliminating Errors in Cryptographic 256776866Skris Computations, J. Cryptology 14 (2001) 2, 101-119, 256876866Skris http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 256976866Skris [Ulf Moeller] 257076866Skris 257176866Skris *) MIPS assembler BIGNUM division bug fix. 257276866Skris [Andy Polyakov] 257376866Skris 257476866Skris *) Disabled incorrect Alpha assembler code. 257576866Skris [Richard Levitte] 257676866Skris 257776866Skris *) Fix PKCS#7 decode routines so they correctly update the length 257876866Skris after reading an EOC for the EXPLICIT tag. 257976866Skris [Steve Henson] 258076866Skris [This change does not apply to 0.9.7.] 258176866Skris 258276866Skris *) Fix bug in PKCS#12 key generation routines. This was triggered 258376866Skris if a 3DES key was generated with a 0 initial byte. Include 258476866Skris PKCS12_BROKEN_KEYGEN compilation option to retain the old 258576866Skris (but broken) behaviour. 258676866Skris [Steve Henson] 258776866Skris 258876866Skris *) Enhance bctest to search for a working bc along $PATH and print 258976866Skris it when found. 259076866Skris [Tim Rice <tim@multitalents.net> via Richard Levitte] 259176866Skris 259276866Skris *) Fix memory leaks in err.c: free err_data string if necessary; 259376866Skris don't write to the wrong index in ERR_set_error_data. 259476866Skris [Bodo Moeller] 259576866Skris 259676866Skris *) Implement ssl23_peek (analogous to ssl23_read), which previously 259776866Skris did not exist. 259876866Skris [Bodo Moeller] 259976866Skris 260076866Skris *) Replace rdtsc with _emit statements for VC++ version 5. 260176866Skris [Jeremy Cooper <jeremy@baymoo.org>] 260276866Skris 260376866Skris *) Make it possible to reuse SSLv2 sessions. 260476866Skris [Richard Levitte] 260576866Skris 260676866Skris *) In copy_email() check for >= 0 as a return value for 260776866Skris X509_NAME_get_index_by_NID() since 0 is a valid index. 260876866Skris [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 260976866Skris 261076866Skris *) Avoid coredump with unsupported or invalid public keys by checking if 261176866Skris X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 261276866Skris PKCS7_verify() fails with non detached data. 261376866Skris [Steve Henson] 261476866Skris 261576866Skris *) Don't use getenv in library functions when run as setuid/setgid. 261676866Skris New function OPENSSL_issetugid(). 261776866Skris [Ulf Moeller] 261876866Skris 261976866Skris *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 262076866Skris due to incorrect handling of multi-threading: 262176866Skris 262276866Skris 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 262376866Skris 262476866Skris 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 262576866Skris 262676866Skris 3. Count how many times MemCheck_off() has been called so that 262776866Skris nested use can be treated correctly. This also avoids 262876866Skris inband-signalling in the previous code (which relied on the 262976866Skris assumption that thread ID 0 is impossible). 263076866Skris [Bodo Moeller] 263176866Skris 263276866Skris *) Add "-rand" option also to s_client and s_server. 263376866Skris [Lutz Jaenicke] 263476866Skris 263576866Skris *) Fix CPU detection on Irix 6.x. 263676866Skris [Kurt Hockenbury <khockenb@stevens-tech.edu> and 263776866Skris "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 263876866Skris 263976866Skris *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 264076866Skris was empty. 264176866Skris [Steve Henson] 264276866Skris [This change does not apply to 0.9.7.] 264376866Skris 264476866Skris *) Use the cached encoding of an X509_NAME structure rather than 264576866Skris copying it. This is apparently the reason for the libsafe "errors" 264676866Skris but the code is actually correct. 264776866Skris [Steve Henson] 264876866Skris 264972613Skris *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 265072613Skris Bleichenbacher's DSA attack. 265176866Skris Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 265276866Skris to be set and top=0 forces the highest bit to be set; top=-1 is new 265376866Skris and leaves the highest bit random. 265476866Skris [Ulf Moeller, Bodo Moeller] 265572613Skris 265672613Skris *) In the NCONF_...-based implementations for CONF_... queries 265772613Skris (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 265872613Skris a temporary CONF structure with the data component set to NULL 265972613Skris (which gives segmentation faults in lh_retrieve). 266072613Skris Instead, use NULL for the CONF pointer in CONF_get_string and 266172613Skris CONF_get_number (which may use environment variables) and directly 266272613Skris return NULL from CONF_get_section. 266372613Skris [Bodo Moeller] 266472613Skris 266572613Skris *) Fix potential buffer overrun for EBCDIC. 266672613Skris [Ulf Moeller] 266772613Skris 266872613Skris *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 266972613Skris keyUsage if basicConstraints absent for a CA. 267072613Skris [Steve Henson] 267172613Skris 267272613Skris *) Make SMIME_write_PKCS7() write mail header values with a format that 267372613Skris is more generally accepted (no spaces before the semicolon), since 267472613Skris some programs can't parse those values properly otherwise. Also make 267572613Skris sure BIO's that break lines after each write do not create invalid 267672613Skris headers. 267772613Skris [Richard Levitte] 267872613Skris 267972613Skris *) Make the CRL encoding routines work with empty SEQUENCE OF. The 268072613Skris macros previously used would not encode an empty SEQUENCE OF 268172613Skris and break the signature. 268272613Skris [Steve Henson] 268376866Skris [This change does not apply to 0.9.7.] 268472613Skris 268572613Skris *) Zero the premaster secret after deriving the master secret in 268672613Skris DH ciphersuites. 268772613Skris [Steve Henson] 268872613Skris 268972613Skris *) Add some EVP_add_digest_alias registrations (as found in 269072613Skris OpenSSL_add_all_digests()) to SSL_library_init() 269172613Skris aka OpenSSL_add_ssl_algorithms(). This provides improved 269272613Skris compatibility with peers using X.509 certificates 269372613Skris with unconventional AlgorithmIdentifier OIDs. 269472613Skris [Bodo Moeller] 269572613Skris 269672613Skris *) Fix for Irix with NO_ASM. 269772613Skris ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 269872613Skris 269972613Skris *) ./config script fixes. 270072613Skris [Ulf Moeller, Richard Levitte] 270172613Skris 270272613Skris *) Fix 'openssl passwd -1'. 270372613Skris [Bodo Moeller] 270472613Skris 270572613Skris *) Change PKCS12_key_gen_asc() so it can cope with non null 270672613Skris terminated strings whose length is passed in the passlen 270772613Skris parameter, for example from PEM callbacks. This was done 270872613Skris by adding an extra length parameter to asc2uni(). 270972613Skris [Steve Henson, reported by <oddissey@samsung.co.kr>] 271072613Skris 271172613Skris *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 271272613Skris call failed, free the DSA structure. 271372613Skris [Bodo Moeller] 271472613Skris 271572613Skris *) Fix to uni2asc() to cope with zero length Unicode strings. 271672613Skris These are present in some PKCS#12 files. 271772613Skris [Steve Henson] 271872613Skris 271972613Skris *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 272072613Skris Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 272172613Skris when writing a 32767 byte record. 272272613Skris [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 272372613Skris 272472613Skris *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 272572613Skris obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 272672613Skris 272772613Skris (RSA objects have a reference count access to which is protected 272872613Skris by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 272972613Skris so they are meant to be shared between threads.) 273072613Skris [Bodo Moeller, Geoff Thorpe; original patch submitted by 273172613Skris "Reddie, Steven" <Steven.Reddie@ca.com>] 273272613Skris 273372613Skris *) Fix a deadlock in CRYPTO_mem_leaks(). 273472613Skris [Bodo Moeller] 273572613Skris 273676866Skris *) Use better test patterns in bntest. 273776866Skris [Ulf M�ller] 273876866Skris 273972613Skris *) rand_win.c fix for Borland C. 274072613Skris [Ulf M�ller] 274172613Skris 274272613Skris *) BN_rshift bugfix for n == 0. 274372613Skris [Bodo Moeller] 274472613Skris 274576866Skris *) Add a 'bctest' script that checks for some known 'bc' bugs 274676866Skris so that 'make test' does not abort just because 'bc' is broken. 274776866Skris [Bodo Moeller] 274876866Skris 274972613Skris *) Store verify_result within SSL_SESSION also for client side to 275072613Skris avoid potential security hole. (Re-used sessions on the client side 275172613Skris always resulted in verify_result==X509_V_OK, not using the original 275272613Skris result of the server certificate verification.) 275372613Skris [Lutz Jaenicke] 275472613Skris 275572613Skris *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 275672613Skris SSL3_RT_APPLICATION_DATA, return 0. 275772613Skris Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 275872613Skris [Bodo Moeller] 275972613Skris 276072613Skris *) Fix SSL_peek: 276172613Skris Both ssl2_peek and ssl3_peek, which were totally broken in earlier 276272613Skris releases, have been re-implemented by renaming the previous 276372613Skris implementations of ssl2_read and ssl3_read to ssl2_read_internal 276472613Skris and ssl3_read_internal, respectively, and adding 'peek' parameters 276572613Skris to them. The new ssl[23]_{read,peek} functions are calls to 276672613Skris ssl[23]_read_internal with the 'peek' flag set appropriately. 276772613Skris A 'peek' parameter has also been added to ssl3_read_bytes, which 276872613Skris does the actual work for ssl3_read_internal. 276972613Skris [Bodo Moeller] 277072613Skris 277176866Skris *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 277276866Skris the method-specific "init()" handler. Also clean up ex_data after 277376866Skris calling the method-specific "finish()" handler. Previously, this was 277476866Skris happening the other way round. 277576866Skris [Geoff Thorpe] 277676866Skris 277772613Skris *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 277872613Skris The previous value, 12, was not always sufficient for BN_mod_exp(). 277972613Skris [Bodo Moeller] 278072613Skris 278176866Skris *) Make sure that shared libraries get the internal name engine with 278276866Skris the full version number and not just 0. This should mark the 278376866Skris shared libraries as not backward compatible. Of course, this should 278476866Skris be changed again when we can guarantee backward binary compatibility. 278576866Skris [Richard Levitte] 278676866Skris 278772613Skris *) Fix typo in get_cert_by_subject() in by_dir.c 278872613Skris [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 278972613Skris 279076866Skris *) Rework the system to generate shared libraries: 279176866Skris 279276866Skris - Make note of the expected extension for the shared libraries and 279376866Skris if there is a need for symbolic links from for example libcrypto.so.0 279476866Skris to libcrypto.so.0.9.7. There is extended info in Configure for 279576866Skris that. 279676866Skris 279776866Skris - Make as few rebuilds of the shared libraries as possible. 279876866Skris 279976866Skris - Still avoid linking the OpenSSL programs with the shared libraries. 280076866Skris 280176866Skris - When installing, install the shared libraries separately from the 280276866Skris static ones. 280376866Skris [Richard Levitte] 280476866Skris 280572613Skris *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 280672613Skris 280772613Skris Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 280872613Skris and not in SSL_clear because the latter is also used by the 280972613Skris accept/connect functions; previously, the settings made by 281072613Skris SSL_set_read_ahead would be lost during the handshake. 281172613Skris [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 281272613Skris 281372613Skris *) Correct util/mkdef.pl to be selective about disabled algorithms. 281472613Skris Previously, it would create entries for disableed algorithms no 281572613Skris matter what. 281672613Skris [Richard Levitte] 281772613Skris 281872613Skris *) Added several new manual pages for SSL_* function. 281972613Skris [Lutz Jaenicke] 282072613Skris 282168651Skris Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 282268651Skris 282368651Skris *) In ssl23_get_client_hello, generate an error message when faced 282468651Skris with an initial SSL 3.0/TLS record that is too small to contain the 282568651Skris first two bytes of the ClientHello message, i.e. client_version. 282668651Skris (Note that this is a pathologic case that probably has never happened 282768651Skris in real life.) The previous approach was to use the version number 282868651Skris from the record header as a substitute; but our protocol choice 282968651Skris should not depend on that one because it is not authenticated 283068651Skris by the Finished messages. 283168651Skris [Bodo Moeller] 283268651Skris 283368651Skris *) More robust randomness gathering functions for Windows. 283468651Skris [Jeffrey Altman <jaltman@columbia.edu>] 283568651Skris 283668651Skris *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 283768651Skris not set then we don't setup the error code for issuer check errors 283868651Skris to avoid possibly overwriting other errors which the callback does 283968651Skris handle. If an application does set the flag then we assume it knows 284068651Skris what it is doing and can handle the new informational codes 284168651Skris appropriately. 284268651Skris [Steve Henson] 284368651Skris 284468651Skris *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 284568651Skris a general "ANY" type, as such it should be able to decode anything 284668651Skris including tagged types. However it didn't check the class so it would 284768651Skris wrongly interpret tagged types in the same way as their universal 284868651Skris counterpart and unknown types were just rejected. Changed so that the 284968651Skris tagged and unknown types are handled in the same way as a SEQUENCE: 285068651Skris that is the encoding is stored intact. There is also a new type 285168651Skris "V_ASN1_OTHER" which is used when the class is not universal, in this 285268651Skris case we have no idea what the actual type is so we just lump them all 285368651Skris together. 285468651Skris [Steve Henson] 285568651Skris 285668651Skris *) On VMS, stdout may very well lead to a file that is written to 285768651Skris in a record-oriented fashion. That means that every write() will 285868651Skris write a separate record, which will be read separately by the 285968651Skris programs trying to read from it. This can be very confusing. 286068651Skris 286168651Skris The solution is to put a BIO filter in the way that will buffer 286268651Skris text until a linefeed is reached, and then write everything a 286368651Skris line at a time, so every record written will be an actual line, 286468651Skris not chunks of lines and not (usually doesn't happen, but I've 286568651Skris seen it once) several lines in one record. BIO_f_linebuffer() is 286668651Skris the answer. 286768651Skris 286868651Skris Currently, it's a VMS-only method, because that's where it has 286968651Skris been tested well enough. 287068651Skris [Richard Levitte] 287168651Skris 287268651Skris *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 287368651Skris it can return incorrect results. 287468651Skris (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 287568651Skris but it was in 0.9.6-beta[12].) 287668651Skris [Bodo Moeller] 287768651Skris 287868651Skris *) Disable the check for content being present when verifying detached 287968651Skris signatures in pk7_smime.c. Some versions of Netscape (wrongly) 288068651Skris include zero length content when signing messages. 288168651Skris [Steve Henson] 288268651Skris 288368651Skris *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 288468651Skris BIO_ctrl (for BIO pairs). 288568651Skris [Bodo M�ller] 288668651Skris 288768651Skris *) Add DSO method for VMS. 288868651Skris [Richard Levitte] 288968651Skris 289068651Skris *) Bug fix: Montgomery multiplication could produce results with the 289168651Skris wrong sign. 289268651Skris [Ulf M�ller] 289368651Skris 289468651Skris *) Add RPM specification openssl.spec and modify it to build three 289568651Skris packages. The default package contains applications, application 289668651Skris documentation and run-time libraries. The devel package contains 289768651Skris include files, static libraries and function documentation. The 289868651Skris doc package contains the contents of the doc directory. The original 289968651Skris openssl.spec was provided by Damien Miller <djm@mindrot.org>. 290068651Skris [Richard Levitte] 290168651Skris 290268651Skris *) Add a large number of documentation files for many SSL routines. 290368651Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 290468651Skris 290568651Skris *) Add a configuration entry for Sony News 4. 290668651Skris [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 290768651Skris 290868651Skris *) Don't set the two most significant bits to one when generating a 290968651Skris random number < q in the DSA library. 291068651Skris [Ulf M�ller] 291168651Skris 291268651Skris *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 291368651Skris behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 291468651Skris the underlying transport is blocking) if a handshake took place. 291568651Skris (The default behaviour is needed by applications such as s_client 291668651Skris and s_server that use select() to determine when to use SSL_read; 291768651Skris but for applications that know in advance when to expect data, it 291868651Skris just makes things more complicated.) 291968651Skris [Bodo Moeller] 292068651Skris 292168651Skris *) Add RAND_egd_bytes(), which gives control over the number of bytes read 292268651Skris from EGD. 292368651Skris [Ben Laurie] 292468651Skris 292568651Skris *) Add a few more EBCDIC conditionals that make `req' and `x509' 292668651Skris work better on such systems. 292768651Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 292868651Skris 292968651Skris *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 293068651Skris Update PKCS12_parse() so it copies the friendlyName and the 293168651Skris keyid to the certificates aux info. 293268651Skris [Steve Henson] 293368651Skris 293468651Skris *) Fix bug in PKCS7_verify() which caused an infinite loop 293568651Skris if there was more than one signature. 293668651Skris [Sven Uszpelkat <su@celocom.de>] 293768651Skris 293868651Skris *) Major change in util/mkdef.pl to include extra information 293968651Skris about each symbol, as well as presentig variables as well 294068651Skris as functions. This change means that there's n more need 294168651Skris to rebuild the .num files when some algorithms are excluded. 294268651Skris [Richard Levitte] 294368651Skris 294468651Skris *) Allow the verify time to be set by an application, 294568651Skris rather than always using the current time. 294668651Skris [Steve Henson] 294768651Skris 294868651Skris *) Phase 2 verify code reorganisation. The certificate 294968651Skris verify code now looks up an issuer certificate by a 295068651Skris number of criteria: subject name, authority key id 295168651Skris and key usage. It also verifies self signed certificates 295268651Skris by the same criteria. The main comparison function is 295368651Skris X509_check_issued() which performs these checks. 295468651Skris 295568651Skris Lot of changes were necessary in order to support this 295668651Skris without completely rewriting the lookup code. 295768651Skris 295868651Skris Authority and subject key identifier are now cached. 295968651Skris 296068651Skris The LHASH 'certs' is X509_STORE has now been replaced 296168651Skris by a STACK_OF(X509_OBJECT). This is mainly because an 296268651Skris LHASH can't store or retrieve multiple objects with 296368651Skris the same hash value. 296468651Skris 296568651Skris As a result various functions (which were all internal 296668651Skris use only) have changed to handle the new X509_STORE 296768651Skris structure. This will break anything that messed round 296868651Skris with X509_STORE internally. 296968651Skris 297068651Skris The functions X509_STORE_add_cert() now checks for an 297168651Skris exact match, rather than just subject name. 297268651Skris 297368651Skris The X509_STORE API doesn't directly support the retrieval 297468651Skris of multiple certificates matching a given criteria, however 297568651Skris this can be worked round by performing a lookup first 297668651Skris (which will fill the cache with candidate certificates) 297768651Skris and then examining the cache for matches. This is probably 297868651Skris the best we can do without throwing out X509_LOOKUP 297968651Skris entirely (maybe later...). 298068651Skris 298168651Skris The X509_VERIFY_CTX structure has been enhanced considerably. 298268651Skris 298368651Skris All certificate lookup operations now go via a get_issuer() 298468651Skris callback. Although this currently uses an X509_STORE it 298568651Skris can be replaced by custom lookups. This is a simple way 298668651Skris to bypass the X509_STORE hackery necessary to make this 298768651Skris work and makes it possible to use more efficient techniques 298868651Skris in future. A very simple version which uses a simple 298968651Skris STACK for its trusted certificate store is also provided 299068651Skris using X509_STORE_CTX_trusted_stack(). 299168651Skris 299268651Skris The verify_cb() and verify() callbacks now have equivalents 299368651Skris in the X509_STORE_CTX structure. 299468651Skris 299568651Skris X509_STORE_CTX also has a 'flags' field which can be used 299668651Skris to customise the verify behaviour. 299768651Skris [Steve Henson] 299868651Skris 299968651Skris *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 300068651Skris excludes S/MIME capabilities. 300168651Skris [Steve Henson] 300268651Skris 300368651Skris *) When a certificate request is read in keep a copy of the 300468651Skris original encoding of the signed data and use it when outputing 300568651Skris again. Signatures then use the original encoding rather than 300668651Skris a decoded, encoded version which may cause problems if the 300768651Skris request is improperly encoded. 300868651Skris [Steve Henson] 300968651Skris 301068651Skris *) For consistency with other BIO_puts implementations, call 301168651Skris buffer_write(b, ...) directly in buffer_puts instead of calling 301268651Skris BIO_write(b, ...). 301368651Skris 301468651Skris In BIO_puts, increment b->num_write as in BIO_write. 301568651Skris [Peter.Sylvester@EdelWeb.fr] 301668651Skris 301768651Skris *) Fix BN_mul_word for the case where the word is 0. (We have to use 301868651Skris BN_zero, we may not return a BIGNUM with an array consisting of 301968651Skris words set to zero.) 302068651Skris [Bodo Moeller] 302168651Skris 302268651Skris *) Avoid calling abort() from within the library when problems are 302368651Skris detected, except if preprocessor symbols have been defined 302468651Skris (such as REF_CHECK, BN_DEBUG etc.). 302568651Skris [Bodo Moeller] 302668651Skris 302768651Skris *) New openssl application 'rsautl'. This utility can be 302868651Skris used for low level RSA operations. DER public key 302968651Skris BIO/fp routines also added. 303068651Skris [Steve Henson] 303168651Skris 303268651Skris *) New Configure entry and patches for compiling on QNX 4. 303368651Skris [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 303468651Skris 303568651Skris *) A demo state-machine implementation was sponsored by 303668651Skris Nuron (http://www.nuron.com/) and is now available in 303768651Skris demos/state_machine. 303868651Skris [Ben Laurie] 303968651Skris 304068651Skris *) New options added to the 'dgst' utility for signature 304168651Skris generation and verification. 304268651Skris [Steve Henson] 304368651Skris 304468651Skris *) Unrecognized PKCS#7 content types are now handled via a 304568651Skris catch all ASN1_TYPE structure. This allows unsupported 304668651Skris types to be stored as a "blob" and an application can 304768651Skris encode and decode it manually. 304868651Skris [Steve Henson] 304968651Skris 305068651Skris *) Fix various signed/unsigned issues to make a_strex.c 305168651Skris compile under VC++. 305268651Skris [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 305368651Skris 305468651Skris *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 305568651Skris length if passed a buffer. ASN1_INTEGER_to_BN failed 305668651Skris if passed a NULL BN and its argument was negative. 305768651Skris [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 305868651Skris 305968651Skris *) Modification to PKCS#7 encoding routines to output definite 306068651Skris length encoding. Since currently the whole structures are in 306168651Skris memory there's not real point in using indefinite length 306268651Skris constructed encoding. However if OpenSSL is compiled with 306368651Skris the flag PKCS7_INDEFINITE_ENCODING the old form is used. 306468651Skris [Steve Henson] 306568651Skris 306668651Skris *) Added BIO_vprintf() and BIO_vsnprintf(). 306768651Skris [Richard Levitte] 306868651Skris 306968651Skris *) Added more prefixes to parse for in the the strings written 307068651Skris through a logging bio, to cover all the levels that are available 307168651Skris through syslog. The prefixes are now: 307268651Skris 307368651Skris PANIC, EMERG, EMR => LOG_EMERG 307468651Skris ALERT, ALR => LOG_ALERT 307568651Skris CRIT, CRI => LOG_CRIT 307668651Skris ERROR, ERR => LOG_ERR 307768651Skris WARNING, WARN, WAR => LOG_WARNING 307868651Skris NOTICE, NOTE, NOT => LOG_NOTICE 307968651Skris INFO, INF => LOG_INFO 308068651Skris DEBUG, DBG => LOG_DEBUG 308168651Skris 308268651Skris and as before, if none of those prefixes are present at the 308368651Skris beginning of the string, LOG_ERR is chosen. 308468651Skris 308568651Skris On Win32, the LOG_* levels are mapped according to this: 308668651Skris 308768651Skris LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 308868651Skris LOG_WARNING => EVENTLOG_WARNING_TYPE 308968651Skris LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 309068651Skris 309168651Skris [Richard Levitte] 309268651Skris 309368651Skris *) Made it possible to reconfigure with just the configuration 309468651Skris argument "reconf" or "reconfigure". The command line arguments 309568651Skris are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 309668651Skris and are retrieved from there when reconfiguring. 309768651Skris [Richard Levitte] 309868651Skris 309968651Skris *) MD4 implemented. 310068651Skris [Assar Westerlund <assar@sics.se>, Richard Levitte] 310168651Skris 310268651Skris *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 310368651Skris [Richard Levitte] 310468651Skris 310568651Skris *) The obj_dat.pl script was messing up the sorting of object 310668651Skris names. The reason was that it compared the quoted version 310768651Skris of strings as a result "OCSP" > "OCSP Signing" because 310868651Skris " > SPACE. Changed script to store unquoted versions of 310968651Skris names and add quotes on output. It was also omitting some 311068651Skris names from the lookup table if they were given a default 311168651Skris value (that is if SN is missing it is given the same 311268651Skris value as LN and vice versa), these are now added on the 311368651Skris grounds that if an object has a name we should be able to 311468651Skris look it up. Finally added warning output when duplicate 311568651Skris short or long names are found. 311668651Skris [Steve Henson] 311768651Skris 311868651Skris *) Changes needed for Tandem NSK. 311968651Skris [Scott Uroff <scott@xypro.com>] 312068651Skris 312168651Skris *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 312268651Skris RSA_padding_check_SSLv23(), special padding was never detected 312368651Skris and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 312468651Skris version rollback attacks was not effective. 312568651Skris 312668651Skris In s23_clnt.c, don't use special rollback-attack detection padding 312768651Skris (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 312868651Skris client; similarly, in s23_srvr.c, don't do the rollback check if 312968651Skris SSL 2.0 is the only protocol enabled in the server. 313068651Skris [Bodo Moeller] 313168651Skris 313268651Skris *) Make it possible to get hexdumps of unprintable data with 'openssl 313368651Skris asn1parse'. By implication, the functions ASN1_parse_dump() and 313468651Skris BIO_dump_indent() are added. 313568651Skris [Richard Levitte] 313668651Skris 313768651Skris *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 313868651Skris these print out strings and name structures based on various 313968651Skris flags including RFC2253 support and proper handling of 314068651Skris multibyte characters. Added options to the 'x509' utility 314168651Skris to allow the various flags to be set. 314268651Skris [Steve Henson] 314368651Skris 314468651Skris *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 314568651Skris Also change the functions X509_cmp_current_time() and 314668651Skris X509_gmtime_adj() work with an ASN1_TIME structure, 314768651Skris this will enable certificates using GeneralizedTime in validity 314868651Skris dates to be checked. 314968651Skris [Steve Henson] 315068651Skris 315168651Skris *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 315268651Skris negative public key encodings) on by default, 315368651Skris NO_NEG_PUBKEY_BUG can be set to disable it. 315468651Skris [Steve Henson] 315568651Skris 315668651Skris *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 315768651Skris content octets. An i2c_ASN1_OBJECT is unnecessary because 315868651Skris the encoding can be trivially obtained from the structure. 315968651Skris [Steve Henson] 316068651Skris 316168651Skris *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 316268651Skris not read locks (CRYPTO_r_[un]lock). 316368651Skris [Bodo Moeller] 316468651Skris 316568651Skris *) A first attempt at creating official support for shared 316668651Skris libraries through configuration. I've kept it so the 316768651Skris default is static libraries only, and the OpenSSL programs 316868651Skris are always statically linked for now, but there are 316968651Skris preparations for dynamic linking in place. 317089837Skris This has been tested on Linux and Tru64. 317168651Skris [Richard Levitte] 317268651Skris 317368651Skris *) Randomness polling function for Win9x, as described in: 317468651Skris Peter Gutmann, Software Generation of Practically Strong 317568651Skris Random Numbers. 317668651Skris [Ulf M�ller] 317768651Skris 317868651Skris *) Fix so PRNG is seeded in req if using an already existing 317968651Skris DSA key. 318068651Skris [Steve Henson] 318168651Skris 318268651Skris *) New options to smime application. -inform and -outform 318368651Skris allow alternative formats for the S/MIME message including 318468651Skris PEM and DER. The -content option allows the content to be 318568651Skris specified separately. This should allow things like Netscape 318668651Skris form signing output easier to verify. 318768651Skris [Steve Henson] 318868651Skris 318968651Skris *) Fix the ASN1 encoding of tags using the 'long form'. 319068651Skris [Steve Henson] 319168651Skris 319268651Skris *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 319368651Skris STRING types. These convert content octets to and from the 319468651Skris underlying type. The actual tag and length octets are 319568651Skris already assumed to have been read in and checked. These 319668651Skris are needed because all other string types have virtually 319768651Skris identical handling apart from the tag. By having versions 319868651Skris of the ASN1 functions that just operate on content octets 319968651Skris IMPLICIT tagging can be handled properly. It also allows 320068651Skris the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 320168651Skris and ASN1_INTEGER are identical apart from the tag. 320268651Skris [Steve Henson] 320368651Skris 320468651Skris *) Change the handling of OID objects as follows: 320568651Skris 320668651Skris - New object identifiers are inserted in objects.txt, following 320768651Skris the syntax given in objects.README. 320868651Skris - objects.pl is used to process obj_mac.num and create a new 320968651Skris obj_mac.h. 321068651Skris - obj_dat.pl is used to create a new obj_dat.h, using the data in 321168651Skris obj_mac.h. 321268651Skris 321368651Skris This is currently kind of a hack, and the perl code in objects.pl 321468651Skris isn't very elegant, but it works as I intended. The simplest way 321568651Skris to check that it worked correctly is to look in obj_dat.h and 321668651Skris check the array nid_objs and make sure the objects haven't moved 321768651Skris around (this is important!). Additions are OK, as well as 321868651Skris consistent name changes. 321968651Skris [Richard Levitte] 322068651Skris 322168651Skris *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 322268651Skris [Bodo Moeller] 322368651Skris 322468651Skris *) Addition of the command line parameter '-rand file' to 'openssl req'. 322568651Skris The given file adds to whatever has already been seeded into the 322668651Skris random pool through the RANDFILE configuration file option or 322768651Skris environment variable, or the default random state file. 322868651Skris [Richard Levitte] 322968651Skris 323068651Skris *) mkstack.pl now sorts each macro group into lexical order. 323168651Skris Previously the output order depended on the order the files 323268651Skris appeared in the directory, resulting in needless rewriting 323368651Skris of safestack.h . 323468651Skris [Steve Henson] 323568651Skris 323668651Skris *) Patches to make OpenSSL compile under Win32 again. Mostly 323768651Skris work arounds for the VC++ problem that it treats func() as 323868651Skris func(void). Also stripped out the parts of mkdef.pl that 323968651Skris added extra typesafe functions: these no longer exist. 324068651Skris [Steve Henson] 324168651Skris 324268651Skris *) Reorganisation of the stack code. The macros are now all 324368651Skris collected in safestack.h . Each macro is defined in terms of 324468651Skris a "stack macro" of the form SKM_<name>(type, a, b). The 324568651Skris DEBUG_SAFESTACK is now handled in terms of function casts, 324668651Skris this has the advantage of retaining type safety without the 324768651Skris use of additional functions. If DEBUG_SAFESTACK is not defined 324868651Skris then the non typesafe macros are used instead. Also modified the 324968651Skris mkstack.pl script to handle the new form. Needs testing to see 325068651Skris if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 325168651Skris the default if no major problems. Similar behaviour for ASN1_SET_OF 325268651Skris and PKCS12_STACK_OF. 325368651Skris [Steve Henson] 325468651Skris 325568651Skris *) When some versions of IIS use the 'NET' form of private key the 325668651Skris key derivation algorithm is different. Normally MD5(password) is 325768651Skris used as a 128 bit RC4 key. In the modified case 325868651Skris MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 325968651Skris new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 326068651Skris as the old Netscape_RSA functions except they have an additional 326168651Skris 'sgckey' parameter which uses the modified algorithm. Also added 326268651Skris an -sgckey command line option to the rsa utility. Thanks to 326368651Skris Adrian Peck <bertie@ncipher.com> for posting details of the modified 326468651Skris algorithm to openssl-dev. 326568651Skris [Steve Henson] 326668651Skris 326768651Skris *) The evp_local.h macros were using 'c.##kname' which resulted in 326868651Skris invalid expansion on some systems (SCO 5.0.5 for example). 326968651Skris Corrected to 'c.kname'. 327068651Skris [Phillip Porch <root@theporch.com>] 327168651Skris 327268651Skris *) New X509_get1_email() and X509_REQ_get1_email() functions that return 327368651Skris a STACK of email addresses from a certificate or request, these look 327468651Skris in the subject name and the subject alternative name extensions and 327568651Skris omit any duplicate addresses. 327668651Skris [Steve Henson] 327768651Skris 327868651Skris *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 327968651Skris This makes DSA verification about 2 % faster. 328068651Skris [Bodo Moeller] 328168651Skris 328268651Skris *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 328368651Skris (meaning that now 2^5 values will be precomputed, which is only 4 KB 328468651Skris plus overhead for 1024 bit moduli). 328568651Skris This makes exponentiations about 0.5 % faster for 1024 bit 328668651Skris exponents (as measured by "openssl speed rsa2048"). 328768651Skris [Bodo Moeller] 328868651Skris 328968651Skris *) Rename memory handling macros to avoid conflicts with other 329068651Skris software: 329168651Skris Malloc => OPENSSL_malloc 329268651Skris Malloc_locked => OPENSSL_malloc_locked 329368651Skris Realloc => OPENSSL_realloc 329468651Skris Free => OPENSSL_free 329568651Skris [Richard Levitte] 329668651Skris 329768651Skris *) New function BN_mod_exp_mont_word for small bases (roughly 15% 329868651Skris faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 329968651Skris [Bodo Moeller] 330068651Skris 330168651Skris *) CygWin32 support. 330268651Skris [John Jarvie <jjarvie@newsguy.com>] 330368651Skris 330468651Skris *) The type-safe stack code has been rejigged. It is now only compiled 330568651Skris in when OpenSSL is configured with the DEBUG_SAFESTACK option and 330668651Skris by default all type-specific stack functions are "#define"d back to 330768651Skris standard stack functions. This results in more streamlined output 330868651Skris but retains the type-safety checking possibilities of the original 330968651Skris approach. 331068651Skris [Geoff Thorpe] 331168651Skris 331268651Skris *) The STACK code has been cleaned up, and certain type declarations 331368651Skris that didn't make a lot of sense have been brought in line. This has 331468651Skris also involved a cleanup of sorts in safestack.h to more correctly 331568651Skris map type-safe stack functions onto their plain stack counterparts. 331668651Skris This work has also resulted in a variety of "const"ifications of 331768651Skris lots of the code, especially "_cmp" operations which should normally 331868651Skris be prototyped with "const" parameters anyway. 331968651Skris [Geoff Thorpe] 332068651Skris 332168651Skris *) When generating bytes for the first time in md_rand.c, 'stir the pool' 332268651Skris by seeding with STATE_SIZE dummy bytes (with zero entropy count). 332368651Skris (The PRNG state consists of two parts, the large pool 'state' and 'md', 332468651Skris where all of 'md' is used each time the PRNG is used, but 'state' 332568651Skris is used only indexed by a cyclic counter. As entropy may not be 332668651Skris well distributed from the beginning, 'md' is important as a 332768651Skris chaining variable. However, the output function chains only half 332868651Skris of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 332968651Skris all of 'md', and seeding with STATE_SIZE dummy bytes will result 333068651Skris in all of 'state' being rewritten, with the new values depending 333168651Skris on virtually all of 'md'. This overcomes the 80 bit limitation.) 333268651Skris [Bodo Moeller] 333368651Skris 333468651Skris *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 333568651Skris the handshake is continued after ssl_verify_cert_chain(); 333668651Skris otherwise, if SSL_VERIFY_NONE is set, remaining error codes 333768651Skris can lead to 'unexplainable' connection aborts later. 333868651Skris [Bodo Moeller; problem tracked down by Lutz Jaenicke] 333968651Skris 334068651Skris *) Major EVP API cipher revision. 334168651Skris Add hooks for extra EVP features. This allows various cipher 334268651Skris parameters to be set in the EVP interface. Support added for variable 334368651Skris key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 334468651Skris setting of RC2 and RC5 parameters. 334568651Skris 334668651Skris Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 334768651Skris ciphers. 334868651Skris 334968651Skris Remove lots of duplicated code from the EVP library. For example *every* 335068651Skris cipher init() function handles the 'iv' in the same way according to the 335168651Skris cipher mode. They also all do nothing if the 'key' parameter is NULL and 335268651Skris for CFB and OFB modes they zero ctx->num. 335368651Skris 335468651Skris New functionality allows removal of S/MIME code RC2 hack. 335568651Skris 335668651Skris Most of the routines have the same form and so can be declared in terms 335768651Skris of macros. 335868651Skris 335968651Skris By shifting this to the top level EVP_CipherInit() it can be removed from 336068651Skris all individual ciphers. If the cipher wants to handle IVs or keys 336168651Skris differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 336268651Skris flags. 336368651Skris 336468651Skris Change lots of functions like EVP_EncryptUpdate() to now return a 336568651Skris value: although software versions of the algorithms cannot fail 336668651Skris any installed hardware versions can. 336768651Skris [Steve Henson] 336868651Skris 336968651Skris *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 337068651Skris this option is set, tolerate broken clients that send the negotiated 337168651Skris protocol version number instead of the requested protocol version 337268651Skris number. 337368651Skris [Bodo Moeller] 337468651Skris 337568651Skris *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 337668651Skris i.e. non-zero for export ciphersuites, zero otherwise. 337768651Skris Previous versions had this flag inverted, inconsistent with 337868651Skris rsa_tmp_cb (..._TMP_RSA_CB). 337968651Skris [Bodo Moeller; problem reported by Amit Chopra] 338068651Skris 338168651Skris *) Add missing DSA library text string. Work around for some IIS 338268651Skris key files with invalid SEQUENCE encoding. 338368651Skris [Steve Henson] 338468651Skris 338568651Skris *) Add a document (doc/standards.txt) that list all kinds of standards 338668651Skris and so on that are implemented in OpenSSL. 338768651Skris [Richard Levitte] 338868651Skris 338968651Skris *) Enhance c_rehash script. Old version would mishandle certificates 339068651Skris with the same subject name hash and wouldn't handle CRLs at all. 339168651Skris Added -fingerprint option to crl utility, to support new c_rehash 339268651Skris features. 339368651Skris [Steve Henson] 339468651Skris 339568651Skris *) Eliminate non-ANSI declarations in crypto.h and stack.h. 339668651Skris [Ulf M�ller] 339768651Skris 339868651Skris *) Fix for SSL server purpose checking. Server checking was 339968651Skris rejecting certificates which had extended key usage present 340068651Skris but no ssl client purpose. 340168651Skris [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 340268651Skris 340368651Skris *) Make PKCS#12 code work with no password. The PKCS#12 spec 340468651Skris is a little unclear about how a blank password is handled. 340568651Skris Since the password in encoded as a BMPString with terminating 340668651Skris double NULL a zero length password would end up as just the 340768651Skris double NULL. However no password at all is different and is 340868651Skris handled differently in the PKCS#12 key generation code. NS 340968651Skris treats a blank password as zero length. MSIE treats it as no 341068651Skris password on export: but it will try both on import. We now do 341168651Skris the same: PKCS12_parse() tries zero length and no password if 341268651Skris the password is set to "" or NULL (NULL is now a valid password: 341368651Skris it wasn't before) as does the pkcs12 application. 341468651Skris [Steve Henson] 341568651Skris 341668651Skris *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 341768651Skris perror when PEM_read_bio_X509_REQ fails, the error message must 341868651Skris be obtained from the error queue. 341968651Skris [Bodo Moeller] 342068651Skris 342168651Skris *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 342268651Skris it in ERR_remove_state if appropriate, and change ERR_get_state 342368651Skris accordingly to avoid race conditions (this is necessary because 342468651Skris thread_hash is no longer constant once set). 342568651Skris [Bodo Moeller] 342668651Skris 342768651Skris *) Bugfix for linux-elf makefile.one. 342868651Skris [Ulf M�ller] 342968651Skris 343068651Skris *) RSA_get_default_method() will now cause a default 343168651Skris RSA_METHOD to be chosen if one doesn't exist already. 343268651Skris Previously this was only set during a call to RSA_new() 343368651Skris or RSA_new_method(NULL) meaning it was possible for 343468651Skris RSA_get_default_method() to return NULL. 343568651Skris [Geoff Thorpe] 343668651Skris 343768651Skris *) Added native name translation to the existing DSO code 343868651Skris that will convert (if the flag to do so is set) filenames 343968651Skris that are sufficiently small and have no path information 344068651Skris into a canonical native form. Eg. "blah" converted to 344168651Skris "libblah.so" or "blah.dll" etc. 344268651Skris [Geoff Thorpe] 344368651Skris 344468651Skris *) New function ERR_error_string_n(e, buf, len) which is like 344568651Skris ERR_error_string(e, buf), but writes at most 'len' bytes 344668651Skris including the 0 terminator. For ERR_error_string_n, 'buf' 344768651Skris may not be NULL. 344868651Skris [Damien Miller <djm@mindrot.org>, Bodo Moeller] 344968651Skris 345068651Skris *) CONF library reworked to become more general. A new CONF 345168651Skris configuration file reader "class" is implemented as well as a 345268651Skris new functions (NCONF_*, for "New CONF") to handle it. The now 345368651Skris old CONF_* functions are still there, but are reimplemented to 345468651Skris work in terms of the new functions. Also, a set of functions 345568651Skris to handle the internal storage of the configuration data is 345668651Skris provided to make it easier to write new configuration file 345768651Skris reader "classes" (I can definitely see something reading a 345868651Skris configuration file in XML format, for example), called _CONF_*, 345968651Skris or "the configuration storage API"... 346068651Skris 346168651Skris The new configuration file reading functions are: 346268651Skris 346368651Skris NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 346468651Skris NCONF_get_section, NCONF_get_string, NCONF_get_numbre 346568651Skris 346668651Skris NCONF_default, NCONF_WIN32 346768651Skris 346868651Skris NCONF_dump_fp, NCONF_dump_bio 346968651Skris 347068651Skris NCONF_default and NCONF_WIN32 are method (or "class") choosers, 347168651Skris NCONF_new creates a new CONF object. This works in the same way 347268651Skris as other interfaces in OpenSSL, like the BIO interface. 347368651Skris NCONF_dump_* dump the internal storage of the configuration file, 347468651Skris which is useful for debugging. All other functions take the same 347568651Skris arguments as the old CONF_* functions wth the exception of the 347668651Skris first that must be a `CONF *' instead of a `LHASH *'. 347768651Skris 347868651Skris To make it easer to use the new classes with the old CONF_* functions, 347968651Skris the function CONF_set_default_method is provided. 348068651Skris [Richard Levitte] 348168651Skris 348268651Skris *) Add '-tls1' option to 'openssl ciphers', which was already 348368651Skris mentioned in the documentation but had not been implemented. 348468651Skris (This option is not yet really useful because even the additional 348568651Skris experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 348668651Skris [Bodo Moeller] 348768651Skris 348868651Skris *) Initial DSO code added into libcrypto for letting OpenSSL (and 348968651Skris OpenSSL-based applications) load shared libraries and bind to 349068651Skris them in a portable way. 349168651Skris [Geoff Thorpe, with contributions from Richard Levitte] 349268651Skris 349359191Skris Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 349459191Skris 349559191Skris *) Make sure _lrotl and _lrotr are only used with MSVC. 349659191Skris 349759191Skris *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 349859191Skris (the default implementation of RAND_status). 349959191Skris 350059191Skris *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 350159191Skris to '-clrext' (= clear extensions), as intended and documented. 350259191Skris [Bodo Moeller; inconsistency pointed out by Michael Attili 350359191Skris <attili@amaxo.com>] 350459191Skris 350559191Skris *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 350659191Skris was larger than the MD block size. 350759191Skris [Steve Henson, pointed out by Yost William <YostW@tce.com>] 350859191Skris 350959191Skris *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 351059191Skris fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 351159191Skris using the passed key: if the passed key was a private key the result 351259191Skris of X509_print(), for example, would be to print out all the private key 351359191Skris components. 351459191Skris [Steve Henson] 351559191Skris 351659191Skris *) des_quad_cksum() byte order bug fix. 351759191Skris [Ulf M�ller, using the problem description in krb4-0.9.7, where 351859191Skris the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 351959191Skris 352059191Skris *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 352159191Skris discouraged. 352259191Skris [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 352359191Skris 352459191Skris *) For easily testing in shell scripts whether some command 352559191Skris 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 352659191Skris returns with exit code 0 iff no command of the given name is available. 352759191Skris 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 352859191Skris the output goes to stdout and nothing is printed to stderr. 352959191Skris Additional arguments are always ignored. 353059191Skris 353159191Skris Since for each cipher there is a command of the same name, 353259191Skris the 'no-cipher' compilation switches can be tested this way. 353359191Skris 353459191Skris ('openssl no-XXX' is not able to detect pseudo-commands such 353559191Skris as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 353659191Skris [Bodo Moeller] 353759191Skris 353859191Skris *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 353959191Skris [Bodo Moeller] 354059191Skris 354159191Skris *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 354259191Skris is set; it will be thrown away anyway because each handshake creates 354359191Skris its own key. 354459191Skris ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 354559191Skris to parameters -- in previous versions (since OpenSSL 0.9.3) the 354659191Skris 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 354759191Skris you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 354859191Skris [Bodo Moeller] 354959191Skris 355059191Skris *) New s_client option -ign_eof: EOF at stdin is ignored, and 355159191Skris 'Q' and 'R' lose their special meanings (quit/renegotiate). 355259191Skris This is part of what -quiet does; unlike -quiet, -ign_eof 355359191Skris does not suppress any output. 355459191Skris [Richard Levitte] 355559191Skris 355659191Skris *) Add compatibility options to the purpose and trust code. The 355759191Skris purpose X509_PURPOSE_ANY is "any purpose" which automatically 355859191Skris accepts a certificate or CA, this was the previous behaviour, 355959191Skris with all the associated security issues. 356059191Skris 356159191Skris X509_TRUST_COMPAT is the old trust behaviour: only and 356259191Skris automatically trust self signed roots in certificate store. A 356359191Skris new trust setting X509_TRUST_DEFAULT is used to specify that 356459191Skris a purpose has no associated trust setting and it should instead 356559191Skris use the value in the default purpose. 356659191Skris [Steve Henson] 356759191Skris 356859191Skris *) Fix the PKCS#8 DSA private key code so it decodes keys again 356959191Skris and fix a memory leak. 357059191Skris [Steve Henson] 357159191Skris 357259191Skris *) In util/mkerr.pl (which implements 'make errors'), preserve 357359191Skris reason strings from the previous version of the .c file, as 357459191Skris the default to have only downcase letters (and digits) in 357559191Skris automatically generated reasons codes is not always appropriate. 357659191Skris [Bodo Moeller] 357759191Skris 357859191Skris *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 357959191Skris using strerror. Previously, ERR_reason_error_string() returned 358059191Skris library names as reason strings for SYSerr; but SYSerr is a special 358159191Skris case where small numbers are errno values, not library numbers. 358259191Skris [Bodo Moeller] 358359191Skris 358459191Skris *) Add '-dsaparam' option to 'openssl dhparam' application. This 358559191Skris converts DSA parameters into DH parameters. (When creating parameters, 358659191Skris DSA_generate_parameters is used.) 358759191Skris [Bodo Moeller] 358859191Skris 358959191Skris *) Include 'length' (recommended exponent length) in C code generated 359059191Skris by 'openssl dhparam -C'. 359159191Skris [Bodo Moeller] 359259191Skris 359359191Skris *) The second argument to set_label in perlasm was already being used 359459191Skris so couldn't be used as a "file scope" flag. Moved to third argument 359559191Skris which was free. 359659191Skris [Steve Henson] 359759191Skris 359859191Skris *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 359959191Skris instead of RAND_bytes for encryption IVs and salts. 360059191Skris [Bodo Moeller] 360159191Skris 360259191Skris *) Include RAND_status() into RAND_METHOD instead of implementing 360359191Skris it only for md_rand.c Otherwise replacing the PRNG by calling 360459191Skris RAND_set_rand_method would be impossible. 360559191Skris [Bodo Moeller] 360659191Skris 360759191Skris *) Don't let DSA_generate_key() enter an infinite loop if the random 360859191Skris number generation fails. 360959191Skris [Bodo Moeller] 361059191Skris 361159191Skris *) New 'rand' application for creating pseudo-random output. 361259191Skris [Bodo Moeller] 361359191Skris 361459191Skris *) Added configuration support for Linux/IA64 361559191Skris [Rolf Haberrecker <rolf@suse.de>] 361659191Skris 361759191Skris *) Assembler module support for Mingw32. 361859191Skris [Ulf M�ller] 361959191Skris 362059191Skris *) Shared library support for HPUX (in shlib/). 362159191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 362259191Skris 362359191Skris *) Shared library support for Solaris gcc. 362459191Skris [Lutz Behnke <behnke@trustcenter.de>] 362559191Skris 362659191Skris Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 362759191Skris 362859191Skris *) PKCS7_encrypt() was adding text MIME headers twice because they 362959191Skris were added manually and by SMIME_crlf_copy(). 363059191Skris [Steve Henson] 363159191Skris 363259191Skris *) In bntest.c don't call BN_rand with zero bits argument. 363359191Skris [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 363459191Skris 363559191Skris *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 363659191Skris case was implemented. This caused BN_div_recp() to fail occasionally. 363759191Skris [Ulf M�ller] 363859191Skris 363959191Skris *) Add an optional second argument to the set_label() in the perl 364059191Skris assembly language builder. If this argument exists and is set 364159191Skris to 1 it signals that the assembler should use a symbol whose 364259191Skris scope is the entire file, not just the current function. This 364359191Skris is needed with MASM which uses the format label:: for this scope. 364459191Skris [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 364559191Skris 364659191Skris *) Change the ASN1 types so they are typedefs by default. Before 364759191Skris almost all types were #define'd to ASN1_STRING which was causing 364859191Skris STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 364959191Skris for example. 365059191Skris [Steve Henson] 365159191Skris 365259191Skris *) Change names of new functions to the new get1/get0 naming 365359191Skris convention: After 'get1', the caller owns a reference count 365459191Skris and has to call ..._free; 'get0' returns a pointer to some 365559191Skris data structure without incrementing reference counters. 365659191Skris (Some of the existing 'get' functions increment a reference 365759191Skris counter, some don't.) 365859191Skris Similarly, 'set1' and 'add1' functions increase reference 365959191Skris counters or duplicate objects. 366059191Skris [Steve Henson] 366159191Skris 366259191Skris *) Allow for the possibility of temp RSA key generation failure: 366359191Skris the code used to assume it always worked and crashed on failure. 366459191Skris [Steve Henson] 366559191Skris 366659191Skris *) Fix potential buffer overrun problem in BIO_printf(). 366759191Skris [Ulf M�ller, using public domain code by Patrick Powell; problem 366859191Skris pointed out by David Sacerdote <das33@cornell.edu>] 366959191Skris 367059191Skris *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 367159191Skris RAND_egd() and RAND_status(). In the command line application, 367259191Skris the EGD socket can be specified like a seed file using RANDFILE 367359191Skris or -rand. 367459191Skris [Ulf M�ller] 367559191Skris 367659191Skris *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 367759191Skris Some CAs (e.g. Verisign) distribute certificates in this form. 367859191Skris [Steve Henson] 367959191Skris 368059191Skris *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 368159191Skris list to exclude them. This means that no special compilation option 368259191Skris is needed to use anonymous DH: it just needs to be included in the 368359191Skris cipher list. 368459191Skris [Steve Henson] 368559191Skris 368659191Skris *) Change the EVP_MD_CTX_type macro so its meaning consistent with 368759191Skris EVP_MD_type. The old functionality is available in a new macro called 368859191Skris EVP_MD_md(). Change code that uses it and update docs. 368959191Skris [Steve Henson] 369059191Skris 369159191Skris *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 369259191Skris where the 'void *' argument is replaced by a function pointer argument. 369359191Skris Previously 'void *' was abused to point to functions, which works on 369459191Skris many platforms, but is not correct. As these functions are usually 369559191Skris called by macros defined in OpenSSL header files, most source code 369659191Skris should work without changes. 369759191Skris [Richard Levitte] 369859191Skris 369959191Skris *) <openssl/opensslconf.h> (which is created by Configure) now contains 370059191Skris sections with information on -D... compiler switches used for 370159191Skris compiling the library so that applications can see them. To enable 370259191Skris one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 370359191Skris must be defined. E.g., 370459191Skris #define OPENSSL_ALGORITHM_DEFINES 370559191Skris #include <openssl/opensslconf.h> 370659191Skris defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 370759191Skris [Richard Levitte, Ulf and Bodo M�ller] 370859191Skris 370959191Skris *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 371059191Skris record layer. 371159191Skris [Bodo Moeller] 371259191Skris 371359191Skris *) Change the 'other' type in certificate aux info to a STACK_OF 371459191Skris X509_ALGOR. Although not an AlgorithmIdentifier as such it has 371559191Skris the required ASN1 format: arbitrary types determined by an OID. 371659191Skris [Steve Henson] 371759191Skris 371859191Skris *) Add some PEM_write_X509_REQ_NEW() functions and a command line 371959191Skris argument to 'req'. This is not because the function is newer or 372059191Skris better than others it just uses the work 'NEW' in the certificate 372159191Skris request header lines. Some software needs this. 372259191Skris [Steve Henson] 372359191Skris 372459191Skris *) Reorganise password command line arguments: now passwords can be 372559191Skris obtained from various sources. Delete the PEM_cb function and make 372659191Skris it the default behaviour: i.e. if the callback is NULL and the 372759191Skris usrdata argument is not NULL interpret it as a null terminated pass 372859191Skris phrase. If usrdata and the callback are NULL then the pass phrase 372959191Skris is prompted for as usual. 373059191Skris [Steve Henson] 373159191Skris 373259191Skris *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 373359191Skris the support is automatically enabled. The resulting binaries will 373459191Skris autodetect the card and use it if present. 373559191Skris [Ben Laurie and Compaq Inc.] 373659191Skris 373759191Skris *) Work around for Netscape hang bug. This sends certificate request 373859191Skris and server done in one record. Since this is perfectly legal in the 373959191Skris SSL/TLS protocol it isn't a "bug" option and is on by default. See 374059191Skris the bugs/SSLv3 entry for more info. 374159191Skris [Steve Henson] 374259191Skris 374359191Skris *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 374459191Skris [Andy Polyakov] 374559191Skris 374659191Skris *) Add -rand argument to smime and pkcs12 applications and read/write 374759191Skris of seed file. 374859191Skris [Steve Henson] 374959191Skris 375059191Skris *) New 'passwd' tool for crypt(3) and apr1 password hashes. 375159191Skris [Bodo Moeller] 375259191Skris 375359191Skris *) Add command line password options to the remaining applications. 375459191Skris [Steve Henson] 375559191Skris 375659191Skris *) Bug fix for BN_div_recp() for numerators with an even number of 375759191Skris bits. 375859191Skris [Ulf M�ller] 375959191Skris 376059191Skris *) More tests in bntest.c, and changed test_bn output. 376159191Skris [Ulf M�ller] 376259191Skris 376359191Skris *) ./config recognizes MacOS X now. 376459191Skris [Andy Polyakov] 376559191Skris 376659191Skris *) Bug fix for BN_div() when the first words of num and divsor are 376759191Skris equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 376859191Skris [Ulf M�ller] 376959191Skris 377059191Skris *) Add support for various broken PKCS#8 formats, and command line 377159191Skris options to produce them. 377259191Skris [Steve Henson] 377359191Skris 377459191Skris *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 377559191Skris get temporary BIGNUMs from a BN_CTX. 377659191Skris [Ulf M�ller] 377759191Skris 377859191Skris *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 377959191Skris for p == 0. 378059191Skris [Ulf M�ller] 378159191Skris 378259191Skris *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 378359191Skris include a #define from the old name to the new. The original intent 378459191Skris was that statically linked binaries could for example just call 378559191Skris SSLeay_add_all_ciphers() to just add ciphers to the table and not 378659191Skris link with digests. This never worked becayse SSLeay_add_all_digests() 378759191Skris and SSLeay_add_all_ciphers() were in the same source file so calling 378859191Skris one would link with the other. They are now in separate source files. 378959191Skris [Steve Henson] 379059191Skris 379159191Skris *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 379259191Skris [Steve Henson] 379359191Skris 379459191Skris *) Use a less unusual form of the Miller-Rabin primality test (it used 379559191Skris a binary algorithm for exponentiation integrated into the Miller-Rabin 379659191Skris loop, our standard modexp algorithms are faster). 379759191Skris [Bodo Moeller] 379859191Skris 379959191Skris *) Support for the EBCDIC character set completed. 380059191Skris [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 380159191Skris 380259191Skris *) Source code cleanups: use const where appropriate, eliminate casts, 380359191Skris use void * instead of char * in lhash. 380459191Skris [Ulf M�ller] 380559191Skris 380659191Skris *) Bugfix: ssl3_send_server_key_exchange was not restartable 380759191Skris (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 380859191Skris this the server could overwrite ephemeral keys that the client 380959191Skris has already seen). 381059191Skris [Bodo Moeller] 381159191Skris 381259191Skris *) Turn DSA_is_prime into a macro that calls BN_is_prime, 381359191Skris using 50 iterations of the Rabin-Miller test. 381459191Skris 381559191Skris DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 381659191Skris iterations of the Rabin-Miller test as required by the appendix 381759191Skris to FIPS PUB 186[-1]) instead of DSA_is_prime. 381859191Skris As BN_is_prime_fasttest includes trial division, DSA parameter 381959191Skris generation becomes much faster. 382059191Skris 382159191Skris This implies a change for the callback functions in DSA_is_prime 382259191Skris and DSA_generate_parameters: The callback function is called once 382359191Skris for each positive witness in the Rabin-Miller test, not just 382459191Skris occasionally in the inner loop; and the parameters to the 382559191Skris callback function now provide an iteration count for the outer 382659191Skris loop rather than for the current invocation of the inner loop. 382759191Skris DSA_generate_parameters additionally can call the callback 382859191Skris function with an 'iteration count' of -1, meaning that a 382959191Skris candidate has passed the trial division test (when q is generated 383059191Skris from an application-provided seed, trial division is skipped). 383159191Skris [Bodo Moeller] 383259191Skris 383359191Skris *) New function BN_is_prime_fasttest that optionally does trial 383459191Skris division before starting the Rabin-Miller test and has 383559191Skris an additional BN_CTX * argument (whereas BN_is_prime always 383659191Skris has to allocate at least one BN_CTX). 383759191Skris 'callback(1, -1, cb_arg)' is called when a number has passed the 383859191Skris trial division stage. 383959191Skris [Bodo Moeller] 384059191Skris 384159191Skris *) Fix for bug in CRL encoding. The validity dates weren't being handled 384259191Skris as ASN1_TIME. 384359191Skris [Steve Henson] 384459191Skris 384559191Skris *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 384659191Skris [Steve Henson] 384759191Skris 384859191Skris *) New function BN_pseudo_rand(). 384959191Skris [Ulf M�ller] 385059191Skris 385159191Skris *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 385259191Skris bignum version of BN_from_montgomery() with the working code from 385359191Skris SSLeay 0.9.0 (the word based version is faster anyway), and clean up 385459191Skris the comments. 385559191Skris [Ulf M�ller] 385659191Skris 385759191Skris *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 385859191Skris made it impossible to use the same SSL_SESSION data structure in 385959191Skris SSL2 clients in multiple threads. 386059191Skris [Bodo Moeller] 386159191Skris 386259191Skris *) The return value of RAND_load_file() no longer counts bytes obtained 386359191Skris by stat(). RAND_load_file(..., -1) is new and uses the complete file 386459191Skris to seed the PRNG (previously an explicit byte count was required). 386559191Skris [Ulf M�ller, Bodo M�ller] 386659191Skris 386759191Skris *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 386859191Skris used (char *) instead of (void *) and had casts all over the place. 386959191Skris [Steve Henson] 387059191Skris 387159191Skris *) Make BN_generate_prime() return NULL on error if ret!=NULL. 387259191Skris [Ulf M�ller] 387359191Skris 387459191Skris *) Retain source code compatibility for BN_prime_checks macro: 387559191Skris BN_is_prime(..., BN_prime_checks, ...) now uses 387659191Skris BN_prime_checks_for_size to determine the appropriate number of 387759191Skris Rabin-Miller iterations. 387859191Skris [Ulf M�ller] 387959191Skris 388059191Skris *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 388159191Skris DH_CHECK_P_NOT_SAFE_PRIME. 388259191Skris (Check if this is true? OpenPGP calls them "strong".) 388359191Skris [Ulf M�ller] 388459191Skris 388559191Skris *) Merge the functionality of "dh" and "gendh" programs into a new program 388659191Skris "dhparam". The old programs are retained for now but will handle DH keys 388759191Skris (instead of parameters) in future. 388859191Skris [Steve Henson] 388959191Skris 389059191Skris *) Make the ciphers, s_server and s_client programs check the return values 389159191Skris when a new cipher list is set. 389259191Skris [Steve Henson] 389359191Skris 389459191Skris *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 389559191Skris ciphers. Before when the 56bit ciphers were enabled the sorting was 389659191Skris wrong. 389759191Skris 389859191Skris The syntax for the cipher sorting has been extended to support sorting by 389959191Skris cipher-strength (using the strength_bits hard coded in the tables). 390059191Skris The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 390159191Skris 390259191Skris Fix a bug in the cipher-command parser: when supplying a cipher command 390359191Skris string with an "undefined" symbol (neither command nor alphanumeric 390459191Skris [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 390559191Skris an error is flagged. 390659191Skris 390759191Skris Due to the strength-sorting extension, the code of the 390859191Skris ssl_create_cipher_list() function was completely rearranged. I hope that 390959191Skris the readability was also increased :-) 391059191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 391159191Skris 391259191Skris *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 391359191Skris for the first serial number and places 2 in the serial number file. This 391459191Skris avoids problems when the root CA is created with serial number zero and 391559191Skris the first user certificate has the same issuer name and serial number 391659191Skris as the root CA. 391759191Skris [Steve Henson] 391859191Skris 391959191Skris *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 392059191Skris the new code. Add documentation for this stuff. 392159191Skris [Steve Henson] 392259191Skris 392359191Skris *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 392459191Skris X509_*() to X509at_*() on the grounds that they don't handle X509 392559191Skris structures and behave in an analagous way to the X509v3 functions: 392659191Skris they shouldn't be called directly but wrapper functions should be used 392759191Skris instead. 392859191Skris 392959191Skris So we also now have some wrapper functions that call the X509at functions 393059191Skris when passed certificate requests. (TO DO: similar things can be done with 393159191Skris PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 393259191Skris things. Some of these need some d2i or i2d and print functionality 393359191Skris because they handle more complex structures.) 393459191Skris [Steve Henson] 393559191Skris 393659191Skris *) Add missing #ifndefs that caused missing symbols when building libssl 393759191Skris as a shared library without RSA. Use #ifndef NO_SSL2 instead of 393859191Skris NO_RSA in ssl/s2*.c. 393959191Skris [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 394059191Skris 394159191Skris *) Precautions against using the PRNG uninitialized: RAND_bytes() now 394259191Skris has a return value which indicates the quality of the random data 394359191Skris (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 394459191Skris error queue. New function RAND_pseudo_bytes() generates output that is 394559191Skris guaranteed to be unique but not unpredictable. RAND_add is like 394659191Skris RAND_seed, but takes an extra argument for an entropy estimate 394759191Skris (RAND_seed always assumes full entropy). 394859191Skris [Ulf M�ller] 394959191Skris 395059191Skris *) Do more iterations of Rabin-Miller probable prime test (specifically, 395159191Skris 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 395259191Skris instead of only 2 for all lengths; see BN_prime_checks_for_size definition 395359191Skris in crypto/bn/bn_prime.c for the complete table). This guarantees a 395459191Skris false-positive rate of at most 2^-80 for random input. 395559191Skris [Bodo Moeller] 395659191Skris 395759191Skris *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 395859191Skris [Bodo Moeller] 395959191Skris 396059191Skris *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 396159191Skris in the 0.9.5 release), this returns the chain 396259191Skris from an X509_CTX structure with a dup of the stack and all 396359191Skris the X509 reference counts upped: so the stack will exist 396459191Skris after X509_CTX_cleanup() has been called. Modify pkcs12.c 396559191Skris to use this. 396659191Skris 396759191Skris Also make SSL_SESSION_print() print out the verify return 396859191Skris code. 396959191Skris [Steve Henson] 397059191Skris 397159191Skris *) Add manpage for the pkcs12 command. Also change the default 397259191Skris behaviour so MAC iteration counts are used unless the new 397359191Skris -nomaciter option is used. This improves file security and 397459191Skris only older versions of MSIE (4.0 for example) need it. 397559191Skris [Steve Henson] 397659191Skris 397759191Skris *) Honor the no-xxx Configure options when creating .DEF files. 397859191Skris [Ulf M�ller] 397959191Skris 398059191Skris *) Add PKCS#10 attributes to field table: challengePassword, 398159191Skris unstructuredName and unstructuredAddress. These are taken from 398259191Skris draft PKCS#9 v2.0 but are compatible with v1.2 provided no 398359191Skris international characters are used. 398459191Skris 398559191Skris More changes to X509_ATTRIBUTE code: allow the setting of types 398659191Skris based on strings. Remove the 'loc' parameter when adding 398759191Skris attributes because these will be a SET OF encoding which is sorted 398859191Skris in ASN1 order. 398959191Skris [Steve Henson] 399059191Skris 399159191Skris *) Initial changes to the 'req' utility to allow request generation 399259191Skris automation. This will allow an application to just generate a template 399359191Skris file containing all the field values and have req construct the 399459191Skris request. 399559191Skris 399659191Skris Initial support for X509_ATTRIBUTE handling. Stacks of these are 399759191Skris used all over the place including certificate requests and PKCS#7 399859191Skris structures. They are currently handled manually where necessary with 399959191Skris some primitive wrappers for PKCS#7. The new functions behave in a 400059191Skris manner analogous to the X509 extension functions: they allow 400159191Skris attributes to be looked up by NID and added. 400259191Skris 400359191Skris Later something similar to the X509V3 code would be desirable to 400459191Skris automatically handle the encoding, decoding and printing of the 400559191Skris more complex types. The string types like challengePassword can 400659191Skris be handled by the string table functions. 400759191Skris 400859191Skris Also modified the multi byte string table handling. Now there is 400959191Skris a 'global mask' which masks out certain types. The table itself 401059191Skris can use the flag STABLE_NO_MASK to ignore the mask setting: this 401159191Skris is useful when for example there is only one permissible type 401259191Skris (as in countryName) and using the mask might result in no valid 401359191Skris types at all. 401459191Skris [Steve Henson] 401559191Skris 401659191Skris *) Clean up 'Finished' handling, and add functions SSL_get_finished and 401759191Skris SSL_get_peer_finished to allow applications to obtain the latest 401859191Skris Finished messages sent to the peer or expected from the peer, 401959191Skris respectively. (SSL_get_peer_finished is usually the Finished message 402059191Skris actually received from the peer, otherwise the protocol will be aborted.) 402159191Skris 402259191Skris As the Finished message are message digests of the complete handshake 402359191Skris (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 402459191Skris be used for external authentication procedures when the authentication 402559191Skris provided by SSL/TLS is not desired or is not enough. 402659191Skris [Bodo Moeller] 402759191Skris 402859191Skris *) Enhanced support for Alpha Linux is added. Now ./config checks if 402959191Skris the host supports BWX extension and if Compaq C is present on the 403059191Skris $PATH. Just exploiting of the BWX extension results in 20-30% 403159191Skris performance kick for some algorithms, e.g. DES and RC4 to mention 403259191Skris a couple. Compaq C in turn generates ~20% faster code for MD5 and 403359191Skris SHA1. 403459191Skris [Andy Polyakov] 403559191Skris 403659191Skris *) Add support for MS "fast SGC". This is arguably a violation of the 403759191Skris SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 403859191Skris weak crypto and after checking the certificate is SGC a second one 403959191Skris with strong crypto. MS SGC stops the first handshake after receiving 404059191Skris the server certificate message and sends a second client hello. Since 404159191Skris a server will typically do all the time consuming operations before 404259191Skris expecting any further messages from the client (server key exchange 404359191Skris is the most expensive) there is little difference between the two. 404459191Skris 404559191Skris To get OpenSSL to support MS SGC we have to permit a second client 404659191Skris hello message after we have sent server done. In addition we have to 404759191Skris reset the MAC if we do get this second client hello. 404859191Skris [Steve Henson] 404959191Skris 405059191Skris *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 405159191Skris if a DER encoded private key is RSA or DSA traditional format. Changed 405259191Skris d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 405359191Skris format DER encoded private key. Newer code should use PKCS#8 format which 405459191Skris has the key type encoded in the ASN1 structure. Added DER private key 405559191Skris support to pkcs8 application. 405659191Skris [Steve Henson] 405759191Skris 405859191Skris *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 405959191Skris ciphersuites has been selected (as required by the SSL 3/TLS 1 406059191Skris specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 406159191Skris is set, we interpret this as a request to violate the specification 406259191Skris (the worst that can happen is a handshake failure, and 'correct' 406359191Skris behaviour would result in a handshake failure anyway). 406459191Skris [Bodo Moeller] 406559191Skris 406659191Skris *) In SSL_CTX_add_session, take into account that there might be multiple 406759191Skris SSL_SESSION structures with the same session ID (e.g. when two threads 406859191Skris concurrently obtain them from an external cache). 406959191Skris The internal cache can handle only one SSL_SESSION with a given ID, 407059191Skris so if there's a conflict, we now throw out the old one to achieve 407159191Skris consistency. 407259191Skris [Bodo Moeller] 407359191Skris 407459191Skris *) Add OIDs for idea and blowfish in CBC mode. This will allow both 407559191Skris to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 407659191Skris some routines that use cipher OIDs: some ciphers do not have OIDs 407759191Skris defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 407859191Skris example. 407959191Skris [Steve Henson] 408059191Skris 408159191Skris *) Simplify the trust setting structure and code. Now we just have 408259191Skris two sequences of OIDs for trusted and rejected settings. These will 408359191Skris typically have values the same as the extended key usage extension 408459191Skris and any application specific purposes. 408559191Skris 408659191Skris The trust checking code now has a default behaviour: it will just 408759191Skris check for an object with the same NID as the passed id. Functions can 408859191Skris be provided to override either the default behaviour or the behaviour 408959191Skris for a given id. SSL client, server and email already have functions 409059191Skris in place for compatibility: they check the NID and also return "trusted" 409159191Skris if the certificate is self signed. 409259191Skris [Steve Henson] 409359191Skris 409459191Skris *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 409559191Skris traditional format into an EVP_PKEY structure. 409659191Skris [Steve Henson] 409759191Skris 409859191Skris *) Add a password callback function PEM_cb() which either prompts for 409959191Skris a password if usr_data is NULL or otherwise assumes it is a null 410059191Skris terminated password. Allow passwords to be passed on command line 410159191Skris environment or config files in a few more utilities. 410259191Skris [Steve Henson] 410359191Skris 410459191Skris *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 410559191Skris keys. Add some short names for PKCS#8 PBE algorithms and allow them 410659191Skris to be specified on the command line for the pkcs8 and pkcs12 utilities. 410759191Skris Update documentation. 410859191Skris [Steve Henson] 410959191Skris 411059191Skris *) Support for ASN1 "NULL" type. This could be handled before by using 411159191Skris ASN1_TYPE but there wasn't any function that would try to read a NULL 411259191Skris and produce an error if it couldn't. For compatibility we also have 411359191Skris ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 411459191Skris don't allocate anything because they don't need to. 411559191Skris [Steve Henson] 411659191Skris 411759191Skris *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 411859191Skris for details. 411959191Skris [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 412059191Skris 412159191Skris *) Rebuild of the memory allocation routines used by OpenSSL code and 412259191Skris possibly others as well. The purpose is to make an interface that 412359191Skris provide hooks so anyone can build a separate set of allocation and 412459191Skris deallocation routines to be used by OpenSSL, for example memory 412559191Skris pool implementations, or something else, which was previously hard 412659191Skris since Malloc(), Realloc() and Free() were defined as macros having 412759191Skris the values malloc, realloc and free, respectively (except for Win32 412859191Skris compilations). The same is provided for memory debugging code. 412959191Skris OpenSSL already comes with functionality to find memory leaks, but 413059191Skris this gives people a chance to debug other memory problems. 413159191Skris 413259191Skris With these changes, a new set of functions and macros have appeared: 413359191Skris 413468651Skris CRYPTO_set_mem_debug_functions() [F] 413568651Skris CRYPTO_get_mem_debug_functions() [F] 413668651Skris CRYPTO_dbg_set_options() [F] 413768651Skris CRYPTO_dbg_get_options() [F] 413868651Skris CRYPTO_malloc_debug_init() [M] 413959191Skris 414059191Skris The memory debug functions are NULL by default, unless the library 414159191Skris is compiled with CRYPTO_MDEBUG or friends is defined. If someone 414259191Skris wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 414359191Skris gives the standard debugging functions that come with OpenSSL) or 414459191Skris CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 414559191Skris provided by the library user) must be used. When the standard 414659191Skris debugging functions are used, CRYPTO_dbg_set_options can be used to 414759191Skris request additional information: 414859191Skris CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 414959191Skris the CRYPTO_MDEBUG_xxx macro when compiling the library. 415059191Skris 415159191Skris Also, things like CRYPTO_set_mem_functions will always give the 415259191Skris expected result (the new set of functions is used for allocation 415359191Skris and deallocation) at all times, regardless of platform and compiler 415459191Skris options. 415559191Skris 415659191Skris To finish it up, some functions that were never use in any other 415759191Skris way than through macros have a new API and new semantic: 415859191Skris 415959191Skris CRYPTO_dbg_malloc() 416059191Skris CRYPTO_dbg_realloc() 416159191Skris CRYPTO_dbg_free() 416259191Skris 416359191Skris All macros of value have retained their old syntax. 416459191Skris [Richard Levitte and Bodo Moeller] 416559191Skris 416659191Skris *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 416759191Skris ordering of SMIMECapabilities wasn't in "strength order" and there 416859191Skris was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 416959191Skris algorithm. 417059191Skris [Steve Henson] 417159191Skris 417259191Skris *) Some ASN1 types with illegal zero length encoding (INTEGER, 417359191Skris ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 417459191Skris [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 417559191Skris 417659191Skris *) Merge in my S/MIME library for OpenSSL. This provides a simple 417759191Skris S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 417859191Skris functionality to handle multipart/signed properly) and a utility 417959191Skris called 'smime' to call all this stuff. This is based on code I 418059191Skris originally wrote for Celo who have kindly allowed it to be 418159191Skris included in OpenSSL. 418259191Skris [Steve Henson] 418359191Skris 418459191Skris *) Add variants des_set_key_checked and des_set_key_unchecked of 418559191Skris des_set_key (aka des_key_sched). Global variable des_check_key 418659191Skris decides which of these is called by des_set_key; this way 418759191Skris des_check_key behaves as it always did, but applications and 418859191Skris the library itself, which was buggy for des_check_key == 1, 418959191Skris have a cleaner way to pick the version they need. 419059191Skris [Bodo Moeller] 419159191Skris 419259191Skris *) New function PKCS12_newpass() which changes the password of a 419359191Skris PKCS12 structure. 419459191Skris [Steve Henson] 419559191Skris 419659191Skris *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 419759191Skris dynamic mix. In both cases the ids can be used as an index into the 419859191Skris table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 419959191Skris functions so they accept a list of the field values and the 420059191Skris application doesn't need to directly manipulate the X509_TRUST 420159191Skris structure. 420259191Skris [Steve Henson] 420359191Skris 420459191Skris *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 420559191Skris need initialising. 420659191Skris [Steve Henson] 420759191Skris 420859191Skris *) Modify the way the V3 extension code looks up extensions. This now 420959191Skris works in a similar way to the object code: we have some "standard" 421059191Skris extensions in a static table which is searched with OBJ_bsearch() 421159191Skris and the application can add dynamic ones if needed. The file 421259191Skris crypto/x509v3/ext_dat.h now has the info: this file needs to be 421359191Skris updated whenever a new extension is added to the core code and kept 421459191Skris in ext_nid order. There is a simple program 'tabtest.c' which checks 421559191Skris this. New extensions are not added too often so this file can readily 421659191Skris be maintained manually. 421759191Skris 421859191Skris There are two big advantages in doing things this way. The extensions 421959191Skris can be looked up immediately and no longer need to be "added" using 422059191Skris X509V3_add_standard_extensions(): this function now does nothing. 422159191Skris [Side note: I get *lots* of email saying the extension code doesn't 422259191Skris work because people forget to call this function] 422359191Skris Also no dynamic allocation is done unless new extensions are added: 422459191Skris so if we don't add custom extensions there is no need to call 422559191Skris X509V3_EXT_cleanup(). 422659191Skris [Steve Henson] 422759191Skris 422859191Skris *) Modify enc utility's salting as follows: make salting the default. Add a 422959191Skris magic header, so unsalted files fail gracefully instead of just decrypting 423059191Skris to garbage. This is because not salting is a big security hole, so people 423159191Skris should be discouraged from doing it. 423259191Skris [Ben Laurie] 423359191Skris 423459191Skris *) Fixes and enhancements to the 'x509' utility. It allowed a message 423559191Skris digest to be passed on the command line but it only used this 423659191Skris parameter when signing a certificate. Modified so all relevant 423759191Skris operations are affected by the digest parameter including the 423859191Skris -fingerprint and -x509toreq options. Also -x509toreq choked if a 423959191Skris DSA key was used because it didn't fix the digest. 424059191Skris [Steve Henson] 424159191Skris 424259191Skris *) Initial certificate chain verify code. Currently tests the untrusted 424359191Skris certificates for consistency with the verify purpose (which is set 424459191Skris when the X509_STORE_CTX structure is set up) and checks the pathlength. 424559191Skris 424659191Skris There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 424759191Skris this is because it will reject chains with invalid extensions whereas 424859191Skris every previous version of OpenSSL and SSLeay made no checks at all. 424959191Skris 425059191Skris Trust code: checks the root CA for the relevant trust settings. Trust 425159191Skris settings have an initial value consistent with the verify purpose: e.g. 425259191Skris if the verify purpose is for SSL client use it expects the CA to be 425359191Skris trusted for SSL client use. However the default value can be changed to 425459191Skris permit custom trust settings: one example of this would be to only trust 425559191Skris certificates from a specific "secure" set of CAs. 425659191Skris 425759191Skris Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 425859191Skris which should be used for version portability: especially since the 425959191Skris verify structure is likely to change more often now. 426059191Skris 426159191Skris SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 426259191Skris to set them. If not set then assume SSL clients will verify SSL servers 426359191Skris and vice versa. 426459191Skris 426559191Skris Two new options to the verify program: -untrusted allows a set of 426659191Skris untrusted certificates to be passed in and -purpose which sets the 426759191Skris intended purpose of the certificate. If a purpose is set then the 426859191Skris new chain verify code is used to check extension consistency. 426959191Skris [Steve Henson] 427059191Skris 427159191Skris *) Support for the authority information access extension. 427259191Skris [Steve Henson] 427359191Skris 427459191Skris *) Modify RSA and DSA PEM read routines to transparently handle 427559191Skris PKCS#8 format private keys. New *_PUBKEY_* functions that handle 427659191Skris public keys in a format compatible with certificate 427759191Skris SubjectPublicKeyInfo structures. Unfortunately there were already 427859191Skris functions called *_PublicKey_* which used various odd formats so 427959191Skris these are retained for compatibility: however the DSA variants were 428059191Skris never in a public release so they have been deleted. Changed dsa/rsa 428159191Skris utilities to handle the new format: note no releases ever handled public 428259191Skris keys so we should be OK. 428359191Skris 428459191Skris The primary motivation for this change is to avoid the same fiasco 428559191Skris that dogs private keys: there are several incompatible private key 428659191Skris formats some of which are standard and some OpenSSL specific and 428759191Skris require various evil hacks to allow partial transparent handling and 428859191Skris even then it doesn't work with DER formats. Given the option anything 428959191Skris other than PKCS#8 should be dumped: but the other formats have to 429059191Skris stay in the name of compatibility. 429159191Skris 429259191Skris With public keys and the benefit of hindsight one standard format 429359191Skris is used which works with EVP_PKEY, RSA or DSA structures: though 429459191Skris it clearly returns an error if you try to read the wrong kind of key. 429559191Skris 429659191Skris Added a -pubkey option to the 'x509' utility to output the public key. 429759191Skris Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 429859191Skris (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 429959191Skris EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 430059191Skris that do the same as the EVP_PKEY_assign_*() except they up the 430159191Skris reference count of the added key (they don't "swallow" the 430259191Skris supplied key). 430359191Skris [Steve Henson] 430459191Skris 430559191Skris *) Fixes to crypto/x509/by_file.c the code to read in certificates and 430659191Skris CRLs would fail if the file contained no certificates or no CRLs: 430759191Skris added a new function to read in both types and return the number 430859191Skris read: this means that if none are read it will be an error. The 430959191Skris DER versions of the certificate and CRL reader would always fail 431059191Skris because it isn't possible to mix certificates and CRLs in DER format 431159191Skris without choking one or the other routine. Changed this to just read 431259191Skris a certificate: this is the best we can do. Also modified the code 431359191Skris in apps/verify.c to take notice of return codes: it was previously 431459191Skris attempting to read in certificates from NULL pointers and ignoring 431559191Skris any errors: this is one reason why the cert and CRL reader seemed 431659191Skris to work. It doesn't check return codes from the default certificate 431759191Skris routines: these may well fail if the certificates aren't installed. 431859191Skris [Steve Henson] 431959191Skris 432059191Skris *) Code to support otherName option in GeneralName. 432159191Skris [Steve Henson] 432259191Skris 432359191Skris *) First update to verify code. Change the verify utility 432459191Skris so it warns if it is passed a self signed certificate: 432559191Skris for consistency with the normal behaviour. X509_verify 432659191Skris has been modified to it will now verify a self signed 432759191Skris certificate if *exactly* the same certificate appears 432859191Skris in the store: it was previously impossible to trust a 432959191Skris single self signed certificate. This means that: 433059191Skris openssl verify ss.pem 433159191Skris now gives a warning about a self signed certificate but 433259191Skris openssl verify -CAfile ss.pem ss.pem 433359191Skris is OK. 433459191Skris [Steve Henson] 433559191Skris 433659191Skris *) For servers, store verify_result in SSL_SESSION data structure 433759191Skris (and add it to external session representation). 433859191Skris This is needed when client certificate verifications fails, 433959191Skris but an application-provided verification callback (set by 434059191Skris SSL_CTX_set_cert_verify_callback) allows accepting the session 434159191Skris anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 434259191Skris but returns 1): When the session is reused, we have to set 434359191Skris ssl->verify_result to the appropriate error code to avoid 434459191Skris security holes. 434559191Skris [Bodo Moeller, problem pointed out by Lutz Jaenicke] 434659191Skris 434759191Skris *) Fix a bug in the new PKCS#7 code: it didn't consider the 434859191Skris case in PKCS7_dataInit() where the signed PKCS7 structure 434959191Skris didn't contain any existing data because it was being created. 435059191Skris [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 435159191Skris 435259191Skris *) Add a salt to the key derivation routines in enc.c. This 435359191Skris forms the first 8 bytes of the encrypted file. Also add a 435459191Skris -S option to allow a salt to be input on the command line. 435559191Skris [Steve Henson] 435659191Skris 435759191Skris *) New function X509_cmp(). Oddly enough there wasn't a function 435859191Skris to compare two certificates. We do this by working out the SHA1 435959191Skris hash and comparing that. X509_cmp() will be needed by the trust 436059191Skris code. 436159191Skris [Steve Henson] 436259191Skris 436359191Skris *) SSL_get1_session() is like SSL_get_session(), but increments 436459191Skris the reference count in the SSL_SESSION returned. 436559191Skris [Geoff Thorpe <geoff@eu.c2.net>] 436659191Skris 436759191Skris *) Fix for 'req': it was adding a null to request attributes. 436859191Skris Also change the X509_LOOKUP and X509_INFO code to handle 436959191Skris certificate auxiliary information. 437059191Skris [Steve Henson] 437159191Skris 437259191Skris *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 437359191Skris the 'enc' command. 437459191Skris [Steve Henson] 437559191Skris 437659191Skris *) Add the possibility to add extra information to the memory leak 437759191Skris detecting output, to form tracebacks, showing from where each 437859191Skris allocation was originated: CRYPTO_push_info("constant string") adds 437959191Skris the string plus current file name and line number to a per-thread 438059191Skris stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 438159191Skris is like calling CYRPTO_pop_info() until the stack is empty. 438259191Skris Also updated memory leak detection code to be multi-thread-safe. 438359191Skris [Richard Levitte] 438459191Skris 438559191Skris *) Add options -text and -noout to pkcs7 utility and delete the 438659191Skris encryption options which never did anything. Update docs. 438759191Skris [Steve Henson] 438859191Skris 438959191Skris *) Add options to some of the utilities to allow the pass phrase 439059191Skris to be included on either the command line (not recommended on 439159191Skris OSes like Unix) or read from the environment. Update the 439259191Skris manpages and fix a few bugs. 439359191Skris [Steve Henson] 439459191Skris 439559191Skris *) Add a few manpages for some of the openssl commands. 439659191Skris [Steve Henson] 439759191Skris 439859191Skris *) Fix the -revoke option in ca. It was freeing up memory twice, 439959191Skris leaking and not finding already revoked certificates. 440059191Skris [Steve Henson] 440159191Skris 440259191Skris *) Extensive changes to support certificate auxiliary information. 440359191Skris This involves the use of X509_CERT_AUX structure and X509_AUX 440459191Skris functions. An X509_AUX function such as PEM_read_X509_AUX() 440559191Skris can still read in a certificate file in the usual way but it 440659191Skris will also read in any additional "auxiliary information". By 440759191Skris doing things this way a fair degree of compatibility can be 440859191Skris retained: existing certificates can have this information added 440959191Skris using the new 'x509' options. 441059191Skris 441159191Skris Current auxiliary information includes an "alias" and some trust 441259191Skris settings. The trust settings will ultimately be used in enhanced 441359191Skris certificate chain verification routines: currently a certificate 441459191Skris can only be trusted if it is self signed and then it is trusted 441559191Skris for all purposes. 441659191Skris [Steve Henson] 441759191Skris 441859191Skris *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 441959191Skris The problem was that one of the replacement routines had not been working 442059191Skris since SSLeay releases. For now the offending routine has been replaced 442159191Skris with non-optimised assembler. Even so, this now gives around 95% 442259191Skris performance improvement for 1024 bit RSA signs. 442359191Skris [Mark Cox] 442459191Skris 442559191Skris *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 442659191Skris handling. Most clients have the effective key size in bits equal to 442759191Skris the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 442859191Skris A few however don't do this and instead use the size of the decrypted key 442959191Skris to determine the RC2 key length and the AlgorithmIdentifier to determine 443059191Skris the effective key length. In this case the effective key length can still 443159191Skris be 40 bits but the key length can be 168 bits for example. This is fixed 443259191Skris by manually forcing an RC2 key into the EVP_PKEY structure because the 443359191Skris EVP code can't currently handle unusual RC2 key sizes: it always assumes 443459191Skris the key length and effective key length are equal. 443559191Skris [Steve Henson] 443659191Skris 443759191Skris *) Add a bunch of functions that should simplify the creation of 443859191Skris X509_NAME structures. Now you should be able to do: 443959191Skris X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 444059191Skris and have it automatically work out the correct field type and fill in 444159191Skris the structures. The more adventurous can try: 444259191Skris X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 444359191Skris and it will (hopefully) work out the correct multibyte encoding. 444459191Skris [Steve Henson] 444559191Skris 444659191Skris *) Change the 'req' utility to use the new field handling and multibyte 444759191Skris copy routines. Before the DN field creation was handled in an ad hoc 444859191Skris way in req, ca, and x509 which was rather broken and didn't support 444959191Skris BMPStrings or UTF8Strings. Since some software doesn't implement 445059191Skris BMPStrings or UTF8Strings yet, they can be enabled using the config file 445159191Skris using the dirstring_type option. See the new comment in the default 445259191Skris openssl.cnf for more info. 445359191Skris [Steve Henson] 445459191Skris 445559191Skris *) Make crypto/rand/md_rand.c more robust: 445659191Skris - Assure unique random numbers after fork(). 445759191Skris - Make sure that concurrent threads access the global counter and 445859191Skris md serializably so that we never lose entropy in them 445959191Skris or use exactly the same state in multiple threads. 446059191Skris Access to the large state is not always serializable because 446159191Skris the additional locking could be a performance killer, and 446259191Skris md should be large enough anyway. 446359191Skris [Bodo Moeller] 446459191Skris 446559191Skris *) New file apps/app_rand.c with commonly needed functionality 446659191Skris for handling the random seed file. 446759191Skris 446859191Skris Use the random seed file in some applications that previously did not: 446959191Skris ca, 447059191Skris dsaparam -genkey (which also ignored its '-rand' option), 447159191Skris s_client, 447259191Skris s_server, 447359191Skris x509 (when signing). 447459191Skris Except on systems with /dev/urandom, it is crucial to have a random 447559191Skris seed file at least for key creation, DSA signing, and for DH exchanges; 447659191Skris for RSA signatures we could do without one. 447759191Skris 447859191Skris gendh and gendsa (unlike genrsa) used to read only the first byte 447959191Skris of each file listed in the '-rand' option. The function as previously 448059191Skris found in genrsa is now in app_rand.c and is used by all programs 448159191Skris that support '-rand'. 448259191Skris [Bodo Moeller] 448359191Skris 448459191Skris *) In RAND_write_file, use mode 0600 for creating files; 448559191Skris don't just chmod when it may be too late. 448659191Skris [Bodo Moeller] 448759191Skris 448859191Skris *) Report an error from X509_STORE_load_locations 448959191Skris when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 449059191Skris [Bill Perry] 449159191Skris 449259191Skris *) New function ASN1_mbstring_copy() this copies a string in either 449359191Skris ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 449459191Skris into an ASN1_STRING type. A mask of permissible types is passed 449559191Skris and it chooses the "minimal" type to use or an error if not type 449659191Skris is suitable. 449759191Skris [Steve Henson] 449859191Skris 449959191Skris *) Add function equivalents to the various macros in asn1.h. The old 450059191Skris macros are retained with an M_ prefix. Code inside the library can 450159191Skris use the M_ macros. External code (including the openssl utility) 450259191Skris should *NOT* in order to be "shared library friendly". 450359191Skris [Steve Henson] 450459191Skris 450559191Skris *) Add various functions that can check a certificate's extensions 450659191Skris to see if it usable for various purposes such as SSL client, 450759191Skris server or S/MIME and CAs of these types. This is currently 450859191Skris VERY EXPERIMENTAL but will ultimately be used for certificate chain 450959191Skris verification. Also added a -purpose flag to x509 utility to 451059191Skris print out all the purposes. 451159191Skris [Steve Henson] 451259191Skris 451359191Skris *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 451459191Skris functions. 451559191Skris [Steve Henson] 451659191Skris 451759191Skris *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 451859191Skris for, obtain and decode and extension and obtain its critical flag. 451959191Skris This allows all the necessary extension code to be handled in a 452059191Skris single function call. 452159191Skris [Steve Henson] 452259191Skris 452359191Skris *) RC4 tune-up featuring 30-40% performance improvement on most RISC 452459191Skris platforms. See crypto/rc4/rc4_enc.c for further details. 452559191Skris [Andy Polyakov] 452659191Skris 452759191Skris *) New -noout option to asn1parse. This causes no output to be produced 452859191Skris its main use is when combined with -strparse and -out to extract data 452959191Skris from a file (which may not be in ASN.1 format). 453059191Skris [Steve Henson] 453159191Skris 453259191Skris *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 453359191Skris when producing the local key id. 453459191Skris [Richard Levitte <levitte@stacken.kth.se>] 453559191Skris 453659191Skris *) New option -dhparam in s_server. This allows a DH parameter file to be 453759191Skris stated explicitly. If it is not stated then it tries the first server 453859191Skris certificate file. The previous behaviour hard coded the filename 453959191Skris "server.pem". 454059191Skris [Steve Henson] 454159191Skris 454259191Skris *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 454359191Skris a public key to be input or output. For example: 454459191Skris openssl rsa -in key.pem -pubout -out pubkey.pem 454559191Skris Also added necessary DSA public key functions to handle this. 454659191Skris [Steve Henson] 454759191Skris 454859191Skris *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 454959191Skris in the message. This was handled by allowing 455059191Skris X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 455159191Skris [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 455259191Skris 455359191Skris *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 455459191Skris to the end of the strings whereas this didn't. This would cause problems 455559191Skris if strings read with d2i_ASN1_bytes() were later modified. 455659191Skris [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 455759191Skris 455859191Skris *) Fix for base64 decode bug. When a base64 bio reads only one line of 455959191Skris data and it contains EOF it will end up returning an error. This is 456059191Skris caused by input 46 bytes long. The cause is due to the way base64 456159191Skris BIOs find the start of base64 encoded data. They do this by trying a 456259191Skris trial decode on each line until they find one that works. When they 456359191Skris do a flag is set and it starts again knowing it can pass all the 456459191Skris data directly through the decoder. Unfortunately it doesn't reset 456559191Skris the context it uses. This means that if EOF is reached an attempt 456659191Skris is made to pass two EOFs through the context and this causes the 456759191Skris resulting error. This can also cause other problems as well. As is 456859191Skris usual with these problems it takes *ages* to find and the fix is 456959191Skris trivial: move one line. 457059191Skris [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 457159191Skris 457259191Skris *) Ugly workaround to get s_client and s_server working under Windows. The 457359191Skris old code wouldn't work because it needed to select() on sockets and the 457459191Skris tty (for keypresses and to see if data could be written). Win32 only 457559191Skris supports select() on sockets so we select() with a 1s timeout on the 457659191Skris sockets and then see if any characters are waiting to be read, if none 457759191Skris are present then we retry, we also assume we can always write data to 457859191Skris the tty. This isn't nice because the code then blocks until we've 457959191Skris received a complete line of data and it is effectively polling the 458059191Skris keyboard at 1s intervals: however it's quite a bit better than not 458159191Skris working at all :-) A dedicated Windows application might handle this 458259191Skris with an event loop for example. 458359191Skris [Steve Henson] 458459191Skris 458559191Skris *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 458659191Skris and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 458759191Skris will be called when RSA_sign() and RSA_verify() are used. This is useful 458859191Skris if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 458959191Skris For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 459059191Skris should *not* be used: RSA_sign() and RSA_verify() must be used instead. 459159191Skris This necessitated the support of an extra signature type NID_md5_sha1 459259191Skris for SSL signatures and modifications to the SSL library to use it instead 459359191Skris of calling RSA_public_decrypt() and RSA_private_encrypt(). 459459191Skris [Steve Henson] 459559191Skris 459659191Skris *) Add new -verify -CAfile and -CApath options to the crl program, these 459759191Skris will lookup a CRL issuers certificate and verify the signature in a 459859191Skris similar way to the verify program. Tidy up the crl program so it 459959191Skris no longer accesses structures directly. Make the ASN1 CRL parsing a bit 460059191Skris less strict. It will now permit CRL extensions even if it is not 460159191Skris a V2 CRL: this will allow it to tolerate some broken CRLs. 460259191Skris [Steve Henson] 460359191Skris 460459191Skris *) Initialize all non-automatic variables each time one of the openssl 460559191Skris sub-programs is started (this is necessary as they may be started 460659191Skris multiple times from the "OpenSSL>" prompt). 460759191Skris [Lennart Bang, Bodo Moeller] 460859191Skris 460959191Skris *) Preliminary compilation option RSA_NULL which disables RSA crypto without 461059191Skris removing all other RSA functionality (this is what NO_RSA does). This 461159191Skris is so (for example) those in the US can disable those operations covered 461259191Skris by the RSA patent while allowing storage and parsing of RSA keys and RSA 461359191Skris key generation. 461459191Skris [Steve Henson] 461559191Skris 461659191Skris *) Non-copying interface to BIO pairs. 461759191Skris (still largely untested) 461859191Skris [Bodo Moeller] 461959191Skris 462059191Skris *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 462159191Skris ASCII string. This was handled independently in various places before. 462259191Skris [Steve Henson] 462359191Skris 462459191Skris *) New functions UTF8_getc() and UTF8_putc() that parse and generate 462559191Skris UTF8 strings a character at a time. 462659191Skris [Steve Henson] 462759191Skris 462859191Skris *) Use client_version from client hello to select the protocol 462959191Skris (s23_srvr.c) and for RSA client key exchange verification 463059191Skris (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 463159191Skris [Bodo Moeller] 463259191Skris 463359191Skris *) Add various utility functions to handle SPKACs, these were previously 463459191Skris handled by poking round in the structure internals. Added new function 463559191Skris NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 463659191Skris print, verify and generate SPKACs. Based on an original idea from 463759191Skris Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 463859191Skris [Steve Henson] 463959191Skris 464059191Skris *) RIPEMD160 is operational on all platforms and is back in 'make test'. 464159191Skris [Andy Polyakov] 464259191Skris 464359191Skris *) Allow the config file extension section to be overwritten on the 464459191Skris command line. Based on an original idea from Massimiliano Pala 464559191Skris <madwolf@comune.modena.it>. The new option is called -extensions 464659191Skris and can be applied to ca, req and x509. Also -reqexts to override 464759191Skris the request extensions in req and -crlexts to override the crl extensions 464859191Skris in ca. 464959191Skris [Steve Henson] 465059191Skris 465159191Skris *) Add new feature to the SPKAC handling in ca. Now you can include 465259191Skris the same field multiple times by preceding it by "XXXX." for example: 465359191Skris 1.OU="Unit name 1" 465459191Skris 2.OU="Unit name 2" 465559191Skris this is the same syntax as used in the req config file. 465659191Skris [Steve Henson] 465759191Skris 465859191Skris *) Allow certificate extensions to be added to certificate requests. These 465959191Skris are specified in a 'req_extensions' option of the req section of the 466059191Skris config file. They can be printed out with the -text option to req but 466159191Skris are otherwise ignored at present. 466259191Skris [Steve Henson] 466359191Skris 466459191Skris *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 466559191Skris data read consists of only the final block it would not decrypted because 466659191Skris EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 466759191Skris A misplaced 'break' also meant the decrypted final block might not be 466859191Skris copied until the next read. 466959191Skris [Steve Henson] 467059191Skris 467159191Skris *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 467259191Skris a few extra parameters to the DH structure: these will be useful if 467359191Skris for example we want the value of 'q' or implement X9.42 DH. 467459191Skris [Steve Henson] 467559191Skris 467659191Skris *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 467759191Skris provides hooks that allow the default DSA functions or functions on a 467859191Skris "per key" basis to be replaced. This allows hardware acceleration and 467959191Skris hardware key storage to be handled without major modification to the 468059191Skris library. Also added low level modexp hooks and CRYPTO_EX structure and 468159191Skris associated functions. 468259191Skris [Steve Henson] 468359191Skris 468459191Skris *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 468559191Skris as "read only": it can't be written to and the buffer it points to will 468659191Skris not be freed. Reading from a read only BIO is much more efficient than 468759191Skris a normal memory BIO. This was added because there are several times when 468859191Skris an area of memory needs to be read from a BIO. The previous method was 468959191Skris to create a memory BIO and write the data to it, this results in two 469059191Skris copies of the data and an O(n^2) reading algorithm. There is a new 469159191Skris function BIO_new_mem_buf() which creates a read only memory BIO from 469259191Skris an area of memory. Also modified the PKCS#7 routines to use read only 469359191Skris memory BIOs. 469459191Skris [Steve Henson] 469559191Skris 469659191Skris *) Bugfix: ssl23_get_client_hello did not work properly when called in 469759191Skris state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 469859191Skris a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 469959191Skris but a retry condition occured while trying to read the rest. 470059191Skris [Bodo Moeller] 470159191Skris 470259191Skris *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 470359191Skris NID_pkcs7_encrypted by default: this was wrong since this should almost 470459191Skris always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 470559191Skris the encrypted data type: this is a more sensible place to put it and it 470659191Skris allows the PKCS#12 code to be tidied up that duplicated this 470759191Skris functionality. 470859191Skris [Steve Henson] 470959191Skris 471059191Skris *) Changed obj_dat.pl script so it takes its input and output files on 471159191Skris the command line. This should avoid shell escape redirection problems 471259191Skris under Win32. 471359191Skris [Steve Henson] 471459191Skris 471559191Skris *) Initial support for certificate extension requests, these are included 471659191Skris in things like Xenroll certificate requests. Included functions to allow 471759191Skris extensions to be obtained and added. 471859191Skris [Steve Henson] 471959191Skris 472059191Skris *) -crlf option to s_client and s_server for sending newlines as 472159191Skris CRLF (as required by many protocols). 472259191Skris [Bodo Moeller] 472359191Skris 472455714Skris Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 472555714Skris 472655714Skris *) Install libRSAglue.a when OpenSSL is built with RSAref. 472755714Skris [Ralf S. Engelschall] 472855714Skris 472955714Skris *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 473055714Skris [Andrija Antonijevic <TheAntony2@bigfoot.com>] 473155714Skris 473255714Skris *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 473355714Skris program. 473455714Skris [Steve Henson] 473555714Skris 473655714Skris *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 473755714Skris DH parameters/keys (q is lost during that conversion, but the resulting 473855714Skris DH parameters contain its length). 473955714Skris 474055714Skris For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 474155714Skris much faster than DH_generate_parameters (which creates parameters 474255714Skris where p = 2*q + 1), and also the smaller q makes DH computations 474355714Skris much more efficient (160-bit exponentiation instead of 1024-bit 474455714Skris exponentiation); so this provides a convenient way to support DHE 474555714Skris ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 474655714Skris utter importance to use 474755714Skris SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 474855714Skris or 474955714Skris SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 475055714Skris when such DH parameters are used, because otherwise small subgroup 475155714Skris attacks may become possible! 475255714Skris [Bodo Moeller] 475355714Skris 475455714Skris *) Avoid memory leak in i2d_DHparams. 475555714Skris [Bodo Moeller] 475655714Skris 475755714Skris *) Allow the -k option to be used more than once in the enc program: 475855714Skris this allows the same encrypted message to be read by multiple recipients. 475955714Skris [Steve Henson] 476055714Skris 476155714Skris *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 476255714Skris an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 476355714Skris it will always use the numerical form of the OID, even if it has a short 476455714Skris or long name. 476555714Skris [Steve Henson] 476655714Skris 476755714Skris *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 476855714Skris method only got called if p,q,dmp1,dmq1,iqmp components were present, 476955714Skris otherwise bn_mod_exp was called. In the case of hardware keys for example 477055714Skris no private key components need be present and it might store extra data 477159191Skris in the RSA structure, which cannot be accessed from bn_mod_exp. 477259191Skris By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 477359191Skris private key operations. 477455714Skris [Steve Henson] 477555714Skris 477655714Skris *) Added support for SPARC Linux. 477755714Skris [Andy Polyakov] 477855714Skris 477955714Skris *) pem_password_cb function type incompatibly changed from 478055714Skris typedef int pem_password_cb(char *buf, int size, int rwflag); 478155714Skris to 478255714Skris ....(char *buf, int size, int rwflag, void *userdata); 478355714Skris so that applications can pass data to their callbacks: 478455714Skris The PEM[_ASN1]_{read,write}... functions and macros now take an 478555714Skris additional void * argument, which is just handed through whenever 478655714Skris the password callback is called. 478759191Skris [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 478855714Skris 478955714Skris New function SSL_CTX_set_default_passwd_cb_userdata. 479055714Skris 479155714Skris Compatibility note: As many C implementations push function arguments 479255714Skris onto the stack in reverse order, the new library version is likely to 479355714Skris interoperate with programs that have been compiled with the old 479455714Skris pem_password_cb definition (PEM_whatever takes some data that 479555714Skris happens to be on the stack as its last argument, and the callback 479655714Skris just ignores this garbage); but there is no guarantee whatsoever that 479755714Skris this will work. 479855714Skris 479955714Skris *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 480055714Skris (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 480155714Skris problems not only on Windows, but also on some Unix platforms. 480255714Skris To avoid problematic command lines, these definitions are now in an 480355714Skris auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 480455714Skris for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 480555714Skris [Bodo Moeller] 480655714Skris 480755714Skris *) MIPS III/IV assembler module is reimplemented. 480855714Skris [Andy Polyakov] 480955714Skris 481055714Skris *) More DES library cleanups: remove references to srand/rand and 481155714Skris delete an unused file. 481255714Skris [Ulf M�ller] 481355714Skris 481455714Skris *) Add support for the the free Netwide assembler (NASM) under Win32, 481555714Skris since not many people have MASM (ml) and it can be hard to obtain. 481655714Skris This is currently experimental but it seems to work OK and pass all 481755714Skris the tests. Check out INSTALL.W32 for info. 481855714Skris [Steve Henson] 481955714Skris 482055714Skris *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 482155714Skris without temporary keys kept an extra copy of the server key, 482255714Skris and connections with temporary keys did not free everything in case 482355714Skris of an error. 482455714Skris [Bodo Moeller] 482555714Skris 482655714Skris *) New function RSA_check_key and new openssl rsa option -check 482755714Skris for verifying the consistency of RSA keys. 482855714Skris [Ulf Moeller, Bodo Moeller] 482955714Skris 483055714Skris *) Various changes to make Win32 compile work: 483155714Skris 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 483255714Skris 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 483355714Skris comparison" warnings. 483455714Skris 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 483555714Skris [Steve Henson] 483655714Skris 483755714Skris *) Add a debugging option to PKCS#5 v2 key generation function: when 483855714Skris you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 483955714Skris derived keys are printed to stderr. 484055714Skris [Steve Henson] 484155714Skris 484255714Skris *) Copy the flags in ASN1_STRING_dup(). 484355714Skris [Roman E. Pavlov <pre@mo.msk.ru>] 484455714Skris 484555714Skris *) The x509 application mishandled signing requests containing DSA 484655714Skris keys when the signing key was also DSA and the parameters didn't match. 484755714Skris 484855714Skris It was supposed to omit the parameters when they matched the signing key: 484955714Skris the verifying software was then supposed to automatically use the CA's 485055714Skris parameters if they were absent from the end user certificate. 485155714Skris 485255714Skris Omitting parameters is no longer recommended. The test was also 485355714Skris the wrong way round! This was probably due to unusual behaviour in 485455714Skris EVP_cmp_parameters() which returns 1 if the parameters match. 485555714Skris This meant that parameters were omitted when they *didn't* match and 485655714Skris the certificate was useless. Certificates signed with 'ca' didn't have 485755714Skris this bug. 485855714Skris [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 485955714Skris 486055714Skris *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 486155714Skris The interface is as follows: 486255714Skris Applications can use 486355714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 486455714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 486555714Skris "off" is now the default. 486655714Skris The library internally uses 486755714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 486855714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 486955714Skris to disable memory-checking temporarily. 487055714Skris 487155714Skris Some inconsistent states that previously were possible (and were 487255714Skris even the default) are now avoided. 487355714Skris 487455714Skris -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 487555714Skris with each memory chunk allocated; this is occasionally more helpful 487655714Skris than just having a counter. 487755714Skris 487855714Skris -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 487955714Skris 488055714Skris -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 488155714Skris extensions. 488255714Skris [Bodo Moeller] 488355714Skris 488455714Skris *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 488555714Skris which largely parallels "options", but is for changing API behaviour, 488655714Skris whereas "options" are about protocol behaviour. 488755714Skris Initial "mode" flags are: 488855714Skris 488955714Skris SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 489055714Skris a single record has been written. 489155714Skris SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 489255714Skris retries use the same buffer location. 489355714Skris (But all of the contents must be 489455714Skris copied!) 489555714Skris [Bodo Moeller] 489655714Skris 489779998Skris *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 489855714Skris worked. 489955714Skris 490055714Skris *) Fix problems with no-hmac etc. 490155714Skris [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 490255714Skris 490355714Skris *) New functions RSA_get_default_method(), RSA_set_method() and 490455714Skris RSA_get_method(). These allows replacement of RSA_METHODs without having 490555714Skris to mess around with the internals of an RSA structure. 490655714Skris [Steve Henson] 490755714Skris 490855714Skris *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 490955714Skris Also really enable memory leak checks in openssl.c and in some 491055714Skris test programs. 491155714Skris [Chad C. Mulligan, Bodo Moeller] 491255714Skris 491355714Skris *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 491455714Skris up the length of negative integers. This has now been simplified to just 491555714Skris store the length when it is first determined and use it later, rather 491655714Skris than trying to keep track of where data is copied and updating it to 491755714Skris point to the end. 491855714Skris [Steve Henson, reported by Brien Wheeler 491955714Skris <bwheeler@authentica-security.com>] 492055714Skris 492155714Skris *) Add a new function PKCS7_signatureVerify. This allows the verification 492255714Skris of a PKCS#7 signature but with the signing certificate passed to the 492355714Skris function itself. This contrasts with PKCS7_dataVerify which assumes the 492455714Skris certificate is present in the PKCS#7 structure. This isn't always the 492555714Skris case: certificates can be omitted from a PKCS#7 structure and be 492655714Skris distributed by "out of band" means (such as a certificate database). 492755714Skris [Steve Henson] 492855714Skris 492955714Skris *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 493055714Skris function prototypes in pem.h, also change util/mkdef.pl to add the 493155714Skris necessary function names. 493255714Skris [Steve Henson] 493355714Skris 493455714Skris *) mk1mf.pl (used by Windows builds) did not properly read the 493555714Skris options set by Configure in the top level Makefile, and Configure 493655714Skris was not even able to write more than one option correctly. 493755714Skris Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 493855714Skris [Bodo Moeller] 493955714Skris 494055714Skris *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 494155714Skris file to be loaded from a BIO or FILE pointer. The BIO version will 494255714Skris for example allow memory BIOs to contain config info. 494355714Skris [Steve Henson] 494455714Skris 494555714Skris *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 494655714Skris Whoever hopes to achieve shared-library compatibility across versions 494755714Skris must use this, not the compile-time macro. 494855714Skris (Exercise 0.9.4: Which is the minimum library version required by 494955714Skris such programs?) 495055714Skris Note: All this applies only to multi-threaded programs, others don't 495155714Skris need locks. 495255714Skris [Bodo Moeller] 495355714Skris 495455714Skris *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 495555714Skris through a BIO pair triggered the default case, i.e. 495655714Skris SSLerr(...,SSL_R_UNKNOWN_STATE). 495755714Skris [Bodo Moeller] 495855714Skris 495955714Skris *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 496055714Skris can use the SSL library even if none of the specific BIOs is 496155714Skris appropriate. 496255714Skris [Bodo Moeller] 496355714Skris 496455714Skris *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 496555714Skris for the encoded length. 496655714Skris [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 496755714Skris 496855714Skris *) Add initial documentation of the X509V3 functions. 496955714Skris [Steve Henson] 497055714Skris 497155714Skris *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 497255714Skris PEM_write_bio_PKCS8PrivateKey() that are equivalent to 497355714Skris PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 497455714Skris secure PKCS#8 private key format with a high iteration count. 497555714Skris [Steve Henson] 497655714Skris 497755714Skris *) Fix determination of Perl interpreter: A perl or perl5 497855714Skris _directory_ in $PATH was also accepted as the interpreter. 497955714Skris [Ralf S. Engelschall] 498055714Skris 498155714Skris *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 498255714Skris wrong with it but it was very old and did things like calling 498355714Skris PEM_ASN1_read() directly and used MD5 for the hash not to mention some 498455714Skris unusual formatting. 498555714Skris [Steve Henson] 498655714Skris 498755714Skris *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 498855714Skris to use the new extension code. 498955714Skris [Steve Henson] 499055714Skris 499155714Skris *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 499255714Skris with macros. This should make it easier to change their form, add extra 499355714Skris arguments etc. Fix a few PEM prototypes which didn't have cipher as a 499455714Skris constant. 499555714Skris [Steve Henson] 499655714Skris 499755714Skris *) Add to configuration table a new entry that can specify an alternative 499855714Skris name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 499955714Skris according to Mark Crispin <MRC@Panda.COM>. 500055714Skris [Bodo Moeller] 500155714Skris 500255714Skris#if 0 500355714Skris *) DES CBC did not update the IV. Weird. 500455714Skris [Ben Laurie] 500555714Skris#else 500655714Skris des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 500755714Skris Changing the behaviour of the former might break existing programs -- 500855714Skris where IV updating is needed, des_ncbc_encrypt can be used. 500955714Skris#endif 501055714Skris 501155714Skris *) When bntest is run from "make test" it drives bc to check its 501255714Skris calculations, as well as internally checking them. If an internal check 501355714Skris fails, it needs to cause bc to give a non-zero result or make test carries 501455714Skris on without noticing the failure. Fixed. 501555714Skris [Ben Laurie] 501655714Skris 501755714Skris *) DES library cleanups. 501855714Skris [Ulf M�ller] 501955714Skris 502055714Skris *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 502155714Skris used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 502255714Skris ciphers. NOTE: although the key derivation function has been verified 502355714Skris against some published test vectors it has not been extensively tested 502455714Skris yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 502555714Skris of v2.0. 502655714Skris [Steve Henson] 502755714Skris 502855714Skris *) Instead of "mkdir -p", which is not fully portable, use new 502955714Skris Perl script "util/mkdir-p.pl". 503055714Skris [Bodo Moeller] 503155714Skris 503255714Skris *) Rewrite the way password based encryption (PBE) is handled. It used to 503355714Skris assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 503455714Skris structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 503555714Skris but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 503655714Skris the 'parameter' field of the AlgorithmIdentifier is passed to the 503755714Skris underlying key generation function so it must do its own ASN1 parsing. 503855714Skris This has also changed the EVP_PBE_CipherInit() function which now has a 503955714Skris 'parameter' argument instead of literal salt and iteration count values 504055714Skris and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 504155714Skris [Steve Henson] 504255714Skris 504355714Skris *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 504455714Skris and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 504555714Skris Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 504655714Skris KEY" because this clashed with PKCS#8 unencrypted string. Since this 504755714Skris value was just used as a "magic string" and not used directly its 504855714Skris value doesn't matter. 504955714Skris [Steve Henson] 505055714Skris 505155714Skris *) Introduce some semblance of const correctness to BN. Shame C doesn't 505255714Skris support mutable. 505355714Skris [Ben Laurie] 505455714Skris 505555714Skris *) "linux-sparc64" configuration (ultrapenguin). 505655714Skris [Ray Miller <ray.miller@oucs.ox.ac.uk>] 505755714Skris "linux-sparc" configuration. 505855714Skris [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 505955714Skris 506055714Skris *) config now generates no-xxx options for missing ciphers. 506155714Skris [Ulf M�ller] 506255714Skris 506355714Skris *) Support the EBCDIC character set (work in progress). 506455714Skris File ebcdic.c not yet included because it has a different license. 506555714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 506655714Skris 506755714Skris *) Support BS2000/OSD-POSIX. 506855714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 506955714Skris 507055714Skris *) Make callbacks for key generation use void * instead of char *. 507155714Skris [Ben Laurie] 507255714Skris 507355714Skris *) Make S/MIME samples compile (not yet tested). 507455714Skris [Ben Laurie] 507555714Skris 507655714Skris *) Additional typesafe stacks. 507755714Skris [Ben Laurie] 507855714Skris 507955714Skris *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 508055714Skris [Bodo Moeller] 508155714Skris 508255714Skris 508355714Skris Changes between 0.9.3 and 0.9.3a [29 May 1999] 508455714Skris 508555714Skris *) New configuration variant "sco5-gcc". 508655714Skris 508755714Skris *) Updated some demos. 508855714Skris [Sean O Riordain, Wade Scholine] 508955714Skris 509055714Skris *) Add missing BIO_free at exit of pkcs12 application. 509155714Skris [Wu Zhigang] 509255714Skris 509355714Skris *) Fix memory leak in conf.c. 509455714Skris [Steve Henson] 509555714Skris 509655714Skris *) Updates for Win32 to assembler version of MD5. 509755714Skris [Steve Henson] 509855714Skris 509955714Skris *) Set #! path to perl in apps/der_chop to where we found it 510055714Skris instead of using a fixed path. 510155714Skris [Bodo Moeller] 510255714Skris 510355714Skris *) SHA library changes for irix64-mips4-cc. 510455714Skris [Andy Polyakov] 510555714Skris 510655714Skris *) Improvements for VMS support. 510755714Skris [Richard Levitte] 510855714Skris 510955714Skris 511055714Skris Changes between 0.9.2b and 0.9.3 [24 May 1999] 511155714Skris 511255714Skris *) Bignum library bug fix. IRIX 6 passes "make test" now! 511355714Skris This also avoids the problems with SC4.2 and unpatched SC5. 511455714Skris [Andy Polyakov <appro@fy.chalmers.se>] 511555714Skris 511655714Skris *) New functions sk_num, sk_value and sk_set to replace the previous macros. 511755714Skris These are required because of the typesafe stack would otherwise break 511855714Skris existing code. If old code used a structure member which used to be STACK 511955714Skris and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 512055714Skris sk_num or sk_value it would produce an error because the num, data members 512155714Skris are not present in STACK_OF. Now it just produces a warning. sk_set 512255714Skris replaces the old method of assigning a value to sk_value 512355714Skris (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 512455714Skris that does this will no longer work (and should use sk_set instead) but 512555714Skris this could be regarded as a "questionable" behaviour anyway. 512655714Skris [Steve Henson] 512755714Skris 512855714Skris *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 512955714Skris correctly handle encrypted S/MIME data. 513055714Skris [Steve Henson] 513155714Skris 513255714Skris *) Change type of various DES function arguments from des_cblock 513355714Skris (which means, in function argument declarations, pointer to char) 513455714Skris to des_cblock * (meaning pointer to array with 8 char elements), 513555714Skris which allows the compiler to do more typechecking; it was like 513655714Skris that back in SSLeay, but with lots of ugly casts. 513755714Skris 513855714Skris Introduce new type const_des_cblock. 513955714Skris [Bodo Moeller] 514055714Skris 514155714Skris *) Reorganise the PKCS#7 library and get rid of some of the more obvious 514255714Skris problems: find RecipientInfo structure that matches recipient certificate 514355714Skris and initialise the ASN1 structures properly based on passed cipher. 514455714Skris [Steve Henson] 514555714Skris 514655714Skris *) Belatedly make the BN tests actually check the results. 514755714Skris [Ben Laurie] 514855714Skris 514955714Skris *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 515055714Skris to and from BNs: it was completely broken. New compilation option 515155714Skris NEG_PUBKEY_BUG to allow for some broken certificates that encode public 515255714Skris key elements as negative integers. 515355714Skris [Steve Henson] 515455714Skris 515555714Skris *) Reorganize and speed up MD5. 515655714Skris [Andy Polyakov <appro@fy.chalmers.se>] 515755714Skris 515855714Skris *) VMS support. 515955714Skris [Richard Levitte <richard@levitte.org>] 516055714Skris 516155714Skris *) New option -out to asn1parse to allow the parsed structure to be 516255714Skris output to a file. This is most useful when combined with the -strparse 516355714Skris option to examine the output of things like OCTET STRINGS. 516455714Skris [Steve Henson] 516555714Skris 516655714Skris *) Make SSL library a little more fool-proof by not requiring any longer 516755714Skris that SSL_set_{accept,connect}_state be called before 516855714Skris SSL_{accept,connect} may be used (SSL_set_..._state is omitted 516955714Skris in many applications because usually everything *appeared* to work as 517055714Skris intended anyway -- now it really works as intended). 517155714Skris [Bodo Moeller] 517255714Skris 517355714Skris *) Move openssl.cnf out of lib/. 517455714Skris [Ulf M�ller] 517555714Skris 517655714Skris *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 517755714Skris -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 517855714Skris -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 517955714Skris [Ralf S. Engelschall] 518055714Skris 518155714Skris *) Various fixes to the EVP and PKCS#7 code. It may now be able to 518255714Skris handle PKCS#7 enveloped data properly. 518355714Skris [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 518455714Skris 518555714Skris *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 518655714Skris copying pointers. The cert_st handling is changed by this in 518755714Skris various ways (and thus what used to be known as ctx->default_cert 518855714Skris is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 518955714Skris any longer when s->cert does not give us what we need). 519055714Skris ssl_cert_instantiate becomes obsolete by this change. 519155714Skris As soon as we've got the new code right (possibly it already is?), 519255714Skris we have solved a couple of bugs of the earlier code where s->cert 519355714Skris was used as if it could not have been shared with other SSL structures. 519455714Skris 519555714Skris Note that using the SSL API in certain dirty ways now will result 519655714Skris in different behaviour than observed with earlier library versions: 519755714Skris Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 519855714Skris does not influence s as it used to. 519955714Skris 520055714Skris In order to clean up things more thoroughly, inside SSL_SESSION 520155714Skris we don't use CERT any longer, but a new structure SESS_CERT 520255714Skris that holds per-session data (if available); currently, this is 520355714Skris the peer's certificate chain and, for clients, the server's certificate 520455714Skris and temporary key. CERT holds only those values that can have 520555714Skris meaningful defaults in an SSL_CTX. 520655714Skris [Bodo Moeller] 520755714Skris 520855714Skris *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 520955714Skris from the internal representation. Various PKCS#7 fixes: remove some 521055714Skris evil casts and set the enc_dig_alg field properly based on the signing 521155714Skris key type. 521255714Skris [Steve Henson] 521355714Skris 521455714Skris *) Allow PKCS#12 password to be set from the command line or the 521555714Skris environment. Let 'ca' get its config file name from the environment 521655714Skris variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 521755714Skris and 'x509'). 521855714Skris [Steve Henson] 521955714Skris 522055714Skris *) Allow certificate policies extension to use an IA5STRING for the 522155714Skris organization field. This is contrary to the PKIX definition but 522255714Skris VeriSign uses it and IE5 only recognises this form. Document 'x509' 522355714Skris extension option. 522455714Skris [Steve Henson] 522555714Skris 522655714Skris *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 522755714Skris without disallowing inline assembler and the like for non-pedantic builds. 522855714Skris [Ben Laurie] 522955714Skris 523055714Skris *) Support Borland C++ builder. 523155714Skris [Janez Jere <jj@void.si>, modified by Ulf M�ller] 523255714Skris 523355714Skris *) Support Mingw32. 523455714Skris [Ulf M�ller] 523555714Skris 523655714Skris *) SHA-1 cleanups and performance enhancements. 523755714Skris [Andy Polyakov <appro@fy.chalmers.se>] 523855714Skris 523955714Skris *) Sparc v8plus assembler for the bignum library. 524055714Skris [Andy Polyakov <appro@fy.chalmers.se>] 524155714Skris 524255714Skris *) Accept any -xxx and +xxx compiler options in Configure. 524355714Skris [Ulf M�ller] 524455714Skris 524555714Skris *) Update HPUX configuration. 524655714Skris [Anonymous] 524755714Skris 524855714Skris *) Add missing sk_<type>_unshift() function to safestack.h 524955714Skris [Ralf S. Engelschall] 525055714Skris 525155714Skris *) New function SSL_CTX_use_certificate_chain_file that sets the 525255714Skris "extra_cert"s in addition to the certificate. (This makes sense 525355714Skris only for "PEM" format files, as chains as a whole are not 525455714Skris DER-encoded.) 525555714Skris [Bodo Moeller] 525655714Skris 525755714Skris *) Support verify_depth from the SSL API. 525855714Skris x509_vfy.c had what can be considered an off-by-one-error: 525955714Skris Its depth (which was not part of the external interface) 526055714Skris was actually counting the number of certificates in a chain; 526155714Skris now it really counts the depth. 526255714Skris [Bodo Moeller] 526355714Skris 526455714Skris *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 526555714Skris instead of X509err, which often resulted in confusing error 526655714Skris messages since the error codes are not globally unique 526755714Skris (e.g. an alleged error in ssl3_accept when a certificate 526855714Skris didn't match the private key). 526955714Skris 527055714Skris *) New function SSL_CTX_set_session_id_context that allows to set a default 527155714Skris value (so that you don't need SSL_set_session_id_context for each 527255714Skris connection using the SSL_CTX). 527355714Skris [Bodo Moeller] 527455714Skris 527555714Skris *) OAEP decoding bug fix. 527655714Skris [Ulf M�ller] 527755714Skris 527855714Skris *) Support INSTALL_PREFIX for package builders, as proposed by 527955714Skris David Harris. 528055714Skris [Bodo Moeller] 528155714Skris 528255714Skris *) New Configure options "threads" and "no-threads". For systems 528355714Skris where the proper compiler options are known (currently Solaris 528455714Skris and Linux), "threads" is the default. 528555714Skris [Bodo Moeller] 528655714Skris 528755714Skris *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 528855714Skris [Bodo Moeller] 528955714Skris 529055714Skris *) Install various scripts to $(OPENSSLDIR)/misc, not to 529155714Skris $(INSTALLTOP)/bin -- they shouldn't clutter directories 529255714Skris such as /usr/local/bin. 529355714Skris [Bodo Moeller] 529455714Skris 529555714Skris *) "make linux-shared" to build shared libraries. 529655714Skris [Niels Poppe <niels@netbox.org>] 529755714Skris 529855714Skris *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 529955714Skris [Ulf M�ller] 530055714Skris 530155714Skris *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 530255714Skris extension adding in x509 utility. 530355714Skris [Steve Henson] 530455714Skris 530555714Skris *) Remove NOPROTO sections and error code comments. 530655714Skris [Ulf M�ller] 530755714Skris 530855714Skris *) Partial rewrite of the DEF file generator to now parse the ANSI 530955714Skris prototypes. 531055714Skris [Steve Henson] 531155714Skris 531255714Skris *) New Configure options --prefix=DIR and --openssldir=DIR. 531355714Skris [Ulf M�ller] 531455714Skris 531555714Skris *) Complete rewrite of the error code script(s). It is all now handled 531655714Skris by one script at the top level which handles error code gathering, 531755714Skris header rewriting and C source file generation. It should be much better 531855714Skris than the old method: it now uses a modified version of Ulf's parser to 531955714Skris read the ANSI prototypes in all header files (thus the old K&R definitions 532055714Skris aren't needed for error creation any more) and do a better job of 532155714Skris translating function codes into names. The old 'ASN1 error code imbedded 532255714Skris in a comment' is no longer necessary and it doesn't use .err files which 532355714Skris have now been deleted. Also the error code call doesn't have to appear all 532455714Skris on one line (which resulted in some large lines...). 532555714Skris [Steve Henson] 532655714Skris 532755714Skris *) Change #include filenames from <foo.h> to <openssl/foo.h>. 532855714Skris [Bodo Moeller] 532955714Skris 533055714Skris *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 533155714Skris 0 (which usually indicates a closed connection), but continue reading. 533255714Skris [Bodo Moeller] 533355714Skris 533455714Skris *) Fix some race conditions. 533555714Skris [Bodo Moeller] 533655714Skris 533755714Skris *) Add support for CRL distribution points extension. Add Certificate 533855714Skris Policies and CRL distribution points documentation. 533955714Skris [Steve Henson] 534055714Skris 534155714Skris *) Move the autogenerated header file parts to crypto/opensslconf.h. 534255714Skris [Ulf M�ller] 534355714Skris 534455714Skris *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 534555714Skris 8 of keying material. Merlin has also confirmed interop with this fix 534655714Skris between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 534755714Skris [Merlin Hughes <merlin@baltimore.ie>] 534855714Skris 534955714Skris *) Fix lots of warnings. 535055714Skris [Richard Levitte <levitte@stacken.kth.se>] 535155714Skris 535255714Skris *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 535355714Skris the directory spec didn't end with a LIST_SEPARATOR_CHAR. 535455714Skris [Richard Levitte <levitte@stacken.kth.se>] 535555714Skris 535655714Skris *) Fix problems with sizeof(long) == 8. 535755714Skris [Andy Polyakov <appro@fy.chalmers.se>] 535855714Skris 535955714Skris *) Change functions to ANSI C. 536055714Skris [Ulf M�ller] 536155714Skris 536255714Skris *) Fix typos in error codes. 536355714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 536455714Skris 536555714Skris *) Remove defunct assembler files from Configure. 536655714Skris [Ulf M�ller] 536755714Skris 536855714Skris *) SPARC v8 assembler BIGNUM implementation. 536955714Skris [Andy Polyakov <appro@fy.chalmers.se>] 537055714Skris 537155714Skris *) Support for Certificate Policies extension: both print and set. 537255714Skris Various additions to support the r2i method this uses. 537355714Skris [Steve Henson] 537455714Skris 537555714Skris *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 537655714Skris return a const string when you are expecting an allocated buffer. 537755714Skris [Ben Laurie] 537855714Skris 537955714Skris *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 538055714Skris types DirectoryString and DisplayText. 538155714Skris [Steve Henson] 538255714Skris 538355714Skris *) Add code to allow r2i extensions to access the configuration database, 538455714Skris add an LHASH database driver and add several ctx helper functions. 538555714Skris [Steve Henson] 538655714Skris 538755714Skris *) Fix an evil bug in bn_expand2() which caused various BN functions to 538855714Skris fail when they extended the size of a BIGNUM. 538955714Skris [Steve Henson] 539055714Skris 539155714Skris *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 539255714Skris support typesafe stack. 539355714Skris [Steve Henson] 539455714Skris 539555714Skris *) Fix typo in SSL_[gs]et_options(). 539655714Skris [Nils Frostberg <nils@medcom.se>] 539755714Skris 539855714Skris *) Delete various functions and files that belonged to the (now obsolete) 539955714Skris old X509V3 handling code. 540055714Skris [Steve Henson] 540155714Skris 540255714Skris *) New Configure option "rsaref". 540355714Skris [Ulf M�ller] 540455714Skris 540555714Skris *) Don't auto-generate pem.h. 540655714Skris [Bodo Moeller] 540755714Skris 540855714Skris *) Introduce type-safe ASN.1 SETs. 540955714Skris [Ben Laurie] 541055714Skris 541155714Skris *) Convert various additional casted stacks to type-safe STACK_OF() variants. 541255714Skris [Ben Laurie, Ralf S. Engelschall, Steve Henson] 541355714Skris 541455714Skris *) Introduce type-safe STACKs. This will almost certainly break lots of code 541555714Skris that links with OpenSSL (well at least cause lots of warnings), but fear 541655714Skris not: the conversion is trivial, and it eliminates loads of evil casts. A 541755714Skris few STACKed things have been converted already. Feel free to convert more. 541855714Skris In the fullness of time, I'll do away with the STACK type altogether. 541955714Skris [Ben Laurie] 542055714Skris 542155714Skris *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 542255714Skris specified in <certfile> by updating the entry in the index.txt file. 542355714Skris This way one no longer has to edit the index.txt file manually for 542455714Skris revoking a certificate. The -revoke option does the gory details now. 542555714Skris [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 542655714Skris 542755714Skris *) Fix `openssl crl -noout -text' combination where `-noout' killed the 542855714Skris `-text' option at all and this way the `-noout -text' combination was 542955714Skris inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 543055714Skris [Ralf S. Engelschall] 543155714Skris 543255714Skris *) Make sure a corresponding plain text error message exists for the 543355714Skris X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 543455714Skris verify callback function determined that a certificate was revoked. 543555714Skris [Ralf S. Engelschall] 543655714Skris 543755714Skris *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 543855714Skris ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 543955714Skris all available cipers including rc5, which was forgotten until now. 544055714Skris In order to let the testing shell script know which algorithms 544155714Skris are available, a new (up to now undocumented) command 544255714Skris "openssl list-cipher-commands" is used. 544355714Skris [Bodo Moeller] 544455714Skris 544555714Skris *) Bugfix: s_client occasionally would sleep in select() when 544655714Skris it should have checked SSL_pending() first. 544755714Skris [Bodo Moeller] 544855714Skris 544955714Skris *) New functions DSA_do_sign and DSA_do_verify to provide access to 545055714Skris the raw DSA values prior to ASN.1 encoding. 545155714Skris [Ulf M�ller] 545255714Skris 545355714Skris *) Tweaks to Configure 545455714Skris [Niels Poppe <niels@netbox.org>] 545555714Skris 545655714Skris *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 545755714Skris yet... 545855714Skris [Steve Henson] 545955714Skris 546055714Skris *) New variables $(RANLIB) and $(PERL) in the Makefiles. 546155714Skris [Ulf M�ller] 546255714Skris 546355714Skris *) New config option to avoid instructions that are illegal on the 80386. 546455714Skris The default code is faster, but requires at least a 486. 546555714Skris [Ulf M�ller] 546655714Skris 546755714Skris *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 546855714Skris SSL2_SERVER_VERSION (not used at all) macros, which are now the 546955714Skris same as SSL2_VERSION anyway. 547055714Skris [Bodo Moeller] 547155714Skris 547255714Skris *) New "-showcerts" option for s_client. 547355714Skris [Bodo Moeller] 547455714Skris 547555714Skris *) Still more PKCS#12 integration. Add pkcs12 application to openssl 547655714Skris application. Various cleanups and fixes. 547755714Skris [Steve Henson] 547855714Skris 547955714Skris *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 548055714Skris modify error routines to work internally. Add error codes and PBE init 548155714Skris to library startup routines. 548255714Skris [Steve Henson] 548355714Skris 548455714Skris *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 548555714Skris packing functions to asn1 and evp. Changed function names and error 548655714Skris codes along the way. 548755714Skris [Steve Henson] 548855714Skris 548955714Skris *) PKCS12 integration: and so it begins... First of several patches to 549055714Skris slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 549155714Skris objects to objects.h 549255714Skris [Steve Henson] 549355714Skris 549455714Skris *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 549555714Skris and display support for Thawte strong extranet extension. 549655714Skris [Steve Henson] 549755714Skris 549855714Skris *) Add LinuxPPC support. 549955714Skris [Jeff Dubrule <igor@pobox.org>] 550055714Skris 550155714Skris *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 550255714Skris bn_div_words in alpha.s. 550355714Skris [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 550455714Skris 550555714Skris *) Make sure the RSA OAEP test is skipped under -DRSAref because 550655714Skris OAEP isn't supported when OpenSSL is built with RSAref. 550755714Skris [Ulf Moeller <ulf@fitug.de>] 550855714Skris 550955714Skris *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 551055714Skris so they no longer are missing under -DNOPROTO. 551155714Skris [Soren S. Jorvang <soren@t.dk>] 551255714Skris 551355714Skris 551455714Skris Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 551555714Skris 551655714Skris *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 551755714Skris doesn't work when the session is reused. Coming soon! 551855714Skris [Ben Laurie] 551955714Skris 552055714Skris *) Fix a security hole, that allows sessions to be reused in the wrong 552155714Skris context thus bypassing client cert protection! All software that uses 552255714Skris client certs and session caches in multiple contexts NEEDS PATCHING to 552355714Skris allow session reuse! A fuller solution is in the works. 552455714Skris [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 552555714Skris 552655714Skris *) Some more source tree cleanups (removed obsolete files 552755714Skris crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 552855714Skris permission on "config" script to be executable) and a fix for the INSTALL 552955714Skris document. 553055714Skris [Ulf Moeller <ulf@fitug.de>] 553155714Skris 553255714Skris *) Remove some legacy and erroneous uses of malloc, free instead of 553355714Skris Malloc, Free. 553455714Skris [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 553555714Skris 553655714Skris *) Make rsa_oaep_test return non-zero on error. 553755714Skris [Ulf Moeller <ulf@fitug.de>] 553855714Skris 553955714Skris *) Add support for native Solaris shared libraries. Configure 554055714Skris solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 554155714Skris if someone would make that last step automatic. 554255714Skris [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 554355714Skris 554455714Skris *) ctx_size was not built with the right compiler during "make links". Fixed. 554555714Skris [Ben Laurie] 554655714Skris 554755714Skris *) Change the meaning of 'ALL' in the cipher list. It now means "everything 554855714Skris except NULL ciphers". This means the default cipher list will no longer 554955714Skris enable NULL ciphers. They need to be specifically enabled e.g. with 555055714Skris the string "DEFAULT:eNULL". 555155714Skris [Steve Henson] 555255714Skris 555355714Skris *) Fix to RSA private encryption routines: if p < q then it would 555455714Skris occasionally produce an invalid result. This will only happen with 555555714Skris externally generated keys because OpenSSL (and SSLeay) ensure p > q. 555655714Skris [Steve Henson] 555755714Skris 555855714Skris *) Be less restrictive and allow also `perl util/perlpath.pl 555955714Skris /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 556055714Skris because this way one can also use an interpreter named `perl5' (which is 556155714Skris usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 556255714Skris installed as `perl'). 556355714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 556455714Skris 556555714Skris *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 556655714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 556755714Skris 556855714Skris *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 556955714Skris advapi32.lib to Win32 build and change the pem test comparision 557055714Skris to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 557155714Skris suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 557255714Skris and crypto/des/ede_cbcm_enc.c. 557355714Skris [Steve Henson] 557455714Skris 557555714Skris *) DES quad checksum was broken on big-endian architectures. Fixed. 557655714Skris [Ben Laurie] 557755714Skris 557855714Skris *) Comment out two functions in bio.h that aren't implemented. Fix up the 557955714Skris Win32 test batch file so it (might) work again. The Win32 test batch file 558055714Skris is horrible: I feel ill.... 558155714Skris [Steve Henson] 558255714Skris 558355714Skris *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 558455714Skris in e_os.h. Audit of header files to check ANSI and non ANSI 558555714Skris sections: 10 functions were absent from non ANSI section and not exported 558655714Skris from Windows DLLs. Fixed up libeay.num for new functions. 558755714Skris [Steve Henson] 558855714Skris 558955714Skris *) Make `openssl version' output lines consistent. 559055714Skris [Ralf S. Engelschall] 559155714Skris 559255714Skris *) Fix Win32 symbol export lists for BIO functions: Added 559355714Skris BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 559455714Skris to ms/libeay{16,32}.def. 559555714Skris [Ralf S. Engelschall] 559655714Skris 559755714Skris *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 559855714Skris fine under Unix and passes some trivial tests I've now added. But the 559955714Skris whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 560055714Skris added to make sure no one expects that this stuff really works in the 560155714Skris OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 560255714Skris up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 560355714Skris openssl_bio.xs. 560455714Skris [Ralf S. Engelschall] 560555714Skris 560655714Skris *) Fix the generation of two part addresses in perl. 560755714Skris [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 560855714Skris 560955714Skris *) Add config entry for Linux on MIPS. 561055714Skris [John Tobey <jtobey@channel1.com>] 561155714Skris 561255714Skris *) Make links whenever Configure is run, unless we are on Windoze. 561355714Skris [Ben Laurie] 561455714Skris 561555714Skris *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 561655714Skris Currently only issuerAltName and AuthorityKeyIdentifier make any sense 561755714Skris in CRLs. 561855714Skris [Steve Henson] 561955714Skris 562055714Skris *) Add a useful kludge to allow package maintainers to specify compiler and 562155714Skris other platforms details on the command line without having to patch the 562255714Skris Configure script everytime: One now can use ``perl Configure 562355714Skris <id>:<details>'', i.e. platform ids are allowed to have details appended 562455714Skris to them (seperated by colons). This is treated as there would be a static 562555714Skris pre-configured entry in Configure's %table under key <id> with value 562655714Skris <details> and ``perl Configure <id>'' is called. So, when you want to 562755714Skris perform a quick test-compile under FreeBSD 3.1 with pgcc and without 562855714Skris assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 562955714Skris now, which overrides the FreeBSD-elf entry on-the-fly. 563055714Skris [Ralf S. Engelschall] 563155714Skris 563255714Skris *) Disable new TLS1 ciphersuites by default: they aren't official yet. 563355714Skris [Ben Laurie] 563455714Skris 563555714Skris *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 563655714Skris on the `perl Configure ...' command line. This way one can compile 563755714Skris OpenSSL libraries with Position Independent Code (PIC) which is needed 563855714Skris for linking it into DSOs. 563955714Skris [Ralf S. Engelschall] 564055714Skris 564155714Skris *) Remarkably, export ciphers were totally broken and no-one had noticed! 564255714Skris Fixed. 564355714Skris [Ben Laurie] 564455714Skris 564555714Skris *) Cleaned up the LICENSE document: The official contact for any license 564655714Skris questions now is the OpenSSL core team under openssl-core@openssl.org. 564755714Skris And add a paragraph about the dual-license situation to make sure people 564855714Skris recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 564955714Skris to the OpenSSL toolkit. 565055714Skris [Ralf S. Engelschall] 565155714Skris 565255714Skris *) General source tree makefile cleanups: Made `making xxx in yyy...' 565355714Skris display consistent in the source tree and replaced `/bin/rm' by `rm'. 565455714Skris Additonally cleaned up the `make links' target: Remove unnecessary 565555714Skris semicolons, subsequent redundant removes, inline point.sh into mklink.sh 565655714Skris to speed processing and no longer clutter the display with confusing 565755714Skris stuff. Instead only the actually done links are displayed. 565855714Skris [Ralf S. Engelschall] 565955714Skris 566055714Skris *) Permit null encryption ciphersuites, used for authentication only. It used 566155714Skris to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 566255714Skris It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 566355714Skris encryption. 566455714Skris [Ben Laurie] 566555714Skris 566655714Skris *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 566755714Skris signed attributes when verifying signatures (this would break them), 566855714Skris the detached data encoding was wrong and public keys obtained using 566955714Skris X509_get_pubkey() weren't freed. 567055714Skris [Steve Henson] 567155714Skris 567255714Skris *) Add text documentation for the BUFFER functions. Also added a work around 567355714Skris to a Win95 console bug. This was triggered by the password read stuff: the 567455714Skris last character typed gets carried over to the next fread(). If you were 567555714Skris generating a new cert request using 'req' for example then the last 567655714Skris character of the passphrase would be CR which would then enter the first 567755714Skris field as blank. 567855714Skris [Steve Henson] 567955714Skris 568055714Skris *) Added the new `Includes OpenSSL Cryptography Software' button as 568155714Skris doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 568255714Skris button and can be used by applications based on OpenSSL to show the 568355714Skris relationship to the OpenSSL project. 568455714Skris [Ralf S. Engelschall] 568555714Skris 568655714Skris *) Remove confusing variables in function signatures in files 568755714Skris ssl/ssl_lib.c and ssl/ssl.h. 568855714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 568955714Skris 569055714Skris *) Don't install bss_file.c under PREFIX/include/ 569155714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 569255714Skris 569355714Skris *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 569455714Skris functions that return function pointers and has support for NT specific 569555714Skris stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 569655714Skris #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 569755714Skris unsigned to signed types: this was killing the Win32 compile. 569855714Skris [Steve Henson] 569955714Skris 570055714Skris *) Add new certificate file to stack functions, 570155714Skris SSL_add_dir_cert_subjects_to_stack() and 570255714Skris SSL_add_file_cert_subjects_to_stack(). These largely supplant 570355714Skris SSL_load_client_CA_file(), and can be used to add multiple certs easily 570455714Skris to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 570555714Skris This means that Apache-SSL and similar packages don't have to mess around 570655714Skris to add as many CAs as they want to the preferred list. 570755714Skris [Ben Laurie] 570855714Skris 570955714Skris *) Experiment with doxygen documentation. Currently only partially applied to 571055714Skris ssl/ssl_lib.c. 571155714Skris See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 571255714Skris openssl.doxy as the configuration file. 571355714Skris [Ben Laurie] 571455714Skris 571555714Skris *) Get rid of remaining C++-style comments which strict C compilers hate. 571655714Skris [Ralf S. Engelschall, pointed out by Carlos Amengual] 571755714Skris 571855714Skris *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 571955714Skris compiled in by default: it has problems with large keys. 572055714Skris [Steve Henson] 572155714Skris 572255714Skris *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 572355714Skris DH private keys and/or callback functions which directly correspond to 572455714Skris their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 572555714Skris is needed for applications which have to configure certificates on a 572655714Skris per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 572755714Skris (e.g. s_server). 572855714Skris For the RSA certificate situation is makes no difference, but 572955714Skris for the DSA certificate situation this fixes the "no shared cipher" 573055714Skris problem where the OpenSSL cipher selection procedure failed because the 573155714Skris temporary keys were not overtaken from the context and the API provided 573255714Skris no way to reconfigure them. 573355714Skris The new functions now let applications reconfigure the stuff and they 573455714Skris are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 573555714Skris SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 573655714Skris non-public-API function ssl_cert_instantiate() is used as a helper 573755714Skris function and also to reduce code redundancy inside ssl_rsa.c. 573855714Skris [Ralf S. Engelschall] 573955714Skris 574055714Skris *) Move s_server -dcert and -dkey options out of the undocumented feature 574155714Skris area because they are useful for the DSA situation and should be 574255714Skris recognized by the users. 574355714Skris [Ralf S. Engelschall] 574455714Skris 574555714Skris *) Fix the cipher decision scheme for export ciphers: the export bits are 574655714Skris *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 574755714Skris SSL_EXP_MASK. So, the original variable has to be used instead of the 574855714Skris already masked variable. 574955714Skris [Richard Levitte <levitte@stacken.kth.se>] 575055714Skris 575155714Skris *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 575255714Skris [Richard Levitte <levitte@stacken.kth.se>] 575355714Skris 575455714Skris *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 575555714Skris from `int' to `unsigned int' because it's a length and initialized by 575655714Skris EVP_DigestFinal() which expects an `unsigned int *'. 575755714Skris [Richard Levitte <levitte@stacken.kth.se>] 575855714Skris 575955714Skris *) Don't hard-code path to Perl interpreter on shebang line of Configure 576055714Skris script. Instead use the usual Shell->Perl transition trick. 576155714Skris [Ralf S. Engelschall] 576255714Skris 576355714Skris *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 576455714Skris (in addition to RSA certificates) to match the behaviour of `openssl dsa 576555714Skris -noout -modulus' as it's already the case for `openssl rsa -noout 576655714Skris -modulus'. For RSA the -modulus is the real "modulus" while for DSA 576755714Skris currently the public key is printed (a decision which was already done by 576855714Skris `openssl dsa -modulus' in the past) which serves a similar purpose. 576955714Skris Additionally the NO_RSA no longer completely removes the whole -modulus 577055714Skris option; it now only avoids using the RSA stuff. Same applies to NO_DSA 577155714Skris now, too. 577255714Skris [Ralf S. Engelschall] 577355714Skris 577455714Skris *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 577555714Skris BIO. See the source (crypto/evp/bio_ok.c) for more info. 577655714Skris [Arne Ansper <arne@ats.cyber.ee>] 577755714Skris 577855714Skris *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 577955714Skris to be added. Now both 'req' and 'ca' can use new objects defined in the 578055714Skris config file. 578155714Skris [Steve Henson] 578255714Skris 578355714Skris *) Add cool BIO that does syslog (or event log on NT). 578455714Skris [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 578555714Skris 578655714Skris *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 578755714Skris TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 578855714Skris TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 578955714Skris Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 579055714Skris [Ben Laurie] 579155714Skris 579255714Skris *) Add preliminary config info for new extension code. 579355714Skris [Steve Henson] 579455714Skris 579555714Skris *) Make RSA_NO_PADDING really use no padding. 579655714Skris [Ulf Moeller <ulf@fitug.de>] 579755714Skris 579855714Skris *) Generate errors when private/public key check is done. 579955714Skris [Ben Laurie] 580055714Skris 580155714Skris *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 580255714Skris for some CRL extensions and new objects added. 580355714Skris [Steve Henson] 580455714Skris 580555714Skris *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 580655714Skris key usage extension and fuller support for authority key id. 580755714Skris [Steve Henson] 580855714Skris 580955714Skris *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 581055714Skris padding method for RSA, which is recommended for new applications in PKCS 581155714Skris #1 v2.0 (RFC 2437, October 1998). 581255714Skris OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 581355714Skris foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 581455714Skris against Bleichbacher's attack on RSA. 581555714Skris [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 581655714Skris Ben Laurie] 581755714Skris 581855714Skris *) Updates to the new SSL compression code 581955714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 582055714Skris 582155714Skris *) Fix so that the version number in the master secret, when passed 582255714Skris via RSA, checks that if TLS was proposed, but we roll back to SSLv3 582355714Skris (because the server will not accept higher), that the version number 582455714Skris is 0x03,0x01, not 0x03,0x00 582555714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 582655714Skris 582755714Skris *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 582855714Skris leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 582955714Skris in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 583055714Skris [Steve Henson] 583155714Skris 583255714Skris *) Support for RAW extensions where an arbitrary extension can be 583355714Skris created by including its DER encoding. See apps/openssl.cnf for 583455714Skris an example. 583555714Skris [Steve Henson] 583655714Skris 583755714Skris *) Make sure latest Perl versions don't interpret some generated C array 583855714Skris code as Perl array code in the crypto/err/err_genc.pl script. 583955714Skris [Lars Weber <3weber@informatik.uni-hamburg.de>] 584055714Skris 584155714Skris *) Modify ms/do_ms.bat to not generate assembly language makefiles since 584255714Skris not many people have the assembler. Various Win32 compilation fixes and 584355714Skris update to the INSTALL.W32 file with (hopefully) more accurate Win32 584455714Skris build instructions. 584555714Skris [Steve Henson] 584655714Skris 584755714Skris *) Modify configure script 'Configure' to automatically create crypto/date.h 584855714Skris file under Win32 and also build pem.h from pem.org. New script 584955714Skris util/mkfiles.pl to create the MINFO file on environments that can't do a 585055714Skris 'make files': perl util/mkfiles.pl >MINFO should work. 585155714Skris [Steve Henson] 585255714Skris 585355714Skris *) Major rework of DES function declarations, in the pursuit of correctness 585455714Skris and purity. As a result, many evil casts evaporated, and some weirdness, 585555714Skris too. You may find this causes warnings in your code. Zapping your evil 585655714Skris casts will probably fix them. Mostly. 585755714Skris [Ben Laurie] 585855714Skris 585955714Skris *) Fix for a typo in asn1.h. Bug fix to object creation script 586055714Skris obj_dat.pl. It considered a zero in an object definition to mean 586155714Skris "end of object": none of the objects in objects.h have any zeros 586255714Skris so it wasn't spotted. 586355714Skris [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 586455714Skris 586555714Skris *) Add support for Triple DES Cipher Block Chaining with Output Feedback 586655714Skris Masking (CBCM). In the absence of test vectors, the best I have been able 586755714Skris to do is check that the decrypt undoes the encrypt, so far. Send me test 586855714Skris vectors if you have them. 586955714Skris [Ben Laurie] 587055714Skris 587155714Skris *) Correct calculation of key length for export ciphers (too much space was 587255714Skris allocated for null ciphers). This has not been tested! 587355714Skris [Ben Laurie] 587455714Skris 587555714Skris *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 587655714Skris message is now correct (it understands "crypto" and "ssl" on its 587755714Skris command line). There is also now an "update" option. This will update 587855714Skris the util/ssleay.num and util/libeay.num files with any new functions. 587955714Skris If you do a: 588055714Skris perl util/mkdef.pl crypto ssl update 588155714Skris it will update them. 588255714Skris [Steve Henson] 588355714Skris 588455714Skris *) Overhauled the Perl interface (perl/*): 588555714Skris - ported BN stuff to OpenSSL's different BN library 588655714Skris - made the perl/ source tree CVS-aware 588755714Skris - renamed the package from SSLeay to OpenSSL (the files still contain 588855714Skris their history because I've copied them in the repository) 588955714Skris - removed obsolete files (the test scripts will be replaced 589055714Skris by better Test::Harness variants in the future) 589155714Skris [Ralf S. Engelschall] 589255714Skris 589355714Skris *) First cut for a very conservative source tree cleanup: 589455714Skris 1. merge various obsolete readme texts into doc/ssleay.txt 589555714Skris where we collect the old documents and readme texts. 589655714Skris 2. remove the first part of files where I'm already sure that we no 589755714Skris longer need them because of three reasons: either they are just temporary 589855714Skris files which were left by Eric or they are preserved original files where 589955714Skris I've verified that the diff is also available in the CVS via "cvs diff 590055714Skris -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 590155714Skris the crypto/md/ stuff). 590255714Skris [Ralf S. Engelschall] 590355714Skris 590455714Skris *) More extension code. Incomplete support for subject and issuer alt 590555714Skris name, issuer and authority key id. Change the i2v function parameters 590655714Skris and add an extra 'crl' parameter in the X509V3_CTX structure: guess 590755714Skris what that's for :-) Fix to ASN1 macro which messed up 590855714Skris IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 590955714Skris [Steve Henson] 591055714Skris 591155714Skris *) Preliminary support for ENUMERATED type. This is largely copied from the 591255714Skris INTEGER code. 591355714Skris [Steve Henson] 591455714Skris 591555714Skris *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 591655714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 591755714Skris 591855714Skris *) Make sure `make rehash' target really finds the `openssl' program. 591955714Skris [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 592055714Skris 592155714Skris *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 592255714Skris like to hear about it if this slows down other processors. 592355714Skris [Ben Laurie] 592455714Skris 592555714Skris *) Add CygWin32 platform information to Configure script. 592655714Skris [Alan Batie <batie@aahz.jf.intel.com>] 592755714Skris 592855714Skris *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 592955714Skris [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 593055714Skris 593155714Skris *) New program nseq to manipulate netscape certificate sequences 593255714Skris [Steve Henson] 593355714Skris 593455714Skris *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 593555714Skris few typos. 593655714Skris [Steve Henson] 593755714Skris 593855714Skris *) Fixes to BN code. Previously the default was to define BN_RECURSION 593955714Skris but the BN code had some problems that would cause failures when 594055714Skris doing certificate verification and some other functions. 594155714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 594255714Skris 594355714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 594455714Skris [Steve Henson] 594555714Skris 594655714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 594755714Skris [Steve Henson] 594855714Skris 594955714Skris *) Add several PKIX and private extended key usage OIDs. 595055714Skris [Steve Henson] 595155714Skris 595255714Skris *) Modify the 'ca' program to handle the new extension code. Modify 595355714Skris openssl.cnf for new extension format, add comments. 595455714Skris [Steve Henson] 595555714Skris 595655714Skris *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 595755714Skris and add a sample to openssl.cnf so req -x509 now adds appropriate 595855714Skris CA extensions. 595955714Skris [Steve Henson] 596055714Skris 596155714Skris *) Continued X509 V3 changes. Add to other makefiles, integrate with the 596255714Skris error code, add initial support to X509_print() and x509 application. 596355714Skris [Steve Henson] 596455714Skris 596555714Skris *) Takes a deep breath and start addding X509 V3 extension support code. Add 596655714Skris files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 596755714Skris stuff is currently isolated and isn't even compiled yet. 596855714Skris [Steve Henson] 596955714Skris 597055714Skris *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 597155714Skris ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 597255714Skris Removed the versions check from X509 routines when loading extensions: 597355714Skris this allows certain broken certificates that don't set the version 597455714Skris properly to be processed. 597555714Skris [Steve Henson] 597655714Skris 597755714Skris *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 597855714Skris Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 597955714Skris can still be regenerated with "make depend". 598055714Skris [Ben Laurie] 598155714Skris 598255714Skris *) Spelling mistake in C version of CAST-128. 598355714Skris [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 598455714Skris 598555714Skris *) Changes to the error generation code. The perl script err-code.pl 598655714Skris now reads in the old error codes and retains the old numbers, only 598755714Skris adding new ones if necessary. It also only changes the .err files if new 598855714Skris codes are added. The makefiles have been modified to only insert errors 598955714Skris when needed (to avoid needlessly modifying header files). This is done 599055714Skris by only inserting errors if the .err file is newer than the auto generated 599155714Skris C file. To rebuild all the error codes from scratch (the old behaviour) 599255714Skris either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 599355714Skris or delete all the .err files. 599455714Skris [Steve Henson] 599555714Skris 599655714Skris *) CAST-128 was incorrectly implemented for short keys. The C version has 599755714Skris been fixed, but is untested. The assembler versions are also fixed, but 599855714Skris new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 599955714Skris to regenerate it if needed. 600055714Skris [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 600155714Skris Hagino <itojun@kame.net>] 600255714Skris 600355714Skris *) File was opened incorrectly in randfile.c. 600455714Skris [Ulf M�ller <ulf@fitug.de>] 600555714Skris 600655714Skris *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 600755714Skris functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 600855714Skris GeneralizedTime. ASN1_TIME is the proper type used in certificates et 600955714Skris al: it's just almost always a UTCTime. Note this patch adds new error 601055714Skris codes so do a "make errors" if there are problems. 601155714Skris [Steve Henson] 601255714Skris 601355714Skris *) Correct Linux 1 recognition in config. 601455714Skris [Ulf M�ller <ulf@fitug.de>] 601555714Skris 601655714Skris *) Remove pointless MD5 hash when using DSA keys in ca. 601755714Skris [Anonymous <nobody@replay.com>] 601855714Skris 601955714Skris *) Generate an error if given an empty string as a cert directory. Also 602055714Skris generate an error if handed NULL (previously returned 0 to indicate an 602155714Skris error, but didn't set one). 602255714Skris [Ben Laurie, reported by Anonymous <nobody@replay.com>] 602355714Skris 602455714Skris *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 602555714Skris [Ben Laurie] 602655714Skris 602755714Skris *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 602855714Skris parameters. This was causing a warning which killed off the Win32 compile. 602955714Skris [Steve Henson] 603055714Skris 603155714Skris *) Remove C++ style comments from crypto/bn/bn_local.h. 603255714Skris [Neil Costigan <neil.costigan@celocom.com>] 603355714Skris 603455714Skris *) The function OBJ_txt2nid was broken. It was supposed to return a nid 603555714Skris based on a text string, looking up short and long names and finally 603655714Skris "dot" format. The "dot" format stuff didn't work. Added new function 603755714Skris OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 603855714Skris OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 603955714Skris OID is not part of the table. 604055714Skris [Steve Henson] 604155714Skris 604255714Skris *) Add prototypes to X509 lookup/verify methods, fixing a bug in 604355714Skris X509_LOOKUP_by_alias(). 604455714Skris [Ben Laurie] 604555714Skris 604655714Skris *) Sort openssl functions by name. 604755714Skris [Ben Laurie] 604855714Skris 604955714Skris *) Get the gendsa program working (hopefully) and add it to app list. Remove 605055714Skris encryption from sample DSA keys (in case anyone is interested the password 605155714Skris was "1234"). 605255714Skris [Steve Henson] 605355714Skris 605455714Skris *) Make _all_ *_free functions accept a NULL pointer. 605555714Skris [Frans Heymans <fheymans@isaserver.be>] 605655714Skris 605755714Skris *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 605855714Skris NULL pointers. 605955714Skris [Anonymous <nobody@replay.com>] 606055714Skris 606155714Skris *) s_server should send the CAfile as acceptable CAs, not its own cert. 606255714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 606355714Skris 606455714Skris *) Don't blow it for numeric -newkey arguments to apps/req. 606555714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 606655714Skris 606755714Skris *) Temp key "for export" tests were wrong in s3_srvr.c. 606855714Skris [Anonymous <nobody@replay.com>] 606955714Skris 607055714Skris *) Add prototype for temp key callback functions 607155714Skris SSL_CTX_set_tmp_{rsa,dh}_callback(). 607255714Skris [Ben Laurie] 607355714Skris 607455714Skris *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 607555714Skris DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 607655714Skris [Steve Henson] 607755714Skris 607855714Skris *) X509_name_add_entry() freed the wrong thing after an error. 607955714Skris [Arne Ansper <arne@ats.cyber.ee>] 608055714Skris 608155714Skris *) rsa_eay.c would attempt to free a NULL context. 608255714Skris [Arne Ansper <arne@ats.cyber.ee>] 608355714Skris 608455714Skris *) BIO_s_socket() had a broken should_retry() on Windoze. 608555714Skris [Arne Ansper <arne@ats.cyber.ee>] 608655714Skris 608755714Skris *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 608855714Skris [Arne Ansper <arne@ats.cyber.ee>] 608955714Skris 609055714Skris *) Make sure the already existing X509_STORE->depth variable is initialized 609155714Skris in X509_STORE_new(), but document the fact that this variable is still 609255714Skris unused in the certificate verification process. 609355714Skris [Ralf S. Engelschall] 609455714Skris 609555714Skris *) Fix the various library and apps files to free up pkeys obtained from 609655714Skris X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 609755714Skris [Steve Henson] 609855714Skris 609955714Skris *) Fix reference counting in X509_PUBKEY_get(). This makes 610055714Skris demos/maurice/example2.c work, amongst others, probably. 610155714Skris [Steve Henson and Ben Laurie] 610255714Skris 610355714Skris *) First cut of a cleanup for apps/. First the `ssleay' program is now named 610455714Skris `openssl' and second, the shortcut symlinks for the `openssl <command>' 610555714Skris are no longer created. This way we have a single and consistent command 610655714Skris line interface `openssl <command>', similar to `cvs <command>'. 610755714Skris [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 610855714Skris 610955714Skris *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 611055714Skris BIT STRING wrapper always have zero unused bits. 611155714Skris [Steve Henson] 611255714Skris 611355714Skris *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 611455714Skris [Steve Henson] 611555714Skris 611655714Skris *) Make the top-level INSTALL documentation easier to understand. 611755714Skris [Paul Sutton] 611855714Skris 611955714Skris *) Makefiles updated to exit if an error occurs in a sub-directory 612055714Skris make (including if user presses ^C) [Paul Sutton] 612155714Skris 612255714Skris *) Make Montgomery context stuff explicit in RSA data structure. 612355714Skris [Ben Laurie] 612455714Skris 612555714Skris *) Fix build order of pem and err to allow for generated pem.h. 612655714Skris [Ben Laurie] 612755714Skris 612855714Skris *) Fix renumbering bug in X509_NAME_delete_entry(). 612955714Skris [Ben Laurie] 613055714Skris 613155714Skris *) Enhanced the err-ins.pl script so it makes the error library number 613255714Skris global and can add a library name. This is needed for external ASN1 and 613355714Skris other error libraries. 613455714Skris [Steve Henson] 613555714Skris 613655714Skris *) Fixed sk_insert which never worked properly. 613755714Skris [Steve Henson] 613855714Skris 613955714Skris *) Fix ASN1 macros so they can handle indefinite length construted 614055714Skris EXPLICIT tags. Some non standard certificates use these: they can now 614155714Skris be read in. 614255714Skris [Steve Henson] 614355714Skris 614455714Skris *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 614555714Skris into a single doc/ssleay.txt bundle. This way the information is still 614655714Skris preserved but no longer messes up this directory. Now it's new room for 614755714Skris the new set of documenation files. 614855714Skris [Ralf S. Engelschall] 614955714Skris 615055714Skris *) SETs were incorrectly DER encoded. This was a major pain, because they 615155714Skris shared code with SEQUENCEs, which aren't coded the same. This means that 615255714Skris almost everything to do with SETs or SEQUENCEs has either changed name or 615355714Skris number of arguments. 615455714Skris [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 615555714Skris 615655714Skris *) Fix test data to work with the above. 615755714Skris [Ben Laurie] 615855714Skris 615955714Skris *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 616055714Skris was already fixed by Eric for 0.9.1 it seems. 616155714Skris [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 616255714Skris 616355714Skris *) Autodetect FreeBSD3. 616455714Skris [Ben Laurie] 616555714Skris 616655714Skris *) Fix various bugs in Configure. This affects the following platforms: 616755714Skris nextstep 616855714Skris ncr-scde 616955714Skris unixware-2.0 617055714Skris unixware-2.0-pentium 617155714Skris sco5-cc. 617255714Skris [Ben Laurie] 617355714Skris 617455714Skris *) Eliminate generated files from CVS. Reorder tests to regenerate files 617555714Skris before they are needed. 617655714Skris [Ben Laurie] 617755714Skris 617855714Skris *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 617955714Skris [Ben Laurie] 618055714Skris 618155714Skris 618255714Skris Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 618355714Skris 618455714Skris *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 618555714Skris changed SSLeay to OpenSSL in version strings. 618655714Skris [Ralf S. Engelschall] 618755714Skris 618855714Skris *) Some fixups to the top-level documents. 618955714Skris [Paul Sutton] 619055714Skris 619155714Skris *) Fixed the nasty bug where rsaref.h was not found under compile-time 619255714Skris because the symlink to include/ was missing. 619355714Skris [Ralf S. Engelschall] 619455714Skris 619555714Skris *) Incorporated the popular no-RSA/DSA-only patches 619655714Skris which allow to compile a RSA-free SSLeay. 619755714Skris [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 619855714Skris 619955714Skris *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 620055714Skris when "ssleay" is still not found. 620155714Skris [Ralf S. Engelschall] 620255714Skris 620355714Skris *) Added more platforms to Configure: Cray T3E, HPUX 11, 620455714Skris [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 620555714Skris 620655714Skris *) Updated the README file. 620755714Skris [Ralf S. Engelschall] 620855714Skris 620955714Skris *) Added various .cvsignore files in the CVS repository subdirs 621055714Skris to make a "cvs update" really silent. 621155714Skris [Ralf S. Engelschall] 621255714Skris 621355714Skris *) Recompiled the error-definition header files and added 621455714Skris missing symbols to the Win32 linker tables. 621555714Skris [Ralf S. Engelschall] 621655714Skris 621755714Skris *) Cleaned up the top-level documents; 621855714Skris o new files: CHANGES and LICENSE 621955714Skris o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 622055714Skris o merged COPYRIGHT into LICENSE 622155714Skris o removed obsolete TODO file 622255714Skris o renamed MICROSOFT to INSTALL.W32 622355714Skris [Ralf S. Engelschall] 622455714Skris 622555714Skris *) Removed dummy files from the 0.9.1b source tree: 622655714Skris crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 622755714Skris crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 622855714Skris crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 622955714Skris crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 623055714Skris util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 623155714Skris [Ralf S. Engelschall] 623255714Skris 623355714Skris *) Added various platform portability fixes. 623455714Skris [Mark J. Cox] 623555714Skris 623655714Skris *) The Genesis of the OpenSSL rpject: 623755714Skris We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 623855714Skris Young and Tim J. Hudson created while they were working for C2Net until 623955714Skris summer 1998. 624055714Skris [The OpenSSL Project] 624155714Skris 624255714Skris 624355714Skris Changes between 0.9.0b and 0.9.1b [not released] 624455714Skris 624555714Skris *) Updated a few CA certificates under certs/ 624655714Skris [Eric A. Young] 624755714Skris 624855714Skris *) Changed some BIGNUM api stuff. 624955714Skris [Eric A. Young] 625055714Skris 625155714Skris *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 625255714Skris DGUX x86, Linux Alpha, etc. 625355714Skris [Eric A. Young] 625455714Skris 625555714Skris *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 625655714Skris RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 625755714Skris available). 625855714Skris [Eric A. Young] 625955714Skris 626055714Skris *) Add -strparse option to asn1pars program which parses nested 626155714Skris binary structures 626255714Skris [Dr Stephen Henson <shenson@bigfoot.com>] 626355714Skris 626455714Skris *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 626555714Skris [Eric A. Young] 626655714Skris 626755714Skris *) DSA fix for "ca" program. 626855714Skris [Eric A. Young] 626955714Skris 627055714Skris *) Added "-genkey" option to "dsaparam" program. 627155714Skris [Eric A. Young] 627255714Skris 627355714Skris *) Added RIPE MD160 (rmd160) message digest. 627455714Skris [Eric A. Young] 627555714Skris 627655714Skris *) Added -a (all) option to "ssleay version" command. 627755714Skris [Eric A. Young] 627855714Skris 627955714Skris *) Added PLATFORM define which is the id given to Configure. 628055714Skris [Eric A. Young] 628155714Skris 628255714Skris *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 628355714Skris [Eric A. Young] 628455714Skris 628555714Skris *) Extended the ASN.1 parser routines. 628655714Skris [Eric A. Young] 628755714Skris 628855714Skris *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 628955714Skris [Eric A. Young] 629055714Skris 629155714Skris *) Added a BN_CTX to the BN library. 629255714Skris [Eric A. Young] 629355714Skris 629455714Skris *) Fixed the weak key values in DES library 629555714Skris [Eric A. Young] 629655714Skris 629755714Skris *) Changed API in EVP library for cipher aliases. 629855714Skris [Eric A. Young] 629955714Skris 630055714Skris *) Added support for RC2/64bit cipher. 630155714Skris [Eric A. Young] 630255714Skris 630355714Skris *) Converted the lhash library to the crypto/mem.c functions. 630455714Skris [Eric A. Young] 630555714Skris 630655714Skris *) Added more recognized ASN.1 object ids. 630755714Skris [Eric A. Young] 630855714Skris 630955714Skris *) Added more RSA padding checks for SSL/TLS. 631055714Skris [Eric A. Young] 631155714Skris 631255714Skris *) Added BIO proxy/filter functionality. 631355714Skris [Eric A. Young] 631455714Skris 631555714Skris *) Added extra_certs to SSL_CTX which can be used 631655714Skris send extra CA certificates to the client in the CA cert chain sending 631755714Skris process. It can be configured with SSL_CTX_add_extra_chain_cert(). 631855714Skris [Eric A. Young] 631955714Skris 632055714Skris *) Now Fortezza is denied in the authentication phase because 632155714Skris this is key exchange mechanism is not supported by SSLeay at all. 632255714Skris [Eric A. Young] 632355714Skris 632455714Skris *) Additional PKCS1 checks. 632555714Skris [Eric A. Young] 632655714Skris 632755714Skris *) Support the string "TLSv1" for all TLS v1 ciphers. 632855714Skris [Eric A. Young] 632955714Skris 633055714Skris *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 633155714Skris ex_data index of the SSL context in the X509_STORE_CTX ex_data. 633255714Skris [Eric A. Young] 633355714Skris 633455714Skris *) Fixed a few memory leaks. 633555714Skris [Eric A. Young] 633655714Skris 633755714Skris *) Fixed various code and comment typos. 633855714Skris [Eric A. Young] 633955714Skris 634055714Skris *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 634155714Skris bytes sent in the client random. 634255714Skris [Edward Bishop <ebishop@spyglass.com>] 634355714Skris 6344