CHANGES revision 72613
155714Skris 255714Skris OpenSSL CHANGES 355714Skris _______________ 455714Skris 572613Skris Changes between 0.9.6 and 0.9.6a [xx XXX 2001] 672613Skris 772613Skris *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 872613Skris Bleichenbacher's DSA attack. 972613Skris [Ulf Moeller] 1072613Skris 1172613Skris *) In the NCONF_...-based implementations for CONF_... queries 1272613Skris (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 1372613Skris a temporary CONF structure with the data component set to NULL 1472613Skris (which gives segmentation faults in lh_retrieve). 1572613Skris Instead, use NULL for the CONF pointer in CONF_get_string and 1672613Skris CONF_get_number (which may use environment variables) and directly 1772613Skris return NULL from CONF_get_section. 1872613Skris [Bodo Moeller] 1972613Skris 2072613Skris *) Fix potential buffer overrun for EBCDIC. 2172613Skris [Ulf Moeller] 2272613Skris 2372613Skris *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 2472613Skris keyUsage if basicConstraints absent for a CA. 2572613Skris [Steve Henson] 2672613Skris 2772613Skris *) Make SMIME_write_PKCS7() write mail header values with a format that 2872613Skris is more generally accepted (no spaces before the semicolon), since 2972613Skris some programs can't parse those values properly otherwise. Also make 3072613Skris sure BIO's that break lines after each write do not create invalid 3172613Skris headers. 3272613Skris [Richard Levitte] 3372613Skris 3472613Skris *) Make the CRL encoding routines work with empty SEQUENCE OF. The 3572613Skris macros previously used would not encode an empty SEQUENCE OF 3672613Skris and break the signature. 3772613Skris [Steve Henson] 3872613Skris 3972613Skris *) Zero the premaster secret after deriving the master secret in 4072613Skris DH ciphersuites. 4172613Skris [Steve Henson] 4272613Skris 4372613Skris *) Add some EVP_add_digest_alias registrations (as found in 4472613Skris OpenSSL_add_all_digests()) to SSL_library_init() 4572613Skris aka OpenSSL_add_ssl_algorithms(). This provides improved 4672613Skris compatibility with peers using X.509 certificates 4772613Skris with unconventional AlgorithmIdentifier OIDs. 4872613Skris [Bodo Moeller] 4972613Skris 5072613Skris *) Fix for Irix with NO_ASM. 5172613Skris ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 5272613Skris 5372613Skris *) ./config script fixes. 5472613Skris [Ulf Moeller, Richard Levitte] 5572613Skris 5672613Skris *) Fix 'openssl passwd -1'. 5772613Skris [Bodo Moeller] 5872613Skris 5972613Skris *) Change PKCS12_key_gen_asc() so it can cope with non null 6072613Skris terminated strings whose length is passed in the passlen 6172613Skris parameter, for example from PEM callbacks. This was done 6272613Skris by adding an extra length parameter to asc2uni(). 6372613Skris [Steve Henson, reported by <oddissey@samsung.co.kr>] 6472613Skris 6572613Skris *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 6672613Skris call failed, free the DSA structure. 6772613Skris [Bodo Moeller] 6872613Skris 6972613Skris *) Fix to uni2asc() to cope with zero length Unicode strings. 7072613Skris These are present in some PKCS#12 files. 7172613Skris [Steve Henson] 7272613Skris 7372613Skris *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 7472613Skris Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 7572613Skris when writing a 32767 byte record. 7672613Skris [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 7772613Skris 7872613Skris *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 7972613Skris obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 8072613Skris 8172613Skris (RSA objects have a reference count access to which is protected 8272613Skris by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 8372613Skris so they are meant to be shared between threads.) 8472613Skris [Bodo Moeller, Geoff Thorpe; original patch submitted by 8572613Skris "Reddie, Steven" <Steven.Reddie@ca.com>] 8672613Skris 8772613Skris *) Fix a deadlock in CRYPTO_mem_leaks(). 8872613Skris [Bodo Moeller] 8972613Skris 9072613Skris *) rand_win.c fix for Borland C. 9172613Skris [Ulf M�ller] 9272613Skris 9372613Skris *) BN_rshift bugfix for n == 0. 9472613Skris [Bodo Moeller] 9572613Skris 9672613Skris *) Store verify_result within SSL_SESSION also for client side to 9772613Skris avoid potential security hole. (Re-used sessions on the client side 9872613Skris always resulted in verify_result==X509_V_OK, not using the original 9972613Skris result of the server certificate verification.) 10072613Skris [Lutz Jaenicke] 10172613Skris 10272613Skris *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 10372613Skris SSL3_RT_APPLICATION_DATA, return 0. 10472613Skris Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 10572613Skris [Bodo Moeller] 10672613Skris 10772613Skris *) Fix SSL_peek: 10872613Skris Both ssl2_peek and ssl3_peek, which were totally broken in earlier 10972613Skris releases, have been re-implemented by renaming the previous 11072613Skris implementations of ssl2_read and ssl3_read to ssl2_read_internal 11172613Skris and ssl3_read_internal, respectively, and adding 'peek' parameters 11272613Skris to them. The new ssl[23]_{read,peek} functions are calls to 11372613Skris ssl[23]_read_internal with the 'peek' flag set appropriately. 11472613Skris A 'peek' parameter has also been added to ssl3_read_bytes, which 11572613Skris does the actual work for ssl3_read_internal. 11672613Skris [Bodo Moeller] 11772613Skris 11872613Skris *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 11972613Skris The previous value, 12, was not always sufficient for BN_mod_exp(). 12072613Skris [Bodo Moeller] 12172613Skris 12272613Skris *) Fix typo in get_cert_by_subject() in by_dir.c 12372613Skris [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 12472613Skris 12572613Skris *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 12672613Skris 12772613Skris Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 12872613Skris and not in SSL_clear because the latter is also used by the 12972613Skris accept/connect functions; previously, the settings made by 13072613Skris SSL_set_read_ahead would be lost during the handshake. 13172613Skris [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 13272613Skris 13372613Skris *) Correct util/mkdef.pl to be selective about disabled algorithms. 13472613Skris Previously, it would create entries for disableed algorithms no 13572613Skris matter what. 13672613Skris [Richard Levitte] 13772613Skris 13872613Skris *) Added several new manual pages for SSL_* function. 13972613Skris [Lutz Jaenicke] 14072613Skris 14168651Skris Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 14268651Skris 14368651Skris *) In ssl23_get_client_hello, generate an error message when faced 14468651Skris with an initial SSL 3.0/TLS record that is too small to contain the 14568651Skris first two bytes of the ClientHello message, i.e. client_version. 14668651Skris (Note that this is a pathologic case that probably has never happened 14768651Skris in real life.) The previous approach was to use the version number 14868651Skris from the record header as a substitute; but our protocol choice 14968651Skris should not depend on that one because it is not authenticated 15068651Skris by the Finished messages. 15168651Skris [Bodo Moeller] 15268651Skris 15368651Skris *) More robust randomness gathering functions for Windows. 15468651Skris [Jeffrey Altman <jaltman@columbia.edu>] 15568651Skris 15668651Skris *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 15768651Skris not set then we don't setup the error code for issuer check errors 15868651Skris to avoid possibly overwriting other errors which the callback does 15968651Skris handle. If an application does set the flag then we assume it knows 16068651Skris what it is doing and can handle the new informational codes 16168651Skris appropriately. 16268651Skris [Steve Henson] 16368651Skris 16468651Skris *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 16568651Skris a general "ANY" type, as such it should be able to decode anything 16668651Skris including tagged types. However it didn't check the class so it would 16768651Skris wrongly interpret tagged types in the same way as their universal 16868651Skris counterpart and unknown types were just rejected. Changed so that the 16968651Skris tagged and unknown types are handled in the same way as a SEQUENCE: 17068651Skris that is the encoding is stored intact. There is also a new type 17168651Skris "V_ASN1_OTHER" which is used when the class is not universal, in this 17268651Skris case we have no idea what the actual type is so we just lump them all 17368651Skris together. 17468651Skris [Steve Henson] 17568651Skris 17668651Skris *) On VMS, stdout may very well lead to a file that is written to 17768651Skris in a record-oriented fashion. That means that every write() will 17868651Skris write a separate record, which will be read separately by the 17968651Skris programs trying to read from it. This can be very confusing. 18068651Skris 18168651Skris The solution is to put a BIO filter in the way that will buffer 18268651Skris text until a linefeed is reached, and then write everything a 18368651Skris line at a time, so every record written will be an actual line, 18468651Skris not chunks of lines and not (usually doesn't happen, but I've 18568651Skris seen it once) several lines in one record. BIO_f_linebuffer() is 18668651Skris the answer. 18768651Skris 18868651Skris Currently, it's a VMS-only method, because that's where it has 18968651Skris been tested well enough. 19068651Skris [Richard Levitte] 19168651Skris 19268651Skris *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 19368651Skris it can return incorrect results. 19468651Skris (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 19568651Skris but it was in 0.9.6-beta[12].) 19668651Skris [Bodo Moeller] 19768651Skris 19868651Skris *) Disable the check for content being present when verifying detached 19968651Skris signatures in pk7_smime.c. Some versions of Netscape (wrongly) 20068651Skris include zero length content when signing messages. 20168651Skris [Steve Henson] 20268651Skris 20368651Skris *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 20468651Skris BIO_ctrl (for BIO pairs). 20568651Skris [Bodo M�ller] 20668651Skris 20768651Skris *) Add DSO method for VMS. 20868651Skris [Richard Levitte] 20968651Skris 21068651Skris *) Bug fix: Montgomery multiplication could produce results with the 21168651Skris wrong sign. 21268651Skris [Ulf M�ller] 21368651Skris 21468651Skris *) Add RPM specification openssl.spec and modify it to build three 21568651Skris packages. The default package contains applications, application 21668651Skris documentation and run-time libraries. The devel package contains 21768651Skris include files, static libraries and function documentation. The 21868651Skris doc package contains the contents of the doc directory. The original 21968651Skris openssl.spec was provided by Damien Miller <djm@mindrot.org>. 22068651Skris [Richard Levitte] 22168651Skris 22268651Skris *) Add a large number of documentation files for many SSL routines. 22368651Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 22468651Skris 22568651Skris *) Add a configuration entry for Sony News 4. 22668651Skris [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 22768651Skris 22868651Skris *) Don't set the two most significant bits to one when generating a 22968651Skris random number < q in the DSA library. 23068651Skris [Ulf M�ller] 23168651Skris 23268651Skris *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 23368651Skris behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 23468651Skris the underlying transport is blocking) if a handshake took place. 23568651Skris (The default behaviour is needed by applications such as s_client 23668651Skris and s_server that use select() to determine when to use SSL_read; 23768651Skris but for applications that know in advance when to expect data, it 23868651Skris just makes things more complicated.) 23968651Skris [Bodo Moeller] 24068651Skris 24168651Skris *) Add RAND_egd_bytes(), which gives control over the number of bytes read 24268651Skris from EGD. 24368651Skris [Ben Laurie] 24468651Skris 24568651Skris *) Add a few more EBCDIC conditionals that make `req' and `x509' 24668651Skris work better on such systems. 24768651Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 24868651Skris 24968651Skris *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 25068651Skris Update PKCS12_parse() so it copies the friendlyName and the 25168651Skris keyid to the certificates aux info. 25268651Skris [Steve Henson] 25368651Skris 25468651Skris *) Fix bug in PKCS7_verify() which caused an infinite loop 25568651Skris if there was more than one signature. 25668651Skris [Sven Uszpelkat <su@celocom.de>] 25768651Skris 25868651Skris *) Major change in util/mkdef.pl to include extra information 25968651Skris about each symbol, as well as presentig variables as well 26068651Skris as functions. This change means that there's n more need 26168651Skris to rebuild the .num files when some algorithms are excluded. 26268651Skris [Richard Levitte] 26368651Skris 26468651Skris *) Allow the verify time to be set by an application, 26568651Skris rather than always using the current time. 26668651Skris [Steve Henson] 26768651Skris 26868651Skris *) Phase 2 verify code reorganisation. The certificate 26968651Skris verify code now looks up an issuer certificate by a 27068651Skris number of criteria: subject name, authority key id 27168651Skris and key usage. It also verifies self signed certificates 27268651Skris by the same criteria. The main comparison function is 27368651Skris X509_check_issued() which performs these checks. 27468651Skris 27568651Skris Lot of changes were necessary in order to support this 27668651Skris without completely rewriting the lookup code. 27768651Skris 27868651Skris Authority and subject key identifier are now cached. 27968651Skris 28068651Skris The LHASH 'certs' is X509_STORE has now been replaced 28168651Skris by a STACK_OF(X509_OBJECT). This is mainly because an 28268651Skris LHASH can't store or retrieve multiple objects with 28368651Skris the same hash value. 28468651Skris 28568651Skris As a result various functions (which were all internal 28668651Skris use only) have changed to handle the new X509_STORE 28768651Skris structure. This will break anything that messed round 28868651Skris with X509_STORE internally. 28968651Skris 29068651Skris The functions X509_STORE_add_cert() now checks for an 29168651Skris exact match, rather than just subject name. 29268651Skris 29368651Skris The X509_STORE API doesn't directly support the retrieval 29468651Skris of multiple certificates matching a given criteria, however 29568651Skris this can be worked round by performing a lookup first 29668651Skris (which will fill the cache with candidate certificates) 29768651Skris and then examining the cache for matches. This is probably 29868651Skris the best we can do without throwing out X509_LOOKUP 29968651Skris entirely (maybe later...). 30068651Skris 30168651Skris The X509_VERIFY_CTX structure has been enhanced considerably. 30268651Skris 30368651Skris All certificate lookup operations now go via a get_issuer() 30468651Skris callback. Although this currently uses an X509_STORE it 30568651Skris can be replaced by custom lookups. This is a simple way 30668651Skris to bypass the X509_STORE hackery necessary to make this 30768651Skris work and makes it possible to use more efficient techniques 30868651Skris in future. A very simple version which uses a simple 30968651Skris STACK for its trusted certificate store is also provided 31068651Skris using X509_STORE_CTX_trusted_stack(). 31168651Skris 31268651Skris The verify_cb() and verify() callbacks now have equivalents 31368651Skris in the X509_STORE_CTX structure. 31468651Skris 31568651Skris X509_STORE_CTX also has a 'flags' field which can be used 31668651Skris to customise the verify behaviour. 31768651Skris [Steve Henson] 31868651Skris 31968651Skris *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 32068651Skris excludes S/MIME capabilities. 32168651Skris [Steve Henson] 32268651Skris 32368651Skris *) When a certificate request is read in keep a copy of the 32468651Skris original encoding of the signed data and use it when outputing 32568651Skris again. Signatures then use the original encoding rather than 32668651Skris a decoded, encoded version which may cause problems if the 32768651Skris request is improperly encoded. 32868651Skris [Steve Henson] 32968651Skris 33068651Skris *) For consistency with other BIO_puts implementations, call 33168651Skris buffer_write(b, ...) directly in buffer_puts instead of calling 33268651Skris BIO_write(b, ...). 33368651Skris 33468651Skris In BIO_puts, increment b->num_write as in BIO_write. 33568651Skris [Peter.Sylvester@EdelWeb.fr] 33668651Skris 33768651Skris *) Fix BN_mul_word for the case where the word is 0. (We have to use 33868651Skris BN_zero, we may not return a BIGNUM with an array consisting of 33968651Skris words set to zero.) 34068651Skris [Bodo Moeller] 34168651Skris 34268651Skris *) Avoid calling abort() from within the library when problems are 34368651Skris detected, except if preprocessor symbols have been defined 34468651Skris (such as REF_CHECK, BN_DEBUG etc.). 34568651Skris [Bodo Moeller] 34668651Skris 34768651Skris *) New openssl application 'rsautl'. This utility can be 34868651Skris used for low level RSA operations. DER public key 34968651Skris BIO/fp routines also added. 35068651Skris [Steve Henson] 35168651Skris 35268651Skris *) New Configure entry and patches for compiling on QNX 4. 35368651Skris [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 35468651Skris 35568651Skris *) A demo state-machine implementation was sponsored by 35668651Skris Nuron (http://www.nuron.com/) and is now available in 35768651Skris demos/state_machine. 35868651Skris [Ben Laurie] 35968651Skris 36068651Skris *) New options added to the 'dgst' utility for signature 36168651Skris generation and verification. 36268651Skris [Steve Henson] 36368651Skris 36468651Skris *) Unrecognized PKCS#7 content types are now handled via a 36568651Skris catch all ASN1_TYPE structure. This allows unsupported 36668651Skris types to be stored as a "blob" and an application can 36768651Skris encode and decode it manually. 36868651Skris [Steve Henson] 36968651Skris 37068651Skris *) Fix various signed/unsigned issues to make a_strex.c 37168651Skris compile under VC++. 37268651Skris [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 37368651Skris 37468651Skris *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 37568651Skris length if passed a buffer. ASN1_INTEGER_to_BN failed 37668651Skris if passed a NULL BN and its argument was negative. 37768651Skris [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 37868651Skris 37968651Skris *) Modification to PKCS#7 encoding routines to output definite 38068651Skris length encoding. Since currently the whole structures are in 38168651Skris memory there's not real point in using indefinite length 38268651Skris constructed encoding. However if OpenSSL is compiled with 38368651Skris the flag PKCS7_INDEFINITE_ENCODING the old form is used. 38468651Skris [Steve Henson] 38568651Skris 38668651Skris *) Added BIO_vprintf() and BIO_vsnprintf(). 38768651Skris [Richard Levitte] 38868651Skris 38968651Skris *) Added more prefixes to parse for in the the strings written 39068651Skris through a logging bio, to cover all the levels that are available 39168651Skris through syslog. The prefixes are now: 39268651Skris 39368651Skris PANIC, EMERG, EMR => LOG_EMERG 39468651Skris ALERT, ALR => LOG_ALERT 39568651Skris CRIT, CRI => LOG_CRIT 39668651Skris ERROR, ERR => LOG_ERR 39768651Skris WARNING, WARN, WAR => LOG_WARNING 39868651Skris NOTICE, NOTE, NOT => LOG_NOTICE 39968651Skris INFO, INF => LOG_INFO 40068651Skris DEBUG, DBG => LOG_DEBUG 40168651Skris 40268651Skris and as before, if none of those prefixes are present at the 40368651Skris beginning of the string, LOG_ERR is chosen. 40468651Skris 40568651Skris On Win32, the LOG_* levels are mapped according to this: 40668651Skris 40768651Skris LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 40868651Skris LOG_WARNING => EVENTLOG_WARNING_TYPE 40968651Skris LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 41068651Skris 41168651Skris [Richard Levitte] 41268651Skris 41368651Skris *) Made it possible to reconfigure with just the configuration 41468651Skris argument "reconf" or "reconfigure". The command line arguments 41568651Skris are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 41668651Skris and are retrieved from there when reconfiguring. 41768651Skris [Richard Levitte] 41868651Skris 41968651Skris *) MD4 implemented. 42068651Skris [Assar Westerlund <assar@sics.se>, Richard Levitte] 42168651Skris 42268651Skris *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 42368651Skris [Richard Levitte] 42468651Skris 42568651Skris *) The obj_dat.pl script was messing up the sorting of object 42668651Skris names. The reason was that it compared the quoted version 42768651Skris of strings as a result "OCSP" > "OCSP Signing" because 42868651Skris " > SPACE. Changed script to store unquoted versions of 42968651Skris names and add quotes on output. It was also omitting some 43068651Skris names from the lookup table if they were given a default 43168651Skris value (that is if SN is missing it is given the same 43268651Skris value as LN and vice versa), these are now added on the 43368651Skris grounds that if an object has a name we should be able to 43468651Skris look it up. Finally added warning output when duplicate 43568651Skris short or long names are found. 43668651Skris [Steve Henson] 43768651Skris 43868651Skris *) Changes needed for Tandem NSK. 43968651Skris [Scott Uroff <scott@xypro.com>] 44068651Skris 44168651Skris *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 44268651Skris RSA_padding_check_SSLv23(), special padding was never detected 44368651Skris and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 44468651Skris version rollback attacks was not effective. 44568651Skris 44668651Skris In s23_clnt.c, don't use special rollback-attack detection padding 44768651Skris (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 44868651Skris client; similarly, in s23_srvr.c, don't do the rollback check if 44968651Skris SSL 2.0 is the only protocol enabled in the server. 45068651Skris [Bodo Moeller] 45168651Skris 45268651Skris *) Make it possible to get hexdumps of unprintable data with 'openssl 45368651Skris asn1parse'. By implication, the functions ASN1_parse_dump() and 45468651Skris BIO_dump_indent() are added. 45568651Skris [Richard Levitte] 45668651Skris 45768651Skris *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 45868651Skris these print out strings and name structures based on various 45968651Skris flags including RFC2253 support and proper handling of 46068651Skris multibyte characters. Added options to the 'x509' utility 46168651Skris to allow the various flags to be set. 46268651Skris [Steve Henson] 46368651Skris 46468651Skris *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 46568651Skris Also change the functions X509_cmp_current_time() and 46668651Skris X509_gmtime_adj() work with an ASN1_TIME structure, 46768651Skris this will enable certificates using GeneralizedTime in validity 46868651Skris dates to be checked. 46968651Skris [Steve Henson] 47068651Skris 47168651Skris *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 47268651Skris negative public key encodings) on by default, 47368651Skris NO_NEG_PUBKEY_BUG can be set to disable it. 47468651Skris [Steve Henson] 47568651Skris 47668651Skris *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 47768651Skris content octets. An i2c_ASN1_OBJECT is unnecessary because 47868651Skris the encoding can be trivially obtained from the structure. 47968651Skris [Steve Henson] 48068651Skris 48168651Skris *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 48268651Skris not read locks (CRYPTO_r_[un]lock). 48368651Skris [Bodo Moeller] 48468651Skris 48568651Skris *) A first attempt at creating official support for shared 48668651Skris libraries through configuration. I've kept it so the 48768651Skris default is static libraries only, and the OpenSSL programs 48868651Skris are always statically linked for now, but there are 48968651Skris preparations for dynamic linking in place. 49068651Skris This has been tested on Linux and True64. 49168651Skris [Richard Levitte] 49268651Skris 49368651Skris *) Randomness polling function for Win9x, as described in: 49468651Skris Peter Gutmann, Software Generation of Practically Strong 49568651Skris Random Numbers. 49668651Skris [Ulf M�ller] 49768651Skris 49868651Skris *) Fix so PRNG is seeded in req if using an already existing 49968651Skris DSA key. 50068651Skris [Steve Henson] 50168651Skris 50268651Skris *) New options to smime application. -inform and -outform 50368651Skris allow alternative formats for the S/MIME message including 50468651Skris PEM and DER. The -content option allows the content to be 50568651Skris specified separately. This should allow things like Netscape 50668651Skris form signing output easier to verify. 50768651Skris [Steve Henson] 50868651Skris 50968651Skris *) Fix the ASN1 encoding of tags using the 'long form'. 51068651Skris [Steve Henson] 51168651Skris 51268651Skris *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 51368651Skris STRING types. These convert content octets to and from the 51468651Skris underlying type. The actual tag and length octets are 51568651Skris already assumed to have been read in and checked. These 51668651Skris are needed because all other string types have virtually 51768651Skris identical handling apart from the tag. By having versions 51868651Skris of the ASN1 functions that just operate on content octets 51968651Skris IMPLICIT tagging can be handled properly. It also allows 52068651Skris the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 52168651Skris and ASN1_INTEGER are identical apart from the tag. 52268651Skris [Steve Henson] 52368651Skris 52468651Skris *) Change the handling of OID objects as follows: 52568651Skris 52668651Skris - New object identifiers are inserted in objects.txt, following 52768651Skris the syntax given in objects.README. 52868651Skris - objects.pl is used to process obj_mac.num and create a new 52968651Skris obj_mac.h. 53068651Skris - obj_dat.pl is used to create a new obj_dat.h, using the data in 53168651Skris obj_mac.h. 53268651Skris 53368651Skris This is currently kind of a hack, and the perl code in objects.pl 53468651Skris isn't very elegant, but it works as I intended. The simplest way 53568651Skris to check that it worked correctly is to look in obj_dat.h and 53668651Skris check the array nid_objs and make sure the objects haven't moved 53768651Skris around (this is important!). Additions are OK, as well as 53868651Skris consistent name changes. 53968651Skris [Richard Levitte] 54068651Skris 54168651Skris *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 54268651Skris [Bodo Moeller] 54368651Skris 54468651Skris *) Addition of the command line parameter '-rand file' to 'openssl req'. 54568651Skris The given file adds to whatever has already been seeded into the 54668651Skris random pool through the RANDFILE configuration file option or 54768651Skris environment variable, or the default random state file. 54868651Skris [Richard Levitte] 54968651Skris 55068651Skris *) mkstack.pl now sorts each macro group into lexical order. 55168651Skris Previously the output order depended on the order the files 55268651Skris appeared in the directory, resulting in needless rewriting 55368651Skris of safestack.h . 55468651Skris [Steve Henson] 55568651Skris 55668651Skris *) Patches to make OpenSSL compile under Win32 again. Mostly 55768651Skris work arounds for the VC++ problem that it treats func() as 55868651Skris func(void). Also stripped out the parts of mkdef.pl that 55968651Skris added extra typesafe functions: these no longer exist. 56068651Skris [Steve Henson] 56168651Skris 56268651Skris *) Reorganisation of the stack code. The macros are now all 56368651Skris collected in safestack.h . Each macro is defined in terms of 56468651Skris a "stack macro" of the form SKM_<name>(type, a, b). The 56568651Skris DEBUG_SAFESTACK is now handled in terms of function casts, 56668651Skris this has the advantage of retaining type safety without the 56768651Skris use of additional functions. If DEBUG_SAFESTACK is not defined 56868651Skris then the non typesafe macros are used instead. Also modified the 56968651Skris mkstack.pl script to handle the new form. Needs testing to see 57068651Skris if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 57168651Skris the default if no major problems. Similar behaviour for ASN1_SET_OF 57268651Skris and PKCS12_STACK_OF. 57368651Skris [Steve Henson] 57468651Skris 57568651Skris *) When some versions of IIS use the 'NET' form of private key the 57668651Skris key derivation algorithm is different. Normally MD5(password) is 57768651Skris used as a 128 bit RC4 key. In the modified case 57868651Skris MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 57968651Skris new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 58068651Skris as the old Netscape_RSA functions except they have an additional 58168651Skris 'sgckey' parameter which uses the modified algorithm. Also added 58268651Skris an -sgckey command line option to the rsa utility. Thanks to 58368651Skris Adrian Peck <bertie@ncipher.com> for posting details of the modified 58468651Skris algorithm to openssl-dev. 58568651Skris [Steve Henson] 58668651Skris 58768651Skris *) The evp_local.h macros were using 'c.##kname' which resulted in 58868651Skris invalid expansion on some systems (SCO 5.0.5 for example). 58968651Skris Corrected to 'c.kname'. 59068651Skris [Phillip Porch <root@theporch.com>] 59168651Skris 59268651Skris *) New X509_get1_email() and X509_REQ_get1_email() functions that return 59368651Skris a STACK of email addresses from a certificate or request, these look 59468651Skris in the subject name and the subject alternative name extensions and 59568651Skris omit any duplicate addresses. 59668651Skris [Steve Henson] 59768651Skris 59868651Skris *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 59968651Skris This makes DSA verification about 2 % faster. 60068651Skris [Bodo Moeller] 60168651Skris 60268651Skris *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 60368651Skris (meaning that now 2^5 values will be precomputed, which is only 4 KB 60468651Skris plus overhead for 1024 bit moduli). 60568651Skris This makes exponentiations about 0.5 % faster for 1024 bit 60668651Skris exponents (as measured by "openssl speed rsa2048"). 60768651Skris [Bodo Moeller] 60868651Skris 60968651Skris *) Rename memory handling macros to avoid conflicts with other 61068651Skris software: 61168651Skris Malloc => OPENSSL_malloc 61268651Skris Malloc_locked => OPENSSL_malloc_locked 61368651Skris Realloc => OPENSSL_realloc 61468651Skris Free => OPENSSL_free 61568651Skris [Richard Levitte] 61668651Skris 61768651Skris *) New function BN_mod_exp_mont_word for small bases (roughly 15% 61868651Skris faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 61968651Skris [Bodo Moeller] 62068651Skris 62168651Skris *) CygWin32 support. 62268651Skris [John Jarvie <jjarvie@newsguy.com>] 62368651Skris 62468651Skris *) The type-safe stack code has been rejigged. It is now only compiled 62568651Skris in when OpenSSL is configured with the DEBUG_SAFESTACK option and 62668651Skris by default all type-specific stack functions are "#define"d back to 62768651Skris standard stack functions. This results in more streamlined output 62868651Skris but retains the type-safety checking possibilities of the original 62968651Skris approach. 63068651Skris [Geoff Thorpe] 63168651Skris 63268651Skris *) The STACK code has been cleaned up, and certain type declarations 63368651Skris that didn't make a lot of sense have been brought in line. This has 63468651Skris also involved a cleanup of sorts in safestack.h to more correctly 63568651Skris map type-safe stack functions onto their plain stack counterparts. 63668651Skris This work has also resulted in a variety of "const"ifications of 63768651Skris lots of the code, especially "_cmp" operations which should normally 63868651Skris be prototyped with "const" parameters anyway. 63968651Skris [Geoff Thorpe] 64068651Skris 64168651Skris *) When generating bytes for the first time in md_rand.c, 'stir the pool' 64268651Skris by seeding with STATE_SIZE dummy bytes (with zero entropy count). 64368651Skris (The PRNG state consists of two parts, the large pool 'state' and 'md', 64468651Skris where all of 'md' is used each time the PRNG is used, but 'state' 64568651Skris is used only indexed by a cyclic counter. As entropy may not be 64668651Skris well distributed from the beginning, 'md' is important as a 64768651Skris chaining variable. However, the output function chains only half 64868651Skris of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 64968651Skris all of 'md', and seeding with STATE_SIZE dummy bytes will result 65068651Skris in all of 'state' being rewritten, with the new values depending 65168651Skris on virtually all of 'md'. This overcomes the 80 bit limitation.) 65268651Skris [Bodo Moeller] 65368651Skris 65468651Skris *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 65568651Skris the handshake is continued after ssl_verify_cert_chain(); 65668651Skris otherwise, if SSL_VERIFY_NONE is set, remaining error codes 65768651Skris can lead to 'unexplainable' connection aborts later. 65868651Skris [Bodo Moeller; problem tracked down by Lutz Jaenicke] 65968651Skris 66068651Skris *) Major EVP API cipher revision. 66168651Skris Add hooks for extra EVP features. This allows various cipher 66268651Skris parameters to be set in the EVP interface. Support added for variable 66368651Skris key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 66468651Skris setting of RC2 and RC5 parameters. 66568651Skris 66668651Skris Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 66768651Skris ciphers. 66868651Skris 66968651Skris Remove lots of duplicated code from the EVP library. For example *every* 67068651Skris cipher init() function handles the 'iv' in the same way according to the 67168651Skris cipher mode. They also all do nothing if the 'key' parameter is NULL and 67268651Skris for CFB and OFB modes they zero ctx->num. 67368651Skris 67468651Skris New functionality allows removal of S/MIME code RC2 hack. 67568651Skris 67668651Skris Most of the routines have the same form and so can be declared in terms 67768651Skris of macros. 67868651Skris 67968651Skris By shifting this to the top level EVP_CipherInit() it can be removed from 68068651Skris all individual ciphers. If the cipher wants to handle IVs or keys 68168651Skris differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 68268651Skris flags. 68368651Skris 68468651Skris Change lots of functions like EVP_EncryptUpdate() to now return a 68568651Skris value: although software versions of the algorithms cannot fail 68668651Skris any installed hardware versions can. 68768651Skris [Steve Henson] 68868651Skris 68968651Skris *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 69068651Skris this option is set, tolerate broken clients that send the negotiated 69168651Skris protocol version number instead of the requested protocol version 69268651Skris number. 69368651Skris [Bodo Moeller] 69468651Skris 69568651Skris *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 69668651Skris i.e. non-zero for export ciphersuites, zero otherwise. 69768651Skris Previous versions had this flag inverted, inconsistent with 69868651Skris rsa_tmp_cb (..._TMP_RSA_CB). 69968651Skris [Bodo Moeller; problem reported by Amit Chopra] 70068651Skris 70168651Skris *) Add missing DSA library text string. Work around for some IIS 70268651Skris key files with invalid SEQUENCE encoding. 70368651Skris [Steve Henson] 70468651Skris 70568651Skris *) Add a document (doc/standards.txt) that list all kinds of standards 70668651Skris and so on that are implemented in OpenSSL. 70768651Skris [Richard Levitte] 70868651Skris 70968651Skris *) Enhance c_rehash script. Old version would mishandle certificates 71068651Skris with the same subject name hash and wouldn't handle CRLs at all. 71168651Skris Added -fingerprint option to crl utility, to support new c_rehash 71268651Skris features. 71368651Skris [Steve Henson] 71468651Skris 71568651Skris *) Eliminate non-ANSI declarations in crypto.h and stack.h. 71668651Skris [Ulf M�ller] 71768651Skris 71868651Skris *) Fix for SSL server purpose checking. Server checking was 71968651Skris rejecting certificates which had extended key usage present 72068651Skris but no ssl client purpose. 72168651Skris [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 72268651Skris 72368651Skris *) Make PKCS#12 code work with no password. The PKCS#12 spec 72468651Skris is a little unclear about how a blank password is handled. 72568651Skris Since the password in encoded as a BMPString with terminating 72668651Skris double NULL a zero length password would end up as just the 72768651Skris double NULL. However no password at all is different and is 72868651Skris handled differently in the PKCS#12 key generation code. NS 72968651Skris treats a blank password as zero length. MSIE treats it as no 73068651Skris password on export: but it will try both on import. We now do 73168651Skris the same: PKCS12_parse() tries zero length and no password if 73268651Skris the password is set to "" or NULL (NULL is now a valid password: 73368651Skris it wasn't before) as does the pkcs12 application. 73468651Skris [Steve Henson] 73568651Skris 73668651Skris *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 73768651Skris perror when PEM_read_bio_X509_REQ fails, the error message must 73868651Skris be obtained from the error queue. 73968651Skris [Bodo Moeller] 74068651Skris 74168651Skris *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 74268651Skris it in ERR_remove_state if appropriate, and change ERR_get_state 74368651Skris accordingly to avoid race conditions (this is necessary because 74468651Skris thread_hash is no longer constant once set). 74568651Skris [Bodo Moeller] 74668651Skris 74768651Skris *) Bugfix for linux-elf makefile.one. 74868651Skris [Ulf M�ller] 74968651Skris 75068651Skris *) RSA_get_default_method() will now cause a default 75168651Skris RSA_METHOD to be chosen if one doesn't exist already. 75268651Skris Previously this was only set during a call to RSA_new() 75368651Skris or RSA_new_method(NULL) meaning it was possible for 75468651Skris RSA_get_default_method() to return NULL. 75568651Skris [Geoff Thorpe] 75668651Skris 75768651Skris *) Added native name translation to the existing DSO code 75868651Skris that will convert (if the flag to do so is set) filenames 75968651Skris that are sufficiently small and have no path information 76068651Skris into a canonical native form. Eg. "blah" converted to 76168651Skris "libblah.so" or "blah.dll" etc. 76268651Skris [Geoff Thorpe] 76368651Skris 76468651Skris *) New function ERR_error_string_n(e, buf, len) which is like 76568651Skris ERR_error_string(e, buf), but writes at most 'len' bytes 76668651Skris including the 0 terminator. For ERR_error_string_n, 'buf' 76768651Skris may not be NULL. 76868651Skris [Damien Miller <djm@mindrot.org>, Bodo Moeller] 76968651Skris 77068651Skris *) CONF library reworked to become more general. A new CONF 77168651Skris configuration file reader "class" is implemented as well as a 77268651Skris new functions (NCONF_*, for "New CONF") to handle it. The now 77368651Skris old CONF_* functions are still there, but are reimplemented to 77468651Skris work in terms of the new functions. Also, a set of functions 77568651Skris to handle the internal storage of the configuration data is 77668651Skris provided to make it easier to write new configuration file 77768651Skris reader "classes" (I can definitely see something reading a 77868651Skris configuration file in XML format, for example), called _CONF_*, 77968651Skris or "the configuration storage API"... 78068651Skris 78168651Skris The new configuration file reading functions are: 78268651Skris 78368651Skris NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 78468651Skris NCONF_get_section, NCONF_get_string, NCONF_get_numbre 78568651Skris 78668651Skris NCONF_default, NCONF_WIN32 78768651Skris 78868651Skris NCONF_dump_fp, NCONF_dump_bio 78968651Skris 79068651Skris NCONF_default and NCONF_WIN32 are method (or "class") choosers, 79168651Skris NCONF_new creates a new CONF object. This works in the same way 79268651Skris as other interfaces in OpenSSL, like the BIO interface. 79368651Skris NCONF_dump_* dump the internal storage of the configuration file, 79468651Skris which is useful for debugging. All other functions take the same 79568651Skris arguments as the old CONF_* functions wth the exception of the 79668651Skris first that must be a `CONF *' instead of a `LHASH *'. 79768651Skris 79868651Skris To make it easer to use the new classes with the old CONF_* functions, 79968651Skris the function CONF_set_default_method is provided. 80068651Skris [Richard Levitte] 80168651Skris 80268651Skris *) Add '-tls1' option to 'openssl ciphers', which was already 80368651Skris mentioned in the documentation but had not been implemented. 80468651Skris (This option is not yet really useful because even the additional 80568651Skris experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 80668651Skris [Bodo Moeller] 80768651Skris 80868651Skris *) Initial DSO code added into libcrypto for letting OpenSSL (and 80968651Skris OpenSSL-based applications) load shared libraries and bind to 81068651Skris them in a portable way. 81168651Skris [Geoff Thorpe, with contributions from Richard Levitte] 81268651Skris 81359191Skris Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 81459191Skris 81559191Skris *) Make sure _lrotl and _lrotr are only used with MSVC. 81659191Skris 81759191Skris *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 81859191Skris (the default implementation of RAND_status). 81959191Skris 82059191Skris *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 82159191Skris to '-clrext' (= clear extensions), as intended and documented. 82259191Skris [Bodo Moeller; inconsistency pointed out by Michael Attili 82359191Skris <attili@amaxo.com>] 82459191Skris 82559191Skris *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 82659191Skris was larger than the MD block size. 82759191Skris [Steve Henson, pointed out by Yost William <YostW@tce.com>] 82859191Skris 82959191Skris *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 83059191Skris fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 83159191Skris using the passed key: if the passed key was a private key the result 83259191Skris of X509_print(), for example, would be to print out all the private key 83359191Skris components. 83459191Skris [Steve Henson] 83559191Skris 83659191Skris *) des_quad_cksum() byte order bug fix. 83759191Skris [Ulf M�ller, using the problem description in krb4-0.9.7, where 83859191Skris the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 83959191Skris 84059191Skris *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 84159191Skris discouraged. 84259191Skris [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 84359191Skris 84459191Skris *) For easily testing in shell scripts whether some command 84559191Skris 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 84659191Skris returns with exit code 0 iff no command of the given name is available. 84759191Skris 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 84859191Skris the output goes to stdout and nothing is printed to stderr. 84959191Skris Additional arguments are always ignored. 85059191Skris 85159191Skris Since for each cipher there is a command of the same name, 85259191Skris the 'no-cipher' compilation switches can be tested this way. 85359191Skris 85459191Skris ('openssl no-XXX' is not able to detect pseudo-commands such 85559191Skris as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 85659191Skris [Bodo Moeller] 85759191Skris 85859191Skris *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 85959191Skris [Bodo Moeller] 86059191Skris 86159191Skris *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 86259191Skris is set; it will be thrown away anyway because each handshake creates 86359191Skris its own key. 86459191Skris ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 86559191Skris to parameters -- in previous versions (since OpenSSL 0.9.3) the 86659191Skris 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 86759191Skris you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 86859191Skris [Bodo Moeller] 86959191Skris 87059191Skris *) New s_client option -ign_eof: EOF at stdin is ignored, and 87159191Skris 'Q' and 'R' lose their special meanings (quit/renegotiate). 87259191Skris This is part of what -quiet does; unlike -quiet, -ign_eof 87359191Skris does not suppress any output. 87459191Skris [Richard Levitte] 87559191Skris 87659191Skris *) Add compatibility options to the purpose and trust code. The 87759191Skris purpose X509_PURPOSE_ANY is "any purpose" which automatically 87859191Skris accepts a certificate or CA, this was the previous behaviour, 87959191Skris with all the associated security issues. 88059191Skris 88159191Skris X509_TRUST_COMPAT is the old trust behaviour: only and 88259191Skris automatically trust self signed roots in certificate store. A 88359191Skris new trust setting X509_TRUST_DEFAULT is used to specify that 88459191Skris a purpose has no associated trust setting and it should instead 88559191Skris use the value in the default purpose. 88659191Skris [Steve Henson] 88759191Skris 88859191Skris *) Fix the PKCS#8 DSA private key code so it decodes keys again 88959191Skris and fix a memory leak. 89059191Skris [Steve Henson] 89159191Skris 89259191Skris *) In util/mkerr.pl (which implements 'make errors'), preserve 89359191Skris reason strings from the previous version of the .c file, as 89459191Skris the default to have only downcase letters (and digits) in 89559191Skris automatically generated reasons codes is not always appropriate. 89659191Skris [Bodo Moeller] 89759191Skris 89859191Skris *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 89959191Skris using strerror. Previously, ERR_reason_error_string() returned 90059191Skris library names as reason strings for SYSerr; but SYSerr is a special 90159191Skris case where small numbers are errno values, not library numbers. 90259191Skris [Bodo Moeller] 90359191Skris 90459191Skris *) Add '-dsaparam' option to 'openssl dhparam' application. This 90559191Skris converts DSA parameters into DH parameters. (When creating parameters, 90659191Skris DSA_generate_parameters is used.) 90759191Skris [Bodo Moeller] 90859191Skris 90959191Skris *) Include 'length' (recommended exponent length) in C code generated 91059191Skris by 'openssl dhparam -C'. 91159191Skris [Bodo Moeller] 91259191Skris 91359191Skris *) The second argument to set_label in perlasm was already being used 91459191Skris so couldn't be used as a "file scope" flag. Moved to third argument 91559191Skris which was free. 91659191Skris [Steve Henson] 91759191Skris 91859191Skris *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 91959191Skris instead of RAND_bytes for encryption IVs and salts. 92059191Skris [Bodo Moeller] 92159191Skris 92259191Skris *) Include RAND_status() into RAND_METHOD instead of implementing 92359191Skris it only for md_rand.c Otherwise replacing the PRNG by calling 92459191Skris RAND_set_rand_method would be impossible. 92559191Skris [Bodo Moeller] 92659191Skris 92759191Skris *) Don't let DSA_generate_key() enter an infinite loop if the random 92859191Skris number generation fails. 92959191Skris [Bodo Moeller] 93059191Skris 93159191Skris *) New 'rand' application for creating pseudo-random output. 93259191Skris [Bodo Moeller] 93359191Skris 93459191Skris *) Added configuration support for Linux/IA64 93559191Skris [Rolf Haberrecker <rolf@suse.de>] 93659191Skris 93759191Skris *) Assembler module support for Mingw32. 93859191Skris [Ulf M�ller] 93959191Skris 94059191Skris *) Shared library support for HPUX (in shlib/). 94159191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 94259191Skris 94359191Skris *) Shared library support for Solaris gcc. 94459191Skris [Lutz Behnke <behnke@trustcenter.de>] 94559191Skris 94659191Skris Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 94759191Skris 94859191Skris *) PKCS7_encrypt() was adding text MIME headers twice because they 94959191Skris were added manually and by SMIME_crlf_copy(). 95059191Skris [Steve Henson] 95159191Skris 95259191Skris *) In bntest.c don't call BN_rand with zero bits argument. 95359191Skris [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 95459191Skris 95559191Skris *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 95659191Skris case was implemented. This caused BN_div_recp() to fail occasionally. 95759191Skris [Ulf M�ller] 95859191Skris 95959191Skris *) Add an optional second argument to the set_label() in the perl 96059191Skris assembly language builder. If this argument exists and is set 96159191Skris to 1 it signals that the assembler should use a symbol whose 96259191Skris scope is the entire file, not just the current function. This 96359191Skris is needed with MASM which uses the format label:: for this scope. 96459191Skris [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 96559191Skris 96659191Skris *) Change the ASN1 types so they are typedefs by default. Before 96759191Skris almost all types were #define'd to ASN1_STRING which was causing 96859191Skris STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 96959191Skris for example. 97059191Skris [Steve Henson] 97159191Skris 97259191Skris *) Change names of new functions to the new get1/get0 naming 97359191Skris convention: After 'get1', the caller owns a reference count 97459191Skris and has to call ..._free; 'get0' returns a pointer to some 97559191Skris data structure without incrementing reference counters. 97659191Skris (Some of the existing 'get' functions increment a reference 97759191Skris counter, some don't.) 97859191Skris Similarly, 'set1' and 'add1' functions increase reference 97959191Skris counters or duplicate objects. 98059191Skris [Steve Henson] 98159191Skris 98259191Skris *) Allow for the possibility of temp RSA key generation failure: 98359191Skris the code used to assume it always worked and crashed on failure. 98459191Skris [Steve Henson] 98559191Skris 98659191Skris *) Fix potential buffer overrun problem in BIO_printf(). 98759191Skris [Ulf M�ller, using public domain code by Patrick Powell; problem 98859191Skris pointed out by David Sacerdote <das33@cornell.edu>] 98959191Skris 99059191Skris *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 99159191Skris RAND_egd() and RAND_status(). In the command line application, 99259191Skris the EGD socket can be specified like a seed file using RANDFILE 99359191Skris or -rand. 99459191Skris [Ulf M�ller] 99559191Skris 99659191Skris *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 99759191Skris Some CAs (e.g. Verisign) distribute certificates in this form. 99859191Skris [Steve Henson] 99959191Skris 100059191Skris *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 100159191Skris list to exclude them. This means that no special compilation option 100259191Skris is needed to use anonymous DH: it just needs to be included in the 100359191Skris cipher list. 100459191Skris [Steve Henson] 100559191Skris 100659191Skris *) Change the EVP_MD_CTX_type macro so its meaning consistent with 100759191Skris EVP_MD_type. The old functionality is available in a new macro called 100859191Skris EVP_MD_md(). Change code that uses it and update docs. 100959191Skris [Steve Henson] 101059191Skris 101159191Skris *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 101259191Skris where the 'void *' argument is replaced by a function pointer argument. 101359191Skris Previously 'void *' was abused to point to functions, which works on 101459191Skris many platforms, but is not correct. As these functions are usually 101559191Skris called by macros defined in OpenSSL header files, most source code 101659191Skris should work without changes. 101759191Skris [Richard Levitte] 101859191Skris 101959191Skris *) <openssl/opensslconf.h> (which is created by Configure) now contains 102059191Skris sections with information on -D... compiler switches used for 102159191Skris compiling the library so that applications can see them. To enable 102259191Skris one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 102359191Skris must be defined. E.g., 102459191Skris #define OPENSSL_ALGORITHM_DEFINES 102559191Skris #include <openssl/opensslconf.h> 102659191Skris defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 102759191Skris [Richard Levitte, Ulf and Bodo M�ller] 102859191Skris 102959191Skris *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 103059191Skris record layer. 103159191Skris [Bodo Moeller] 103259191Skris 103359191Skris *) Change the 'other' type in certificate aux info to a STACK_OF 103459191Skris X509_ALGOR. Although not an AlgorithmIdentifier as such it has 103559191Skris the required ASN1 format: arbitrary types determined by an OID. 103659191Skris [Steve Henson] 103759191Skris 103859191Skris *) Add some PEM_write_X509_REQ_NEW() functions and a command line 103959191Skris argument to 'req'. This is not because the function is newer or 104059191Skris better than others it just uses the work 'NEW' in the certificate 104159191Skris request header lines. Some software needs this. 104259191Skris [Steve Henson] 104359191Skris 104459191Skris *) Reorganise password command line arguments: now passwords can be 104559191Skris obtained from various sources. Delete the PEM_cb function and make 104659191Skris it the default behaviour: i.e. if the callback is NULL and the 104759191Skris usrdata argument is not NULL interpret it as a null terminated pass 104859191Skris phrase. If usrdata and the callback are NULL then the pass phrase 104959191Skris is prompted for as usual. 105059191Skris [Steve Henson] 105159191Skris 105259191Skris *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 105359191Skris the support is automatically enabled. The resulting binaries will 105459191Skris autodetect the card and use it if present. 105559191Skris [Ben Laurie and Compaq Inc.] 105659191Skris 105759191Skris *) Work around for Netscape hang bug. This sends certificate request 105859191Skris and server done in one record. Since this is perfectly legal in the 105959191Skris SSL/TLS protocol it isn't a "bug" option and is on by default. See 106059191Skris the bugs/SSLv3 entry for more info. 106159191Skris [Steve Henson] 106259191Skris 106359191Skris *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 106459191Skris [Andy Polyakov] 106559191Skris 106659191Skris *) Add -rand argument to smime and pkcs12 applications and read/write 106759191Skris of seed file. 106859191Skris [Steve Henson] 106959191Skris 107059191Skris *) New 'passwd' tool for crypt(3) and apr1 password hashes. 107159191Skris [Bodo Moeller] 107259191Skris 107359191Skris *) Add command line password options to the remaining applications. 107459191Skris [Steve Henson] 107559191Skris 107659191Skris *) Bug fix for BN_div_recp() for numerators with an even number of 107759191Skris bits. 107859191Skris [Ulf M�ller] 107959191Skris 108059191Skris *) More tests in bntest.c, and changed test_bn output. 108159191Skris [Ulf M�ller] 108259191Skris 108359191Skris *) ./config recognizes MacOS X now. 108459191Skris [Andy Polyakov] 108559191Skris 108659191Skris *) Bug fix for BN_div() when the first words of num and divsor are 108759191Skris equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 108859191Skris [Ulf M�ller] 108959191Skris 109059191Skris *) Add support for various broken PKCS#8 formats, and command line 109159191Skris options to produce them. 109259191Skris [Steve Henson] 109359191Skris 109459191Skris *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 109559191Skris get temporary BIGNUMs from a BN_CTX. 109659191Skris [Ulf M�ller] 109759191Skris 109859191Skris *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 109959191Skris for p == 0. 110059191Skris [Ulf M�ller] 110159191Skris 110259191Skris *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 110359191Skris include a #define from the old name to the new. The original intent 110459191Skris was that statically linked binaries could for example just call 110559191Skris SSLeay_add_all_ciphers() to just add ciphers to the table and not 110659191Skris link with digests. This never worked becayse SSLeay_add_all_digests() 110759191Skris and SSLeay_add_all_ciphers() were in the same source file so calling 110859191Skris one would link with the other. They are now in separate source files. 110959191Skris [Steve Henson] 111059191Skris 111159191Skris *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 111259191Skris [Steve Henson] 111359191Skris 111459191Skris *) Use a less unusual form of the Miller-Rabin primality test (it used 111559191Skris a binary algorithm for exponentiation integrated into the Miller-Rabin 111659191Skris loop, our standard modexp algorithms are faster). 111759191Skris [Bodo Moeller] 111859191Skris 111959191Skris *) Support for the EBCDIC character set completed. 112059191Skris [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 112159191Skris 112259191Skris *) Source code cleanups: use const where appropriate, eliminate casts, 112359191Skris use void * instead of char * in lhash. 112459191Skris [Ulf M�ller] 112559191Skris 112659191Skris *) Bugfix: ssl3_send_server_key_exchange was not restartable 112759191Skris (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 112859191Skris this the server could overwrite ephemeral keys that the client 112959191Skris has already seen). 113059191Skris [Bodo Moeller] 113159191Skris 113259191Skris *) Turn DSA_is_prime into a macro that calls BN_is_prime, 113359191Skris using 50 iterations of the Rabin-Miller test. 113459191Skris 113559191Skris DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 113659191Skris iterations of the Rabin-Miller test as required by the appendix 113759191Skris to FIPS PUB 186[-1]) instead of DSA_is_prime. 113859191Skris As BN_is_prime_fasttest includes trial division, DSA parameter 113959191Skris generation becomes much faster. 114059191Skris 114159191Skris This implies a change for the callback functions in DSA_is_prime 114259191Skris and DSA_generate_parameters: The callback function is called once 114359191Skris for each positive witness in the Rabin-Miller test, not just 114459191Skris occasionally in the inner loop; and the parameters to the 114559191Skris callback function now provide an iteration count for the outer 114659191Skris loop rather than for the current invocation of the inner loop. 114759191Skris DSA_generate_parameters additionally can call the callback 114859191Skris function with an 'iteration count' of -1, meaning that a 114959191Skris candidate has passed the trial division test (when q is generated 115059191Skris from an application-provided seed, trial division is skipped). 115159191Skris [Bodo Moeller] 115259191Skris 115359191Skris *) New function BN_is_prime_fasttest that optionally does trial 115459191Skris division before starting the Rabin-Miller test and has 115559191Skris an additional BN_CTX * argument (whereas BN_is_prime always 115659191Skris has to allocate at least one BN_CTX). 115759191Skris 'callback(1, -1, cb_arg)' is called when a number has passed the 115859191Skris trial division stage. 115959191Skris [Bodo Moeller] 116059191Skris 116159191Skris *) Fix for bug in CRL encoding. The validity dates weren't being handled 116259191Skris as ASN1_TIME. 116359191Skris [Steve Henson] 116459191Skris 116559191Skris *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 116659191Skris [Steve Henson] 116759191Skris 116859191Skris *) New function BN_pseudo_rand(). 116959191Skris [Ulf M�ller] 117059191Skris 117159191Skris *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 117259191Skris bignum version of BN_from_montgomery() with the working code from 117359191Skris SSLeay 0.9.0 (the word based version is faster anyway), and clean up 117459191Skris the comments. 117559191Skris [Ulf M�ller] 117659191Skris 117759191Skris *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 117859191Skris made it impossible to use the same SSL_SESSION data structure in 117959191Skris SSL2 clients in multiple threads. 118059191Skris [Bodo Moeller] 118159191Skris 118259191Skris *) The return value of RAND_load_file() no longer counts bytes obtained 118359191Skris by stat(). RAND_load_file(..., -1) is new and uses the complete file 118459191Skris to seed the PRNG (previously an explicit byte count was required). 118559191Skris [Ulf M�ller, Bodo M�ller] 118659191Skris 118759191Skris *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 118859191Skris used (char *) instead of (void *) and had casts all over the place. 118959191Skris [Steve Henson] 119059191Skris 119159191Skris *) Make BN_generate_prime() return NULL on error if ret!=NULL. 119259191Skris [Ulf M�ller] 119359191Skris 119459191Skris *) Retain source code compatibility for BN_prime_checks macro: 119559191Skris BN_is_prime(..., BN_prime_checks, ...) now uses 119659191Skris BN_prime_checks_for_size to determine the appropriate number of 119759191Skris Rabin-Miller iterations. 119859191Skris [Ulf M�ller] 119959191Skris 120059191Skris *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 120159191Skris DH_CHECK_P_NOT_SAFE_PRIME. 120259191Skris (Check if this is true? OpenPGP calls them "strong".) 120359191Skris [Ulf M�ller] 120459191Skris 120559191Skris *) Merge the functionality of "dh" and "gendh" programs into a new program 120659191Skris "dhparam". The old programs are retained for now but will handle DH keys 120759191Skris (instead of parameters) in future. 120859191Skris [Steve Henson] 120959191Skris 121059191Skris *) Make the ciphers, s_server and s_client programs check the return values 121159191Skris when a new cipher list is set. 121259191Skris [Steve Henson] 121359191Skris 121459191Skris *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 121559191Skris ciphers. Before when the 56bit ciphers were enabled the sorting was 121659191Skris wrong. 121759191Skris 121859191Skris The syntax for the cipher sorting has been extended to support sorting by 121959191Skris cipher-strength (using the strength_bits hard coded in the tables). 122059191Skris The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 122159191Skris 122259191Skris Fix a bug in the cipher-command parser: when supplying a cipher command 122359191Skris string with an "undefined" symbol (neither command nor alphanumeric 122459191Skris [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 122559191Skris an error is flagged. 122659191Skris 122759191Skris Due to the strength-sorting extension, the code of the 122859191Skris ssl_create_cipher_list() function was completely rearranged. I hope that 122959191Skris the readability was also increased :-) 123059191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 123159191Skris 123259191Skris *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 123359191Skris for the first serial number and places 2 in the serial number file. This 123459191Skris avoids problems when the root CA is created with serial number zero and 123559191Skris the first user certificate has the same issuer name and serial number 123659191Skris as the root CA. 123759191Skris [Steve Henson] 123859191Skris 123959191Skris *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 124059191Skris the new code. Add documentation for this stuff. 124159191Skris [Steve Henson] 124259191Skris 124359191Skris *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 124459191Skris X509_*() to X509at_*() on the grounds that they don't handle X509 124559191Skris structures and behave in an analagous way to the X509v3 functions: 124659191Skris they shouldn't be called directly but wrapper functions should be used 124759191Skris instead. 124859191Skris 124959191Skris So we also now have some wrapper functions that call the X509at functions 125059191Skris when passed certificate requests. (TO DO: similar things can be done with 125159191Skris PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 125259191Skris things. Some of these need some d2i or i2d and print functionality 125359191Skris because they handle more complex structures.) 125459191Skris [Steve Henson] 125559191Skris 125659191Skris *) Add missing #ifndefs that caused missing symbols when building libssl 125759191Skris as a shared library without RSA. Use #ifndef NO_SSL2 instead of 125859191Skris NO_RSA in ssl/s2*.c. 125959191Skris [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 126059191Skris 126159191Skris *) Precautions against using the PRNG uninitialized: RAND_bytes() now 126259191Skris has a return value which indicates the quality of the random data 126359191Skris (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 126459191Skris error queue. New function RAND_pseudo_bytes() generates output that is 126559191Skris guaranteed to be unique but not unpredictable. RAND_add is like 126659191Skris RAND_seed, but takes an extra argument for an entropy estimate 126759191Skris (RAND_seed always assumes full entropy). 126859191Skris [Ulf M�ller] 126959191Skris 127059191Skris *) Do more iterations of Rabin-Miller probable prime test (specifically, 127159191Skris 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 127259191Skris instead of only 2 for all lengths; see BN_prime_checks_for_size definition 127359191Skris in crypto/bn/bn_prime.c for the complete table). This guarantees a 127459191Skris false-positive rate of at most 2^-80 for random input. 127559191Skris [Bodo Moeller] 127659191Skris 127759191Skris *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 127859191Skris [Bodo Moeller] 127959191Skris 128059191Skris *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 128159191Skris in the 0.9.5 release), this returns the chain 128259191Skris from an X509_CTX structure with a dup of the stack and all 128359191Skris the X509 reference counts upped: so the stack will exist 128459191Skris after X509_CTX_cleanup() has been called. Modify pkcs12.c 128559191Skris to use this. 128659191Skris 128759191Skris Also make SSL_SESSION_print() print out the verify return 128859191Skris code. 128959191Skris [Steve Henson] 129059191Skris 129159191Skris *) Add manpage for the pkcs12 command. Also change the default 129259191Skris behaviour so MAC iteration counts are used unless the new 129359191Skris -nomaciter option is used. This improves file security and 129459191Skris only older versions of MSIE (4.0 for example) need it. 129559191Skris [Steve Henson] 129659191Skris 129759191Skris *) Honor the no-xxx Configure options when creating .DEF files. 129859191Skris [Ulf M�ller] 129959191Skris 130059191Skris *) Add PKCS#10 attributes to field table: challengePassword, 130159191Skris unstructuredName and unstructuredAddress. These are taken from 130259191Skris draft PKCS#9 v2.0 but are compatible with v1.2 provided no 130359191Skris international characters are used. 130459191Skris 130559191Skris More changes to X509_ATTRIBUTE code: allow the setting of types 130659191Skris based on strings. Remove the 'loc' parameter when adding 130759191Skris attributes because these will be a SET OF encoding which is sorted 130859191Skris in ASN1 order. 130959191Skris [Steve Henson] 131059191Skris 131159191Skris *) Initial changes to the 'req' utility to allow request generation 131259191Skris automation. This will allow an application to just generate a template 131359191Skris file containing all the field values and have req construct the 131459191Skris request. 131559191Skris 131659191Skris Initial support for X509_ATTRIBUTE handling. Stacks of these are 131759191Skris used all over the place including certificate requests and PKCS#7 131859191Skris structures. They are currently handled manually where necessary with 131959191Skris some primitive wrappers for PKCS#7. The new functions behave in a 132059191Skris manner analogous to the X509 extension functions: they allow 132159191Skris attributes to be looked up by NID and added. 132259191Skris 132359191Skris Later something similar to the X509V3 code would be desirable to 132459191Skris automatically handle the encoding, decoding and printing of the 132559191Skris more complex types. The string types like challengePassword can 132659191Skris be handled by the string table functions. 132759191Skris 132859191Skris Also modified the multi byte string table handling. Now there is 132959191Skris a 'global mask' which masks out certain types. The table itself 133059191Skris can use the flag STABLE_NO_MASK to ignore the mask setting: this 133159191Skris is useful when for example there is only one permissible type 133259191Skris (as in countryName) and using the mask might result in no valid 133359191Skris types at all. 133459191Skris [Steve Henson] 133559191Skris 133659191Skris *) Clean up 'Finished' handling, and add functions SSL_get_finished and 133759191Skris SSL_get_peer_finished to allow applications to obtain the latest 133859191Skris Finished messages sent to the peer or expected from the peer, 133959191Skris respectively. (SSL_get_peer_finished is usually the Finished message 134059191Skris actually received from the peer, otherwise the protocol will be aborted.) 134159191Skris 134259191Skris As the Finished message are message digests of the complete handshake 134359191Skris (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 134459191Skris be used for external authentication procedures when the authentication 134559191Skris provided by SSL/TLS is not desired or is not enough. 134659191Skris [Bodo Moeller] 134759191Skris 134859191Skris *) Enhanced support for Alpha Linux is added. Now ./config checks if 134959191Skris the host supports BWX extension and if Compaq C is present on the 135059191Skris $PATH. Just exploiting of the BWX extension results in 20-30% 135159191Skris performance kick for some algorithms, e.g. DES and RC4 to mention 135259191Skris a couple. Compaq C in turn generates ~20% faster code for MD5 and 135359191Skris SHA1. 135459191Skris [Andy Polyakov] 135559191Skris 135659191Skris *) Add support for MS "fast SGC". This is arguably a violation of the 135759191Skris SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 135859191Skris weak crypto and after checking the certificate is SGC a second one 135959191Skris with strong crypto. MS SGC stops the first handshake after receiving 136059191Skris the server certificate message and sends a second client hello. Since 136159191Skris a server will typically do all the time consuming operations before 136259191Skris expecting any further messages from the client (server key exchange 136359191Skris is the most expensive) there is little difference between the two. 136459191Skris 136559191Skris To get OpenSSL to support MS SGC we have to permit a second client 136659191Skris hello message after we have sent server done. In addition we have to 136759191Skris reset the MAC if we do get this second client hello. 136859191Skris [Steve Henson] 136959191Skris 137059191Skris *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 137159191Skris if a DER encoded private key is RSA or DSA traditional format. Changed 137259191Skris d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 137359191Skris format DER encoded private key. Newer code should use PKCS#8 format which 137459191Skris has the key type encoded in the ASN1 structure. Added DER private key 137559191Skris support to pkcs8 application. 137659191Skris [Steve Henson] 137759191Skris 137859191Skris *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 137959191Skris ciphersuites has been selected (as required by the SSL 3/TLS 1 138059191Skris specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 138159191Skris is set, we interpret this as a request to violate the specification 138259191Skris (the worst that can happen is a handshake failure, and 'correct' 138359191Skris behaviour would result in a handshake failure anyway). 138459191Skris [Bodo Moeller] 138559191Skris 138659191Skris *) In SSL_CTX_add_session, take into account that there might be multiple 138759191Skris SSL_SESSION structures with the same session ID (e.g. when two threads 138859191Skris concurrently obtain them from an external cache). 138959191Skris The internal cache can handle only one SSL_SESSION with a given ID, 139059191Skris so if there's a conflict, we now throw out the old one to achieve 139159191Skris consistency. 139259191Skris [Bodo Moeller] 139359191Skris 139459191Skris *) Add OIDs for idea and blowfish in CBC mode. This will allow both 139559191Skris to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 139659191Skris some routines that use cipher OIDs: some ciphers do not have OIDs 139759191Skris defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 139859191Skris example. 139959191Skris [Steve Henson] 140059191Skris 140159191Skris *) Simplify the trust setting structure and code. Now we just have 140259191Skris two sequences of OIDs for trusted and rejected settings. These will 140359191Skris typically have values the same as the extended key usage extension 140459191Skris and any application specific purposes. 140559191Skris 140659191Skris The trust checking code now has a default behaviour: it will just 140759191Skris check for an object with the same NID as the passed id. Functions can 140859191Skris be provided to override either the default behaviour or the behaviour 140959191Skris for a given id. SSL client, server and email already have functions 141059191Skris in place for compatibility: they check the NID and also return "trusted" 141159191Skris if the certificate is self signed. 141259191Skris [Steve Henson] 141359191Skris 141459191Skris *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 141559191Skris traditional format into an EVP_PKEY structure. 141659191Skris [Steve Henson] 141759191Skris 141859191Skris *) Add a password callback function PEM_cb() which either prompts for 141959191Skris a password if usr_data is NULL or otherwise assumes it is a null 142059191Skris terminated password. Allow passwords to be passed on command line 142159191Skris environment or config files in a few more utilities. 142259191Skris [Steve Henson] 142359191Skris 142459191Skris *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 142559191Skris keys. Add some short names for PKCS#8 PBE algorithms and allow them 142659191Skris to be specified on the command line for the pkcs8 and pkcs12 utilities. 142759191Skris Update documentation. 142859191Skris [Steve Henson] 142959191Skris 143059191Skris *) Support for ASN1 "NULL" type. This could be handled before by using 143159191Skris ASN1_TYPE but there wasn't any function that would try to read a NULL 143259191Skris and produce an error if it couldn't. For compatibility we also have 143359191Skris ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 143459191Skris don't allocate anything because they don't need to. 143559191Skris [Steve Henson] 143659191Skris 143759191Skris *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 143859191Skris for details. 143959191Skris [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 144059191Skris 144159191Skris *) Rebuild of the memory allocation routines used by OpenSSL code and 144259191Skris possibly others as well. The purpose is to make an interface that 144359191Skris provide hooks so anyone can build a separate set of allocation and 144459191Skris deallocation routines to be used by OpenSSL, for example memory 144559191Skris pool implementations, or something else, which was previously hard 144659191Skris since Malloc(), Realloc() and Free() were defined as macros having 144759191Skris the values malloc, realloc and free, respectively (except for Win32 144859191Skris compilations). The same is provided for memory debugging code. 144959191Skris OpenSSL already comes with functionality to find memory leaks, but 145059191Skris this gives people a chance to debug other memory problems. 145159191Skris 145259191Skris With these changes, a new set of functions and macros have appeared: 145359191Skris 145468651Skris CRYPTO_set_mem_debug_functions() [F] 145568651Skris CRYPTO_get_mem_debug_functions() [F] 145668651Skris CRYPTO_dbg_set_options() [F] 145768651Skris CRYPTO_dbg_get_options() [F] 145868651Skris CRYPTO_malloc_debug_init() [M] 145959191Skris 146059191Skris The memory debug functions are NULL by default, unless the library 146159191Skris is compiled with CRYPTO_MDEBUG or friends is defined. If someone 146259191Skris wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 146359191Skris gives the standard debugging functions that come with OpenSSL) or 146459191Skris CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 146559191Skris provided by the library user) must be used. When the standard 146659191Skris debugging functions are used, CRYPTO_dbg_set_options can be used to 146759191Skris request additional information: 146859191Skris CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 146959191Skris the CRYPTO_MDEBUG_xxx macro when compiling the library. 147059191Skris 147159191Skris Also, things like CRYPTO_set_mem_functions will always give the 147259191Skris expected result (the new set of functions is used for allocation 147359191Skris and deallocation) at all times, regardless of platform and compiler 147459191Skris options. 147559191Skris 147659191Skris To finish it up, some functions that were never use in any other 147759191Skris way than through macros have a new API and new semantic: 147859191Skris 147959191Skris CRYPTO_dbg_malloc() 148059191Skris CRYPTO_dbg_realloc() 148159191Skris CRYPTO_dbg_free() 148259191Skris 148359191Skris All macros of value have retained their old syntax. 148459191Skris [Richard Levitte and Bodo Moeller] 148559191Skris 148659191Skris *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 148759191Skris ordering of SMIMECapabilities wasn't in "strength order" and there 148859191Skris was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 148959191Skris algorithm. 149059191Skris [Steve Henson] 149159191Skris 149259191Skris *) Some ASN1 types with illegal zero length encoding (INTEGER, 149359191Skris ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 149459191Skris [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 149559191Skris 149659191Skris *) Merge in my S/MIME library for OpenSSL. This provides a simple 149759191Skris S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 149859191Skris functionality to handle multipart/signed properly) and a utility 149959191Skris called 'smime' to call all this stuff. This is based on code I 150059191Skris originally wrote for Celo who have kindly allowed it to be 150159191Skris included in OpenSSL. 150259191Skris [Steve Henson] 150359191Skris 150459191Skris *) Add variants des_set_key_checked and des_set_key_unchecked of 150559191Skris des_set_key (aka des_key_sched). Global variable des_check_key 150659191Skris decides which of these is called by des_set_key; this way 150759191Skris des_check_key behaves as it always did, but applications and 150859191Skris the library itself, which was buggy for des_check_key == 1, 150959191Skris have a cleaner way to pick the version they need. 151059191Skris [Bodo Moeller] 151159191Skris 151259191Skris *) New function PKCS12_newpass() which changes the password of a 151359191Skris PKCS12 structure. 151459191Skris [Steve Henson] 151559191Skris 151659191Skris *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 151759191Skris dynamic mix. In both cases the ids can be used as an index into the 151859191Skris table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 151959191Skris functions so they accept a list of the field values and the 152059191Skris application doesn't need to directly manipulate the X509_TRUST 152159191Skris structure. 152259191Skris [Steve Henson] 152359191Skris 152459191Skris *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 152559191Skris need initialising. 152659191Skris [Steve Henson] 152759191Skris 152859191Skris *) Modify the way the V3 extension code looks up extensions. This now 152959191Skris works in a similar way to the object code: we have some "standard" 153059191Skris extensions in a static table which is searched with OBJ_bsearch() 153159191Skris and the application can add dynamic ones if needed. The file 153259191Skris crypto/x509v3/ext_dat.h now has the info: this file needs to be 153359191Skris updated whenever a new extension is added to the core code and kept 153459191Skris in ext_nid order. There is a simple program 'tabtest.c' which checks 153559191Skris this. New extensions are not added too often so this file can readily 153659191Skris be maintained manually. 153759191Skris 153859191Skris There are two big advantages in doing things this way. The extensions 153959191Skris can be looked up immediately and no longer need to be "added" using 154059191Skris X509V3_add_standard_extensions(): this function now does nothing. 154159191Skris [Side note: I get *lots* of email saying the extension code doesn't 154259191Skris work because people forget to call this function] 154359191Skris Also no dynamic allocation is done unless new extensions are added: 154459191Skris so if we don't add custom extensions there is no need to call 154559191Skris X509V3_EXT_cleanup(). 154659191Skris [Steve Henson] 154759191Skris 154859191Skris *) Modify enc utility's salting as follows: make salting the default. Add a 154959191Skris magic header, so unsalted files fail gracefully instead of just decrypting 155059191Skris to garbage. This is because not salting is a big security hole, so people 155159191Skris should be discouraged from doing it. 155259191Skris [Ben Laurie] 155359191Skris 155459191Skris *) Fixes and enhancements to the 'x509' utility. It allowed a message 155559191Skris digest to be passed on the command line but it only used this 155659191Skris parameter when signing a certificate. Modified so all relevant 155759191Skris operations are affected by the digest parameter including the 155859191Skris -fingerprint and -x509toreq options. Also -x509toreq choked if a 155959191Skris DSA key was used because it didn't fix the digest. 156059191Skris [Steve Henson] 156159191Skris 156259191Skris *) Initial certificate chain verify code. Currently tests the untrusted 156359191Skris certificates for consistency with the verify purpose (which is set 156459191Skris when the X509_STORE_CTX structure is set up) and checks the pathlength. 156559191Skris 156659191Skris There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 156759191Skris this is because it will reject chains with invalid extensions whereas 156859191Skris every previous version of OpenSSL and SSLeay made no checks at all. 156959191Skris 157059191Skris Trust code: checks the root CA for the relevant trust settings. Trust 157159191Skris settings have an initial value consistent with the verify purpose: e.g. 157259191Skris if the verify purpose is for SSL client use it expects the CA to be 157359191Skris trusted for SSL client use. However the default value can be changed to 157459191Skris permit custom trust settings: one example of this would be to only trust 157559191Skris certificates from a specific "secure" set of CAs. 157659191Skris 157759191Skris Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 157859191Skris which should be used for version portability: especially since the 157959191Skris verify structure is likely to change more often now. 158059191Skris 158159191Skris SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 158259191Skris to set them. If not set then assume SSL clients will verify SSL servers 158359191Skris and vice versa. 158459191Skris 158559191Skris Two new options to the verify program: -untrusted allows a set of 158659191Skris untrusted certificates to be passed in and -purpose which sets the 158759191Skris intended purpose of the certificate. If a purpose is set then the 158859191Skris new chain verify code is used to check extension consistency. 158959191Skris [Steve Henson] 159059191Skris 159159191Skris *) Support for the authority information access extension. 159259191Skris [Steve Henson] 159359191Skris 159459191Skris *) Modify RSA and DSA PEM read routines to transparently handle 159559191Skris PKCS#8 format private keys. New *_PUBKEY_* functions that handle 159659191Skris public keys in a format compatible with certificate 159759191Skris SubjectPublicKeyInfo structures. Unfortunately there were already 159859191Skris functions called *_PublicKey_* which used various odd formats so 159959191Skris these are retained for compatibility: however the DSA variants were 160059191Skris never in a public release so they have been deleted. Changed dsa/rsa 160159191Skris utilities to handle the new format: note no releases ever handled public 160259191Skris keys so we should be OK. 160359191Skris 160459191Skris The primary motivation for this change is to avoid the same fiasco 160559191Skris that dogs private keys: there are several incompatible private key 160659191Skris formats some of which are standard and some OpenSSL specific and 160759191Skris require various evil hacks to allow partial transparent handling and 160859191Skris even then it doesn't work with DER formats. Given the option anything 160959191Skris other than PKCS#8 should be dumped: but the other formats have to 161059191Skris stay in the name of compatibility. 161159191Skris 161259191Skris With public keys and the benefit of hindsight one standard format 161359191Skris is used which works with EVP_PKEY, RSA or DSA structures: though 161459191Skris it clearly returns an error if you try to read the wrong kind of key. 161559191Skris 161659191Skris Added a -pubkey option to the 'x509' utility to output the public key. 161759191Skris Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 161859191Skris (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 161959191Skris EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 162059191Skris that do the same as the EVP_PKEY_assign_*() except they up the 162159191Skris reference count of the added key (they don't "swallow" the 162259191Skris supplied key). 162359191Skris [Steve Henson] 162459191Skris 162559191Skris *) Fixes to crypto/x509/by_file.c the code to read in certificates and 162659191Skris CRLs would fail if the file contained no certificates or no CRLs: 162759191Skris added a new function to read in both types and return the number 162859191Skris read: this means that if none are read it will be an error. The 162959191Skris DER versions of the certificate and CRL reader would always fail 163059191Skris because it isn't possible to mix certificates and CRLs in DER format 163159191Skris without choking one or the other routine. Changed this to just read 163259191Skris a certificate: this is the best we can do. Also modified the code 163359191Skris in apps/verify.c to take notice of return codes: it was previously 163459191Skris attempting to read in certificates from NULL pointers and ignoring 163559191Skris any errors: this is one reason why the cert and CRL reader seemed 163659191Skris to work. It doesn't check return codes from the default certificate 163759191Skris routines: these may well fail if the certificates aren't installed. 163859191Skris [Steve Henson] 163959191Skris 164059191Skris *) Code to support otherName option in GeneralName. 164159191Skris [Steve Henson] 164259191Skris 164359191Skris *) First update to verify code. Change the verify utility 164459191Skris so it warns if it is passed a self signed certificate: 164559191Skris for consistency with the normal behaviour. X509_verify 164659191Skris has been modified to it will now verify a self signed 164759191Skris certificate if *exactly* the same certificate appears 164859191Skris in the store: it was previously impossible to trust a 164959191Skris single self signed certificate. This means that: 165059191Skris openssl verify ss.pem 165159191Skris now gives a warning about a self signed certificate but 165259191Skris openssl verify -CAfile ss.pem ss.pem 165359191Skris is OK. 165459191Skris [Steve Henson] 165559191Skris 165659191Skris *) For servers, store verify_result in SSL_SESSION data structure 165759191Skris (and add it to external session representation). 165859191Skris This is needed when client certificate verifications fails, 165959191Skris but an application-provided verification callback (set by 166059191Skris SSL_CTX_set_cert_verify_callback) allows accepting the session 166159191Skris anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 166259191Skris but returns 1): When the session is reused, we have to set 166359191Skris ssl->verify_result to the appropriate error code to avoid 166459191Skris security holes. 166559191Skris [Bodo Moeller, problem pointed out by Lutz Jaenicke] 166659191Skris 166759191Skris *) Fix a bug in the new PKCS#7 code: it didn't consider the 166859191Skris case in PKCS7_dataInit() where the signed PKCS7 structure 166959191Skris didn't contain any existing data because it was being created. 167059191Skris [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 167159191Skris 167259191Skris *) Add a salt to the key derivation routines in enc.c. This 167359191Skris forms the first 8 bytes of the encrypted file. Also add a 167459191Skris -S option to allow a salt to be input on the command line. 167559191Skris [Steve Henson] 167659191Skris 167759191Skris *) New function X509_cmp(). Oddly enough there wasn't a function 167859191Skris to compare two certificates. We do this by working out the SHA1 167959191Skris hash and comparing that. X509_cmp() will be needed by the trust 168059191Skris code. 168159191Skris [Steve Henson] 168259191Skris 168359191Skris *) SSL_get1_session() is like SSL_get_session(), but increments 168459191Skris the reference count in the SSL_SESSION returned. 168559191Skris [Geoff Thorpe <geoff@eu.c2.net>] 168659191Skris 168759191Skris *) Fix for 'req': it was adding a null to request attributes. 168859191Skris Also change the X509_LOOKUP and X509_INFO code to handle 168959191Skris certificate auxiliary information. 169059191Skris [Steve Henson] 169159191Skris 169259191Skris *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 169359191Skris the 'enc' command. 169459191Skris [Steve Henson] 169559191Skris 169659191Skris *) Add the possibility to add extra information to the memory leak 169759191Skris detecting output, to form tracebacks, showing from where each 169859191Skris allocation was originated: CRYPTO_push_info("constant string") adds 169959191Skris the string plus current file name and line number to a per-thread 170059191Skris stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 170159191Skris is like calling CYRPTO_pop_info() until the stack is empty. 170259191Skris Also updated memory leak detection code to be multi-thread-safe. 170359191Skris [Richard Levitte] 170459191Skris 170559191Skris *) Add options -text and -noout to pkcs7 utility and delete the 170659191Skris encryption options which never did anything. Update docs. 170759191Skris [Steve Henson] 170859191Skris 170959191Skris *) Add options to some of the utilities to allow the pass phrase 171059191Skris to be included on either the command line (not recommended on 171159191Skris OSes like Unix) or read from the environment. Update the 171259191Skris manpages and fix a few bugs. 171359191Skris [Steve Henson] 171459191Skris 171559191Skris *) Add a few manpages for some of the openssl commands. 171659191Skris [Steve Henson] 171759191Skris 171859191Skris *) Fix the -revoke option in ca. It was freeing up memory twice, 171959191Skris leaking and not finding already revoked certificates. 172059191Skris [Steve Henson] 172159191Skris 172259191Skris *) Extensive changes to support certificate auxiliary information. 172359191Skris This involves the use of X509_CERT_AUX structure and X509_AUX 172459191Skris functions. An X509_AUX function such as PEM_read_X509_AUX() 172559191Skris can still read in a certificate file in the usual way but it 172659191Skris will also read in any additional "auxiliary information". By 172759191Skris doing things this way a fair degree of compatibility can be 172859191Skris retained: existing certificates can have this information added 172959191Skris using the new 'x509' options. 173059191Skris 173159191Skris Current auxiliary information includes an "alias" and some trust 173259191Skris settings. The trust settings will ultimately be used in enhanced 173359191Skris certificate chain verification routines: currently a certificate 173459191Skris can only be trusted if it is self signed and then it is trusted 173559191Skris for all purposes. 173659191Skris [Steve Henson] 173759191Skris 173859191Skris *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 173959191Skris The problem was that one of the replacement routines had not been working 174059191Skris since SSLeay releases. For now the offending routine has been replaced 174159191Skris with non-optimised assembler. Even so, this now gives around 95% 174259191Skris performance improvement for 1024 bit RSA signs. 174359191Skris [Mark Cox] 174459191Skris 174559191Skris *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 174659191Skris handling. Most clients have the effective key size in bits equal to 174759191Skris the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 174859191Skris A few however don't do this and instead use the size of the decrypted key 174959191Skris to determine the RC2 key length and the AlgorithmIdentifier to determine 175059191Skris the effective key length. In this case the effective key length can still 175159191Skris be 40 bits but the key length can be 168 bits for example. This is fixed 175259191Skris by manually forcing an RC2 key into the EVP_PKEY structure because the 175359191Skris EVP code can't currently handle unusual RC2 key sizes: it always assumes 175459191Skris the key length and effective key length are equal. 175559191Skris [Steve Henson] 175659191Skris 175759191Skris *) Add a bunch of functions that should simplify the creation of 175859191Skris X509_NAME structures. Now you should be able to do: 175959191Skris X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 176059191Skris and have it automatically work out the correct field type and fill in 176159191Skris the structures. The more adventurous can try: 176259191Skris X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 176359191Skris and it will (hopefully) work out the correct multibyte encoding. 176459191Skris [Steve Henson] 176559191Skris 176659191Skris *) Change the 'req' utility to use the new field handling and multibyte 176759191Skris copy routines. Before the DN field creation was handled in an ad hoc 176859191Skris way in req, ca, and x509 which was rather broken and didn't support 176959191Skris BMPStrings or UTF8Strings. Since some software doesn't implement 177059191Skris BMPStrings or UTF8Strings yet, they can be enabled using the config file 177159191Skris using the dirstring_type option. See the new comment in the default 177259191Skris openssl.cnf for more info. 177359191Skris [Steve Henson] 177459191Skris 177559191Skris *) Make crypto/rand/md_rand.c more robust: 177659191Skris - Assure unique random numbers after fork(). 177759191Skris - Make sure that concurrent threads access the global counter and 177859191Skris md serializably so that we never lose entropy in them 177959191Skris or use exactly the same state in multiple threads. 178059191Skris Access to the large state is not always serializable because 178159191Skris the additional locking could be a performance killer, and 178259191Skris md should be large enough anyway. 178359191Skris [Bodo Moeller] 178459191Skris 178559191Skris *) New file apps/app_rand.c with commonly needed functionality 178659191Skris for handling the random seed file. 178759191Skris 178859191Skris Use the random seed file in some applications that previously did not: 178959191Skris ca, 179059191Skris dsaparam -genkey (which also ignored its '-rand' option), 179159191Skris s_client, 179259191Skris s_server, 179359191Skris x509 (when signing). 179459191Skris Except on systems with /dev/urandom, it is crucial to have a random 179559191Skris seed file at least for key creation, DSA signing, and for DH exchanges; 179659191Skris for RSA signatures we could do without one. 179759191Skris 179859191Skris gendh and gendsa (unlike genrsa) used to read only the first byte 179959191Skris of each file listed in the '-rand' option. The function as previously 180059191Skris found in genrsa is now in app_rand.c and is used by all programs 180159191Skris that support '-rand'. 180259191Skris [Bodo Moeller] 180359191Skris 180459191Skris *) In RAND_write_file, use mode 0600 for creating files; 180559191Skris don't just chmod when it may be too late. 180659191Skris [Bodo Moeller] 180759191Skris 180859191Skris *) Report an error from X509_STORE_load_locations 180959191Skris when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 181059191Skris [Bill Perry] 181159191Skris 181259191Skris *) New function ASN1_mbstring_copy() this copies a string in either 181359191Skris ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 181459191Skris into an ASN1_STRING type. A mask of permissible types is passed 181559191Skris and it chooses the "minimal" type to use or an error if not type 181659191Skris is suitable. 181759191Skris [Steve Henson] 181859191Skris 181959191Skris *) Add function equivalents to the various macros in asn1.h. The old 182059191Skris macros are retained with an M_ prefix. Code inside the library can 182159191Skris use the M_ macros. External code (including the openssl utility) 182259191Skris should *NOT* in order to be "shared library friendly". 182359191Skris [Steve Henson] 182459191Skris 182559191Skris *) Add various functions that can check a certificate's extensions 182659191Skris to see if it usable for various purposes such as SSL client, 182759191Skris server or S/MIME and CAs of these types. This is currently 182859191Skris VERY EXPERIMENTAL but will ultimately be used for certificate chain 182959191Skris verification. Also added a -purpose flag to x509 utility to 183059191Skris print out all the purposes. 183159191Skris [Steve Henson] 183259191Skris 183359191Skris *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 183459191Skris functions. 183559191Skris [Steve Henson] 183659191Skris 183759191Skris *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 183859191Skris for, obtain and decode and extension and obtain its critical flag. 183959191Skris This allows all the necessary extension code to be handled in a 184059191Skris single function call. 184159191Skris [Steve Henson] 184259191Skris 184359191Skris *) RC4 tune-up featuring 30-40% performance improvement on most RISC 184459191Skris platforms. See crypto/rc4/rc4_enc.c for further details. 184559191Skris [Andy Polyakov] 184659191Skris 184759191Skris *) New -noout option to asn1parse. This causes no output to be produced 184859191Skris its main use is when combined with -strparse and -out to extract data 184959191Skris from a file (which may not be in ASN.1 format). 185059191Skris [Steve Henson] 185159191Skris 185259191Skris *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 185359191Skris when producing the local key id. 185459191Skris [Richard Levitte <levitte@stacken.kth.se>] 185559191Skris 185659191Skris *) New option -dhparam in s_server. This allows a DH parameter file to be 185759191Skris stated explicitly. If it is not stated then it tries the first server 185859191Skris certificate file. The previous behaviour hard coded the filename 185959191Skris "server.pem". 186059191Skris [Steve Henson] 186159191Skris 186259191Skris *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 186359191Skris a public key to be input or output. For example: 186459191Skris openssl rsa -in key.pem -pubout -out pubkey.pem 186559191Skris Also added necessary DSA public key functions to handle this. 186659191Skris [Steve Henson] 186759191Skris 186859191Skris *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 186959191Skris in the message. This was handled by allowing 187059191Skris X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 187159191Skris [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 187259191Skris 187359191Skris *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 187459191Skris to the end of the strings whereas this didn't. This would cause problems 187559191Skris if strings read with d2i_ASN1_bytes() were later modified. 187659191Skris [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 187759191Skris 187859191Skris *) Fix for base64 decode bug. When a base64 bio reads only one line of 187959191Skris data and it contains EOF it will end up returning an error. This is 188059191Skris caused by input 46 bytes long. The cause is due to the way base64 188159191Skris BIOs find the start of base64 encoded data. They do this by trying a 188259191Skris trial decode on each line until they find one that works. When they 188359191Skris do a flag is set and it starts again knowing it can pass all the 188459191Skris data directly through the decoder. Unfortunately it doesn't reset 188559191Skris the context it uses. This means that if EOF is reached an attempt 188659191Skris is made to pass two EOFs through the context and this causes the 188759191Skris resulting error. This can also cause other problems as well. As is 188859191Skris usual with these problems it takes *ages* to find and the fix is 188959191Skris trivial: move one line. 189059191Skris [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 189159191Skris 189259191Skris *) Ugly workaround to get s_client and s_server working under Windows. The 189359191Skris old code wouldn't work because it needed to select() on sockets and the 189459191Skris tty (for keypresses and to see if data could be written). Win32 only 189559191Skris supports select() on sockets so we select() with a 1s timeout on the 189659191Skris sockets and then see if any characters are waiting to be read, if none 189759191Skris are present then we retry, we also assume we can always write data to 189859191Skris the tty. This isn't nice because the code then blocks until we've 189959191Skris received a complete line of data and it is effectively polling the 190059191Skris keyboard at 1s intervals: however it's quite a bit better than not 190159191Skris working at all :-) A dedicated Windows application might handle this 190259191Skris with an event loop for example. 190359191Skris [Steve Henson] 190459191Skris 190559191Skris *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 190659191Skris and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 190759191Skris will be called when RSA_sign() and RSA_verify() are used. This is useful 190859191Skris if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 190959191Skris For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 191059191Skris should *not* be used: RSA_sign() and RSA_verify() must be used instead. 191159191Skris This necessitated the support of an extra signature type NID_md5_sha1 191259191Skris for SSL signatures and modifications to the SSL library to use it instead 191359191Skris of calling RSA_public_decrypt() and RSA_private_encrypt(). 191459191Skris [Steve Henson] 191559191Skris 191659191Skris *) Add new -verify -CAfile and -CApath options to the crl program, these 191759191Skris will lookup a CRL issuers certificate and verify the signature in a 191859191Skris similar way to the verify program. Tidy up the crl program so it 191959191Skris no longer accesses structures directly. Make the ASN1 CRL parsing a bit 192059191Skris less strict. It will now permit CRL extensions even if it is not 192159191Skris a V2 CRL: this will allow it to tolerate some broken CRLs. 192259191Skris [Steve Henson] 192359191Skris 192459191Skris *) Initialize all non-automatic variables each time one of the openssl 192559191Skris sub-programs is started (this is necessary as they may be started 192659191Skris multiple times from the "OpenSSL>" prompt). 192759191Skris [Lennart Bang, Bodo Moeller] 192859191Skris 192959191Skris *) Preliminary compilation option RSA_NULL which disables RSA crypto without 193059191Skris removing all other RSA functionality (this is what NO_RSA does). This 193159191Skris is so (for example) those in the US can disable those operations covered 193259191Skris by the RSA patent while allowing storage and parsing of RSA keys and RSA 193359191Skris key generation. 193459191Skris [Steve Henson] 193559191Skris 193659191Skris *) Non-copying interface to BIO pairs. 193759191Skris (still largely untested) 193859191Skris [Bodo Moeller] 193959191Skris 194059191Skris *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 194159191Skris ASCII string. This was handled independently in various places before. 194259191Skris [Steve Henson] 194359191Skris 194459191Skris *) New functions UTF8_getc() and UTF8_putc() that parse and generate 194559191Skris UTF8 strings a character at a time. 194659191Skris [Steve Henson] 194759191Skris 194859191Skris *) Use client_version from client hello to select the protocol 194959191Skris (s23_srvr.c) and for RSA client key exchange verification 195059191Skris (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 195159191Skris [Bodo Moeller] 195259191Skris 195359191Skris *) Add various utility functions to handle SPKACs, these were previously 195459191Skris handled by poking round in the structure internals. Added new function 195559191Skris NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 195659191Skris print, verify and generate SPKACs. Based on an original idea from 195759191Skris Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 195859191Skris [Steve Henson] 195959191Skris 196059191Skris *) RIPEMD160 is operational on all platforms and is back in 'make test'. 196159191Skris [Andy Polyakov] 196259191Skris 196359191Skris *) Allow the config file extension section to be overwritten on the 196459191Skris command line. Based on an original idea from Massimiliano Pala 196559191Skris <madwolf@comune.modena.it>. The new option is called -extensions 196659191Skris and can be applied to ca, req and x509. Also -reqexts to override 196759191Skris the request extensions in req and -crlexts to override the crl extensions 196859191Skris in ca. 196959191Skris [Steve Henson] 197059191Skris 197159191Skris *) Add new feature to the SPKAC handling in ca. Now you can include 197259191Skris the same field multiple times by preceding it by "XXXX." for example: 197359191Skris 1.OU="Unit name 1" 197459191Skris 2.OU="Unit name 2" 197559191Skris this is the same syntax as used in the req config file. 197659191Skris [Steve Henson] 197759191Skris 197859191Skris *) Allow certificate extensions to be added to certificate requests. These 197959191Skris are specified in a 'req_extensions' option of the req section of the 198059191Skris config file. They can be printed out with the -text option to req but 198159191Skris are otherwise ignored at present. 198259191Skris [Steve Henson] 198359191Skris 198459191Skris *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 198559191Skris data read consists of only the final block it would not decrypted because 198659191Skris EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 198759191Skris A misplaced 'break' also meant the decrypted final block might not be 198859191Skris copied until the next read. 198959191Skris [Steve Henson] 199059191Skris 199159191Skris *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 199259191Skris a few extra parameters to the DH structure: these will be useful if 199359191Skris for example we want the value of 'q' or implement X9.42 DH. 199459191Skris [Steve Henson] 199559191Skris 199659191Skris *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 199759191Skris provides hooks that allow the default DSA functions or functions on a 199859191Skris "per key" basis to be replaced. This allows hardware acceleration and 199959191Skris hardware key storage to be handled without major modification to the 200059191Skris library. Also added low level modexp hooks and CRYPTO_EX structure and 200159191Skris associated functions. 200259191Skris [Steve Henson] 200359191Skris 200459191Skris *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 200559191Skris as "read only": it can't be written to and the buffer it points to will 200659191Skris not be freed. Reading from a read only BIO is much more efficient than 200759191Skris a normal memory BIO. This was added because there are several times when 200859191Skris an area of memory needs to be read from a BIO. The previous method was 200959191Skris to create a memory BIO and write the data to it, this results in two 201059191Skris copies of the data and an O(n^2) reading algorithm. There is a new 201159191Skris function BIO_new_mem_buf() which creates a read only memory BIO from 201259191Skris an area of memory. Also modified the PKCS#7 routines to use read only 201359191Skris memory BIOs. 201459191Skris [Steve Henson] 201559191Skris 201659191Skris *) Bugfix: ssl23_get_client_hello did not work properly when called in 201759191Skris state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 201859191Skris a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 201959191Skris but a retry condition occured while trying to read the rest. 202059191Skris [Bodo Moeller] 202159191Skris 202259191Skris *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 202359191Skris NID_pkcs7_encrypted by default: this was wrong since this should almost 202459191Skris always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 202559191Skris the encrypted data type: this is a more sensible place to put it and it 202659191Skris allows the PKCS#12 code to be tidied up that duplicated this 202759191Skris functionality. 202859191Skris [Steve Henson] 202959191Skris 203059191Skris *) Changed obj_dat.pl script so it takes its input and output files on 203159191Skris the command line. This should avoid shell escape redirection problems 203259191Skris under Win32. 203359191Skris [Steve Henson] 203459191Skris 203559191Skris *) Initial support for certificate extension requests, these are included 203659191Skris in things like Xenroll certificate requests. Included functions to allow 203759191Skris extensions to be obtained and added. 203859191Skris [Steve Henson] 203959191Skris 204059191Skris *) -crlf option to s_client and s_server for sending newlines as 204159191Skris CRLF (as required by many protocols). 204259191Skris [Bodo Moeller] 204359191Skris 204455714Skris Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 204555714Skris 204655714Skris *) Install libRSAglue.a when OpenSSL is built with RSAref. 204755714Skris [Ralf S. Engelschall] 204855714Skris 204955714Skris *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 205055714Skris [Andrija Antonijevic <TheAntony2@bigfoot.com>] 205155714Skris 205255714Skris *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 205355714Skris program. 205455714Skris [Steve Henson] 205555714Skris 205655714Skris *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 205755714Skris DH parameters/keys (q is lost during that conversion, but the resulting 205855714Skris DH parameters contain its length). 205955714Skris 206055714Skris For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 206155714Skris much faster than DH_generate_parameters (which creates parameters 206255714Skris where p = 2*q + 1), and also the smaller q makes DH computations 206355714Skris much more efficient (160-bit exponentiation instead of 1024-bit 206455714Skris exponentiation); so this provides a convenient way to support DHE 206555714Skris ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 206655714Skris utter importance to use 206755714Skris SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 206855714Skris or 206955714Skris SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 207055714Skris when such DH parameters are used, because otherwise small subgroup 207155714Skris attacks may become possible! 207255714Skris [Bodo Moeller] 207355714Skris 207455714Skris *) Avoid memory leak in i2d_DHparams. 207555714Skris [Bodo Moeller] 207655714Skris 207755714Skris *) Allow the -k option to be used more than once in the enc program: 207855714Skris this allows the same encrypted message to be read by multiple recipients. 207955714Skris [Steve Henson] 208055714Skris 208155714Skris *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 208255714Skris an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 208355714Skris it will always use the numerical form of the OID, even if it has a short 208455714Skris or long name. 208555714Skris [Steve Henson] 208655714Skris 208755714Skris *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 208855714Skris method only got called if p,q,dmp1,dmq1,iqmp components were present, 208955714Skris otherwise bn_mod_exp was called. In the case of hardware keys for example 209055714Skris no private key components need be present and it might store extra data 209159191Skris in the RSA structure, which cannot be accessed from bn_mod_exp. 209259191Skris By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 209359191Skris private key operations. 209455714Skris [Steve Henson] 209555714Skris 209655714Skris *) Added support for SPARC Linux. 209755714Skris [Andy Polyakov] 209855714Skris 209955714Skris *) pem_password_cb function type incompatibly changed from 210055714Skris typedef int pem_password_cb(char *buf, int size, int rwflag); 210155714Skris to 210255714Skris ....(char *buf, int size, int rwflag, void *userdata); 210355714Skris so that applications can pass data to their callbacks: 210455714Skris The PEM[_ASN1]_{read,write}... functions and macros now take an 210555714Skris additional void * argument, which is just handed through whenever 210655714Skris the password callback is called. 210759191Skris [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 210855714Skris 210955714Skris New function SSL_CTX_set_default_passwd_cb_userdata. 211055714Skris 211155714Skris Compatibility note: As many C implementations push function arguments 211255714Skris onto the stack in reverse order, the new library version is likely to 211355714Skris interoperate with programs that have been compiled with the old 211455714Skris pem_password_cb definition (PEM_whatever takes some data that 211555714Skris happens to be on the stack as its last argument, and the callback 211655714Skris just ignores this garbage); but there is no guarantee whatsoever that 211755714Skris this will work. 211855714Skris 211955714Skris *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 212055714Skris (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 212155714Skris problems not only on Windows, but also on some Unix platforms. 212255714Skris To avoid problematic command lines, these definitions are now in an 212355714Skris auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 212455714Skris for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 212555714Skris [Bodo Moeller] 212655714Skris 212755714Skris *) MIPS III/IV assembler module is reimplemented. 212855714Skris [Andy Polyakov] 212955714Skris 213055714Skris *) More DES library cleanups: remove references to srand/rand and 213155714Skris delete an unused file. 213255714Skris [Ulf M�ller] 213355714Skris 213455714Skris *) Add support for the the free Netwide assembler (NASM) under Win32, 213555714Skris since not many people have MASM (ml) and it can be hard to obtain. 213655714Skris This is currently experimental but it seems to work OK and pass all 213755714Skris the tests. Check out INSTALL.W32 for info. 213855714Skris [Steve Henson] 213955714Skris 214055714Skris *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 214155714Skris without temporary keys kept an extra copy of the server key, 214255714Skris and connections with temporary keys did not free everything in case 214355714Skris of an error. 214455714Skris [Bodo Moeller] 214555714Skris 214655714Skris *) New function RSA_check_key and new openssl rsa option -check 214755714Skris for verifying the consistency of RSA keys. 214855714Skris [Ulf Moeller, Bodo Moeller] 214955714Skris 215055714Skris *) Various changes to make Win32 compile work: 215155714Skris 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 215255714Skris 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 215355714Skris comparison" warnings. 215455714Skris 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 215555714Skris [Steve Henson] 215655714Skris 215755714Skris *) Add a debugging option to PKCS#5 v2 key generation function: when 215855714Skris you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 215955714Skris derived keys are printed to stderr. 216055714Skris [Steve Henson] 216155714Skris 216255714Skris *) Copy the flags in ASN1_STRING_dup(). 216355714Skris [Roman E. Pavlov <pre@mo.msk.ru>] 216455714Skris 216555714Skris *) The x509 application mishandled signing requests containing DSA 216655714Skris keys when the signing key was also DSA and the parameters didn't match. 216755714Skris 216855714Skris It was supposed to omit the parameters when they matched the signing key: 216955714Skris the verifying software was then supposed to automatically use the CA's 217055714Skris parameters if they were absent from the end user certificate. 217155714Skris 217255714Skris Omitting parameters is no longer recommended. The test was also 217355714Skris the wrong way round! This was probably due to unusual behaviour in 217455714Skris EVP_cmp_parameters() which returns 1 if the parameters match. 217555714Skris This meant that parameters were omitted when they *didn't* match and 217655714Skris the certificate was useless. Certificates signed with 'ca' didn't have 217755714Skris this bug. 217855714Skris [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 217955714Skris 218055714Skris *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 218155714Skris The interface is as follows: 218255714Skris Applications can use 218355714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 218455714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 218555714Skris "off" is now the default. 218655714Skris The library internally uses 218755714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 218855714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 218955714Skris to disable memory-checking temporarily. 219055714Skris 219155714Skris Some inconsistent states that previously were possible (and were 219255714Skris even the default) are now avoided. 219355714Skris 219455714Skris -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 219555714Skris with each memory chunk allocated; this is occasionally more helpful 219655714Skris than just having a counter. 219755714Skris 219855714Skris -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 219955714Skris 220055714Skris -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 220155714Skris extensions. 220255714Skris [Bodo Moeller] 220355714Skris 220455714Skris *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 220555714Skris which largely parallels "options", but is for changing API behaviour, 220655714Skris whereas "options" are about protocol behaviour. 220755714Skris Initial "mode" flags are: 220855714Skris 220955714Skris SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 221055714Skris a single record has been written. 221155714Skris SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 221255714Skris retries use the same buffer location. 221355714Skris (But all of the contents must be 221455714Skris copied!) 221555714Skris [Bodo Moeller] 221655714Skris 221755714Skris *) Bugfix: SSL_set_mode ignored its parameter, only SSL_CTX_set_mode 221855714Skris worked. 221955714Skris 222055714Skris *) Fix problems with no-hmac etc. 222155714Skris [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 222255714Skris 222355714Skris *) New functions RSA_get_default_method(), RSA_set_method() and 222455714Skris RSA_get_method(). These allows replacement of RSA_METHODs without having 222555714Skris to mess around with the internals of an RSA structure. 222655714Skris [Steve Henson] 222755714Skris 222855714Skris *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 222955714Skris Also really enable memory leak checks in openssl.c and in some 223055714Skris test programs. 223155714Skris [Chad C. Mulligan, Bodo Moeller] 223255714Skris 223355714Skris *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 223455714Skris up the length of negative integers. This has now been simplified to just 223555714Skris store the length when it is first determined and use it later, rather 223655714Skris than trying to keep track of where data is copied and updating it to 223755714Skris point to the end. 223855714Skris [Steve Henson, reported by Brien Wheeler 223955714Skris <bwheeler@authentica-security.com>] 224055714Skris 224155714Skris *) Add a new function PKCS7_signatureVerify. This allows the verification 224255714Skris of a PKCS#7 signature but with the signing certificate passed to the 224355714Skris function itself. This contrasts with PKCS7_dataVerify which assumes the 224455714Skris certificate is present in the PKCS#7 structure. This isn't always the 224555714Skris case: certificates can be omitted from a PKCS#7 structure and be 224655714Skris distributed by "out of band" means (such as a certificate database). 224755714Skris [Steve Henson] 224855714Skris 224955714Skris *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 225055714Skris function prototypes in pem.h, also change util/mkdef.pl to add the 225155714Skris necessary function names. 225255714Skris [Steve Henson] 225355714Skris 225455714Skris *) mk1mf.pl (used by Windows builds) did not properly read the 225555714Skris options set by Configure in the top level Makefile, and Configure 225655714Skris was not even able to write more than one option correctly. 225755714Skris Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 225855714Skris [Bodo Moeller] 225955714Skris 226055714Skris *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 226155714Skris file to be loaded from a BIO or FILE pointer. The BIO version will 226255714Skris for example allow memory BIOs to contain config info. 226355714Skris [Steve Henson] 226455714Skris 226555714Skris *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 226655714Skris Whoever hopes to achieve shared-library compatibility across versions 226755714Skris must use this, not the compile-time macro. 226855714Skris (Exercise 0.9.4: Which is the minimum library version required by 226955714Skris such programs?) 227055714Skris Note: All this applies only to multi-threaded programs, others don't 227155714Skris need locks. 227255714Skris [Bodo Moeller] 227355714Skris 227455714Skris *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 227555714Skris through a BIO pair triggered the default case, i.e. 227655714Skris SSLerr(...,SSL_R_UNKNOWN_STATE). 227755714Skris [Bodo Moeller] 227855714Skris 227955714Skris *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 228055714Skris can use the SSL library even if none of the specific BIOs is 228155714Skris appropriate. 228255714Skris [Bodo Moeller] 228355714Skris 228455714Skris *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 228555714Skris for the encoded length. 228655714Skris [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 228755714Skris 228855714Skris *) Add initial documentation of the X509V3 functions. 228955714Skris [Steve Henson] 229055714Skris 229155714Skris *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 229255714Skris PEM_write_bio_PKCS8PrivateKey() that are equivalent to 229355714Skris PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 229455714Skris secure PKCS#8 private key format with a high iteration count. 229555714Skris [Steve Henson] 229655714Skris 229755714Skris *) Fix determination of Perl interpreter: A perl or perl5 229855714Skris _directory_ in $PATH was also accepted as the interpreter. 229955714Skris [Ralf S. Engelschall] 230055714Skris 230155714Skris *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 230255714Skris wrong with it but it was very old and did things like calling 230355714Skris PEM_ASN1_read() directly and used MD5 for the hash not to mention some 230455714Skris unusual formatting. 230555714Skris [Steve Henson] 230655714Skris 230755714Skris *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 230855714Skris to use the new extension code. 230955714Skris [Steve Henson] 231055714Skris 231155714Skris *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 231255714Skris with macros. This should make it easier to change their form, add extra 231355714Skris arguments etc. Fix a few PEM prototypes which didn't have cipher as a 231455714Skris constant. 231555714Skris [Steve Henson] 231655714Skris 231755714Skris *) Add to configuration table a new entry that can specify an alternative 231855714Skris name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 231955714Skris according to Mark Crispin <MRC@Panda.COM>. 232055714Skris [Bodo Moeller] 232155714Skris 232255714Skris#if 0 232355714Skris *) DES CBC did not update the IV. Weird. 232455714Skris [Ben Laurie] 232555714Skris#else 232655714Skris des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 232755714Skris Changing the behaviour of the former might break existing programs -- 232855714Skris where IV updating is needed, des_ncbc_encrypt can be used. 232955714Skris#endif 233055714Skris 233155714Skris *) When bntest is run from "make test" it drives bc to check its 233255714Skris calculations, as well as internally checking them. If an internal check 233355714Skris fails, it needs to cause bc to give a non-zero result or make test carries 233455714Skris on without noticing the failure. Fixed. 233555714Skris [Ben Laurie] 233655714Skris 233755714Skris *) DES library cleanups. 233855714Skris [Ulf M�ller] 233955714Skris 234055714Skris *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 234155714Skris used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 234255714Skris ciphers. NOTE: although the key derivation function has been verified 234355714Skris against some published test vectors it has not been extensively tested 234455714Skris yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 234555714Skris of v2.0. 234655714Skris [Steve Henson] 234755714Skris 234855714Skris *) Instead of "mkdir -p", which is not fully portable, use new 234955714Skris Perl script "util/mkdir-p.pl". 235055714Skris [Bodo Moeller] 235155714Skris 235255714Skris *) Rewrite the way password based encryption (PBE) is handled. It used to 235355714Skris assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 235455714Skris structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 235555714Skris but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 235655714Skris the 'parameter' field of the AlgorithmIdentifier is passed to the 235755714Skris underlying key generation function so it must do its own ASN1 parsing. 235855714Skris This has also changed the EVP_PBE_CipherInit() function which now has a 235955714Skris 'parameter' argument instead of literal salt and iteration count values 236055714Skris and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 236155714Skris [Steve Henson] 236255714Skris 236355714Skris *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 236455714Skris and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 236555714Skris Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 236655714Skris KEY" because this clashed with PKCS#8 unencrypted string. Since this 236755714Skris value was just used as a "magic string" and not used directly its 236855714Skris value doesn't matter. 236955714Skris [Steve Henson] 237055714Skris 237155714Skris *) Introduce some semblance of const correctness to BN. Shame C doesn't 237255714Skris support mutable. 237355714Skris [Ben Laurie] 237455714Skris 237555714Skris *) "linux-sparc64" configuration (ultrapenguin). 237655714Skris [Ray Miller <ray.miller@oucs.ox.ac.uk>] 237755714Skris "linux-sparc" configuration. 237855714Skris [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 237955714Skris 238055714Skris *) config now generates no-xxx options for missing ciphers. 238155714Skris [Ulf M�ller] 238255714Skris 238355714Skris *) Support the EBCDIC character set (work in progress). 238455714Skris File ebcdic.c not yet included because it has a different license. 238555714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 238655714Skris 238755714Skris *) Support BS2000/OSD-POSIX. 238855714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 238955714Skris 239055714Skris *) Make callbacks for key generation use void * instead of char *. 239155714Skris [Ben Laurie] 239255714Skris 239355714Skris *) Make S/MIME samples compile (not yet tested). 239455714Skris [Ben Laurie] 239555714Skris 239655714Skris *) Additional typesafe stacks. 239755714Skris [Ben Laurie] 239855714Skris 239955714Skris *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 240055714Skris [Bodo Moeller] 240155714Skris 240255714Skris 240355714Skris Changes between 0.9.3 and 0.9.3a [29 May 1999] 240455714Skris 240555714Skris *) New configuration variant "sco5-gcc". 240655714Skris 240755714Skris *) Updated some demos. 240855714Skris [Sean O Riordain, Wade Scholine] 240955714Skris 241055714Skris *) Add missing BIO_free at exit of pkcs12 application. 241155714Skris [Wu Zhigang] 241255714Skris 241355714Skris *) Fix memory leak in conf.c. 241455714Skris [Steve Henson] 241555714Skris 241655714Skris *) Updates for Win32 to assembler version of MD5. 241755714Skris [Steve Henson] 241855714Skris 241955714Skris *) Set #! path to perl in apps/der_chop to where we found it 242055714Skris instead of using a fixed path. 242155714Skris [Bodo Moeller] 242255714Skris 242355714Skris *) SHA library changes for irix64-mips4-cc. 242455714Skris [Andy Polyakov] 242555714Skris 242655714Skris *) Improvements for VMS support. 242755714Skris [Richard Levitte] 242855714Skris 242955714Skris 243055714Skris Changes between 0.9.2b and 0.9.3 [24 May 1999] 243155714Skris 243255714Skris *) Bignum library bug fix. IRIX 6 passes "make test" now! 243355714Skris This also avoids the problems with SC4.2 and unpatched SC5. 243455714Skris [Andy Polyakov <appro@fy.chalmers.se>] 243555714Skris 243655714Skris *) New functions sk_num, sk_value and sk_set to replace the previous macros. 243755714Skris These are required because of the typesafe stack would otherwise break 243855714Skris existing code. If old code used a structure member which used to be STACK 243955714Skris and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 244055714Skris sk_num or sk_value it would produce an error because the num, data members 244155714Skris are not present in STACK_OF. Now it just produces a warning. sk_set 244255714Skris replaces the old method of assigning a value to sk_value 244355714Skris (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 244455714Skris that does this will no longer work (and should use sk_set instead) but 244555714Skris this could be regarded as a "questionable" behaviour anyway. 244655714Skris [Steve Henson] 244755714Skris 244855714Skris *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 244955714Skris correctly handle encrypted S/MIME data. 245055714Skris [Steve Henson] 245155714Skris 245255714Skris *) Change type of various DES function arguments from des_cblock 245355714Skris (which means, in function argument declarations, pointer to char) 245455714Skris to des_cblock * (meaning pointer to array with 8 char elements), 245555714Skris which allows the compiler to do more typechecking; it was like 245655714Skris that back in SSLeay, but with lots of ugly casts. 245755714Skris 245855714Skris Introduce new type const_des_cblock. 245955714Skris [Bodo Moeller] 246055714Skris 246155714Skris *) Reorganise the PKCS#7 library and get rid of some of the more obvious 246255714Skris problems: find RecipientInfo structure that matches recipient certificate 246355714Skris and initialise the ASN1 structures properly based on passed cipher. 246455714Skris [Steve Henson] 246555714Skris 246655714Skris *) Belatedly make the BN tests actually check the results. 246755714Skris [Ben Laurie] 246855714Skris 246955714Skris *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 247055714Skris to and from BNs: it was completely broken. New compilation option 247155714Skris NEG_PUBKEY_BUG to allow for some broken certificates that encode public 247255714Skris key elements as negative integers. 247355714Skris [Steve Henson] 247455714Skris 247555714Skris *) Reorganize and speed up MD5. 247655714Skris [Andy Polyakov <appro@fy.chalmers.se>] 247755714Skris 247855714Skris *) VMS support. 247955714Skris [Richard Levitte <richard@levitte.org>] 248055714Skris 248155714Skris *) New option -out to asn1parse to allow the parsed structure to be 248255714Skris output to a file. This is most useful when combined with the -strparse 248355714Skris option to examine the output of things like OCTET STRINGS. 248455714Skris [Steve Henson] 248555714Skris 248655714Skris *) Make SSL library a little more fool-proof by not requiring any longer 248755714Skris that SSL_set_{accept,connect}_state be called before 248855714Skris SSL_{accept,connect} may be used (SSL_set_..._state is omitted 248955714Skris in many applications because usually everything *appeared* to work as 249055714Skris intended anyway -- now it really works as intended). 249155714Skris [Bodo Moeller] 249255714Skris 249355714Skris *) Move openssl.cnf out of lib/. 249455714Skris [Ulf M�ller] 249555714Skris 249655714Skris *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 249755714Skris -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 249855714Skris -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 249955714Skris [Ralf S. Engelschall] 250055714Skris 250155714Skris *) Various fixes to the EVP and PKCS#7 code. It may now be able to 250255714Skris handle PKCS#7 enveloped data properly. 250355714Skris [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 250455714Skris 250555714Skris *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 250655714Skris copying pointers. The cert_st handling is changed by this in 250755714Skris various ways (and thus what used to be known as ctx->default_cert 250855714Skris is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 250955714Skris any longer when s->cert does not give us what we need). 251055714Skris ssl_cert_instantiate becomes obsolete by this change. 251155714Skris As soon as we've got the new code right (possibly it already is?), 251255714Skris we have solved a couple of bugs of the earlier code where s->cert 251355714Skris was used as if it could not have been shared with other SSL structures. 251455714Skris 251555714Skris Note that using the SSL API in certain dirty ways now will result 251655714Skris in different behaviour than observed with earlier library versions: 251755714Skris Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 251855714Skris does not influence s as it used to. 251955714Skris 252055714Skris In order to clean up things more thoroughly, inside SSL_SESSION 252155714Skris we don't use CERT any longer, but a new structure SESS_CERT 252255714Skris that holds per-session data (if available); currently, this is 252355714Skris the peer's certificate chain and, for clients, the server's certificate 252455714Skris and temporary key. CERT holds only those values that can have 252555714Skris meaningful defaults in an SSL_CTX. 252655714Skris [Bodo Moeller] 252755714Skris 252855714Skris *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 252955714Skris from the internal representation. Various PKCS#7 fixes: remove some 253055714Skris evil casts and set the enc_dig_alg field properly based on the signing 253155714Skris key type. 253255714Skris [Steve Henson] 253355714Skris 253455714Skris *) Allow PKCS#12 password to be set from the command line or the 253555714Skris environment. Let 'ca' get its config file name from the environment 253655714Skris variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 253755714Skris and 'x509'). 253855714Skris [Steve Henson] 253955714Skris 254055714Skris *) Allow certificate policies extension to use an IA5STRING for the 254155714Skris organization field. This is contrary to the PKIX definition but 254255714Skris VeriSign uses it and IE5 only recognises this form. Document 'x509' 254355714Skris extension option. 254455714Skris [Steve Henson] 254555714Skris 254655714Skris *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 254755714Skris without disallowing inline assembler and the like for non-pedantic builds. 254855714Skris [Ben Laurie] 254955714Skris 255055714Skris *) Support Borland C++ builder. 255155714Skris [Janez Jere <jj@void.si>, modified by Ulf M�ller] 255255714Skris 255355714Skris *) Support Mingw32. 255455714Skris [Ulf M�ller] 255555714Skris 255655714Skris *) SHA-1 cleanups and performance enhancements. 255755714Skris [Andy Polyakov <appro@fy.chalmers.se>] 255855714Skris 255955714Skris *) Sparc v8plus assembler for the bignum library. 256055714Skris [Andy Polyakov <appro@fy.chalmers.se>] 256155714Skris 256255714Skris *) Accept any -xxx and +xxx compiler options in Configure. 256355714Skris [Ulf M�ller] 256455714Skris 256555714Skris *) Update HPUX configuration. 256655714Skris [Anonymous] 256755714Skris 256855714Skris *) Add missing sk_<type>_unshift() function to safestack.h 256955714Skris [Ralf S. Engelschall] 257055714Skris 257155714Skris *) New function SSL_CTX_use_certificate_chain_file that sets the 257255714Skris "extra_cert"s in addition to the certificate. (This makes sense 257355714Skris only for "PEM" format files, as chains as a whole are not 257455714Skris DER-encoded.) 257555714Skris [Bodo Moeller] 257655714Skris 257755714Skris *) Support verify_depth from the SSL API. 257855714Skris x509_vfy.c had what can be considered an off-by-one-error: 257955714Skris Its depth (which was not part of the external interface) 258055714Skris was actually counting the number of certificates in a chain; 258155714Skris now it really counts the depth. 258255714Skris [Bodo Moeller] 258355714Skris 258455714Skris *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 258555714Skris instead of X509err, which often resulted in confusing error 258655714Skris messages since the error codes are not globally unique 258755714Skris (e.g. an alleged error in ssl3_accept when a certificate 258855714Skris didn't match the private key). 258955714Skris 259055714Skris *) New function SSL_CTX_set_session_id_context that allows to set a default 259155714Skris value (so that you don't need SSL_set_session_id_context for each 259255714Skris connection using the SSL_CTX). 259355714Skris [Bodo Moeller] 259455714Skris 259555714Skris *) OAEP decoding bug fix. 259655714Skris [Ulf M�ller] 259755714Skris 259855714Skris *) Support INSTALL_PREFIX for package builders, as proposed by 259955714Skris David Harris. 260055714Skris [Bodo Moeller] 260155714Skris 260255714Skris *) New Configure options "threads" and "no-threads". For systems 260355714Skris where the proper compiler options are known (currently Solaris 260455714Skris and Linux), "threads" is the default. 260555714Skris [Bodo Moeller] 260655714Skris 260755714Skris *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 260855714Skris [Bodo Moeller] 260955714Skris 261055714Skris *) Install various scripts to $(OPENSSLDIR)/misc, not to 261155714Skris $(INSTALLTOP)/bin -- they shouldn't clutter directories 261255714Skris such as /usr/local/bin. 261355714Skris [Bodo Moeller] 261455714Skris 261555714Skris *) "make linux-shared" to build shared libraries. 261655714Skris [Niels Poppe <niels@netbox.org>] 261755714Skris 261855714Skris *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 261955714Skris [Ulf M�ller] 262055714Skris 262155714Skris *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 262255714Skris extension adding in x509 utility. 262355714Skris [Steve Henson] 262455714Skris 262555714Skris *) Remove NOPROTO sections and error code comments. 262655714Skris [Ulf M�ller] 262755714Skris 262855714Skris *) Partial rewrite of the DEF file generator to now parse the ANSI 262955714Skris prototypes. 263055714Skris [Steve Henson] 263155714Skris 263255714Skris *) New Configure options --prefix=DIR and --openssldir=DIR. 263355714Skris [Ulf M�ller] 263455714Skris 263555714Skris *) Complete rewrite of the error code script(s). It is all now handled 263655714Skris by one script at the top level which handles error code gathering, 263755714Skris header rewriting and C source file generation. It should be much better 263855714Skris than the old method: it now uses a modified version of Ulf's parser to 263955714Skris read the ANSI prototypes in all header files (thus the old K&R definitions 264055714Skris aren't needed for error creation any more) and do a better job of 264155714Skris translating function codes into names. The old 'ASN1 error code imbedded 264255714Skris in a comment' is no longer necessary and it doesn't use .err files which 264355714Skris have now been deleted. Also the error code call doesn't have to appear all 264455714Skris on one line (which resulted in some large lines...). 264555714Skris [Steve Henson] 264655714Skris 264755714Skris *) Change #include filenames from <foo.h> to <openssl/foo.h>. 264855714Skris [Bodo Moeller] 264955714Skris 265055714Skris *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 265155714Skris 0 (which usually indicates a closed connection), but continue reading. 265255714Skris [Bodo Moeller] 265355714Skris 265455714Skris *) Fix some race conditions. 265555714Skris [Bodo Moeller] 265655714Skris 265755714Skris *) Add support for CRL distribution points extension. Add Certificate 265855714Skris Policies and CRL distribution points documentation. 265955714Skris [Steve Henson] 266055714Skris 266155714Skris *) Move the autogenerated header file parts to crypto/opensslconf.h. 266255714Skris [Ulf M�ller] 266355714Skris 266455714Skris *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 266555714Skris 8 of keying material. Merlin has also confirmed interop with this fix 266655714Skris between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 266755714Skris [Merlin Hughes <merlin@baltimore.ie>] 266855714Skris 266955714Skris *) Fix lots of warnings. 267055714Skris [Richard Levitte <levitte@stacken.kth.se>] 267155714Skris 267255714Skris *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 267355714Skris the directory spec didn't end with a LIST_SEPARATOR_CHAR. 267455714Skris [Richard Levitte <levitte@stacken.kth.se>] 267555714Skris 267655714Skris *) Fix problems with sizeof(long) == 8. 267755714Skris [Andy Polyakov <appro@fy.chalmers.se>] 267855714Skris 267955714Skris *) Change functions to ANSI C. 268055714Skris [Ulf M�ller] 268155714Skris 268255714Skris *) Fix typos in error codes. 268355714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 268455714Skris 268555714Skris *) Remove defunct assembler files from Configure. 268655714Skris [Ulf M�ller] 268755714Skris 268855714Skris *) SPARC v8 assembler BIGNUM implementation. 268955714Skris [Andy Polyakov <appro@fy.chalmers.se>] 269055714Skris 269155714Skris *) Support for Certificate Policies extension: both print and set. 269255714Skris Various additions to support the r2i method this uses. 269355714Skris [Steve Henson] 269455714Skris 269555714Skris *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 269655714Skris return a const string when you are expecting an allocated buffer. 269755714Skris [Ben Laurie] 269855714Skris 269955714Skris *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 270055714Skris types DirectoryString and DisplayText. 270155714Skris [Steve Henson] 270255714Skris 270355714Skris *) Add code to allow r2i extensions to access the configuration database, 270455714Skris add an LHASH database driver and add several ctx helper functions. 270555714Skris [Steve Henson] 270655714Skris 270755714Skris *) Fix an evil bug in bn_expand2() which caused various BN functions to 270855714Skris fail when they extended the size of a BIGNUM. 270955714Skris [Steve Henson] 271055714Skris 271155714Skris *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 271255714Skris support typesafe stack. 271355714Skris [Steve Henson] 271455714Skris 271555714Skris *) Fix typo in SSL_[gs]et_options(). 271655714Skris [Nils Frostberg <nils@medcom.se>] 271755714Skris 271855714Skris *) Delete various functions and files that belonged to the (now obsolete) 271955714Skris old X509V3 handling code. 272055714Skris [Steve Henson] 272155714Skris 272255714Skris *) New Configure option "rsaref". 272355714Skris [Ulf M�ller] 272455714Skris 272555714Skris *) Don't auto-generate pem.h. 272655714Skris [Bodo Moeller] 272755714Skris 272855714Skris *) Introduce type-safe ASN.1 SETs. 272955714Skris [Ben Laurie] 273055714Skris 273155714Skris *) Convert various additional casted stacks to type-safe STACK_OF() variants. 273255714Skris [Ben Laurie, Ralf S. Engelschall, Steve Henson] 273355714Skris 273455714Skris *) Introduce type-safe STACKs. This will almost certainly break lots of code 273555714Skris that links with OpenSSL (well at least cause lots of warnings), but fear 273655714Skris not: the conversion is trivial, and it eliminates loads of evil casts. A 273755714Skris few STACKed things have been converted already. Feel free to convert more. 273855714Skris In the fullness of time, I'll do away with the STACK type altogether. 273955714Skris [Ben Laurie] 274055714Skris 274155714Skris *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 274255714Skris specified in <certfile> by updating the entry in the index.txt file. 274355714Skris This way one no longer has to edit the index.txt file manually for 274455714Skris revoking a certificate. The -revoke option does the gory details now. 274555714Skris [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 274655714Skris 274755714Skris *) Fix `openssl crl -noout -text' combination where `-noout' killed the 274855714Skris `-text' option at all and this way the `-noout -text' combination was 274955714Skris inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 275055714Skris [Ralf S. Engelschall] 275155714Skris 275255714Skris *) Make sure a corresponding plain text error message exists for the 275355714Skris X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 275455714Skris verify callback function determined that a certificate was revoked. 275555714Skris [Ralf S. Engelschall] 275655714Skris 275755714Skris *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 275855714Skris ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 275955714Skris all available cipers including rc5, which was forgotten until now. 276055714Skris In order to let the testing shell script know which algorithms 276155714Skris are available, a new (up to now undocumented) command 276255714Skris "openssl list-cipher-commands" is used. 276355714Skris [Bodo Moeller] 276455714Skris 276555714Skris *) Bugfix: s_client occasionally would sleep in select() when 276655714Skris it should have checked SSL_pending() first. 276755714Skris [Bodo Moeller] 276855714Skris 276955714Skris *) New functions DSA_do_sign and DSA_do_verify to provide access to 277055714Skris the raw DSA values prior to ASN.1 encoding. 277155714Skris [Ulf M�ller] 277255714Skris 277355714Skris *) Tweaks to Configure 277455714Skris [Niels Poppe <niels@netbox.org>] 277555714Skris 277655714Skris *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 277755714Skris yet... 277855714Skris [Steve Henson] 277955714Skris 278055714Skris *) New variables $(RANLIB) and $(PERL) in the Makefiles. 278155714Skris [Ulf M�ller] 278255714Skris 278355714Skris *) New config option to avoid instructions that are illegal on the 80386. 278455714Skris The default code is faster, but requires at least a 486. 278555714Skris [Ulf M�ller] 278655714Skris 278755714Skris *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 278855714Skris SSL2_SERVER_VERSION (not used at all) macros, which are now the 278955714Skris same as SSL2_VERSION anyway. 279055714Skris [Bodo Moeller] 279155714Skris 279255714Skris *) New "-showcerts" option for s_client. 279355714Skris [Bodo Moeller] 279455714Skris 279555714Skris *) Still more PKCS#12 integration. Add pkcs12 application to openssl 279655714Skris application. Various cleanups and fixes. 279755714Skris [Steve Henson] 279855714Skris 279955714Skris *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 280055714Skris modify error routines to work internally. Add error codes and PBE init 280155714Skris to library startup routines. 280255714Skris [Steve Henson] 280355714Skris 280455714Skris *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 280555714Skris packing functions to asn1 and evp. Changed function names and error 280655714Skris codes along the way. 280755714Skris [Steve Henson] 280855714Skris 280955714Skris *) PKCS12 integration: and so it begins... First of several patches to 281055714Skris slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 281155714Skris objects to objects.h 281255714Skris [Steve Henson] 281355714Skris 281455714Skris *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 281555714Skris and display support for Thawte strong extranet extension. 281655714Skris [Steve Henson] 281755714Skris 281855714Skris *) Add LinuxPPC support. 281955714Skris [Jeff Dubrule <igor@pobox.org>] 282055714Skris 282155714Skris *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 282255714Skris bn_div_words in alpha.s. 282355714Skris [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 282455714Skris 282555714Skris *) Make sure the RSA OAEP test is skipped under -DRSAref because 282655714Skris OAEP isn't supported when OpenSSL is built with RSAref. 282755714Skris [Ulf Moeller <ulf@fitug.de>] 282855714Skris 282955714Skris *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 283055714Skris so they no longer are missing under -DNOPROTO. 283155714Skris [Soren S. Jorvang <soren@t.dk>] 283255714Skris 283355714Skris 283455714Skris Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 283555714Skris 283655714Skris *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 283755714Skris doesn't work when the session is reused. Coming soon! 283855714Skris [Ben Laurie] 283955714Skris 284055714Skris *) Fix a security hole, that allows sessions to be reused in the wrong 284155714Skris context thus bypassing client cert protection! All software that uses 284255714Skris client certs and session caches in multiple contexts NEEDS PATCHING to 284355714Skris allow session reuse! A fuller solution is in the works. 284455714Skris [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 284555714Skris 284655714Skris *) Some more source tree cleanups (removed obsolete files 284755714Skris crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 284855714Skris permission on "config" script to be executable) and a fix for the INSTALL 284955714Skris document. 285055714Skris [Ulf Moeller <ulf@fitug.de>] 285155714Skris 285255714Skris *) Remove some legacy and erroneous uses of malloc, free instead of 285355714Skris Malloc, Free. 285455714Skris [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 285555714Skris 285655714Skris *) Make rsa_oaep_test return non-zero on error. 285755714Skris [Ulf Moeller <ulf@fitug.de>] 285855714Skris 285955714Skris *) Add support for native Solaris shared libraries. Configure 286055714Skris solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 286155714Skris if someone would make that last step automatic. 286255714Skris [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 286355714Skris 286455714Skris *) ctx_size was not built with the right compiler during "make links". Fixed. 286555714Skris [Ben Laurie] 286655714Skris 286755714Skris *) Change the meaning of 'ALL' in the cipher list. It now means "everything 286855714Skris except NULL ciphers". This means the default cipher list will no longer 286955714Skris enable NULL ciphers. They need to be specifically enabled e.g. with 287055714Skris the string "DEFAULT:eNULL". 287155714Skris [Steve Henson] 287255714Skris 287355714Skris *) Fix to RSA private encryption routines: if p < q then it would 287455714Skris occasionally produce an invalid result. This will only happen with 287555714Skris externally generated keys because OpenSSL (and SSLeay) ensure p > q. 287655714Skris [Steve Henson] 287755714Skris 287855714Skris *) Be less restrictive and allow also `perl util/perlpath.pl 287955714Skris /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 288055714Skris because this way one can also use an interpreter named `perl5' (which is 288155714Skris usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 288255714Skris installed as `perl'). 288355714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 288455714Skris 288555714Skris *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 288655714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 288755714Skris 288855714Skris *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 288955714Skris advapi32.lib to Win32 build and change the pem test comparision 289055714Skris to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 289155714Skris suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 289255714Skris and crypto/des/ede_cbcm_enc.c. 289355714Skris [Steve Henson] 289455714Skris 289555714Skris *) DES quad checksum was broken on big-endian architectures. Fixed. 289655714Skris [Ben Laurie] 289755714Skris 289855714Skris *) Comment out two functions in bio.h that aren't implemented. Fix up the 289955714Skris Win32 test batch file so it (might) work again. The Win32 test batch file 290055714Skris is horrible: I feel ill.... 290155714Skris [Steve Henson] 290255714Skris 290355714Skris *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 290455714Skris in e_os.h. Audit of header files to check ANSI and non ANSI 290555714Skris sections: 10 functions were absent from non ANSI section and not exported 290655714Skris from Windows DLLs. Fixed up libeay.num for new functions. 290755714Skris [Steve Henson] 290855714Skris 290955714Skris *) Make `openssl version' output lines consistent. 291055714Skris [Ralf S. Engelschall] 291155714Skris 291255714Skris *) Fix Win32 symbol export lists for BIO functions: Added 291355714Skris BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 291455714Skris to ms/libeay{16,32}.def. 291555714Skris [Ralf S. Engelschall] 291655714Skris 291755714Skris *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 291855714Skris fine under Unix and passes some trivial tests I've now added. But the 291955714Skris whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 292055714Skris added to make sure no one expects that this stuff really works in the 292155714Skris OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 292255714Skris up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 292355714Skris openssl_bio.xs. 292455714Skris [Ralf S. Engelschall] 292555714Skris 292655714Skris *) Fix the generation of two part addresses in perl. 292755714Skris [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 292855714Skris 292955714Skris *) Add config entry for Linux on MIPS. 293055714Skris [John Tobey <jtobey@channel1.com>] 293155714Skris 293255714Skris *) Make links whenever Configure is run, unless we are on Windoze. 293355714Skris [Ben Laurie] 293455714Skris 293555714Skris *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 293655714Skris Currently only issuerAltName and AuthorityKeyIdentifier make any sense 293755714Skris in CRLs. 293855714Skris [Steve Henson] 293955714Skris 294055714Skris *) Add a useful kludge to allow package maintainers to specify compiler and 294155714Skris other platforms details on the command line without having to patch the 294255714Skris Configure script everytime: One now can use ``perl Configure 294355714Skris <id>:<details>'', i.e. platform ids are allowed to have details appended 294455714Skris to them (seperated by colons). This is treated as there would be a static 294555714Skris pre-configured entry in Configure's %table under key <id> with value 294655714Skris <details> and ``perl Configure <id>'' is called. So, when you want to 294755714Skris perform a quick test-compile under FreeBSD 3.1 with pgcc and without 294855714Skris assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 294955714Skris now, which overrides the FreeBSD-elf entry on-the-fly. 295055714Skris [Ralf S. Engelschall] 295155714Skris 295255714Skris *) Disable new TLS1 ciphersuites by default: they aren't official yet. 295355714Skris [Ben Laurie] 295455714Skris 295555714Skris *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 295655714Skris on the `perl Configure ...' command line. This way one can compile 295755714Skris OpenSSL libraries with Position Independent Code (PIC) which is needed 295855714Skris for linking it into DSOs. 295955714Skris [Ralf S. Engelschall] 296055714Skris 296155714Skris *) Remarkably, export ciphers were totally broken and no-one had noticed! 296255714Skris Fixed. 296355714Skris [Ben Laurie] 296455714Skris 296555714Skris *) Cleaned up the LICENSE document: The official contact for any license 296655714Skris questions now is the OpenSSL core team under openssl-core@openssl.org. 296755714Skris And add a paragraph about the dual-license situation to make sure people 296855714Skris recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 296955714Skris to the OpenSSL toolkit. 297055714Skris [Ralf S. Engelschall] 297155714Skris 297255714Skris *) General source tree makefile cleanups: Made `making xxx in yyy...' 297355714Skris display consistent in the source tree and replaced `/bin/rm' by `rm'. 297455714Skris Additonally cleaned up the `make links' target: Remove unnecessary 297555714Skris semicolons, subsequent redundant removes, inline point.sh into mklink.sh 297655714Skris to speed processing and no longer clutter the display with confusing 297755714Skris stuff. Instead only the actually done links are displayed. 297855714Skris [Ralf S. Engelschall] 297955714Skris 298055714Skris *) Permit null encryption ciphersuites, used for authentication only. It used 298155714Skris to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 298255714Skris It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 298355714Skris encryption. 298455714Skris [Ben Laurie] 298555714Skris 298655714Skris *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 298755714Skris signed attributes when verifying signatures (this would break them), 298855714Skris the detached data encoding was wrong and public keys obtained using 298955714Skris X509_get_pubkey() weren't freed. 299055714Skris [Steve Henson] 299155714Skris 299255714Skris *) Add text documentation for the BUFFER functions. Also added a work around 299355714Skris to a Win95 console bug. This was triggered by the password read stuff: the 299455714Skris last character typed gets carried over to the next fread(). If you were 299555714Skris generating a new cert request using 'req' for example then the last 299655714Skris character of the passphrase would be CR which would then enter the first 299755714Skris field as blank. 299855714Skris [Steve Henson] 299955714Skris 300055714Skris *) Added the new `Includes OpenSSL Cryptography Software' button as 300155714Skris doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 300255714Skris button and can be used by applications based on OpenSSL to show the 300355714Skris relationship to the OpenSSL project. 300455714Skris [Ralf S. Engelschall] 300555714Skris 300655714Skris *) Remove confusing variables in function signatures in files 300755714Skris ssl/ssl_lib.c and ssl/ssl.h. 300855714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 300955714Skris 301055714Skris *) Don't install bss_file.c under PREFIX/include/ 301155714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 301255714Skris 301355714Skris *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 301455714Skris functions that return function pointers and has support for NT specific 301555714Skris stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 301655714Skris #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 301755714Skris unsigned to signed types: this was killing the Win32 compile. 301855714Skris [Steve Henson] 301955714Skris 302055714Skris *) Add new certificate file to stack functions, 302155714Skris SSL_add_dir_cert_subjects_to_stack() and 302255714Skris SSL_add_file_cert_subjects_to_stack(). These largely supplant 302355714Skris SSL_load_client_CA_file(), and can be used to add multiple certs easily 302455714Skris to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 302555714Skris This means that Apache-SSL and similar packages don't have to mess around 302655714Skris to add as many CAs as they want to the preferred list. 302755714Skris [Ben Laurie] 302855714Skris 302955714Skris *) Experiment with doxygen documentation. Currently only partially applied to 303055714Skris ssl/ssl_lib.c. 303155714Skris See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 303255714Skris openssl.doxy as the configuration file. 303355714Skris [Ben Laurie] 303455714Skris 303555714Skris *) Get rid of remaining C++-style comments which strict C compilers hate. 303655714Skris [Ralf S. Engelschall, pointed out by Carlos Amengual] 303755714Skris 303855714Skris *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 303955714Skris compiled in by default: it has problems with large keys. 304055714Skris [Steve Henson] 304155714Skris 304255714Skris *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 304355714Skris DH private keys and/or callback functions which directly correspond to 304455714Skris their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 304555714Skris is needed for applications which have to configure certificates on a 304655714Skris per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 304755714Skris (e.g. s_server). 304855714Skris For the RSA certificate situation is makes no difference, but 304955714Skris for the DSA certificate situation this fixes the "no shared cipher" 305055714Skris problem where the OpenSSL cipher selection procedure failed because the 305155714Skris temporary keys were not overtaken from the context and the API provided 305255714Skris no way to reconfigure them. 305355714Skris The new functions now let applications reconfigure the stuff and they 305455714Skris are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 305555714Skris SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 305655714Skris non-public-API function ssl_cert_instantiate() is used as a helper 305755714Skris function and also to reduce code redundancy inside ssl_rsa.c. 305855714Skris [Ralf S. Engelschall] 305955714Skris 306055714Skris *) Move s_server -dcert and -dkey options out of the undocumented feature 306155714Skris area because they are useful for the DSA situation and should be 306255714Skris recognized by the users. 306355714Skris [Ralf S. Engelschall] 306455714Skris 306555714Skris *) Fix the cipher decision scheme for export ciphers: the export bits are 306655714Skris *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 306755714Skris SSL_EXP_MASK. So, the original variable has to be used instead of the 306855714Skris already masked variable. 306955714Skris [Richard Levitte <levitte@stacken.kth.se>] 307055714Skris 307155714Skris *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 307255714Skris [Richard Levitte <levitte@stacken.kth.se>] 307355714Skris 307455714Skris *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 307555714Skris from `int' to `unsigned int' because it's a length and initialized by 307655714Skris EVP_DigestFinal() which expects an `unsigned int *'. 307755714Skris [Richard Levitte <levitte@stacken.kth.se>] 307855714Skris 307955714Skris *) Don't hard-code path to Perl interpreter on shebang line of Configure 308055714Skris script. Instead use the usual Shell->Perl transition trick. 308155714Skris [Ralf S. Engelschall] 308255714Skris 308355714Skris *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 308455714Skris (in addition to RSA certificates) to match the behaviour of `openssl dsa 308555714Skris -noout -modulus' as it's already the case for `openssl rsa -noout 308655714Skris -modulus'. For RSA the -modulus is the real "modulus" while for DSA 308755714Skris currently the public key is printed (a decision which was already done by 308855714Skris `openssl dsa -modulus' in the past) which serves a similar purpose. 308955714Skris Additionally the NO_RSA no longer completely removes the whole -modulus 309055714Skris option; it now only avoids using the RSA stuff. Same applies to NO_DSA 309155714Skris now, too. 309255714Skris [Ralf S. Engelschall] 309355714Skris 309455714Skris *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 309555714Skris BIO. See the source (crypto/evp/bio_ok.c) for more info. 309655714Skris [Arne Ansper <arne@ats.cyber.ee>] 309755714Skris 309855714Skris *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 309955714Skris to be added. Now both 'req' and 'ca' can use new objects defined in the 310055714Skris config file. 310155714Skris [Steve Henson] 310255714Skris 310355714Skris *) Add cool BIO that does syslog (or event log on NT). 310455714Skris [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 310555714Skris 310655714Skris *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 310755714Skris TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 310855714Skris TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 310955714Skris Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 311055714Skris [Ben Laurie] 311155714Skris 311255714Skris *) Add preliminary config info for new extension code. 311355714Skris [Steve Henson] 311455714Skris 311555714Skris *) Make RSA_NO_PADDING really use no padding. 311655714Skris [Ulf Moeller <ulf@fitug.de>] 311755714Skris 311855714Skris *) Generate errors when private/public key check is done. 311955714Skris [Ben Laurie] 312055714Skris 312155714Skris *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 312255714Skris for some CRL extensions and new objects added. 312355714Skris [Steve Henson] 312455714Skris 312555714Skris *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 312655714Skris key usage extension and fuller support for authority key id. 312755714Skris [Steve Henson] 312855714Skris 312955714Skris *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 313055714Skris padding method for RSA, which is recommended for new applications in PKCS 313155714Skris #1 v2.0 (RFC 2437, October 1998). 313255714Skris OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 313355714Skris foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 313455714Skris against Bleichbacher's attack on RSA. 313555714Skris [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 313655714Skris Ben Laurie] 313755714Skris 313855714Skris *) Updates to the new SSL compression code 313955714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 314055714Skris 314155714Skris *) Fix so that the version number in the master secret, when passed 314255714Skris via RSA, checks that if TLS was proposed, but we roll back to SSLv3 314355714Skris (because the server will not accept higher), that the version number 314455714Skris is 0x03,0x01, not 0x03,0x00 314555714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 314655714Skris 314755714Skris *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 314855714Skris leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 314955714Skris in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 315055714Skris [Steve Henson] 315155714Skris 315255714Skris *) Support for RAW extensions where an arbitrary extension can be 315355714Skris created by including its DER encoding. See apps/openssl.cnf for 315455714Skris an example. 315555714Skris [Steve Henson] 315655714Skris 315755714Skris *) Make sure latest Perl versions don't interpret some generated C array 315855714Skris code as Perl array code in the crypto/err/err_genc.pl script. 315955714Skris [Lars Weber <3weber@informatik.uni-hamburg.de>] 316055714Skris 316155714Skris *) Modify ms/do_ms.bat to not generate assembly language makefiles since 316255714Skris not many people have the assembler. Various Win32 compilation fixes and 316355714Skris update to the INSTALL.W32 file with (hopefully) more accurate Win32 316455714Skris build instructions. 316555714Skris [Steve Henson] 316655714Skris 316755714Skris *) Modify configure script 'Configure' to automatically create crypto/date.h 316855714Skris file under Win32 and also build pem.h from pem.org. New script 316955714Skris util/mkfiles.pl to create the MINFO file on environments that can't do a 317055714Skris 'make files': perl util/mkfiles.pl >MINFO should work. 317155714Skris [Steve Henson] 317255714Skris 317355714Skris *) Major rework of DES function declarations, in the pursuit of correctness 317455714Skris and purity. As a result, many evil casts evaporated, and some weirdness, 317555714Skris too. You may find this causes warnings in your code. Zapping your evil 317655714Skris casts will probably fix them. Mostly. 317755714Skris [Ben Laurie] 317855714Skris 317955714Skris *) Fix for a typo in asn1.h. Bug fix to object creation script 318055714Skris obj_dat.pl. It considered a zero in an object definition to mean 318155714Skris "end of object": none of the objects in objects.h have any zeros 318255714Skris so it wasn't spotted. 318355714Skris [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 318455714Skris 318555714Skris *) Add support for Triple DES Cipher Block Chaining with Output Feedback 318655714Skris Masking (CBCM). In the absence of test vectors, the best I have been able 318755714Skris to do is check that the decrypt undoes the encrypt, so far. Send me test 318855714Skris vectors if you have them. 318955714Skris [Ben Laurie] 319055714Skris 319155714Skris *) Correct calculation of key length for export ciphers (too much space was 319255714Skris allocated for null ciphers). This has not been tested! 319355714Skris [Ben Laurie] 319455714Skris 319555714Skris *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 319655714Skris message is now correct (it understands "crypto" and "ssl" on its 319755714Skris command line). There is also now an "update" option. This will update 319855714Skris the util/ssleay.num and util/libeay.num files with any new functions. 319955714Skris If you do a: 320055714Skris perl util/mkdef.pl crypto ssl update 320155714Skris it will update them. 320255714Skris [Steve Henson] 320355714Skris 320455714Skris *) Overhauled the Perl interface (perl/*): 320555714Skris - ported BN stuff to OpenSSL's different BN library 320655714Skris - made the perl/ source tree CVS-aware 320755714Skris - renamed the package from SSLeay to OpenSSL (the files still contain 320855714Skris their history because I've copied them in the repository) 320955714Skris - removed obsolete files (the test scripts will be replaced 321055714Skris by better Test::Harness variants in the future) 321155714Skris [Ralf S. Engelschall] 321255714Skris 321355714Skris *) First cut for a very conservative source tree cleanup: 321455714Skris 1. merge various obsolete readme texts into doc/ssleay.txt 321555714Skris where we collect the old documents and readme texts. 321655714Skris 2. remove the first part of files where I'm already sure that we no 321755714Skris longer need them because of three reasons: either they are just temporary 321855714Skris files which were left by Eric or they are preserved original files where 321955714Skris I've verified that the diff is also available in the CVS via "cvs diff 322055714Skris -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 322155714Skris the crypto/md/ stuff). 322255714Skris [Ralf S. Engelschall] 322355714Skris 322455714Skris *) More extension code. Incomplete support for subject and issuer alt 322555714Skris name, issuer and authority key id. Change the i2v function parameters 322655714Skris and add an extra 'crl' parameter in the X509V3_CTX structure: guess 322755714Skris what that's for :-) Fix to ASN1 macro which messed up 322855714Skris IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 322955714Skris [Steve Henson] 323055714Skris 323155714Skris *) Preliminary support for ENUMERATED type. This is largely copied from the 323255714Skris INTEGER code. 323355714Skris [Steve Henson] 323455714Skris 323555714Skris *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 323655714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 323755714Skris 323855714Skris *) Make sure `make rehash' target really finds the `openssl' program. 323955714Skris [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 324055714Skris 324155714Skris *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 324255714Skris like to hear about it if this slows down other processors. 324355714Skris [Ben Laurie] 324455714Skris 324555714Skris *) Add CygWin32 platform information to Configure script. 324655714Skris [Alan Batie <batie@aahz.jf.intel.com>] 324755714Skris 324855714Skris *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 324955714Skris [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 325055714Skris 325155714Skris *) New program nseq to manipulate netscape certificate sequences 325255714Skris [Steve Henson] 325355714Skris 325455714Skris *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 325555714Skris few typos. 325655714Skris [Steve Henson] 325755714Skris 325855714Skris *) Fixes to BN code. Previously the default was to define BN_RECURSION 325955714Skris but the BN code had some problems that would cause failures when 326055714Skris doing certificate verification and some other functions. 326155714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 326255714Skris 326355714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 326455714Skris [Steve Henson] 326555714Skris 326655714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 326755714Skris [Steve Henson] 326855714Skris 326955714Skris *) Add several PKIX and private extended key usage OIDs. 327055714Skris [Steve Henson] 327155714Skris 327255714Skris *) Modify the 'ca' program to handle the new extension code. Modify 327355714Skris openssl.cnf for new extension format, add comments. 327455714Skris [Steve Henson] 327555714Skris 327655714Skris *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 327755714Skris and add a sample to openssl.cnf so req -x509 now adds appropriate 327855714Skris CA extensions. 327955714Skris [Steve Henson] 328055714Skris 328155714Skris *) Continued X509 V3 changes. Add to other makefiles, integrate with the 328255714Skris error code, add initial support to X509_print() and x509 application. 328355714Skris [Steve Henson] 328455714Skris 328555714Skris *) Takes a deep breath and start addding X509 V3 extension support code. Add 328655714Skris files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 328755714Skris stuff is currently isolated and isn't even compiled yet. 328855714Skris [Steve Henson] 328955714Skris 329055714Skris *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 329155714Skris ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 329255714Skris Removed the versions check from X509 routines when loading extensions: 329355714Skris this allows certain broken certificates that don't set the version 329455714Skris properly to be processed. 329555714Skris [Steve Henson] 329655714Skris 329755714Skris *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 329855714Skris Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 329955714Skris can still be regenerated with "make depend". 330055714Skris [Ben Laurie] 330155714Skris 330255714Skris *) Spelling mistake in C version of CAST-128. 330355714Skris [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 330455714Skris 330555714Skris *) Changes to the error generation code. The perl script err-code.pl 330655714Skris now reads in the old error codes and retains the old numbers, only 330755714Skris adding new ones if necessary. It also only changes the .err files if new 330855714Skris codes are added. The makefiles have been modified to only insert errors 330955714Skris when needed (to avoid needlessly modifying header files). This is done 331055714Skris by only inserting errors if the .err file is newer than the auto generated 331155714Skris C file. To rebuild all the error codes from scratch (the old behaviour) 331255714Skris either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 331355714Skris or delete all the .err files. 331455714Skris [Steve Henson] 331555714Skris 331655714Skris *) CAST-128 was incorrectly implemented for short keys. The C version has 331755714Skris been fixed, but is untested. The assembler versions are also fixed, but 331855714Skris new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 331955714Skris to regenerate it if needed. 332055714Skris [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 332155714Skris Hagino <itojun@kame.net>] 332255714Skris 332355714Skris *) File was opened incorrectly in randfile.c. 332455714Skris [Ulf M�ller <ulf@fitug.de>] 332555714Skris 332655714Skris *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 332755714Skris functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 332855714Skris GeneralizedTime. ASN1_TIME is the proper type used in certificates et 332955714Skris al: it's just almost always a UTCTime. Note this patch adds new error 333055714Skris codes so do a "make errors" if there are problems. 333155714Skris [Steve Henson] 333255714Skris 333355714Skris *) Correct Linux 1 recognition in config. 333455714Skris [Ulf M�ller <ulf@fitug.de>] 333555714Skris 333655714Skris *) Remove pointless MD5 hash when using DSA keys in ca. 333755714Skris [Anonymous <nobody@replay.com>] 333855714Skris 333955714Skris *) Generate an error if given an empty string as a cert directory. Also 334055714Skris generate an error if handed NULL (previously returned 0 to indicate an 334155714Skris error, but didn't set one). 334255714Skris [Ben Laurie, reported by Anonymous <nobody@replay.com>] 334355714Skris 334455714Skris *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 334555714Skris [Ben Laurie] 334655714Skris 334755714Skris *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 334855714Skris parameters. This was causing a warning which killed off the Win32 compile. 334955714Skris [Steve Henson] 335055714Skris 335155714Skris *) Remove C++ style comments from crypto/bn/bn_local.h. 335255714Skris [Neil Costigan <neil.costigan@celocom.com>] 335355714Skris 335455714Skris *) The function OBJ_txt2nid was broken. It was supposed to return a nid 335555714Skris based on a text string, looking up short and long names and finally 335655714Skris "dot" format. The "dot" format stuff didn't work. Added new function 335755714Skris OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 335855714Skris OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 335955714Skris OID is not part of the table. 336055714Skris [Steve Henson] 336155714Skris 336255714Skris *) Add prototypes to X509 lookup/verify methods, fixing a bug in 336355714Skris X509_LOOKUP_by_alias(). 336455714Skris [Ben Laurie] 336555714Skris 336655714Skris *) Sort openssl functions by name. 336755714Skris [Ben Laurie] 336855714Skris 336955714Skris *) Get the gendsa program working (hopefully) and add it to app list. Remove 337055714Skris encryption from sample DSA keys (in case anyone is interested the password 337155714Skris was "1234"). 337255714Skris [Steve Henson] 337355714Skris 337455714Skris *) Make _all_ *_free functions accept a NULL pointer. 337555714Skris [Frans Heymans <fheymans@isaserver.be>] 337655714Skris 337755714Skris *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 337855714Skris NULL pointers. 337955714Skris [Anonymous <nobody@replay.com>] 338055714Skris 338155714Skris *) s_server should send the CAfile as acceptable CAs, not its own cert. 338255714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 338355714Skris 338455714Skris *) Don't blow it for numeric -newkey arguments to apps/req. 338555714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 338655714Skris 338755714Skris *) Temp key "for export" tests were wrong in s3_srvr.c. 338855714Skris [Anonymous <nobody@replay.com>] 338955714Skris 339055714Skris *) Add prototype for temp key callback functions 339155714Skris SSL_CTX_set_tmp_{rsa,dh}_callback(). 339255714Skris [Ben Laurie] 339355714Skris 339455714Skris *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 339555714Skris DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 339655714Skris [Steve Henson] 339755714Skris 339855714Skris *) X509_name_add_entry() freed the wrong thing after an error. 339955714Skris [Arne Ansper <arne@ats.cyber.ee>] 340055714Skris 340155714Skris *) rsa_eay.c would attempt to free a NULL context. 340255714Skris [Arne Ansper <arne@ats.cyber.ee>] 340355714Skris 340455714Skris *) BIO_s_socket() had a broken should_retry() on Windoze. 340555714Skris [Arne Ansper <arne@ats.cyber.ee>] 340655714Skris 340755714Skris *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 340855714Skris [Arne Ansper <arne@ats.cyber.ee>] 340955714Skris 341055714Skris *) Make sure the already existing X509_STORE->depth variable is initialized 341155714Skris in X509_STORE_new(), but document the fact that this variable is still 341255714Skris unused in the certificate verification process. 341355714Skris [Ralf S. Engelschall] 341455714Skris 341555714Skris *) Fix the various library and apps files to free up pkeys obtained from 341655714Skris X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 341755714Skris [Steve Henson] 341855714Skris 341955714Skris *) Fix reference counting in X509_PUBKEY_get(). This makes 342055714Skris demos/maurice/example2.c work, amongst others, probably. 342155714Skris [Steve Henson and Ben Laurie] 342255714Skris 342355714Skris *) First cut of a cleanup for apps/. First the `ssleay' program is now named 342455714Skris `openssl' and second, the shortcut symlinks for the `openssl <command>' 342555714Skris are no longer created. This way we have a single and consistent command 342655714Skris line interface `openssl <command>', similar to `cvs <command>'. 342755714Skris [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 342855714Skris 342955714Skris *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 343055714Skris BIT STRING wrapper always have zero unused bits. 343155714Skris [Steve Henson] 343255714Skris 343355714Skris *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 343455714Skris [Steve Henson] 343555714Skris 343655714Skris *) Make the top-level INSTALL documentation easier to understand. 343755714Skris [Paul Sutton] 343855714Skris 343955714Skris *) Makefiles updated to exit if an error occurs in a sub-directory 344055714Skris make (including if user presses ^C) [Paul Sutton] 344155714Skris 344255714Skris *) Make Montgomery context stuff explicit in RSA data structure. 344355714Skris [Ben Laurie] 344455714Skris 344555714Skris *) Fix build order of pem and err to allow for generated pem.h. 344655714Skris [Ben Laurie] 344755714Skris 344855714Skris *) Fix renumbering bug in X509_NAME_delete_entry(). 344955714Skris [Ben Laurie] 345055714Skris 345155714Skris *) Enhanced the err-ins.pl script so it makes the error library number 345255714Skris global and can add a library name. This is needed for external ASN1 and 345355714Skris other error libraries. 345455714Skris [Steve Henson] 345555714Skris 345655714Skris *) Fixed sk_insert which never worked properly. 345755714Skris [Steve Henson] 345855714Skris 345955714Skris *) Fix ASN1 macros so they can handle indefinite length construted 346055714Skris EXPLICIT tags. Some non standard certificates use these: they can now 346155714Skris be read in. 346255714Skris [Steve Henson] 346355714Skris 346455714Skris *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 346555714Skris into a single doc/ssleay.txt bundle. This way the information is still 346655714Skris preserved but no longer messes up this directory. Now it's new room for 346755714Skris the new set of documenation files. 346855714Skris [Ralf S. Engelschall] 346955714Skris 347055714Skris *) SETs were incorrectly DER encoded. This was a major pain, because they 347155714Skris shared code with SEQUENCEs, which aren't coded the same. This means that 347255714Skris almost everything to do with SETs or SEQUENCEs has either changed name or 347355714Skris number of arguments. 347455714Skris [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 347555714Skris 347655714Skris *) Fix test data to work with the above. 347755714Skris [Ben Laurie] 347855714Skris 347955714Skris *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 348055714Skris was already fixed by Eric for 0.9.1 it seems. 348155714Skris [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 348255714Skris 348355714Skris *) Autodetect FreeBSD3. 348455714Skris [Ben Laurie] 348555714Skris 348655714Skris *) Fix various bugs in Configure. This affects the following platforms: 348755714Skris nextstep 348855714Skris ncr-scde 348955714Skris unixware-2.0 349055714Skris unixware-2.0-pentium 349155714Skris sco5-cc. 349255714Skris [Ben Laurie] 349355714Skris 349455714Skris *) Eliminate generated files from CVS. Reorder tests to regenerate files 349555714Skris before they are needed. 349655714Skris [Ben Laurie] 349755714Skris 349855714Skris *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 349955714Skris [Ben Laurie] 350055714Skris 350155714Skris 350255714Skris Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 350355714Skris 350455714Skris *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 350555714Skris changed SSLeay to OpenSSL in version strings. 350655714Skris [Ralf S. Engelschall] 350755714Skris 350855714Skris *) Some fixups to the top-level documents. 350955714Skris [Paul Sutton] 351055714Skris 351155714Skris *) Fixed the nasty bug where rsaref.h was not found under compile-time 351255714Skris because the symlink to include/ was missing. 351355714Skris [Ralf S. Engelschall] 351455714Skris 351555714Skris *) Incorporated the popular no-RSA/DSA-only patches 351655714Skris which allow to compile a RSA-free SSLeay. 351755714Skris [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 351855714Skris 351955714Skris *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 352055714Skris when "ssleay" is still not found. 352155714Skris [Ralf S. Engelschall] 352255714Skris 352355714Skris *) Added more platforms to Configure: Cray T3E, HPUX 11, 352455714Skris [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 352555714Skris 352655714Skris *) Updated the README file. 352755714Skris [Ralf S. Engelschall] 352855714Skris 352955714Skris *) Added various .cvsignore files in the CVS repository subdirs 353055714Skris to make a "cvs update" really silent. 353155714Skris [Ralf S. Engelschall] 353255714Skris 353355714Skris *) Recompiled the error-definition header files and added 353455714Skris missing symbols to the Win32 linker tables. 353555714Skris [Ralf S. Engelschall] 353655714Skris 353755714Skris *) Cleaned up the top-level documents; 353855714Skris o new files: CHANGES and LICENSE 353955714Skris o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 354055714Skris o merged COPYRIGHT into LICENSE 354155714Skris o removed obsolete TODO file 354255714Skris o renamed MICROSOFT to INSTALL.W32 354355714Skris [Ralf S. Engelschall] 354455714Skris 354555714Skris *) Removed dummy files from the 0.9.1b source tree: 354655714Skris crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 354755714Skris crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 354855714Skris crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 354955714Skris crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 355055714Skris util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 355155714Skris [Ralf S. Engelschall] 355255714Skris 355355714Skris *) Added various platform portability fixes. 355455714Skris [Mark J. Cox] 355555714Skris 355655714Skris *) The Genesis of the OpenSSL rpject: 355755714Skris We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 355855714Skris Young and Tim J. Hudson created while they were working for C2Net until 355955714Skris summer 1998. 356055714Skris [The OpenSSL Project] 356155714Skris 356255714Skris 356355714Skris Changes between 0.9.0b and 0.9.1b [not released] 356455714Skris 356555714Skris *) Updated a few CA certificates under certs/ 356655714Skris [Eric A. Young] 356755714Skris 356855714Skris *) Changed some BIGNUM api stuff. 356955714Skris [Eric A. Young] 357055714Skris 357155714Skris *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 357255714Skris DGUX x86, Linux Alpha, etc. 357355714Skris [Eric A. Young] 357455714Skris 357555714Skris *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 357655714Skris RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 357755714Skris available). 357855714Skris [Eric A. Young] 357955714Skris 358055714Skris *) Add -strparse option to asn1pars program which parses nested 358155714Skris binary structures 358255714Skris [Dr Stephen Henson <shenson@bigfoot.com>] 358355714Skris 358455714Skris *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 358555714Skris [Eric A. Young] 358655714Skris 358755714Skris *) DSA fix for "ca" program. 358855714Skris [Eric A. Young] 358955714Skris 359055714Skris *) Added "-genkey" option to "dsaparam" program. 359155714Skris [Eric A. Young] 359255714Skris 359355714Skris *) Added RIPE MD160 (rmd160) message digest. 359455714Skris [Eric A. Young] 359555714Skris 359655714Skris *) Added -a (all) option to "ssleay version" command. 359755714Skris [Eric A. Young] 359855714Skris 359955714Skris *) Added PLATFORM define which is the id given to Configure. 360055714Skris [Eric A. Young] 360155714Skris 360255714Skris *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 360355714Skris [Eric A. Young] 360455714Skris 360555714Skris *) Extended the ASN.1 parser routines. 360655714Skris [Eric A. Young] 360755714Skris 360855714Skris *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 360955714Skris [Eric A. Young] 361055714Skris 361155714Skris *) Added a BN_CTX to the BN library. 361255714Skris [Eric A. Young] 361355714Skris 361455714Skris *) Fixed the weak key values in DES library 361555714Skris [Eric A. Young] 361655714Skris 361755714Skris *) Changed API in EVP library for cipher aliases. 361855714Skris [Eric A. Young] 361955714Skris 362055714Skris *) Added support for RC2/64bit cipher. 362155714Skris [Eric A. Young] 362255714Skris 362355714Skris *) Converted the lhash library to the crypto/mem.c functions. 362455714Skris [Eric A. Young] 362555714Skris 362655714Skris *) Added more recognized ASN.1 object ids. 362755714Skris [Eric A. Young] 362855714Skris 362955714Skris *) Added more RSA padding checks for SSL/TLS. 363055714Skris [Eric A. Young] 363155714Skris 363255714Skris *) Added BIO proxy/filter functionality. 363355714Skris [Eric A. Young] 363455714Skris 363555714Skris *) Added extra_certs to SSL_CTX which can be used 363655714Skris send extra CA certificates to the client in the CA cert chain sending 363755714Skris process. It can be configured with SSL_CTX_add_extra_chain_cert(). 363855714Skris [Eric A. Young] 363955714Skris 364055714Skris *) Now Fortezza is denied in the authentication phase because 364155714Skris this is key exchange mechanism is not supported by SSLeay at all. 364255714Skris [Eric A. Young] 364355714Skris 364455714Skris *) Additional PKCS1 checks. 364555714Skris [Eric A. Young] 364655714Skris 364755714Skris *) Support the string "TLSv1" for all TLS v1 ciphers. 364855714Skris [Eric A. Young] 364955714Skris 365055714Skris *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 365155714Skris ex_data index of the SSL context in the X509_STORE_CTX ex_data. 365255714Skris [Eric A. Young] 365355714Skris 365455714Skris *) Fixed a few memory leaks. 365555714Skris [Eric A. Young] 365655714Skris 365755714Skris *) Fixed various code and comment typos. 365855714Skris [Eric A. Young] 365955714Skris 366055714Skris *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 366155714Skris bytes sent in the client random. 366255714Skris [Edward Bishop <ebishop@spyglass.com>] 366355714Skris 3664