CHANGES revision 76866
155714Skris 255714Skris OpenSSL CHANGES 355714Skris _______________ 455714Skris 576866Skris Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 672613Skris 776866Skris *) Fix a couple of memory leaks in PKCS7_dataDecode() 876866Skris [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 976866Skris 1076866Skris *) Change Configure and Makefiles to provide EXE_EXT, which will contain 1176866Skris the default extension for executables, if any. Also, make the perl 1276866Skris scripts that use symlink() to test if it really exists and use "cp" 1376866Skris if it doesn't. All this made OpenSSL compilable and installable in 1476866Skris CygWin. 1576866Skris [Richard Levitte] 1676866Skris 1776866Skris *) Fix for asn1_GetSequence() for indefinite length constructed data. 1876866Skris If SEQUENCE is length is indefinite just set c->slen to the total 1976866Skris amount of data available. 2076866Skris [Steve Henson, reported by shige@FreeBSD.org] 2176866Skris [This change does not apply to 0.9.7.] 2276866Skris 2376866Skris *) Change bctest to avoid here-documents inside command substitution 2476866Skris (workaround for FreeBSD /bin/sh bug). 2576866Skris For compatibility with Ultrix, avoid shell functions (introduced 2676866Skris in the bctest version that searches along $PATH). 2776866Skris [Bodo Moeller] 2876866Skris 2976866Skris *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 3076866Skris with des_encrypt() defined on some operating systems, like Solaris 3176866Skris and UnixWare. 3276866Skris [Richard Levitte] 3376866Skris 3476866Skris *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 3576866Skris On the Importance of Eliminating Errors in Cryptographic 3676866Skris Computations, J. Cryptology 14 (2001) 2, 101-119, 3776866Skris http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 3876866Skris [Ulf Moeller] 3976866Skris 4076866Skris *) MIPS assembler BIGNUM division bug fix. 4176866Skris [Andy Polyakov] 4276866Skris 4376866Skris *) Disabled incorrect Alpha assembler code. 4476866Skris [Richard Levitte] 4576866Skris 4676866Skris *) Fix PKCS#7 decode routines so they correctly update the length 4776866Skris after reading an EOC for the EXPLICIT tag. 4876866Skris [Steve Henson] 4976866Skris [This change does not apply to 0.9.7.] 5076866Skris 5176866Skris *) Fix bug in PKCS#12 key generation routines. This was triggered 5276866Skris if a 3DES key was generated with a 0 initial byte. Include 5376866Skris PKCS12_BROKEN_KEYGEN compilation option to retain the old 5476866Skris (but broken) behaviour. 5576866Skris [Steve Henson] 5676866Skris 5776866Skris *) Enhance bctest to search for a working bc along $PATH and print 5876866Skris it when found. 5976866Skris [Tim Rice <tim@multitalents.net> via Richard Levitte] 6076866Skris 6176866Skris *) Fix memory leaks in err.c: free err_data string if necessary; 6276866Skris don't write to the wrong index in ERR_set_error_data. 6376866Skris [Bodo Moeller] 6476866Skris 6576866Skris *) Implement ssl23_peek (analogous to ssl23_read), which previously 6676866Skris did not exist. 6776866Skris [Bodo Moeller] 6876866Skris 6976866Skris *) Replace rdtsc with _emit statements for VC++ version 5. 7076866Skris [Jeremy Cooper <jeremy@baymoo.org>] 7176866Skris 7276866Skris *) Make it possible to reuse SSLv2 sessions. 7376866Skris [Richard Levitte] 7476866Skris 7576866Skris *) In copy_email() check for >= 0 as a return value for 7676866Skris X509_NAME_get_index_by_NID() since 0 is a valid index. 7776866Skris [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 7876866Skris 7976866Skris *) Avoid coredump with unsupported or invalid public keys by checking if 8076866Skris X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 8176866Skris PKCS7_verify() fails with non detached data. 8276866Skris [Steve Henson] 8376866Skris 8476866Skris *) Don't use getenv in library functions when run as setuid/setgid. 8576866Skris New function OPENSSL_issetugid(). 8676866Skris [Ulf Moeller] 8776866Skris 8876866Skris *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 8976866Skris due to incorrect handling of multi-threading: 9076866Skris 9176866Skris 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 9276866Skris 9376866Skris 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 9476866Skris 9576866Skris 3. Count how many times MemCheck_off() has been called so that 9676866Skris nested use can be treated correctly. This also avoids 9776866Skris inband-signalling in the previous code (which relied on the 9876866Skris assumption that thread ID 0 is impossible). 9976866Skris [Bodo Moeller] 10076866Skris 10176866Skris *) Add "-rand" option also to s_client and s_server. 10276866Skris [Lutz Jaenicke] 10376866Skris 10476866Skris *) Fix CPU detection on Irix 6.x. 10576866Skris [Kurt Hockenbury <khockenb@stevens-tech.edu> and 10676866Skris "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 10776866Skris 10876866Skris *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 10976866Skris was empty. 11076866Skris [Steve Henson] 11176866Skris [This change does not apply to 0.9.7.] 11276866Skris 11376866Skris *) Use the cached encoding of an X509_NAME structure rather than 11476866Skris copying it. This is apparently the reason for the libsafe "errors" 11576866Skris but the code is actually correct. 11676866Skris [Steve Henson] 11776866Skris 11872613Skris *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 11972613Skris Bleichenbacher's DSA attack. 12076866Skris Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 12176866Skris to be set and top=0 forces the highest bit to be set; top=-1 is new 12276866Skris and leaves the highest bit random. 12376866Skris [Ulf Moeller, Bodo Moeller] 12472613Skris 12572613Skris *) In the NCONF_...-based implementations for CONF_... queries 12672613Skris (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 12772613Skris a temporary CONF structure with the data component set to NULL 12872613Skris (which gives segmentation faults in lh_retrieve). 12972613Skris Instead, use NULL for the CONF pointer in CONF_get_string and 13072613Skris CONF_get_number (which may use environment variables) and directly 13172613Skris return NULL from CONF_get_section. 13272613Skris [Bodo Moeller] 13372613Skris 13472613Skris *) Fix potential buffer overrun for EBCDIC. 13572613Skris [Ulf Moeller] 13672613Skris 13772613Skris *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 13872613Skris keyUsage if basicConstraints absent for a CA. 13972613Skris [Steve Henson] 14072613Skris 14172613Skris *) Make SMIME_write_PKCS7() write mail header values with a format that 14272613Skris is more generally accepted (no spaces before the semicolon), since 14372613Skris some programs can't parse those values properly otherwise. Also make 14472613Skris sure BIO's that break lines after each write do not create invalid 14572613Skris headers. 14672613Skris [Richard Levitte] 14772613Skris 14872613Skris *) Make the CRL encoding routines work with empty SEQUENCE OF. The 14972613Skris macros previously used would not encode an empty SEQUENCE OF 15072613Skris and break the signature. 15172613Skris [Steve Henson] 15276866Skris [This change does not apply to 0.9.7.] 15372613Skris 15472613Skris *) Zero the premaster secret after deriving the master secret in 15572613Skris DH ciphersuites. 15672613Skris [Steve Henson] 15772613Skris 15872613Skris *) Add some EVP_add_digest_alias registrations (as found in 15972613Skris OpenSSL_add_all_digests()) to SSL_library_init() 16072613Skris aka OpenSSL_add_ssl_algorithms(). This provides improved 16172613Skris compatibility with peers using X.509 certificates 16272613Skris with unconventional AlgorithmIdentifier OIDs. 16372613Skris [Bodo Moeller] 16472613Skris 16572613Skris *) Fix for Irix with NO_ASM. 16672613Skris ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 16772613Skris 16872613Skris *) ./config script fixes. 16972613Skris [Ulf Moeller, Richard Levitte] 17072613Skris 17172613Skris *) Fix 'openssl passwd -1'. 17272613Skris [Bodo Moeller] 17372613Skris 17472613Skris *) Change PKCS12_key_gen_asc() so it can cope with non null 17572613Skris terminated strings whose length is passed in the passlen 17672613Skris parameter, for example from PEM callbacks. This was done 17772613Skris by adding an extra length parameter to asc2uni(). 17872613Skris [Steve Henson, reported by <oddissey@samsung.co.kr>] 17972613Skris 18072613Skris *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 18172613Skris call failed, free the DSA structure. 18272613Skris [Bodo Moeller] 18372613Skris 18472613Skris *) Fix to uni2asc() to cope with zero length Unicode strings. 18572613Skris These are present in some PKCS#12 files. 18672613Skris [Steve Henson] 18772613Skris 18872613Skris *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 18972613Skris Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 19072613Skris when writing a 32767 byte record. 19172613Skris [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 19272613Skris 19372613Skris *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 19472613Skris obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 19572613Skris 19672613Skris (RSA objects have a reference count access to which is protected 19772613Skris by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 19872613Skris so they are meant to be shared between threads.) 19972613Skris [Bodo Moeller, Geoff Thorpe; original patch submitted by 20072613Skris "Reddie, Steven" <Steven.Reddie@ca.com>] 20172613Skris 20272613Skris *) Fix a deadlock in CRYPTO_mem_leaks(). 20372613Skris [Bodo Moeller] 20472613Skris 20576866Skris *) Use better test patterns in bntest. 20676866Skris [Ulf M�ller] 20776866Skris 20872613Skris *) rand_win.c fix for Borland C. 20972613Skris [Ulf M�ller] 21072613Skris 21172613Skris *) BN_rshift bugfix for n == 0. 21272613Skris [Bodo Moeller] 21372613Skris 21476866Skris *) Add a 'bctest' script that checks for some known 'bc' bugs 21576866Skris so that 'make test' does not abort just because 'bc' is broken. 21676866Skris [Bodo Moeller] 21776866Skris 21872613Skris *) Store verify_result within SSL_SESSION also for client side to 21972613Skris avoid potential security hole. (Re-used sessions on the client side 22072613Skris always resulted in verify_result==X509_V_OK, not using the original 22172613Skris result of the server certificate verification.) 22272613Skris [Lutz Jaenicke] 22372613Skris 22472613Skris *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 22572613Skris SSL3_RT_APPLICATION_DATA, return 0. 22672613Skris Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 22772613Skris [Bodo Moeller] 22872613Skris 22972613Skris *) Fix SSL_peek: 23072613Skris Both ssl2_peek and ssl3_peek, which were totally broken in earlier 23172613Skris releases, have been re-implemented by renaming the previous 23272613Skris implementations of ssl2_read and ssl3_read to ssl2_read_internal 23372613Skris and ssl3_read_internal, respectively, and adding 'peek' parameters 23472613Skris to them. The new ssl[23]_{read,peek} functions are calls to 23572613Skris ssl[23]_read_internal with the 'peek' flag set appropriately. 23672613Skris A 'peek' parameter has also been added to ssl3_read_bytes, which 23772613Skris does the actual work for ssl3_read_internal. 23872613Skris [Bodo Moeller] 23972613Skris 24076866Skris *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 24176866Skris the method-specific "init()" handler. Also clean up ex_data after 24276866Skris calling the method-specific "finish()" handler. Previously, this was 24376866Skris happening the other way round. 24476866Skris [Geoff Thorpe] 24576866Skris 24672613Skris *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 24772613Skris The previous value, 12, was not always sufficient for BN_mod_exp(). 24872613Skris [Bodo Moeller] 24972613Skris 25076866Skris *) Make sure that shared libraries get the internal name engine with 25176866Skris the full version number and not just 0. This should mark the 25276866Skris shared libraries as not backward compatible. Of course, this should 25376866Skris be changed again when we can guarantee backward binary compatibility. 25476866Skris [Richard Levitte] 25576866Skris 25672613Skris *) Fix typo in get_cert_by_subject() in by_dir.c 25772613Skris [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 25872613Skris 25976866Skris *) Rework the system to generate shared libraries: 26076866Skris 26176866Skris - Make note of the expected extension for the shared libraries and 26276866Skris if there is a need for symbolic links from for example libcrypto.so.0 26376866Skris to libcrypto.so.0.9.7. There is extended info in Configure for 26476866Skris that. 26576866Skris 26676866Skris - Make as few rebuilds of the shared libraries as possible. 26776866Skris 26876866Skris - Still avoid linking the OpenSSL programs with the shared libraries. 26976866Skris 27076866Skris - When installing, install the shared libraries separately from the 27176866Skris static ones. 27276866Skris [Richard Levitte] 27376866Skris 27472613Skris *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 27572613Skris 27672613Skris Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 27772613Skris and not in SSL_clear because the latter is also used by the 27872613Skris accept/connect functions; previously, the settings made by 27972613Skris SSL_set_read_ahead would be lost during the handshake. 28072613Skris [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 28172613Skris 28272613Skris *) Correct util/mkdef.pl to be selective about disabled algorithms. 28372613Skris Previously, it would create entries for disableed algorithms no 28472613Skris matter what. 28572613Skris [Richard Levitte] 28672613Skris 28772613Skris *) Added several new manual pages for SSL_* function. 28872613Skris [Lutz Jaenicke] 28972613Skris 29068651Skris Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 29168651Skris 29268651Skris *) In ssl23_get_client_hello, generate an error message when faced 29368651Skris with an initial SSL 3.0/TLS record that is too small to contain the 29468651Skris first two bytes of the ClientHello message, i.e. client_version. 29568651Skris (Note that this is a pathologic case that probably has never happened 29668651Skris in real life.) The previous approach was to use the version number 29768651Skris from the record header as a substitute; but our protocol choice 29868651Skris should not depend on that one because it is not authenticated 29968651Skris by the Finished messages. 30068651Skris [Bodo Moeller] 30168651Skris 30268651Skris *) More robust randomness gathering functions for Windows. 30368651Skris [Jeffrey Altman <jaltman@columbia.edu>] 30468651Skris 30568651Skris *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 30668651Skris not set then we don't setup the error code for issuer check errors 30768651Skris to avoid possibly overwriting other errors which the callback does 30868651Skris handle. If an application does set the flag then we assume it knows 30968651Skris what it is doing and can handle the new informational codes 31068651Skris appropriately. 31168651Skris [Steve Henson] 31268651Skris 31368651Skris *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 31468651Skris a general "ANY" type, as such it should be able to decode anything 31568651Skris including tagged types. However it didn't check the class so it would 31668651Skris wrongly interpret tagged types in the same way as their universal 31768651Skris counterpart and unknown types were just rejected. Changed so that the 31868651Skris tagged and unknown types are handled in the same way as a SEQUENCE: 31968651Skris that is the encoding is stored intact. There is also a new type 32068651Skris "V_ASN1_OTHER" which is used when the class is not universal, in this 32168651Skris case we have no idea what the actual type is so we just lump them all 32268651Skris together. 32368651Skris [Steve Henson] 32468651Skris 32568651Skris *) On VMS, stdout may very well lead to a file that is written to 32668651Skris in a record-oriented fashion. That means that every write() will 32768651Skris write a separate record, which will be read separately by the 32868651Skris programs trying to read from it. This can be very confusing. 32968651Skris 33068651Skris The solution is to put a BIO filter in the way that will buffer 33168651Skris text until a linefeed is reached, and then write everything a 33268651Skris line at a time, so every record written will be an actual line, 33368651Skris not chunks of lines and not (usually doesn't happen, but I've 33468651Skris seen it once) several lines in one record. BIO_f_linebuffer() is 33568651Skris the answer. 33668651Skris 33768651Skris Currently, it's a VMS-only method, because that's where it has 33868651Skris been tested well enough. 33968651Skris [Richard Levitte] 34068651Skris 34168651Skris *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 34268651Skris it can return incorrect results. 34368651Skris (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 34468651Skris but it was in 0.9.6-beta[12].) 34568651Skris [Bodo Moeller] 34668651Skris 34768651Skris *) Disable the check for content being present when verifying detached 34868651Skris signatures in pk7_smime.c. Some versions of Netscape (wrongly) 34968651Skris include zero length content when signing messages. 35068651Skris [Steve Henson] 35168651Skris 35268651Skris *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 35368651Skris BIO_ctrl (for BIO pairs). 35468651Skris [Bodo M�ller] 35568651Skris 35668651Skris *) Add DSO method for VMS. 35768651Skris [Richard Levitte] 35868651Skris 35968651Skris *) Bug fix: Montgomery multiplication could produce results with the 36068651Skris wrong sign. 36168651Skris [Ulf M�ller] 36268651Skris 36368651Skris *) Add RPM specification openssl.spec and modify it to build three 36468651Skris packages. The default package contains applications, application 36568651Skris documentation and run-time libraries. The devel package contains 36668651Skris include files, static libraries and function documentation. The 36768651Skris doc package contains the contents of the doc directory. The original 36868651Skris openssl.spec was provided by Damien Miller <djm@mindrot.org>. 36968651Skris [Richard Levitte] 37068651Skris 37168651Skris *) Add a large number of documentation files for many SSL routines. 37268651Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 37368651Skris 37468651Skris *) Add a configuration entry for Sony News 4. 37568651Skris [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 37668651Skris 37768651Skris *) Don't set the two most significant bits to one when generating a 37868651Skris random number < q in the DSA library. 37968651Skris [Ulf M�ller] 38068651Skris 38168651Skris *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 38268651Skris behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 38368651Skris the underlying transport is blocking) if a handshake took place. 38468651Skris (The default behaviour is needed by applications such as s_client 38568651Skris and s_server that use select() to determine when to use SSL_read; 38668651Skris but for applications that know in advance when to expect data, it 38768651Skris just makes things more complicated.) 38868651Skris [Bodo Moeller] 38968651Skris 39068651Skris *) Add RAND_egd_bytes(), which gives control over the number of bytes read 39168651Skris from EGD. 39268651Skris [Ben Laurie] 39368651Skris 39468651Skris *) Add a few more EBCDIC conditionals that make `req' and `x509' 39568651Skris work better on such systems. 39668651Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 39768651Skris 39868651Skris *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 39968651Skris Update PKCS12_parse() so it copies the friendlyName and the 40068651Skris keyid to the certificates aux info. 40168651Skris [Steve Henson] 40268651Skris 40368651Skris *) Fix bug in PKCS7_verify() which caused an infinite loop 40468651Skris if there was more than one signature. 40568651Skris [Sven Uszpelkat <su@celocom.de>] 40668651Skris 40768651Skris *) Major change in util/mkdef.pl to include extra information 40868651Skris about each symbol, as well as presentig variables as well 40968651Skris as functions. This change means that there's n more need 41068651Skris to rebuild the .num files when some algorithms are excluded. 41168651Skris [Richard Levitte] 41268651Skris 41368651Skris *) Allow the verify time to be set by an application, 41468651Skris rather than always using the current time. 41568651Skris [Steve Henson] 41668651Skris 41768651Skris *) Phase 2 verify code reorganisation. The certificate 41868651Skris verify code now looks up an issuer certificate by a 41968651Skris number of criteria: subject name, authority key id 42068651Skris and key usage. It also verifies self signed certificates 42168651Skris by the same criteria. The main comparison function is 42268651Skris X509_check_issued() which performs these checks. 42368651Skris 42468651Skris Lot of changes were necessary in order to support this 42568651Skris without completely rewriting the lookup code. 42668651Skris 42768651Skris Authority and subject key identifier are now cached. 42868651Skris 42968651Skris The LHASH 'certs' is X509_STORE has now been replaced 43068651Skris by a STACK_OF(X509_OBJECT). This is mainly because an 43168651Skris LHASH can't store or retrieve multiple objects with 43268651Skris the same hash value. 43368651Skris 43468651Skris As a result various functions (which were all internal 43568651Skris use only) have changed to handle the new X509_STORE 43668651Skris structure. This will break anything that messed round 43768651Skris with X509_STORE internally. 43868651Skris 43968651Skris The functions X509_STORE_add_cert() now checks for an 44068651Skris exact match, rather than just subject name. 44168651Skris 44268651Skris The X509_STORE API doesn't directly support the retrieval 44368651Skris of multiple certificates matching a given criteria, however 44468651Skris this can be worked round by performing a lookup first 44568651Skris (which will fill the cache with candidate certificates) 44668651Skris and then examining the cache for matches. This is probably 44768651Skris the best we can do without throwing out X509_LOOKUP 44868651Skris entirely (maybe later...). 44968651Skris 45068651Skris The X509_VERIFY_CTX structure has been enhanced considerably. 45168651Skris 45268651Skris All certificate lookup operations now go via a get_issuer() 45368651Skris callback. Although this currently uses an X509_STORE it 45468651Skris can be replaced by custom lookups. This is a simple way 45568651Skris to bypass the X509_STORE hackery necessary to make this 45668651Skris work and makes it possible to use more efficient techniques 45768651Skris in future. A very simple version which uses a simple 45868651Skris STACK for its trusted certificate store is also provided 45968651Skris using X509_STORE_CTX_trusted_stack(). 46068651Skris 46168651Skris The verify_cb() and verify() callbacks now have equivalents 46268651Skris in the X509_STORE_CTX structure. 46368651Skris 46468651Skris X509_STORE_CTX also has a 'flags' field which can be used 46568651Skris to customise the verify behaviour. 46668651Skris [Steve Henson] 46768651Skris 46868651Skris *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 46968651Skris excludes S/MIME capabilities. 47068651Skris [Steve Henson] 47168651Skris 47268651Skris *) When a certificate request is read in keep a copy of the 47368651Skris original encoding of the signed data and use it when outputing 47468651Skris again. Signatures then use the original encoding rather than 47568651Skris a decoded, encoded version which may cause problems if the 47668651Skris request is improperly encoded. 47768651Skris [Steve Henson] 47868651Skris 47968651Skris *) For consistency with other BIO_puts implementations, call 48068651Skris buffer_write(b, ...) directly in buffer_puts instead of calling 48168651Skris BIO_write(b, ...). 48268651Skris 48368651Skris In BIO_puts, increment b->num_write as in BIO_write. 48468651Skris [Peter.Sylvester@EdelWeb.fr] 48568651Skris 48668651Skris *) Fix BN_mul_word for the case where the word is 0. (We have to use 48768651Skris BN_zero, we may not return a BIGNUM with an array consisting of 48868651Skris words set to zero.) 48968651Skris [Bodo Moeller] 49068651Skris 49168651Skris *) Avoid calling abort() from within the library when problems are 49268651Skris detected, except if preprocessor symbols have been defined 49368651Skris (such as REF_CHECK, BN_DEBUG etc.). 49468651Skris [Bodo Moeller] 49568651Skris 49668651Skris *) New openssl application 'rsautl'. This utility can be 49768651Skris used for low level RSA operations. DER public key 49868651Skris BIO/fp routines also added. 49968651Skris [Steve Henson] 50068651Skris 50168651Skris *) New Configure entry and patches for compiling on QNX 4. 50268651Skris [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 50368651Skris 50468651Skris *) A demo state-machine implementation was sponsored by 50568651Skris Nuron (http://www.nuron.com/) and is now available in 50668651Skris demos/state_machine. 50768651Skris [Ben Laurie] 50868651Skris 50968651Skris *) New options added to the 'dgst' utility for signature 51068651Skris generation and verification. 51168651Skris [Steve Henson] 51268651Skris 51368651Skris *) Unrecognized PKCS#7 content types are now handled via a 51468651Skris catch all ASN1_TYPE structure. This allows unsupported 51568651Skris types to be stored as a "blob" and an application can 51668651Skris encode and decode it manually. 51768651Skris [Steve Henson] 51868651Skris 51968651Skris *) Fix various signed/unsigned issues to make a_strex.c 52068651Skris compile under VC++. 52168651Skris [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 52268651Skris 52368651Skris *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 52468651Skris length if passed a buffer. ASN1_INTEGER_to_BN failed 52568651Skris if passed a NULL BN and its argument was negative. 52668651Skris [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 52768651Skris 52868651Skris *) Modification to PKCS#7 encoding routines to output definite 52968651Skris length encoding. Since currently the whole structures are in 53068651Skris memory there's not real point in using indefinite length 53168651Skris constructed encoding. However if OpenSSL is compiled with 53268651Skris the flag PKCS7_INDEFINITE_ENCODING the old form is used. 53368651Skris [Steve Henson] 53468651Skris 53568651Skris *) Added BIO_vprintf() and BIO_vsnprintf(). 53668651Skris [Richard Levitte] 53768651Skris 53868651Skris *) Added more prefixes to parse for in the the strings written 53968651Skris through a logging bio, to cover all the levels that are available 54068651Skris through syslog. The prefixes are now: 54168651Skris 54268651Skris PANIC, EMERG, EMR => LOG_EMERG 54368651Skris ALERT, ALR => LOG_ALERT 54468651Skris CRIT, CRI => LOG_CRIT 54568651Skris ERROR, ERR => LOG_ERR 54668651Skris WARNING, WARN, WAR => LOG_WARNING 54768651Skris NOTICE, NOTE, NOT => LOG_NOTICE 54868651Skris INFO, INF => LOG_INFO 54968651Skris DEBUG, DBG => LOG_DEBUG 55068651Skris 55168651Skris and as before, if none of those prefixes are present at the 55268651Skris beginning of the string, LOG_ERR is chosen. 55368651Skris 55468651Skris On Win32, the LOG_* levels are mapped according to this: 55568651Skris 55668651Skris LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 55768651Skris LOG_WARNING => EVENTLOG_WARNING_TYPE 55868651Skris LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 55968651Skris 56068651Skris [Richard Levitte] 56168651Skris 56268651Skris *) Made it possible to reconfigure with just the configuration 56368651Skris argument "reconf" or "reconfigure". The command line arguments 56468651Skris are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 56568651Skris and are retrieved from there when reconfiguring. 56668651Skris [Richard Levitte] 56768651Skris 56868651Skris *) MD4 implemented. 56968651Skris [Assar Westerlund <assar@sics.se>, Richard Levitte] 57068651Skris 57168651Skris *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 57268651Skris [Richard Levitte] 57368651Skris 57468651Skris *) The obj_dat.pl script was messing up the sorting of object 57568651Skris names. The reason was that it compared the quoted version 57668651Skris of strings as a result "OCSP" > "OCSP Signing" because 57768651Skris " > SPACE. Changed script to store unquoted versions of 57868651Skris names and add quotes on output. It was also omitting some 57968651Skris names from the lookup table if they were given a default 58068651Skris value (that is if SN is missing it is given the same 58168651Skris value as LN and vice versa), these are now added on the 58268651Skris grounds that if an object has a name we should be able to 58368651Skris look it up. Finally added warning output when duplicate 58468651Skris short or long names are found. 58568651Skris [Steve Henson] 58668651Skris 58768651Skris *) Changes needed for Tandem NSK. 58868651Skris [Scott Uroff <scott@xypro.com>] 58968651Skris 59068651Skris *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 59168651Skris RSA_padding_check_SSLv23(), special padding was never detected 59268651Skris and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 59368651Skris version rollback attacks was not effective. 59468651Skris 59568651Skris In s23_clnt.c, don't use special rollback-attack detection padding 59668651Skris (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 59768651Skris client; similarly, in s23_srvr.c, don't do the rollback check if 59868651Skris SSL 2.0 is the only protocol enabled in the server. 59968651Skris [Bodo Moeller] 60068651Skris 60168651Skris *) Make it possible to get hexdumps of unprintable data with 'openssl 60268651Skris asn1parse'. By implication, the functions ASN1_parse_dump() and 60368651Skris BIO_dump_indent() are added. 60468651Skris [Richard Levitte] 60568651Skris 60668651Skris *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 60768651Skris these print out strings and name structures based on various 60868651Skris flags including RFC2253 support and proper handling of 60968651Skris multibyte characters. Added options to the 'x509' utility 61068651Skris to allow the various flags to be set. 61168651Skris [Steve Henson] 61268651Skris 61368651Skris *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 61468651Skris Also change the functions X509_cmp_current_time() and 61568651Skris X509_gmtime_adj() work with an ASN1_TIME structure, 61668651Skris this will enable certificates using GeneralizedTime in validity 61768651Skris dates to be checked. 61868651Skris [Steve Henson] 61968651Skris 62068651Skris *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 62168651Skris negative public key encodings) on by default, 62268651Skris NO_NEG_PUBKEY_BUG can be set to disable it. 62368651Skris [Steve Henson] 62468651Skris 62568651Skris *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 62668651Skris content octets. An i2c_ASN1_OBJECT is unnecessary because 62768651Skris the encoding can be trivially obtained from the structure. 62868651Skris [Steve Henson] 62968651Skris 63068651Skris *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 63168651Skris not read locks (CRYPTO_r_[un]lock). 63268651Skris [Bodo Moeller] 63368651Skris 63468651Skris *) A first attempt at creating official support for shared 63568651Skris libraries through configuration. I've kept it so the 63668651Skris default is static libraries only, and the OpenSSL programs 63768651Skris are always statically linked for now, but there are 63868651Skris preparations for dynamic linking in place. 63968651Skris This has been tested on Linux and True64. 64068651Skris [Richard Levitte] 64168651Skris 64268651Skris *) Randomness polling function for Win9x, as described in: 64368651Skris Peter Gutmann, Software Generation of Practically Strong 64468651Skris Random Numbers. 64568651Skris [Ulf M�ller] 64668651Skris 64768651Skris *) Fix so PRNG is seeded in req if using an already existing 64868651Skris DSA key. 64968651Skris [Steve Henson] 65068651Skris 65168651Skris *) New options to smime application. -inform and -outform 65268651Skris allow alternative formats for the S/MIME message including 65368651Skris PEM and DER. The -content option allows the content to be 65468651Skris specified separately. This should allow things like Netscape 65568651Skris form signing output easier to verify. 65668651Skris [Steve Henson] 65768651Skris 65868651Skris *) Fix the ASN1 encoding of tags using the 'long form'. 65968651Skris [Steve Henson] 66068651Skris 66168651Skris *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 66268651Skris STRING types. These convert content octets to and from the 66368651Skris underlying type. The actual tag and length octets are 66468651Skris already assumed to have been read in and checked. These 66568651Skris are needed because all other string types have virtually 66668651Skris identical handling apart from the tag. By having versions 66768651Skris of the ASN1 functions that just operate on content octets 66868651Skris IMPLICIT tagging can be handled properly. It also allows 66968651Skris the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 67068651Skris and ASN1_INTEGER are identical apart from the tag. 67168651Skris [Steve Henson] 67268651Skris 67368651Skris *) Change the handling of OID objects as follows: 67468651Skris 67568651Skris - New object identifiers are inserted in objects.txt, following 67668651Skris the syntax given in objects.README. 67768651Skris - objects.pl is used to process obj_mac.num and create a new 67868651Skris obj_mac.h. 67968651Skris - obj_dat.pl is used to create a new obj_dat.h, using the data in 68068651Skris obj_mac.h. 68168651Skris 68268651Skris This is currently kind of a hack, and the perl code in objects.pl 68368651Skris isn't very elegant, but it works as I intended. The simplest way 68468651Skris to check that it worked correctly is to look in obj_dat.h and 68568651Skris check the array nid_objs and make sure the objects haven't moved 68668651Skris around (this is important!). Additions are OK, as well as 68768651Skris consistent name changes. 68868651Skris [Richard Levitte] 68968651Skris 69068651Skris *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 69168651Skris [Bodo Moeller] 69268651Skris 69368651Skris *) Addition of the command line parameter '-rand file' to 'openssl req'. 69468651Skris The given file adds to whatever has already been seeded into the 69568651Skris random pool through the RANDFILE configuration file option or 69668651Skris environment variable, or the default random state file. 69768651Skris [Richard Levitte] 69868651Skris 69968651Skris *) mkstack.pl now sorts each macro group into lexical order. 70068651Skris Previously the output order depended on the order the files 70168651Skris appeared in the directory, resulting in needless rewriting 70268651Skris of safestack.h . 70368651Skris [Steve Henson] 70468651Skris 70568651Skris *) Patches to make OpenSSL compile under Win32 again. Mostly 70668651Skris work arounds for the VC++ problem that it treats func() as 70768651Skris func(void). Also stripped out the parts of mkdef.pl that 70868651Skris added extra typesafe functions: these no longer exist. 70968651Skris [Steve Henson] 71068651Skris 71168651Skris *) Reorganisation of the stack code. The macros are now all 71268651Skris collected in safestack.h . Each macro is defined in terms of 71368651Skris a "stack macro" of the form SKM_<name>(type, a, b). The 71468651Skris DEBUG_SAFESTACK is now handled in terms of function casts, 71568651Skris this has the advantage of retaining type safety without the 71668651Skris use of additional functions. If DEBUG_SAFESTACK is not defined 71768651Skris then the non typesafe macros are used instead. Also modified the 71868651Skris mkstack.pl script to handle the new form. Needs testing to see 71968651Skris if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 72068651Skris the default if no major problems. Similar behaviour for ASN1_SET_OF 72168651Skris and PKCS12_STACK_OF. 72268651Skris [Steve Henson] 72368651Skris 72468651Skris *) When some versions of IIS use the 'NET' form of private key the 72568651Skris key derivation algorithm is different. Normally MD5(password) is 72668651Skris used as a 128 bit RC4 key. In the modified case 72768651Skris MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 72868651Skris new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 72968651Skris as the old Netscape_RSA functions except they have an additional 73068651Skris 'sgckey' parameter which uses the modified algorithm. Also added 73168651Skris an -sgckey command line option to the rsa utility. Thanks to 73268651Skris Adrian Peck <bertie@ncipher.com> for posting details of the modified 73368651Skris algorithm to openssl-dev. 73468651Skris [Steve Henson] 73568651Skris 73668651Skris *) The evp_local.h macros were using 'c.##kname' which resulted in 73768651Skris invalid expansion on some systems (SCO 5.0.5 for example). 73868651Skris Corrected to 'c.kname'. 73968651Skris [Phillip Porch <root@theporch.com>] 74068651Skris 74168651Skris *) New X509_get1_email() and X509_REQ_get1_email() functions that return 74268651Skris a STACK of email addresses from a certificate or request, these look 74368651Skris in the subject name and the subject alternative name extensions and 74468651Skris omit any duplicate addresses. 74568651Skris [Steve Henson] 74668651Skris 74768651Skris *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 74868651Skris This makes DSA verification about 2 % faster. 74968651Skris [Bodo Moeller] 75068651Skris 75168651Skris *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 75268651Skris (meaning that now 2^5 values will be precomputed, which is only 4 KB 75368651Skris plus overhead for 1024 bit moduli). 75468651Skris This makes exponentiations about 0.5 % faster for 1024 bit 75568651Skris exponents (as measured by "openssl speed rsa2048"). 75668651Skris [Bodo Moeller] 75768651Skris 75868651Skris *) Rename memory handling macros to avoid conflicts with other 75968651Skris software: 76068651Skris Malloc => OPENSSL_malloc 76168651Skris Malloc_locked => OPENSSL_malloc_locked 76268651Skris Realloc => OPENSSL_realloc 76368651Skris Free => OPENSSL_free 76468651Skris [Richard Levitte] 76568651Skris 76668651Skris *) New function BN_mod_exp_mont_word for small bases (roughly 15% 76768651Skris faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 76868651Skris [Bodo Moeller] 76968651Skris 77068651Skris *) CygWin32 support. 77168651Skris [John Jarvie <jjarvie@newsguy.com>] 77268651Skris 77368651Skris *) The type-safe stack code has been rejigged. It is now only compiled 77468651Skris in when OpenSSL is configured with the DEBUG_SAFESTACK option and 77568651Skris by default all type-specific stack functions are "#define"d back to 77668651Skris standard stack functions. This results in more streamlined output 77768651Skris but retains the type-safety checking possibilities of the original 77868651Skris approach. 77968651Skris [Geoff Thorpe] 78068651Skris 78168651Skris *) The STACK code has been cleaned up, and certain type declarations 78268651Skris that didn't make a lot of sense have been brought in line. This has 78368651Skris also involved a cleanup of sorts in safestack.h to more correctly 78468651Skris map type-safe stack functions onto their plain stack counterparts. 78568651Skris This work has also resulted in a variety of "const"ifications of 78668651Skris lots of the code, especially "_cmp" operations which should normally 78768651Skris be prototyped with "const" parameters anyway. 78868651Skris [Geoff Thorpe] 78968651Skris 79068651Skris *) When generating bytes for the first time in md_rand.c, 'stir the pool' 79168651Skris by seeding with STATE_SIZE dummy bytes (with zero entropy count). 79268651Skris (The PRNG state consists of two parts, the large pool 'state' and 'md', 79368651Skris where all of 'md' is used each time the PRNG is used, but 'state' 79468651Skris is used only indexed by a cyclic counter. As entropy may not be 79568651Skris well distributed from the beginning, 'md' is important as a 79668651Skris chaining variable. However, the output function chains only half 79768651Skris of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 79868651Skris all of 'md', and seeding with STATE_SIZE dummy bytes will result 79968651Skris in all of 'state' being rewritten, with the new values depending 80068651Skris on virtually all of 'md'. This overcomes the 80 bit limitation.) 80168651Skris [Bodo Moeller] 80268651Skris 80368651Skris *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 80468651Skris the handshake is continued after ssl_verify_cert_chain(); 80568651Skris otherwise, if SSL_VERIFY_NONE is set, remaining error codes 80668651Skris can lead to 'unexplainable' connection aborts later. 80768651Skris [Bodo Moeller; problem tracked down by Lutz Jaenicke] 80868651Skris 80968651Skris *) Major EVP API cipher revision. 81068651Skris Add hooks for extra EVP features. This allows various cipher 81168651Skris parameters to be set in the EVP interface. Support added for variable 81268651Skris key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 81368651Skris setting of RC2 and RC5 parameters. 81468651Skris 81568651Skris Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 81668651Skris ciphers. 81768651Skris 81868651Skris Remove lots of duplicated code from the EVP library. For example *every* 81968651Skris cipher init() function handles the 'iv' in the same way according to the 82068651Skris cipher mode. They also all do nothing if the 'key' parameter is NULL and 82168651Skris for CFB and OFB modes they zero ctx->num. 82268651Skris 82368651Skris New functionality allows removal of S/MIME code RC2 hack. 82468651Skris 82568651Skris Most of the routines have the same form and so can be declared in terms 82668651Skris of macros. 82768651Skris 82868651Skris By shifting this to the top level EVP_CipherInit() it can be removed from 82968651Skris all individual ciphers. If the cipher wants to handle IVs or keys 83068651Skris differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 83168651Skris flags. 83268651Skris 83368651Skris Change lots of functions like EVP_EncryptUpdate() to now return a 83468651Skris value: although software versions of the algorithms cannot fail 83568651Skris any installed hardware versions can. 83668651Skris [Steve Henson] 83768651Skris 83868651Skris *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 83968651Skris this option is set, tolerate broken clients that send the negotiated 84068651Skris protocol version number instead of the requested protocol version 84168651Skris number. 84268651Skris [Bodo Moeller] 84368651Skris 84468651Skris *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 84568651Skris i.e. non-zero for export ciphersuites, zero otherwise. 84668651Skris Previous versions had this flag inverted, inconsistent with 84768651Skris rsa_tmp_cb (..._TMP_RSA_CB). 84868651Skris [Bodo Moeller; problem reported by Amit Chopra] 84968651Skris 85068651Skris *) Add missing DSA library text string. Work around for some IIS 85168651Skris key files with invalid SEQUENCE encoding. 85268651Skris [Steve Henson] 85368651Skris 85468651Skris *) Add a document (doc/standards.txt) that list all kinds of standards 85568651Skris and so on that are implemented in OpenSSL. 85668651Skris [Richard Levitte] 85768651Skris 85868651Skris *) Enhance c_rehash script. Old version would mishandle certificates 85968651Skris with the same subject name hash and wouldn't handle CRLs at all. 86068651Skris Added -fingerprint option to crl utility, to support new c_rehash 86168651Skris features. 86268651Skris [Steve Henson] 86368651Skris 86468651Skris *) Eliminate non-ANSI declarations in crypto.h and stack.h. 86568651Skris [Ulf M�ller] 86668651Skris 86768651Skris *) Fix for SSL server purpose checking. Server checking was 86868651Skris rejecting certificates which had extended key usage present 86968651Skris but no ssl client purpose. 87068651Skris [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 87168651Skris 87268651Skris *) Make PKCS#12 code work with no password. The PKCS#12 spec 87368651Skris is a little unclear about how a blank password is handled. 87468651Skris Since the password in encoded as a BMPString with terminating 87568651Skris double NULL a zero length password would end up as just the 87668651Skris double NULL. However no password at all is different and is 87768651Skris handled differently in the PKCS#12 key generation code. NS 87868651Skris treats a blank password as zero length. MSIE treats it as no 87968651Skris password on export: but it will try both on import. We now do 88068651Skris the same: PKCS12_parse() tries zero length and no password if 88168651Skris the password is set to "" or NULL (NULL is now a valid password: 88268651Skris it wasn't before) as does the pkcs12 application. 88368651Skris [Steve Henson] 88468651Skris 88568651Skris *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 88668651Skris perror when PEM_read_bio_X509_REQ fails, the error message must 88768651Skris be obtained from the error queue. 88868651Skris [Bodo Moeller] 88968651Skris 89068651Skris *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 89168651Skris it in ERR_remove_state if appropriate, and change ERR_get_state 89268651Skris accordingly to avoid race conditions (this is necessary because 89368651Skris thread_hash is no longer constant once set). 89468651Skris [Bodo Moeller] 89568651Skris 89668651Skris *) Bugfix for linux-elf makefile.one. 89768651Skris [Ulf M�ller] 89868651Skris 89968651Skris *) RSA_get_default_method() will now cause a default 90068651Skris RSA_METHOD to be chosen if one doesn't exist already. 90168651Skris Previously this was only set during a call to RSA_new() 90268651Skris or RSA_new_method(NULL) meaning it was possible for 90368651Skris RSA_get_default_method() to return NULL. 90468651Skris [Geoff Thorpe] 90568651Skris 90668651Skris *) Added native name translation to the existing DSO code 90768651Skris that will convert (if the flag to do so is set) filenames 90868651Skris that are sufficiently small and have no path information 90968651Skris into a canonical native form. Eg. "blah" converted to 91068651Skris "libblah.so" or "blah.dll" etc. 91168651Skris [Geoff Thorpe] 91268651Skris 91368651Skris *) New function ERR_error_string_n(e, buf, len) which is like 91468651Skris ERR_error_string(e, buf), but writes at most 'len' bytes 91568651Skris including the 0 terminator. For ERR_error_string_n, 'buf' 91668651Skris may not be NULL. 91768651Skris [Damien Miller <djm@mindrot.org>, Bodo Moeller] 91868651Skris 91968651Skris *) CONF library reworked to become more general. A new CONF 92068651Skris configuration file reader "class" is implemented as well as a 92168651Skris new functions (NCONF_*, for "New CONF") to handle it. The now 92268651Skris old CONF_* functions are still there, but are reimplemented to 92368651Skris work in terms of the new functions. Also, a set of functions 92468651Skris to handle the internal storage of the configuration data is 92568651Skris provided to make it easier to write new configuration file 92668651Skris reader "classes" (I can definitely see something reading a 92768651Skris configuration file in XML format, for example), called _CONF_*, 92868651Skris or "the configuration storage API"... 92968651Skris 93068651Skris The new configuration file reading functions are: 93168651Skris 93268651Skris NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 93368651Skris NCONF_get_section, NCONF_get_string, NCONF_get_numbre 93468651Skris 93568651Skris NCONF_default, NCONF_WIN32 93668651Skris 93768651Skris NCONF_dump_fp, NCONF_dump_bio 93868651Skris 93968651Skris NCONF_default and NCONF_WIN32 are method (or "class") choosers, 94068651Skris NCONF_new creates a new CONF object. This works in the same way 94168651Skris as other interfaces in OpenSSL, like the BIO interface. 94268651Skris NCONF_dump_* dump the internal storage of the configuration file, 94368651Skris which is useful for debugging. All other functions take the same 94468651Skris arguments as the old CONF_* functions wth the exception of the 94568651Skris first that must be a `CONF *' instead of a `LHASH *'. 94668651Skris 94768651Skris To make it easer to use the new classes with the old CONF_* functions, 94868651Skris the function CONF_set_default_method is provided. 94968651Skris [Richard Levitte] 95068651Skris 95168651Skris *) Add '-tls1' option to 'openssl ciphers', which was already 95268651Skris mentioned in the documentation but had not been implemented. 95368651Skris (This option is not yet really useful because even the additional 95468651Skris experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 95568651Skris [Bodo Moeller] 95668651Skris 95768651Skris *) Initial DSO code added into libcrypto for letting OpenSSL (and 95868651Skris OpenSSL-based applications) load shared libraries and bind to 95968651Skris them in a portable way. 96068651Skris [Geoff Thorpe, with contributions from Richard Levitte] 96168651Skris 96259191Skris Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 96359191Skris 96459191Skris *) Make sure _lrotl and _lrotr are only used with MSVC. 96559191Skris 96659191Skris *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 96759191Skris (the default implementation of RAND_status). 96859191Skris 96959191Skris *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 97059191Skris to '-clrext' (= clear extensions), as intended and documented. 97159191Skris [Bodo Moeller; inconsistency pointed out by Michael Attili 97259191Skris <attili@amaxo.com>] 97359191Skris 97459191Skris *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 97559191Skris was larger than the MD block size. 97659191Skris [Steve Henson, pointed out by Yost William <YostW@tce.com>] 97759191Skris 97859191Skris *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 97959191Skris fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 98059191Skris using the passed key: if the passed key was a private key the result 98159191Skris of X509_print(), for example, would be to print out all the private key 98259191Skris components. 98359191Skris [Steve Henson] 98459191Skris 98559191Skris *) des_quad_cksum() byte order bug fix. 98659191Skris [Ulf M�ller, using the problem description in krb4-0.9.7, where 98759191Skris the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 98859191Skris 98959191Skris *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 99059191Skris discouraged. 99159191Skris [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 99259191Skris 99359191Skris *) For easily testing in shell scripts whether some command 99459191Skris 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 99559191Skris returns with exit code 0 iff no command of the given name is available. 99659191Skris 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 99759191Skris the output goes to stdout and nothing is printed to stderr. 99859191Skris Additional arguments are always ignored. 99959191Skris 100059191Skris Since for each cipher there is a command of the same name, 100159191Skris the 'no-cipher' compilation switches can be tested this way. 100259191Skris 100359191Skris ('openssl no-XXX' is not able to detect pseudo-commands such 100459191Skris as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 100559191Skris [Bodo Moeller] 100659191Skris 100759191Skris *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 100859191Skris [Bodo Moeller] 100959191Skris 101059191Skris *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 101159191Skris is set; it will be thrown away anyway because each handshake creates 101259191Skris its own key. 101359191Skris ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 101459191Skris to parameters -- in previous versions (since OpenSSL 0.9.3) the 101559191Skris 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 101659191Skris you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 101759191Skris [Bodo Moeller] 101859191Skris 101959191Skris *) New s_client option -ign_eof: EOF at stdin is ignored, and 102059191Skris 'Q' and 'R' lose their special meanings (quit/renegotiate). 102159191Skris This is part of what -quiet does; unlike -quiet, -ign_eof 102259191Skris does not suppress any output. 102359191Skris [Richard Levitte] 102459191Skris 102559191Skris *) Add compatibility options to the purpose and trust code. The 102659191Skris purpose X509_PURPOSE_ANY is "any purpose" which automatically 102759191Skris accepts a certificate or CA, this was the previous behaviour, 102859191Skris with all the associated security issues. 102959191Skris 103059191Skris X509_TRUST_COMPAT is the old trust behaviour: only and 103159191Skris automatically trust self signed roots in certificate store. A 103259191Skris new trust setting X509_TRUST_DEFAULT is used to specify that 103359191Skris a purpose has no associated trust setting and it should instead 103459191Skris use the value in the default purpose. 103559191Skris [Steve Henson] 103659191Skris 103759191Skris *) Fix the PKCS#8 DSA private key code so it decodes keys again 103859191Skris and fix a memory leak. 103959191Skris [Steve Henson] 104059191Skris 104159191Skris *) In util/mkerr.pl (which implements 'make errors'), preserve 104259191Skris reason strings from the previous version of the .c file, as 104359191Skris the default to have only downcase letters (and digits) in 104459191Skris automatically generated reasons codes is not always appropriate. 104559191Skris [Bodo Moeller] 104659191Skris 104759191Skris *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 104859191Skris using strerror. Previously, ERR_reason_error_string() returned 104959191Skris library names as reason strings for SYSerr; but SYSerr is a special 105059191Skris case where small numbers are errno values, not library numbers. 105159191Skris [Bodo Moeller] 105259191Skris 105359191Skris *) Add '-dsaparam' option to 'openssl dhparam' application. This 105459191Skris converts DSA parameters into DH parameters. (When creating parameters, 105559191Skris DSA_generate_parameters is used.) 105659191Skris [Bodo Moeller] 105759191Skris 105859191Skris *) Include 'length' (recommended exponent length) in C code generated 105959191Skris by 'openssl dhparam -C'. 106059191Skris [Bodo Moeller] 106159191Skris 106259191Skris *) The second argument to set_label in perlasm was already being used 106359191Skris so couldn't be used as a "file scope" flag. Moved to third argument 106459191Skris which was free. 106559191Skris [Steve Henson] 106659191Skris 106759191Skris *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 106859191Skris instead of RAND_bytes for encryption IVs and salts. 106959191Skris [Bodo Moeller] 107059191Skris 107159191Skris *) Include RAND_status() into RAND_METHOD instead of implementing 107259191Skris it only for md_rand.c Otherwise replacing the PRNG by calling 107359191Skris RAND_set_rand_method would be impossible. 107459191Skris [Bodo Moeller] 107559191Skris 107659191Skris *) Don't let DSA_generate_key() enter an infinite loop if the random 107759191Skris number generation fails. 107859191Skris [Bodo Moeller] 107959191Skris 108059191Skris *) New 'rand' application for creating pseudo-random output. 108159191Skris [Bodo Moeller] 108259191Skris 108359191Skris *) Added configuration support for Linux/IA64 108459191Skris [Rolf Haberrecker <rolf@suse.de>] 108559191Skris 108659191Skris *) Assembler module support for Mingw32. 108759191Skris [Ulf M�ller] 108859191Skris 108959191Skris *) Shared library support for HPUX (in shlib/). 109059191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 109159191Skris 109259191Skris *) Shared library support for Solaris gcc. 109359191Skris [Lutz Behnke <behnke@trustcenter.de>] 109459191Skris 109559191Skris Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 109659191Skris 109759191Skris *) PKCS7_encrypt() was adding text MIME headers twice because they 109859191Skris were added manually and by SMIME_crlf_copy(). 109959191Skris [Steve Henson] 110059191Skris 110159191Skris *) In bntest.c don't call BN_rand with zero bits argument. 110259191Skris [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 110359191Skris 110459191Skris *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 110559191Skris case was implemented. This caused BN_div_recp() to fail occasionally. 110659191Skris [Ulf M�ller] 110759191Skris 110859191Skris *) Add an optional second argument to the set_label() in the perl 110959191Skris assembly language builder. If this argument exists and is set 111059191Skris to 1 it signals that the assembler should use a symbol whose 111159191Skris scope is the entire file, not just the current function. This 111259191Skris is needed with MASM which uses the format label:: for this scope. 111359191Skris [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 111459191Skris 111559191Skris *) Change the ASN1 types so they are typedefs by default. Before 111659191Skris almost all types were #define'd to ASN1_STRING which was causing 111759191Skris STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 111859191Skris for example. 111959191Skris [Steve Henson] 112059191Skris 112159191Skris *) Change names of new functions to the new get1/get0 naming 112259191Skris convention: After 'get1', the caller owns a reference count 112359191Skris and has to call ..._free; 'get0' returns a pointer to some 112459191Skris data structure without incrementing reference counters. 112559191Skris (Some of the existing 'get' functions increment a reference 112659191Skris counter, some don't.) 112759191Skris Similarly, 'set1' and 'add1' functions increase reference 112859191Skris counters or duplicate objects. 112959191Skris [Steve Henson] 113059191Skris 113159191Skris *) Allow for the possibility of temp RSA key generation failure: 113259191Skris the code used to assume it always worked and crashed on failure. 113359191Skris [Steve Henson] 113459191Skris 113559191Skris *) Fix potential buffer overrun problem in BIO_printf(). 113659191Skris [Ulf M�ller, using public domain code by Patrick Powell; problem 113759191Skris pointed out by David Sacerdote <das33@cornell.edu>] 113859191Skris 113959191Skris *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 114059191Skris RAND_egd() and RAND_status(). In the command line application, 114159191Skris the EGD socket can be specified like a seed file using RANDFILE 114259191Skris or -rand. 114359191Skris [Ulf M�ller] 114459191Skris 114559191Skris *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 114659191Skris Some CAs (e.g. Verisign) distribute certificates in this form. 114759191Skris [Steve Henson] 114859191Skris 114959191Skris *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 115059191Skris list to exclude them. This means that no special compilation option 115159191Skris is needed to use anonymous DH: it just needs to be included in the 115259191Skris cipher list. 115359191Skris [Steve Henson] 115459191Skris 115559191Skris *) Change the EVP_MD_CTX_type macro so its meaning consistent with 115659191Skris EVP_MD_type. The old functionality is available in a new macro called 115759191Skris EVP_MD_md(). Change code that uses it and update docs. 115859191Skris [Steve Henson] 115959191Skris 116059191Skris *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 116159191Skris where the 'void *' argument is replaced by a function pointer argument. 116259191Skris Previously 'void *' was abused to point to functions, which works on 116359191Skris many platforms, but is not correct. As these functions are usually 116459191Skris called by macros defined in OpenSSL header files, most source code 116559191Skris should work without changes. 116659191Skris [Richard Levitte] 116759191Skris 116859191Skris *) <openssl/opensslconf.h> (which is created by Configure) now contains 116959191Skris sections with information on -D... compiler switches used for 117059191Skris compiling the library so that applications can see them. To enable 117159191Skris one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 117259191Skris must be defined. E.g., 117359191Skris #define OPENSSL_ALGORITHM_DEFINES 117459191Skris #include <openssl/opensslconf.h> 117559191Skris defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 117659191Skris [Richard Levitte, Ulf and Bodo M�ller] 117759191Skris 117859191Skris *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 117959191Skris record layer. 118059191Skris [Bodo Moeller] 118159191Skris 118259191Skris *) Change the 'other' type in certificate aux info to a STACK_OF 118359191Skris X509_ALGOR. Although not an AlgorithmIdentifier as such it has 118459191Skris the required ASN1 format: arbitrary types determined by an OID. 118559191Skris [Steve Henson] 118659191Skris 118759191Skris *) Add some PEM_write_X509_REQ_NEW() functions and a command line 118859191Skris argument to 'req'. This is not because the function is newer or 118959191Skris better than others it just uses the work 'NEW' in the certificate 119059191Skris request header lines. Some software needs this. 119159191Skris [Steve Henson] 119259191Skris 119359191Skris *) Reorganise password command line arguments: now passwords can be 119459191Skris obtained from various sources. Delete the PEM_cb function and make 119559191Skris it the default behaviour: i.e. if the callback is NULL and the 119659191Skris usrdata argument is not NULL interpret it as a null terminated pass 119759191Skris phrase. If usrdata and the callback are NULL then the pass phrase 119859191Skris is prompted for as usual. 119959191Skris [Steve Henson] 120059191Skris 120159191Skris *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 120259191Skris the support is automatically enabled. The resulting binaries will 120359191Skris autodetect the card and use it if present. 120459191Skris [Ben Laurie and Compaq Inc.] 120559191Skris 120659191Skris *) Work around for Netscape hang bug. This sends certificate request 120759191Skris and server done in one record. Since this is perfectly legal in the 120859191Skris SSL/TLS protocol it isn't a "bug" option and is on by default. See 120959191Skris the bugs/SSLv3 entry for more info. 121059191Skris [Steve Henson] 121159191Skris 121259191Skris *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 121359191Skris [Andy Polyakov] 121459191Skris 121559191Skris *) Add -rand argument to smime and pkcs12 applications and read/write 121659191Skris of seed file. 121759191Skris [Steve Henson] 121859191Skris 121959191Skris *) New 'passwd' tool for crypt(3) and apr1 password hashes. 122059191Skris [Bodo Moeller] 122159191Skris 122259191Skris *) Add command line password options to the remaining applications. 122359191Skris [Steve Henson] 122459191Skris 122559191Skris *) Bug fix for BN_div_recp() for numerators with an even number of 122659191Skris bits. 122759191Skris [Ulf M�ller] 122859191Skris 122959191Skris *) More tests in bntest.c, and changed test_bn output. 123059191Skris [Ulf M�ller] 123159191Skris 123259191Skris *) ./config recognizes MacOS X now. 123359191Skris [Andy Polyakov] 123459191Skris 123559191Skris *) Bug fix for BN_div() when the first words of num and divsor are 123659191Skris equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 123759191Skris [Ulf M�ller] 123859191Skris 123959191Skris *) Add support for various broken PKCS#8 formats, and command line 124059191Skris options to produce them. 124159191Skris [Steve Henson] 124259191Skris 124359191Skris *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 124459191Skris get temporary BIGNUMs from a BN_CTX. 124559191Skris [Ulf M�ller] 124659191Skris 124759191Skris *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 124859191Skris for p == 0. 124959191Skris [Ulf M�ller] 125059191Skris 125159191Skris *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 125259191Skris include a #define from the old name to the new. The original intent 125359191Skris was that statically linked binaries could for example just call 125459191Skris SSLeay_add_all_ciphers() to just add ciphers to the table and not 125559191Skris link with digests. This never worked becayse SSLeay_add_all_digests() 125659191Skris and SSLeay_add_all_ciphers() were in the same source file so calling 125759191Skris one would link with the other. They are now in separate source files. 125859191Skris [Steve Henson] 125959191Skris 126059191Skris *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 126159191Skris [Steve Henson] 126259191Skris 126359191Skris *) Use a less unusual form of the Miller-Rabin primality test (it used 126459191Skris a binary algorithm for exponentiation integrated into the Miller-Rabin 126559191Skris loop, our standard modexp algorithms are faster). 126659191Skris [Bodo Moeller] 126759191Skris 126859191Skris *) Support for the EBCDIC character set completed. 126959191Skris [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 127059191Skris 127159191Skris *) Source code cleanups: use const where appropriate, eliminate casts, 127259191Skris use void * instead of char * in lhash. 127359191Skris [Ulf M�ller] 127459191Skris 127559191Skris *) Bugfix: ssl3_send_server_key_exchange was not restartable 127659191Skris (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 127759191Skris this the server could overwrite ephemeral keys that the client 127859191Skris has already seen). 127959191Skris [Bodo Moeller] 128059191Skris 128159191Skris *) Turn DSA_is_prime into a macro that calls BN_is_prime, 128259191Skris using 50 iterations of the Rabin-Miller test. 128359191Skris 128459191Skris DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 128559191Skris iterations of the Rabin-Miller test as required by the appendix 128659191Skris to FIPS PUB 186[-1]) instead of DSA_is_prime. 128759191Skris As BN_is_prime_fasttest includes trial division, DSA parameter 128859191Skris generation becomes much faster. 128959191Skris 129059191Skris This implies a change for the callback functions in DSA_is_prime 129159191Skris and DSA_generate_parameters: The callback function is called once 129259191Skris for each positive witness in the Rabin-Miller test, not just 129359191Skris occasionally in the inner loop; and the parameters to the 129459191Skris callback function now provide an iteration count for the outer 129559191Skris loop rather than for the current invocation of the inner loop. 129659191Skris DSA_generate_parameters additionally can call the callback 129759191Skris function with an 'iteration count' of -1, meaning that a 129859191Skris candidate has passed the trial division test (when q is generated 129959191Skris from an application-provided seed, trial division is skipped). 130059191Skris [Bodo Moeller] 130159191Skris 130259191Skris *) New function BN_is_prime_fasttest that optionally does trial 130359191Skris division before starting the Rabin-Miller test and has 130459191Skris an additional BN_CTX * argument (whereas BN_is_prime always 130559191Skris has to allocate at least one BN_CTX). 130659191Skris 'callback(1, -1, cb_arg)' is called when a number has passed the 130759191Skris trial division stage. 130859191Skris [Bodo Moeller] 130959191Skris 131059191Skris *) Fix for bug in CRL encoding. The validity dates weren't being handled 131159191Skris as ASN1_TIME. 131259191Skris [Steve Henson] 131359191Skris 131459191Skris *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 131559191Skris [Steve Henson] 131659191Skris 131759191Skris *) New function BN_pseudo_rand(). 131859191Skris [Ulf M�ller] 131959191Skris 132059191Skris *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 132159191Skris bignum version of BN_from_montgomery() with the working code from 132259191Skris SSLeay 0.9.0 (the word based version is faster anyway), and clean up 132359191Skris the comments. 132459191Skris [Ulf M�ller] 132559191Skris 132659191Skris *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 132759191Skris made it impossible to use the same SSL_SESSION data structure in 132859191Skris SSL2 clients in multiple threads. 132959191Skris [Bodo Moeller] 133059191Skris 133159191Skris *) The return value of RAND_load_file() no longer counts bytes obtained 133259191Skris by stat(). RAND_load_file(..., -1) is new and uses the complete file 133359191Skris to seed the PRNG (previously an explicit byte count was required). 133459191Skris [Ulf M�ller, Bodo M�ller] 133559191Skris 133659191Skris *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 133759191Skris used (char *) instead of (void *) and had casts all over the place. 133859191Skris [Steve Henson] 133959191Skris 134059191Skris *) Make BN_generate_prime() return NULL on error if ret!=NULL. 134159191Skris [Ulf M�ller] 134259191Skris 134359191Skris *) Retain source code compatibility for BN_prime_checks macro: 134459191Skris BN_is_prime(..., BN_prime_checks, ...) now uses 134559191Skris BN_prime_checks_for_size to determine the appropriate number of 134659191Skris Rabin-Miller iterations. 134759191Skris [Ulf M�ller] 134859191Skris 134959191Skris *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 135059191Skris DH_CHECK_P_NOT_SAFE_PRIME. 135159191Skris (Check if this is true? OpenPGP calls them "strong".) 135259191Skris [Ulf M�ller] 135359191Skris 135459191Skris *) Merge the functionality of "dh" and "gendh" programs into a new program 135559191Skris "dhparam". The old programs are retained for now but will handle DH keys 135659191Skris (instead of parameters) in future. 135759191Skris [Steve Henson] 135859191Skris 135959191Skris *) Make the ciphers, s_server and s_client programs check the return values 136059191Skris when a new cipher list is set. 136159191Skris [Steve Henson] 136259191Skris 136359191Skris *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 136459191Skris ciphers. Before when the 56bit ciphers were enabled the sorting was 136559191Skris wrong. 136659191Skris 136759191Skris The syntax for the cipher sorting has been extended to support sorting by 136859191Skris cipher-strength (using the strength_bits hard coded in the tables). 136959191Skris The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 137059191Skris 137159191Skris Fix a bug in the cipher-command parser: when supplying a cipher command 137259191Skris string with an "undefined" symbol (neither command nor alphanumeric 137359191Skris [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 137459191Skris an error is flagged. 137559191Skris 137659191Skris Due to the strength-sorting extension, the code of the 137759191Skris ssl_create_cipher_list() function was completely rearranged. I hope that 137859191Skris the readability was also increased :-) 137959191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 138059191Skris 138159191Skris *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 138259191Skris for the first serial number and places 2 in the serial number file. This 138359191Skris avoids problems when the root CA is created with serial number zero and 138459191Skris the first user certificate has the same issuer name and serial number 138559191Skris as the root CA. 138659191Skris [Steve Henson] 138759191Skris 138859191Skris *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 138959191Skris the new code. Add documentation for this stuff. 139059191Skris [Steve Henson] 139159191Skris 139259191Skris *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 139359191Skris X509_*() to X509at_*() on the grounds that they don't handle X509 139459191Skris structures and behave in an analagous way to the X509v3 functions: 139559191Skris they shouldn't be called directly but wrapper functions should be used 139659191Skris instead. 139759191Skris 139859191Skris So we also now have some wrapper functions that call the X509at functions 139959191Skris when passed certificate requests. (TO DO: similar things can be done with 140059191Skris PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 140159191Skris things. Some of these need some d2i or i2d and print functionality 140259191Skris because they handle more complex structures.) 140359191Skris [Steve Henson] 140459191Skris 140559191Skris *) Add missing #ifndefs that caused missing symbols when building libssl 140659191Skris as a shared library without RSA. Use #ifndef NO_SSL2 instead of 140759191Skris NO_RSA in ssl/s2*.c. 140859191Skris [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 140959191Skris 141059191Skris *) Precautions against using the PRNG uninitialized: RAND_bytes() now 141159191Skris has a return value which indicates the quality of the random data 141259191Skris (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 141359191Skris error queue. New function RAND_pseudo_bytes() generates output that is 141459191Skris guaranteed to be unique but not unpredictable. RAND_add is like 141559191Skris RAND_seed, but takes an extra argument for an entropy estimate 141659191Skris (RAND_seed always assumes full entropy). 141759191Skris [Ulf M�ller] 141859191Skris 141959191Skris *) Do more iterations of Rabin-Miller probable prime test (specifically, 142059191Skris 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 142159191Skris instead of only 2 for all lengths; see BN_prime_checks_for_size definition 142259191Skris in crypto/bn/bn_prime.c for the complete table). This guarantees a 142359191Skris false-positive rate of at most 2^-80 for random input. 142459191Skris [Bodo Moeller] 142559191Skris 142659191Skris *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 142759191Skris [Bodo Moeller] 142859191Skris 142959191Skris *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 143059191Skris in the 0.9.5 release), this returns the chain 143159191Skris from an X509_CTX structure with a dup of the stack and all 143259191Skris the X509 reference counts upped: so the stack will exist 143359191Skris after X509_CTX_cleanup() has been called. Modify pkcs12.c 143459191Skris to use this. 143559191Skris 143659191Skris Also make SSL_SESSION_print() print out the verify return 143759191Skris code. 143859191Skris [Steve Henson] 143959191Skris 144059191Skris *) Add manpage for the pkcs12 command. Also change the default 144159191Skris behaviour so MAC iteration counts are used unless the new 144259191Skris -nomaciter option is used. This improves file security and 144359191Skris only older versions of MSIE (4.0 for example) need it. 144459191Skris [Steve Henson] 144559191Skris 144659191Skris *) Honor the no-xxx Configure options when creating .DEF files. 144759191Skris [Ulf M�ller] 144859191Skris 144959191Skris *) Add PKCS#10 attributes to field table: challengePassword, 145059191Skris unstructuredName and unstructuredAddress. These are taken from 145159191Skris draft PKCS#9 v2.0 but are compatible with v1.2 provided no 145259191Skris international characters are used. 145359191Skris 145459191Skris More changes to X509_ATTRIBUTE code: allow the setting of types 145559191Skris based on strings. Remove the 'loc' parameter when adding 145659191Skris attributes because these will be a SET OF encoding which is sorted 145759191Skris in ASN1 order. 145859191Skris [Steve Henson] 145959191Skris 146059191Skris *) Initial changes to the 'req' utility to allow request generation 146159191Skris automation. This will allow an application to just generate a template 146259191Skris file containing all the field values and have req construct the 146359191Skris request. 146459191Skris 146559191Skris Initial support for X509_ATTRIBUTE handling. Stacks of these are 146659191Skris used all over the place including certificate requests and PKCS#7 146759191Skris structures. They are currently handled manually where necessary with 146859191Skris some primitive wrappers for PKCS#7. The new functions behave in a 146959191Skris manner analogous to the X509 extension functions: they allow 147059191Skris attributes to be looked up by NID and added. 147159191Skris 147259191Skris Later something similar to the X509V3 code would be desirable to 147359191Skris automatically handle the encoding, decoding and printing of the 147459191Skris more complex types. The string types like challengePassword can 147559191Skris be handled by the string table functions. 147659191Skris 147759191Skris Also modified the multi byte string table handling. Now there is 147859191Skris a 'global mask' which masks out certain types. The table itself 147959191Skris can use the flag STABLE_NO_MASK to ignore the mask setting: this 148059191Skris is useful when for example there is only one permissible type 148159191Skris (as in countryName) and using the mask might result in no valid 148259191Skris types at all. 148359191Skris [Steve Henson] 148459191Skris 148559191Skris *) Clean up 'Finished' handling, and add functions SSL_get_finished and 148659191Skris SSL_get_peer_finished to allow applications to obtain the latest 148759191Skris Finished messages sent to the peer or expected from the peer, 148859191Skris respectively. (SSL_get_peer_finished is usually the Finished message 148959191Skris actually received from the peer, otherwise the protocol will be aborted.) 149059191Skris 149159191Skris As the Finished message are message digests of the complete handshake 149259191Skris (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 149359191Skris be used for external authentication procedures when the authentication 149459191Skris provided by SSL/TLS is not desired or is not enough. 149559191Skris [Bodo Moeller] 149659191Skris 149759191Skris *) Enhanced support for Alpha Linux is added. Now ./config checks if 149859191Skris the host supports BWX extension and if Compaq C is present on the 149959191Skris $PATH. Just exploiting of the BWX extension results in 20-30% 150059191Skris performance kick for some algorithms, e.g. DES and RC4 to mention 150159191Skris a couple. Compaq C in turn generates ~20% faster code for MD5 and 150259191Skris SHA1. 150359191Skris [Andy Polyakov] 150459191Skris 150559191Skris *) Add support for MS "fast SGC". This is arguably a violation of the 150659191Skris SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 150759191Skris weak crypto and after checking the certificate is SGC a second one 150859191Skris with strong crypto. MS SGC stops the first handshake after receiving 150959191Skris the server certificate message and sends a second client hello. Since 151059191Skris a server will typically do all the time consuming operations before 151159191Skris expecting any further messages from the client (server key exchange 151259191Skris is the most expensive) there is little difference between the two. 151359191Skris 151459191Skris To get OpenSSL to support MS SGC we have to permit a second client 151559191Skris hello message after we have sent server done. In addition we have to 151659191Skris reset the MAC if we do get this second client hello. 151759191Skris [Steve Henson] 151859191Skris 151959191Skris *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 152059191Skris if a DER encoded private key is RSA or DSA traditional format. Changed 152159191Skris d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 152259191Skris format DER encoded private key. Newer code should use PKCS#8 format which 152359191Skris has the key type encoded in the ASN1 structure. Added DER private key 152459191Skris support to pkcs8 application. 152559191Skris [Steve Henson] 152659191Skris 152759191Skris *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 152859191Skris ciphersuites has been selected (as required by the SSL 3/TLS 1 152959191Skris specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 153059191Skris is set, we interpret this as a request to violate the specification 153159191Skris (the worst that can happen is a handshake failure, and 'correct' 153259191Skris behaviour would result in a handshake failure anyway). 153359191Skris [Bodo Moeller] 153459191Skris 153559191Skris *) In SSL_CTX_add_session, take into account that there might be multiple 153659191Skris SSL_SESSION structures with the same session ID (e.g. when two threads 153759191Skris concurrently obtain them from an external cache). 153859191Skris The internal cache can handle only one SSL_SESSION with a given ID, 153959191Skris so if there's a conflict, we now throw out the old one to achieve 154059191Skris consistency. 154159191Skris [Bodo Moeller] 154259191Skris 154359191Skris *) Add OIDs for idea and blowfish in CBC mode. This will allow both 154459191Skris to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 154559191Skris some routines that use cipher OIDs: some ciphers do not have OIDs 154659191Skris defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 154759191Skris example. 154859191Skris [Steve Henson] 154959191Skris 155059191Skris *) Simplify the trust setting structure and code. Now we just have 155159191Skris two sequences of OIDs for trusted and rejected settings. These will 155259191Skris typically have values the same as the extended key usage extension 155359191Skris and any application specific purposes. 155459191Skris 155559191Skris The trust checking code now has a default behaviour: it will just 155659191Skris check for an object with the same NID as the passed id. Functions can 155759191Skris be provided to override either the default behaviour or the behaviour 155859191Skris for a given id. SSL client, server and email already have functions 155959191Skris in place for compatibility: they check the NID and also return "trusted" 156059191Skris if the certificate is self signed. 156159191Skris [Steve Henson] 156259191Skris 156359191Skris *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 156459191Skris traditional format into an EVP_PKEY structure. 156559191Skris [Steve Henson] 156659191Skris 156759191Skris *) Add a password callback function PEM_cb() which either prompts for 156859191Skris a password if usr_data is NULL or otherwise assumes it is a null 156959191Skris terminated password. Allow passwords to be passed on command line 157059191Skris environment or config files in a few more utilities. 157159191Skris [Steve Henson] 157259191Skris 157359191Skris *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 157459191Skris keys. Add some short names for PKCS#8 PBE algorithms and allow them 157559191Skris to be specified on the command line for the pkcs8 and pkcs12 utilities. 157659191Skris Update documentation. 157759191Skris [Steve Henson] 157859191Skris 157959191Skris *) Support for ASN1 "NULL" type. This could be handled before by using 158059191Skris ASN1_TYPE but there wasn't any function that would try to read a NULL 158159191Skris and produce an error if it couldn't. For compatibility we also have 158259191Skris ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 158359191Skris don't allocate anything because they don't need to. 158459191Skris [Steve Henson] 158559191Skris 158659191Skris *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 158759191Skris for details. 158859191Skris [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 158959191Skris 159059191Skris *) Rebuild of the memory allocation routines used by OpenSSL code and 159159191Skris possibly others as well. The purpose is to make an interface that 159259191Skris provide hooks so anyone can build a separate set of allocation and 159359191Skris deallocation routines to be used by OpenSSL, for example memory 159459191Skris pool implementations, or something else, which was previously hard 159559191Skris since Malloc(), Realloc() and Free() were defined as macros having 159659191Skris the values malloc, realloc and free, respectively (except for Win32 159759191Skris compilations). The same is provided for memory debugging code. 159859191Skris OpenSSL already comes with functionality to find memory leaks, but 159959191Skris this gives people a chance to debug other memory problems. 160059191Skris 160159191Skris With these changes, a new set of functions and macros have appeared: 160259191Skris 160368651Skris CRYPTO_set_mem_debug_functions() [F] 160468651Skris CRYPTO_get_mem_debug_functions() [F] 160568651Skris CRYPTO_dbg_set_options() [F] 160668651Skris CRYPTO_dbg_get_options() [F] 160768651Skris CRYPTO_malloc_debug_init() [M] 160859191Skris 160959191Skris The memory debug functions are NULL by default, unless the library 161059191Skris is compiled with CRYPTO_MDEBUG or friends is defined. If someone 161159191Skris wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 161259191Skris gives the standard debugging functions that come with OpenSSL) or 161359191Skris CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 161459191Skris provided by the library user) must be used. When the standard 161559191Skris debugging functions are used, CRYPTO_dbg_set_options can be used to 161659191Skris request additional information: 161759191Skris CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 161859191Skris the CRYPTO_MDEBUG_xxx macro when compiling the library. 161959191Skris 162059191Skris Also, things like CRYPTO_set_mem_functions will always give the 162159191Skris expected result (the new set of functions is used for allocation 162259191Skris and deallocation) at all times, regardless of platform and compiler 162359191Skris options. 162459191Skris 162559191Skris To finish it up, some functions that were never use in any other 162659191Skris way than through macros have a new API and new semantic: 162759191Skris 162859191Skris CRYPTO_dbg_malloc() 162959191Skris CRYPTO_dbg_realloc() 163059191Skris CRYPTO_dbg_free() 163159191Skris 163259191Skris All macros of value have retained their old syntax. 163359191Skris [Richard Levitte and Bodo Moeller] 163459191Skris 163559191Skris *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 163659191Skris ordering of SMIMECapabilities wasn't in "strength order" and there 163759191Skris was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 163859191Skris algorithm. 163959191Skris [Steve Henson] 164059191Skris 164159191Skris *) Some ASN1 types with illegal zero length encoding (INTEGER, 164259191Skris ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 164359191Skris [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 164459191Skris 164559191Skris *) Merge in my S/MIME library for OpenSSL. This provides a simple 164659191Skris S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 164759191Skris functionality to handle multipart/signed properly) and a utility 164859191Skris called 'smime' to call all this stuff. This is based on code I 164959191Skris originally wrote for Celo who have kindly allowed it to be 165059191Skris included in OpenSSL. 165159191Skris [Steve Henson] 165259191Skris 165359191Skris *) Add variants des_set_key_checked and des_set_key_unchecked of 165459191Skris des_set_key (aka des_key_sched). Global variable des_check_key 165559191Skris decides which of these is called by des_set_key; this way 165659191Skris des_check_key behaves as it always did, but applications and 165759191Skris the library itself, which was buggy for des_check_key == 1, 165859191Skris have a cleaner way to pick the version they need. 165959191Skris [Bodo Moeller] 166059191Skris 166159191Skris *) New function PKCS12_newpass() which changes the password of a 166259191Skris PKCS12 structure. 166359191Skris [Steve Henson] 166459191Skris 166559191Skris *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 166659191Skris dynamic mix. In both cases the ids can be used as an index into the 166759191Skris table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 166859191Skris functions so they accept a list of the field values and the 166959191Skris application doesn't need to directly manipulate the X509_TRUST 167059191Skris structure. 167159191Skris [Steve Henson] 167259191Skris 167359191Skris *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 167459191Skris need initialising. 167559191Skris [Steve Henson] 167659191Skris 167759191Skris *) Modify the way the V3 extension code looks up extensions. This now 167859191Skris works in a similar way to the object code: we have some "standard" 167959191Skris extensions in a static table which is searched with OBJ_bsearch() 168059191Skris and the application can add dynamic ones if needed. The file 168159191Skris crypto/x509v3/ext_dat.h now has the info: this file needs to be 168259191Skris updated whenever a new extension is added to the core code and kept 168359191Skris in ext_nid order. There is a simple program 'tabtest.c' which checks 168459191Skris this. New extensions are not added too often so this file can readily 168559191Skris be maintained manually. 168659191Skris 168759191Skris There are two big advantages in doing things this way. The extensions 168859191Skris can be looked up immediately and no longer need to be "added" using 168959191Skris X509V3_add_standard_extensions(): this function now does nothing. 169059191Skris [Side note: I get *lots* of email saying the extension code doesn't 169159191Skris work because people forget to call this function] 169259191Skris Also no dynamic allocation is done unless new extensions are added: 169359191Skris so if we don't add custom extensions there is no need to call 169459191Skris X509V3_EXT_cleanup(). 169559191Skris [Steve Henson] 169659191Skris 169759191Skris *) Modify enc utility's salting as follows: make salting the default. Add a 169859191Skris magic header, so unsalted files fail gracefully instead of just decrypting 169959191Skris to garbage. This is because not salting is a big security hole, so people 170059191Skris should be discouraged from doing it. 170159191Skris [Ben Laurie] 170259191Skris 170359191Skris *) Fixes and enhancements to the 'x509' utility. It allowed a message 170459191Skris digest to be passed on the command line but it only used this 170559191Skris parameter when signing a certificate. Modified so all relevant 170659191Skris operations are affected by the digest parameter including the 170759191Skris -fingerprint and -x509toreq options. Also -x509toreq choked if a 170859191Skris DSA key was used because it didn't fix the digest. 170959191Skris [Steve Henson] 171059191Skris 171159191Skris *) Initial certificate chain verify code. Currently tests the untrusted 171259191Skris certificates for consistency with the verify purpose (which is set 171359191Skris when the X509_STORE_CTX structure is set up) and checks the pathlength. 171459191Skris 171559191Skris There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 171659191Skris this is because it will reject chains with invalid extensions whereas 171759191Skris every previous version of OpenSSL and SSLeay made no checks at all. 171859191Skris 171959191Skris Trust code: checks the root CA for the relevant trust settings. Trust 172059191Skris settings have an initial value consistent with the verify purpose: e.g. 172159191Skris if the verify purpose is for SSL client use it expects the CA to be 172259191Skris trusted for SSL client use. However the default value can be changed to 172359191Skris permit custom trust settings: one example of this would be to only trust 172459191Skris certificates from a specific "secure" set of CAs. 172559191Skris 172659191Skris Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 172759191Skris which should be used for version portability: especially since the 172859191Skris verify structure is likely to change more often now. 172959191Skris 173059191Skris SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 173159191Skris to set them. If not set then assume SSL clients will verify SSL servers 173259191Skris and vice versa. 173359191Skris 173459191Skris Two new options to the verify program: -untrusted allows a set of 173559191Skris untrusted certificates to be passed in and -purpose which sets the 173659191Skris intended purpose of the certificate. If a purpose is set then the 173759191Skris new chain verify code is used to check extension consistency. 173859191Skris [Steve Henson] 173959191Skris 174059191Skris *) Support for the authority information access extension. 174159191Skris [Steve Henson] 174259191Skris 174359191Skris *) Modify RSA and DSA PEM read routines to transparently handle 174459191Skris PKCS#8 format private keys. New *_PUBKEY_* functions that handle 174559191Skris public keys in a format compatible with certificate 174659191Skris SubjectPublicKeyInfo structures. Unfortunately there were already 174759191Skris functions called *_PublicKey_* which used various odd formats so 174859191Skris these are retained for compatibility: however the DSA variants were 174959191Skris never in a public release so they have been deleted. Changed dsa/rsa 175059191Skris utilities to handle the new format: note no releases ever handled public 175159191Skris keys so we should be OK. 175259191Skris 175359191Skris The primary motivation for this change is to avoid the same fiasco 175459191Skris that dogs private keys: there are several incompatible private key 175559191Skris formats some of which are standard and some OpenSSL specific and 175659191Skris require various evil hacks to allow partial transparent handling and 175759191Skris even then it doesn't work with DER formats. Given the option anything 175859191Skris other than PKCS#8 should be dumped: but the other formats have to 175959191Skris stay in the name of compatibility. 176059191Skris 176159191Skris With public keys and the benefit of hindsight one standard format 176259191Skris is used which works with EVP_PKEY, RSA or DSA structures: though 176359191Skris it clearly returns an error if you try to read the wrong kind of key. 176459191Skris 176559191Skris Added a -pubkey option to the 'x509' utility to output the public key. 176659191Skris Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 176759191Skris (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 176859191Skris EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 176959191Skris that do the same as the EVP_PKEY_assign_*() except they up the 177059191Skris reference count of the added key (they don't "swallow" the 177159191Skris supplied key). 177259191Skris [Steve Henson] 177359191Skris 177459191Skris *) Fixes to crypto/x509/by_file.c the code to read in certificates and 177559191Skris CRLs would fail if the file contained no certificates or no CRLs: 177659191Skris added a new function to read in both types and return the number 177759191Skris read: this means that if none are read it will be an error. The 177859191Skris DER versions of the certificate and CRL reader would always fail 177959191Skris because it isn't possible to mix certificates and CRLs in DER format 178059191Skris without choking one or the other routine. Changed this to just read 178159191Skris a certificate: this is the best we can do. Also modified the code 178259191Skris in apps/verify.c to take notice of return codes: it was previously 178359191Skris attempting to read in certificates from NULL pointers and ignoring 178459191Skris any errors: this is one reason why the cert and CRL reader seemed 178559191Skris to work. It doesn't check return codes from the default certificate 178659191Skris routines: these may well fail if the certificates aren't installed. 178759191Skris [Steve Henson] 178859191Skris 178959191Skris *) Code to support otherName option in GeneralName. 179059191Skris [Steve Henson] 179159191Skris 179259191Skris *) First update to verify code. Change the verify utility 179359191Skris so it warns if it is passed a self signed certificate: 179459191Skris for consistency with the normal behaviour. X509_verify 179559191Skris has been modified to it will now verify a self signed 179659191Skris certificate if *exactly* the same certificate appears 179759191Skris in the store: it was previously impossible to trust a 179859191Skris single self signed certificate. This means that: 179959191Skris openssl verify ss.pem 180059191Skris now gives a warning about a self signed certificate but 180159191Skris openssl verify -CAfile ss.pem ss.pem 180259191Skris is OK. 180359191Skris [Steve Henson] 180459191Skris 180559191Skris *) For servers, store verify_result in SSL_SESSION data structure 180659191Skris (and add it to external session representation). 180759191Skris This is needed when client certificate verifications fails, 180859191Skris but an application-provided verification callback (set by 180959191Skris SSL_CTX_set_cert_verify_callback) allows accepting the session 181059191Skris anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 181159191Skris but returns 1): When the session is reused, we have to set 181259191Skris ssl->verify_result to the appropriate error code to avoid 181359191Skris security holes. 181459191Skris [Bodo Moeller, problem pointed out by Lutz Jaenicke] 181559191Skris 181659191Skris *) Fix a bug in the new PKCS#7 code: it didn't consider the 181759191Skris case in PKCS7_dataInit() where the signed PKCS7 structure 181859191Skris didn't contain any existing data because it was being created. 181959191Skris [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 182059191Skris 182159191Skris *) Add a salt to the key derivation routines in enc.c. This 182259191Skris forms the first 8 bytes of the encrypted file. Also add a 182359191Skris -S option to allow a salt to be input on the command line. 182459191Skris [Steve Henson] 182559191Skris 182659191Skris *) New function X509_cmp(). Oddly enough there wasn't a function 182759191Skris to compare two certificates. We do this by working out the SHA1 182859191Skris hash and comparing that. X509_cmp() will be needed by the trust 182959191Skris code. 183059191Skris [Steve Henson] 183159191Skris 183259191Skris *) SSL_get1_session() is like SSL_get_session(), but increments 183359191Skris the reference count in the SSL_SESSION returned. 183459191Skris [Geoff Thorpe <geoff@eu.c2.net>] 183559191Skris 183659191Skris *) Fix for 'req': it was adding a null to request attributes. 183759191Skris Also change the X509_LOOKUP and X509_INFO code to handle 183859191Skris certificate auxiliary information. 183959191Skris [Steve Henson] 184059191Skris 184159191Skris *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 184259191Skris the 'enc' command. 184359191Skris [Steve Henson] 184459191Skris 184559191Skris *) Add the possibility to add extra information to the memory leak 184659191Skris detecting output, to form tracebacks, showing from where each 184759191Skris allocation was originated: CRYPTO_push_info("constant string") adds 184859191Skris the string plus current file name and line number to a per-thread 184959191Skris stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 185059191Skris is like calling CYRPTO_pop_info() until the stack is empty. 185159191Skris Also updated memory leak detection code to be multi-thread-safe. 185259191Skris [Richard Levitte] 185359191Skris 185459191Skris *) Add options -text and -noout to pkcs7 utility and delete the 185559191Skris encryption options which never did anything. Update docs. 185659191Skris [Steve Henson] 185759191Skris 185859191Skris *) Add options to some of the utilities to allow the pass phrase 185959191Skris to be included on either the command line (not recommended on 186059191Skris OSes like Unix) or read from the environment. Update the 186159191Skris manpages and fix a few bugs. 186259191Skris [Steve Henson] 186359191Skris 186459191Skris *) Add a few manpages for some of the openssl commands. 186559191Skris [Steve Henson] 186659191Skris 186759191Skris *) Fix the -revoke option in ca. It was freeing up memory twice, 186859191Skris leaking and not finding already revoked certificates. 186959191Skris [Steve Henson] 187059191Skris 187159191Skris *) Extensive changes to support certificate auxiliary information. 187259191Skris This involves the use of X509_CERT_AUX structure and X509_AUX 187359191Skris functions. An X509_AUX function such as PEM_read_X509_AUX() 187459191Skris can still read in a certificate file in the usual way but it 187559191Skris will also read in any additional "auxiliary information". By 187659191Skris doing things this way a fair degree of compatibility can be 187759191Skris retained: existing certificates can have this information added 187859191Skris using the new 'x509' options. 187959191Skris 188059191Skris Current auxiliary information includes an "alias" and some trust 188159191Skris settings. The trust settings will ultimately be used in enhanced 188259191Skris certificate chain verification routines: currently a certificate 188359191Skris can only be trusted if it is self signed and then it is trusted 188459191Skris for all purposes. 188559191Skris [Steve Henson] 188659191Skris 188759191Skris *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 188859191Skris The problem was that one of the replacement routines had not been working 188959191Skris since SSLeay releases. For now the offending routine has been replaced 189059191Skris with non-optimised assembler. Even so, this now gives around 95% 189159191Skris performance improvement for 1024 bit RSA signs. 189259191Skris [Mark Cox] 189359191Skris 189459191Skris *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 189559191Skris handling. Most clients have the effective key size in bits equal to 189659191Skris the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 189759191Skris A few however don't do this and instead use the size of the decrypted key 189859191Skris to determine the RC2 key length and the AlgorithmIdentifier to determine 189959191Skris the effective key length. In this case the effective key length can still 190059191Skris be 40 bits but the key length can be 168 bits for example. This is fixed 190159191Skris by manually forcing an RC2 key into the EVP_PKEY structure because the 190259191Skris EVP code can't currently handle unusual RC2 key sizes: it always assumes 190359191Skris the key length and effective key length are equal. 190459191Skris [Steve Henson] 190559191Skris 190659191Skris *) Add a bunch of functions that should simplify the creation of 190759191Skris X509_NAME structures. Now you should be able to do: 190859191Skris X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 190959191Skris and have it automatically work out the correct field type and fill in 191059191Skris the structures. The more adventurous can try: 191159191Skris X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 191259191Skris and it will (hopefully) work out the correct multibyte encoding. 191359191Skris [Steve Henson] 191459191Skris 191559191Skris *) Change the 'req' utility to use the new field handling and multibyte 191659191Skris copy routines. Before the DN field creation was handled in an ad hoc 191759191Skris way in req, ca, and x509 which was rather broken and didn't support 191859191Skris BMPStrings or UTF8Strings. Since some software doesn't implement 191959191Skris BMPStrings or UTF8Strings yet, they can be enabled using the config file 192059191Skris using the dirstring_type option. See the new comment in the default 192159191Skris openssl.cnf for more info. 192259191Skris [Steve Henson] 192359191Skris 192459191Skris *) Make crypto/rand/md_rand.c more robust: 192559191Skris - Assure unique random numbers after fork(). 192659191Skris - Make sure that concurrent threads access the global counter and 192759191Skris md serializably so that we never lose entropy in them 192859191Skris or use exactly the same state in multiple threads. 192959191Skris Access to the large state is not always serializable because 193059191Skris the additional locking could be a performance killer, and 193159191Skris md should be large enough anyway. 193259191Skris [Bodo Moeller] 193359191Skris 193459191Skris *) New file apps/app_rand.c with commonly needed functionality 193559191Skris for handling the random seed file. 193659191Skris 193759191Skris Use the random seed file in some applications that previously did not: 193859191Skris ca, 193959191Skris dsaparam -genkey (which also ignored its '-rand' option), 194059191Skris s_client, 194159191Skris s_server, 194259191Skris x509 (when signing). 194359191Skris Except on systems with /dev/urandom, it is crucial to have a random 194459191Skris seed file at least for key creation, DSA signing, and for DH exchanges; 194559191Skris for RSA signatures we could do without one. 194659191Skris 194759191Skris gendh and gendsa (unlike genrsa) used to read only the first byte 194859191Skris of each file listed in the '-rand' option. The function as previously 194959191Skris found in genrsa is now in app_rand.c and is used by all programs 195059191Skris that support '-rand'. 195159191Skris [Bodo Moeller] 195259191Skris 195359191Skris *) In RAND_write_file, use mode 0600 for creating files; 195459191Skris don't just chmod when it may be too late. 195559191Skris [Bodo Moeller] 195659191Skris 195759191Skris *) Report an error from X509_STORE_load_locations 195859191Skris when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 195959191Skris [Bill Perry] 196059191Skris 196159191Skris *) New function ASN1_mbstring_copy() this copies a string in either 196259191Skris ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 196359191Skris into an ASN1_STRING type. A mask of permissible types is passed 196459191Skris and it chooses the "minimal" type to use or an error if not type 196559191Skris is suitable. 196659191Skris [Steve Henson] 196759191Skris 196859191Skris *) Add function equivalents to the various macros in asn1.h. The old 196959191Skris macros are retained with an M_ prefix. Code inside the library can 197059191Skris use the M_ macros. External code (including the openssl utility) 197159191Skris should *NOT* in order to be "shared library friendly". 197259191Skris [Steve Henson] 197359191Skris 197459191Skris *) Add various functions that can check a certificate's extensions 197559191Skris to see if it usable for various purposes such as SSL client, 197659191Skris server or S/MIME and CAs of these types. This is currently 197759191Skris VERY EXPERIMENTAL but will ultimately be used for certificate chain 197859191Skris verification. Also added a -purpose flag to x509 utility to 197959191Skris print out all the purposes. 198059191Skris [Steve Henson] 198159191Skris 198259191Skris *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 198359191Skris functions. 198459191Skris [Steve Henson] 198559191Skris 198659191Skris *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 198759191Skris for, obtain and decode and extension and obtain its critical flag. 198859191Skris This allows all the necessary extension code to be handled in a 198959191Skris single function call. 199059191Skris [Steve Henson] 199159191Skris 199259191Skris *) RC4 tune-up featuring 30-40% performance improvement on most RISC 199359191Skris platforms. See crypto/rc4/rc4_enc.c for further details. 199459191Skris [Andy Polyakov] 199559191Skris 199659191Skris *) New -noout option to asn1parse. This causes no output to be produced 199759191Skris its main use is when combined with -strparse and -out to extract data 199859191Skris from a file (which may not be in ASN.1 format). 199959191Skris [Steve Henson] 200059191Skris 200159191Skris *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 200259191Skris when producing the local key id. 200359191Skris [Richard Levitte <levitte@stacken.kth.se>] 200459191Skris 200559191Skris *) New option -dhparam in s_server. This allows a DH parameter file to be 200659191Skris stated explicitly. If it is not stated then it tries the first server 200759191Skris certificate file. The previous behaviour hard coded the filename 200859191Skris "server.pem". 200959191Skris [Steve Henson] 201059191Skris 201159191Skris *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 201259191Skris a public key to be input or output. For example: 201359191Skris openssl rsa -in key.pem -pubout -out pubkey.pem 201459191Skris Also added necessary DSA public key functions to handle this. 201559191Skris [Steve Henson] 201659191Skris 201759191Skris *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 201859191Skris in the message. This was handled by allowing 201959191Skris X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 202059191Skris [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 202159191Skris 202259191Skris *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 202359191Skris to the end of the strings whereas this didn't. This would cause problems 202459191Skris if strings read with d2i_ASN1_bytes() were later modified. 202559191Skris [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 202659191Skris 202759191Skris *) Fix for base64 decode bug. When a base64 bio reads only one line of 202859191Skris data and it contains EOF it will end up returning an error. This is 202959191Skris caused by input 46 bytes long. The cause is due to the way base64 203059191Skris BIOs find the start of base64 encoded data. They do this by trying a 203159191Skris trial decode on each line until they find one that works. When they 203259191Skris do a flag is set and it starts again knowing it can pass all the 203359191Skris data directly through the decoder. Unfortunately it doesn't reset 203459191Skris the context it uses. This means that if EOF is reached an attempt 203559191Skris is made to pass two EOFs through the context and this causes the 203659191Skris resulting error. This can also cause other problems as well. As is 203759191Skris usual with these problems it takes *ages* to find and the fix is 203859191Skris trivial: move one line. 203959191Skris [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 204059191Skris 204159191Skris *) Ugly workaround to get s_client and s_server working under Windows. The 204259191Skris old code wouldn't work because it needed to select() on sockets and the 204359191Skris tty (for keypresses and to see if data could be written). Win32 only 204459191Skris supports select() on sockets so we select() with a 1s timeout on the 204559191Skris sockets and then see if any characters are waiting to be read, if none 204659191Skris are present then we retry, we also assume we can always write data to 204759191Skris the tty. This isn't nice because the code then blocks until we've 204859191Skris received a complete line of data and it is effectively polling the 204959191Skris keyboard at 1s intervals: however it's quite a bit better than not 205059191Skris working at all :-) A dedicated Windows application might handle this 205159191Skris with an event loop for example. 205259191Skris [Steve Henson] 205359191Skris 205459191Skris *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 205559191Skris and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 205659191Skris will be called when RSA_sign() and RSA_verify() are used. This is useful 205759191Skris if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 205859191Skris For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 205959191Skris should *not* be used: RSA_sign() and RSA_verify() must be used instead. 206059191Skris This necessitated the support of an extra signature type NID_md5_sha1 206159191Skris for SSL signatures and modifications to the SSL library to use it instead 206259191Skris of calling RSA_public_decrypt() and RSA_private_encrypt(). 206359191Skris [Steve Henson] 206459191Skris 206559191Skris *) Add new -verify -CAfile and -CApath options to the crl program, these 206659191Skris will lookup a CRL issuers certificate and verify the signature in a 206759191Skris similar way to the verify program. Tidy up the crl program so it 206859191Skris no longer accesses structures directly. Make the ASN1 CRL parsing a bit 206959191Skris less strict. It will now permit CRL extensions even if it is not 207059191Skris a V2 CRL: this will allow it to tolerate some broken CRLs. 207159191Skris [Steve Henson] 207259191Skris 207359191Skris *) Initialize all non-automatic variables each time one of the openssl 207459191Skris sub-programs is started (this is necessary as they may be started 207559191Skris multiple times from the "OpenSSL>" prompt). 207659191Skris [Lennart Bang, Bodo Moeller] 207759191Skris 207859191Skris *) Preliminary compilation option RSA_NULL which disables RSA crypto without 207959191Skris removing all other RSA functionality (this is what NO_RSA does). This 208059191Skris is so (for example) those in the US can disable those operations covered 208159191Skris by the RSA patent while allowing storage and parsing of RSA keys and RSA 208259191Skris key generation. 208359191Skris [Steve Henson] 208459191Skris 208559191Skris *) Non-copying interface to BIO pairs. 208659191Skris (still largely untested) 208759191Skris [Bodo Moeller] 208859191Skris 208959191Skris *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 209059191Skris ASCII string. This was handled independently in various places before. 209159191Skris [Steve Henson] 209259191Skris 209359191Skris *) New functions UTF8_getc() and UTF8_putc() that parse and generate 209459191Skris UTF8 strings a character at a time. 209559191Skris [Steve Henson] 209659191Skris 209759191Skris *) Use client_version from client hello to select the protocol 209859191Skris (s23_srvr.c) and for RSA client key exchange verification 209959191Skris (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 210059191Skris [Bodo Moeller] 210159191Skris 210259191Skris *) Add various utility functions to handle SPKACs, these were previously 210359191Skris handled by poking round in the structure internals. Added new function 210459191Skris NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 210559191Skris print, verify and generate SPKACs. Based on an original idea from 210659191Skris Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 210759191Skris [Steve Henson] 210859191Skris 210959191Skris *) RIPEMD160 is operational on all platforms and is back in 'make test'. 211059191Skris [Andy Polyakov] 211159191Skris 211259191Skris *) Allow the config file extension section to be overwritten on the 211359191Skris command line. Based on an original idea from Massimiliano Pala 211459191Skris <madwolf@comune.modena.it>. The new option is called -extensions 211559191Skris and can be applied to ca, req and x509. Also -reqexts to override 211659191Skris the request extensions in req and -crlexts to override the crl extensions 211759191Skris in ca. 211859191Skris [Steve Henson] 211959191Skris 212059191Skris *) Add new feature to the SPKAC handling in ca. Now you can include 212159191Skris the same field multiple times by preceding it by "XXXX." for example: 212259191Skris 1.OU="Unit name 1" 212359191Skris 2.OU="Unit name 2" 212459191Skris this is the same syntax as used in the req config file. 212559191Skris [Steve Henson] 212659191Skris 212759191Skris *) Allow certificate extensions to be added to certificate requests. These 212859191Skris are specified in a 'req_extensions' option of the req section of the 212959191Skris config file. They can be printed out with the -text option to req but 213059191Skris are otherwise ignored at present. 213159191Skris [Steve Henson] 213259191Skris 213359191Skris *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 213459191Skris data read consists of only the final block it would not decrypted because 213559191Skris EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 213659191Skris A misplaced 'break' also meant the decrypted final block might not be 213759191Skris copied until the next read. 213859191Skris [Steve Henson] 213959191Skris 214059191Skris *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 214159191Skris a few extra parameters to the DH structure: these will be useful if 214259191Skris for example we want the value of 'q' or implement X9.42 DH. 214359191Skris [Steve Henson] 214459191Skris 214559191Skris *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 214659191Skris provides hooks that allow the default DSA functions or functions on a 214759191Skris "per key" basis to be replaced. This allows hardware acceleration and 214859191Skris hardware key storage to be handled without major modification to the 214959191Skris library. Also added low level modexp hooks and CRYPTO_EX structure and 215059191Skris associated functions. 215159191Skris [Steve Henson] 215259191Skris 215359191Skris *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 215459191Skris as "read only": it can't be written to and the buffer it points to will 215559191Skris not be freed. Reading from a read only BIO is much more efficient than 215659191Skris a normal memory BIO. This was added because there are several times when 215759191Skris an area of memory needs to be read from a BIO. The previous method was 215859191Skris to create a memory BIO and write the data to it, this results in two 215959191Skris copies of the data and an O(n^2) reading algorithm. There is a new 216059191Skris function BIO_new_mem_buf() which creates a read only memory BIO from 216159191Skris an area of memory. Also modified the PKCS#7 routines to use read only 216259191Skris memory BIOs. 216359191Skris [Steve Henson] 216459191Skris 216559191Skris *) Bugfix: ssl23_get_client_hello did not work properly when called in 216659191Skris state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 216759191Skris a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 216859191Skris but a retry condition occured while trying to read the rest. 216959191Skris [Bodo Moeller] 217059191Skris 217159191Skris *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 217259191Skris NID_pkcs7_encrypted by default: this was wrong since this should almost 217359191Skris always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 217459191Skris the encrypted data type: this is a more sensible place to put it and it 217559191Skris allows the PKCS#12 code to be tidied up that duplicated this 217659191Skris functionality. 217759191Skris [Steve Henson] 217859191Skris 217959191Skris *) Changed obj_dat.pl script so it takes its input and output files on 218059191Skris the command line. This should avoid shell escape redirection problems 218159191Skris under Win32. 218259191Skris [Steve Henson] 218359191Skris 218459191Skris *) Initial support for certificate extension requests, these are included 218559191Skris in things like Xenroll certificate requests. Included functions to allow 218659191Skris extensions to be obtained and added. 218759191Skris [Steve Henson] 218859191Skris 218959191Skris *) -crlf option to s_client and s_server for sending newlines as 219059191Skris CRLF (as required by many protocols). 219159191Skris [Bodo Moeller] 219259191Skris 219355714Skris Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 219455714Skris 219555714Skris *) Install libRSAglue.a when OpenSSL is built with RSAref. 219655714Skris [Ralf S. Engelschall] 219755714Skris 219855714Skris *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 219955714Skris [Andrija Antonijevic <TheAntony2@bigfoot.com>] 220055714Skris 220155714Skris *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 220255714Skris program. 220355714Skris [Steve Henson] 220455714Skris 220555714Skris *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 220655714Skris DH parameters/keys (q is lost during that conversion, but the resulting 220755714Skris DH parameters contain its length). 220855714Skris 220955714Skris For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 221055714Skris much faster than DH_generate_parameters (which creates parameters 221155714Skris where p = 2*q + 1), and also the smaller q makes DH computations 221255714Skris much more efficient (160-bit exponentiation instead of 1024-bit 221355714Skris exponentiation); so this provides a convenient way to support DHE 221455714Skris ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 221555714Skris utter importance to use 221655714Skris SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 221755714Skris or 221855714Skris SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 221955714Skris when such DH parameters are used, because otherwise small subgroup 222055714Skris attacks may become possible! 222155714Skris [Bodo Moeller] 222255714Skris 222355714Skris *) Avoid memory leak in i2d_DHparams. 222455714Skris [Bodo Moeller] 222555714Skris 222655714Skris *) Allow the -k option to be used more than once in the enc program: 222755714Skris this allows the same encrypted message to be read by multiple recipients. 222855714Skris [Steve Henson] 222955714Skris 223055714Skris *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 223155714Skris an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 223255714Skris it will always use the numerical form of the OID, even if it has a short 223355714Skris or long name. 223455714Skris [Steve Henson] 223555714Skris 223655714Skris *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 223755714Skris method only got called if p,q,dmp1,dmq1,iqmp components were present, 223855714Skris otherwise bn_mod_exp was called. In the case of hardware keys for example 223955714Skris no private key components need be present and it might store extra data 224059191Skris in the RSA structure, which cannot be accessed from bn_mod_exp. 224159191Skris By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 224259191Skris private key operations. 224355714Skris [Steve Henson] 224455714Skris 224555714Skris *) Added support for SPARC Linux. 224655714Skris [Andy Polyakov] 224755714Skris 224855714Skris *) pem_password_cb function type incompatibly changed from 224955714Skris typedef int pem_password_cb(char *buf, int size, int rwflag); 225055714Skris to 225155714Skris ....(char *buf, int size, int rwflag, void *userdata); 225255714Skris so that applications can pass data to their callbacks: 225355714Skris The PEM[_ASN1]_{read,write}... functions and macros now take an 225455714Skris additional void * argument, which is just handed through whenever 225555714Skris the password callback is called. 225659191Skris [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 225755714Skris 225855714Skris New function SSL_CTX_set_default_passwd_cb_userdata. 225955714Skris 226055714Skris Compatibility note: As many C implementations push function arguments 226155714Skris onto the stack in reverse order, the new library version is likely to 226255714Skris interoperate with programs that have been compiled with the old 226355714Skris pem_password_cb definition (PEM_whatever takes some data that 226455714Skris happens to be on the stack as its last argument, and the callback 226555714Skris just ignores this garbage); but there is no guarantee whatsoever that 226655714Skris this will work. 226755714Skris 226855714Skris *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 226955714Skris (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 227055714Skris problems not only on Windows, but also on some Unix platforms. 227155714Skris To avoid problematic command lines, these definitions are now in an 227255714Skris auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 227355714Skris for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 227455714Skris [Bodo Moeller] 227555714Skris 227655714Skris *) MIPS III/IV assembler module is reimplemented. 227755714Skris [Andy Polyakov] 227855714Skris 227955714Skris *) More DES library cleanups: remove references to srand/rand and 228055714Skris delete an unused file. 228155714Skris [Ulf M�ller] 228255714Skris 228355714Skris *) Add support for the the free Netwide assembler (NASM) under Win32, 228455714Skris since not many people have MASM (ml) and it can be hard to obtain. 228555714Skris This is currently experimental but it seems to work OK and pass all 228655714Skris the tests. Check out INSTALL.W32 for info. 228755714Skris [Steve Henson] 228855714Skris 228955714Skris *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 229055714Skris without temporary keys kept an extra copy of the server key, 229155714Skris and connections with temporary keys did not free everything in case 229255714Skris of an error. 229355714Skris [Bodo Moeller] 229455714Skris 229555714Skris *) New function RSA_check_key and new openssl rsa option -check 229655714Skris for verifying the consistency of RSA keys. 229755714Skris [Ulf Moeller, Bodo Moeller] 229855714Skris 229955714Skris *) Various changes to make Win32 compile work: 230055714Skris 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 230155714Skris 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 230255714Skris comparison" warnings. 230355714Skris 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 230455714Skris [Steve Henson] 230555714Skris 230655714Skris *) Add a debugging option to PKCS#5 v2 key generation function: when 230755714Skris you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 230855714Skris derived keys are printed to stderr. 230955714Skris [Steve Henson] 231055714Skris 231155714Skris *) Copy the flags in ASN1_STRING_dup(). 231255714Skris [Roman E. Pavlov <pre@mo.msk.ru>] 231355714Skris 231455714Skris *) The x509 application mishandled signing requests containing DSA 231555714Skris keys when the signing key was also DSA and the parameters didn't match. 231655714Skris 231755714Skris It was supposed to omit the parameters when they matched the signing key: 231855714Skris the verifying software was then supposed to automatically use the CA's 231955714Skris parameters if they were absent from the end user certificate. 232055714Skris 232155714Skris Omitting parameters is no longer recommended. The test was also 232255714Skris the wrong way round! This was probably due to unusual behaviour in 232355714Skris EVP_cmp_parameters() which returns 1 if the parameters match. 232455714Skris This meant that parameters were omitted when they *didn't* match and 232555714Skris the certificate was useless. Certificates signed with 'ca' didn't have 232655714Skris this bug. 232755714Skris [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 232855714Skris 232955714Skris *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 233055714Skris The interface is as follows: 233155714Skris Applications can use 233255714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 233355714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 233455714Skris "off" is now the default. 233555714Skris The library internally uses 233655714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 233755714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 233855714Skris to disable memory-checking temporarily. 233955714Skris 234055714Skris Some inconsistent states that previously were possible (and were 234155714Skris even the default) are now avoided. 234255714Skris 234355714Skris -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 234455714Skris with each memory chunk allocated; this is occasionally more helpful 234555714Skris than just having a counter. 234655714Skris 234755714Skris -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 234855714Skris 234955714Skris -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 235055714Skris extensions. 235155714Skris [Bodo Moeller] 235255714Skris 235355714Skris *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 235455714Skris which largely parallels "options", but is for changing API behaviour, 235555714Skris whereas "options" are about protocol behaviour. 235655714Skris Initial "mode" flags are: 235755714Skris 235855714Skris SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 235955714Skris a single record has been written. 236055714Skris SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 236155714Skris retries use the same buffer location. 236255714Skris (But all of the contents must be 236355714Skris copied!) 236455714Skris [Bodo Moeller] 236555714Skris 236655714Skris *) Bugfix: SSL_set_mode ignored its parameter, only SSL_CTX_set_mode 236755714Skris worked. 236855714Skris 236955714Skris *) Fix problems with no-hmac etc. 237055714Skris [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 237155714Skris 237255714Skris *) New functions RSA_get_default_method(), RSA_set_method() and 237355714Skris RSA_get_method(). These allows replacement of RSA_METHODs without having 237455714Skris to mess around with the internals of an RSA structure. 237555714Skris [Steve Henson] 237655714Skris 237755714Skris *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 237855714Skris Also really enable memory leak checks in openssl.c and in some 237955714Skris test programs. 238055714Skris [Chad C. Mulligan, Bodo Moeller] 238155714Skris 238255714Skris *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 238355714Skris up the length of negative integers. This has now been simplified to just 238455714Skris store the length when it is first determined and use it later, rather 238555714Skris than trying to keep track of where data is copied and updating it to 238655714Skris point to the end. 238755714Skris [Steve Henson, reported by Brien Wheeler 238855714Skris <bwheeler@authentica-security.com>] 238955714Skris 239055714Skris *) Add a new function PKCS7_signatureVerify. This allows the verification 239155714Skris of a PKCS#7 signature but with the signing certificate passed to the 239255714Skris function itself. This contrasts with PKCS7_dataVerify which assumes the 239355714Skris certificate is present in the PKCS#7 structure. This isn't always the 239455714Skris case: certificates can be omitted from a PKCS#7 structure and be 239555714Skris distributed by "out of band" means (such as a certificate database). 239655714Skris [Steve Henson] 239755714Skris 239855714Skris *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 239955714Skris function prototypes in pem.h, also change util/mkdef.pl to add the 240055714Skris necessary function names. 240155714Skris [Steve Henson] 240255714Skris 240355714Skris *) mk1mf.pl (used by Windows builds) did not properly read the 240455714Skris options set by Configure in the top level Makefile, and Configure 240555714Skris was not even able to write more than one option correctly. 240655714Skris Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 240755714Skris [Bodo Moeller] 240855714Skris 240955714Skris *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 241055714Skris file to be loaded from a BIO or FILE pointer. The BIO version will 241155714Skris for example allow memory BIOs to contain config info. 241255714Skris [Steve Henson] 241355714Skris 241455714Skris *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 241555714Skris Whoever hopes to achieve shared-library compatibility across versions 241655714Skris must use this, not the compile-time macro. 241755714Skris (Exercise 0.9.4: Which is the minimum library version required by 241855714Skris such programs?) 241955714Skris Note: All this applies only to multi-threaded programs, others don't 242055714Skris need locks. 242155714Skris [Bodo Moeller] 242255714Skris 242355714Skris *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 242455714Skris through a BIO pair triggered the default case, i.e. 242555714Skris SSLerr(...,SSL_R_UNKNOWN_STATE). 242655714Skris [Bodo Moeller] 242755714Skris 242855714Skris *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 242955714Skris can use the SSL library even if none of the specific BIOs is 243055714Skris appropriate. 243155714Skris [Bodo Moeller] 243255714Skris 243355714Skris *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 243455714Skris for the encoded length. 243555714Skris [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 243655714Skris 243755714Skris *) Add initial documentation of the X509V3 functions. 243855714Skris [Steve Henson] 243955714Skris 244055714Skris *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 244155714Skris PEM_write_bio_PKCS8PrivateKey() that are equivalent to 244255714Skris PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 244355714Skris secure PKCS#8 private key format with a high iteration count. 244455714Skris [Steve Henson] 244555714Skris 244655714Skris *) Fix determination of Perl interpreter: A perl or perl5 244755714Skris _directory_ in $PATH was also accepted as the interpreter. 244855714Skris [Ralf S. Engelschall] 244955714Skris 245055714Skris *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 245155714Skris wrong with it but it was very old and did things like calling 245255714Skris PEM_ASN1_read() directly and used MD5 for the hash not to mention some 245355714Skris unusual formatting. 245455714Skris [Steve Henson] 245555714Skris 245655714Skris *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 245755714Skris to use the new extension code. 245855714Skris [Steve Henson] 245955714Skris 246055714Skris *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 246155714Skris with macros. This should make it easier to change their form, add extra 246255714Skris arguments etc. Fix a few PEM prototypes which didn't have cipher as a 246355714Skris constant. 246455714Skris [Steve Henson] 246555714Skris 246655714Skris *) Add to configuration table a new entry that can specify an alternative 246755714Skris name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 246855714Skris according to Mark Crispin <MRC@Panda.COM>. 246955714Skris [Bodo Moeller] 247055714Skris 247155714Skris#if 0 247255714Skris *) DES CBC did not update the IV. Weird. 247355714Skris [Ben Laurie] 247455714Skris#else 247555714Skris des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 247655714Skris Changing the behaviour of the former might break existing programs -- 247755714Skris where IV updating is needed, des_ncbc_encrypt can be used. 247855714Skris#endif 247955714Skris 248055714Skris *) When bntest is run from "make test" it drives bc to check its 248155714Skris calculations, as well as internally checking them. If an internal check 248255714Skris fails, it needs to cause bc to give a non-zero result or make test carries 248355714Skris on without noticing the failure. Fixed. 248455714Skris [Ben Laurie] 248555714Skris 248655714Skris *) DES library cleanups. 248755714Skris [Ulf M�ller] 248855714Skris 248955714Skris *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 249055714Skris used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 249155714Skris ciphers. NOTE: although the key derivation function has been verified 249255714Skris against some published test vectors it has not been extensively tested 249355714Skris yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 249455714Skris of v2.0. 249555714Skris [Steve Henson] 249655714Skris 249755714Skris *) Instead of "mkdir -p", which is not fully portable, use new 249855714Skris Perl script "util/mkdir-p.pl". 249955714Skris [Bodo Moeller] 250055714Skris 250155714Skris *) Rewrite the way password based encryption (PBE) is handled. It used to 250255714Skris assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 250355714Skris structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 250455714Skris but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 250555714Skris the 'parameter' field of the AlgorithmIdentifier is passed to the 250655714Skris underlying key generation function so it must do its own ASN1 parsing. 250755714Skris This has also changed the EVP_PBE_CipherInit() function which now has a 250855714Skris 'parameter' argument instead of literal salt and iteration count values 250955714Skris and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 251055714Skris [Steve Henson] 251155714Skris 251255714Skris *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 251355714Skris and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 251455714Skris Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 251555714Skris KEY" because this clashed with PKCS#8 unencrypted string. Since this 251655714Skris value was just used as a "magic string" and not used directly its 251755714Skris value doesn't matter. 251855714Skris [Steve Henson] 251955714Skris 252055714Skris *) Introduce some semblance of const correctness to BN. Shame C doesn't 252155714Skris support mutable. 252255714Skris [Ben Laurie] 252355714Skris 252455714Skris *) "linux-sparc64" configuration (ultrapenguin). 252555714Skris [Ray Miller <ray.miller@oucs.ox.ac.uk>] 252655714Skris "linux-sparc" configuration. 252755714Skris [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 252855714Skris 252955714Skris *) config now generates no-xxx options for missing ciphers. 253055714Skris [Ulf M�ller] 253155714Skris 253255714Skris *) Support the EBCDIC character set (work in progress). 253355714Skris File ebcdic.c not yet included because it has a different license. 253455714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 253555714Skris 253655714Skris *) Support BS2000/OSD-POSIX. 253755714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 253855714Skris 253955714Skris *) Make callbacks for key generation use void * instead of char *. 254055714Skris [Ben Laurie] 254155714Skris 254255714Skris *) Make S/MIME samples compile (not yet tested). 254355714Skris [Ben Laurie] 254455714Skris 254555714Skris *) Additional typesafe stacks. 254655714Skris [Ben Laurie] 254755714Skris 254855714Skris *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 254955714Skris [Bodo Moeller] 255055714Skris 255155714Skris 255255714Skris Changes between 0.9.3 and 0.9.3a [29 May 1999] 255355714Skris 255455714Skris *) New configuration variant "sco5-gcc". 255555714Skris 255655714Skris *) Updated some demos. 255755714Skris [Sean O Riordain, Wade Scholine] 255855714Skris 255955714Skris *) Add missing BIO_free at exit of pkcs12 application. 256055714Skris [Wu Zhigang] 256155714Skris 256255714Skris *) Fix memory leak in conf.c. 256355714Skris [Steve Henson] 256455714Skris 256555714Skris *) Updates for Win32 to assembler version of MD5. 256655714Skris [Steve Henson] 256755714Skris 256855714Skris *) Set #! path to perl in apps/der_chop to where we found it 256955714Skris instead of using a fixed path. 257055714Skris [Bodo Moeller] 257155714Skris 257255714Skris *) SHA library changes for irix64-mips4-cc. 257355714Skris [Andy Polyakov] 257455714Skris 257555714Skris *) Improvements for VMS support. 257655714Skris [Richard Levitte] 257755714Skris 257855714Skris 257955714Skris Changes between 0.9.2b and 0.9.3 [24 May 1999] 258055714Skris 258155714Skris *) Bignum library bug fix. IRIX 6 passes "make test" now! 258255714Skris This also avoids the problems with SC4.2 and unpatched SC5. 258355714Skris [Andy Polyakov <appro@fy.chalmers.se>] 258455714Skris 258555714Skris *) New functions sk_num, sk_value and sk_set to replace the previous macros. 258655714Skris These are required because of the typesafe stack would otherwise break 258755714Skris existing code. If old code used a structure member which used to be STACK 258855714Skris and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 258955714Skris sk_num or sk_value it would produce an error because the num, data members 259055714Skris are not present in STACK_OF. Now it just produces a warning. sk_set 259155714Skris replaces the old method of assigning a value to sk_value 259255714Skris (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 259355714Skris that does this will no longer work (and should use sk_set instead) but 259455714Skris this could be regarded as a "questionable" behaviour anyway. 259555714Skris [Steve Henson] 259655714Skris 259755714Skris *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 259855714Skris correctly handle encrypted S/MIME data. 259955714Skris [Steve Henson] 260055714Skris 260155714Skris *) Change type of various DES function arguments from des_cblock 260255714Skris (which means, in function argument declarations, pointer to char) 260355714Skris to des_cblock * (meaning pointer to array with 8 char elements), 260455714Skris which allows the compiler to do more typechecking; it was like 260555714Skris that back in SSLeay, but with lots of ugly casts. 260655714Skris 260755714Skris Introduce new type const_des_cblock. 260855714Skris [Bodo Moeller] 260955714Skris 261055714Skris *) Reorganise the PKCS#7 library and get rid of some of the more obvious 261155714Skris problems: find RecipientInfo structure that matches recipient certificate 261255714Skris and initialise the ASN1 structures properly based on passed cipher. 261355714Skris [Steve Henson] 261455714Skris 261555714Skris *) Belatedly make the BN tests actually check the results. 261655714Skris [Ben Laurie] 261755714Skris 261855714Skris *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 261955714Skris to and from BNs: it was completely broken. New compilation option 262055714Skris NEG_PUBKEY_BUG to allow for some broken certificates that encode public 262155714Skris key elements as negative integers. 262255714Skris [Steve Henson] 262355714Skris 262455714Skris *) Reorganize and speed up MD5. 262555714Skris [Andy Polyakov <appro@fy.chalmers.se>] 262655714Skris 262755714Skris *) VMS support. 262855714Skris [Richard Levitte <richard@levitte.org>] 262955714Skris 263055714Skris *) New option -out to asn1parse to allow the parsed structure to be 263155714Skris output to a file. This is most useful when combined with the -strparse 263255714Skris option to examine the output of things like OCTET STRINGS. 263355714Skris [Steve Henson] 263455714Skris 263555714Skris *) Make SSL library a little more fool-proof by not requiring any longer 263655714Skris that SSL_set_{accept,connect}_state be called before 263755714Skris SSL_{accept,connect} may be used (SSL_set_..._state is omitted 263855714Skris in many applications because usually everything *appeared* to work as 263955714Skris intended anyway -- now it really works as intended). 264055714Skris [Bodo Moeller] 264155714Skris 264255714Skris *) Move openssl.cnf out of lib/. 264355714Skris [Ulf M�ller] 264455714Skris 264555714Skris *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 264655714Skris -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 264755714Skris -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 264855714Skris [Ralf S. Engelschall] 264955714Skris 265055714Skris *) Various fixes to the EVP and PKCS#7 code. It may now be able to 265155714Skris handle PKCS#7 enveloped data properly. 265255714Skris [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 265355714Skris 265455714Skris *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 265555714Skris copying pointers. The cert_st handling is changed by this in 265655714Skris various ways (and thus what used to be known as ctx->default_cert 265755714Skris is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 265855714Skris any longer when s->cert does not give us what we need). 265955714Skris ssl_cert_instantiate becomes obsolete by this change. 266055714Skris As soon as we've got the new code right (possibly it already is?), 266155714Skris we have solved a couple of bugs of the earlier code where s->cert 266255714Skris was used as if it could not have been shared with other SSL structures. 266355714Skris 266455714Skris Note that using the SSL API in certain dirty ways now will result 266555714Skris in different behaviour than observed with earlier library versions: 266655714Skris Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 266755714Skris does not influence s as it used to. 266855714Skris 266955714Skris In order to clean up things more thoroughly, inside SSL_SESSION 267055714Skris we don't use CERT any longer, but a new structure SESS_CERT 267155714Skris that holds per-session data (if available); currently, this is 267255714Skris the peer's certificate chain and, for clients, the server's certificate 267355714Skris and temporary key. CERT holds only those values that can have 267455714Skris meaningful defaults in an SSL_CTX. 267555714Skris [Bodo Moeller] 267655714Skris 267755714Skris *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 267855714Skris from the internal representation. Various PKCS#7 fixes: remove some 267955714Skris evil casts and set the enc_dig_alg field properly based on the signing 268055714Skris key type. 268155714Skris [Steve Henson] 268255714Skris 268355714Skris *) Allow PKCS#12 password to be set from the command line or the 268455714Skris environment. Let 'ca' get its config file name from the environment 268555714Skris variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 268655714Skris and 'x509'). 268755714Skris [Steve Henson] 268855714Skris 268955714Skris *) Allow certificate policies extension to use an IA5STRING for the 269055714Skris organization field. This is contrary to the PKIX definition but 269155714Skris VeriSign uses it and IE5 only recognises this form. Document 'x509' 269255714Skris extension option. 269355714Skris [Steve Henson] 269455714Skris 269555714Skris *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 269655714Skris without disallowing inline assembler and the like for non-pedantic builds. 269755714Skris [Ben Laurie] 269855714Skris 269955714Skris *) Support Borland C++ builder. 270055714Skris [Janez Jere <jj@void.si>, modified by Ulf M�ller] 270155714Skris 270255714Skris *) Support Mingw32. 270355714Skris [Ulf M�ller] 270455714Skris 270555714Skris *) SHA-1 cleanups and performance enhancements. 270655714Skris [Andy Polyakov <appro@fy.chalmers.se>] 270755714Skris 270855714Skris *) Sparc v8plus assembler for the bignum library. 270955714Skris [Andy Polyakov <appro@fy.chalmers.se>] 271055714Skris 271155714Skris *) Accept any -xxx and +xxx compiler options in Configure. 271255714Skris [Ulf M�ller] 271355714Skris 271455714Skris *) Update HPUX configuration. 271555714Skris [Anonymous] 271655714Skris 271755714Skris *) Add missing sk_<type>_unshift() function to safestack.h 271855714Skris [Ralf S. Engelschall] 271955714Skris 272055714Skris *) New function SSL_CTX_use_certificate_chain_file that sets the 272155714Skris "extra_cert"s in addition to the certificate. (This makes sense 272255714Skris only for "PEM" format files, as chains as a whole are not 272355714Skris DER-encoded.) 272455714Skris [Bodo Moeller] 272555714Skris 272655714Skris *) Support verify_depth from the SSL API. 272755714Skris x509_vfy.c had what can be considered an off-by-one-error: 272855714Skris Its depth (which was not part of the external interface) 272955714Skris was actually counting the number of certificates in a chain; 273055714Skris now it really counts the depth. 273155714Skris [Bodo Moeller] 273255714Skris 273355714Skris *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 273455714Skris instead of X509err, which often resulted in confusing error 273555714Skris messages since the error codes are not globally unique 273655714Skris (e.g. an alleged error in ssl3_accept when a certificate 273755714Skris didn't match the private key). 273855714Skris 273955714Skris *) New function SSL_CTX_set_session_id_context that allows to set a default 274055714Skris value (so that you don't need SSL_set_session_id_context for each 274155714Skris connection using the SSL_CTX). 274255714Skris [Bodo Moeller] 274355714Skris 274455714Skris *) OAEP decoding bug fix. 274555714Skris [Ulf M�ller] 274655714Skris 274755714Skris *) Support INSTALL_PREFIX for package builders, as proposed by 274855714Skris David Harris. 274955714Skris [Bodo Moeller] 275055714Skris 275155714Skris *) New Configure options "threads" and "no-threads". For systems 275255714Skris where the proper compiler options are known (currently Solaris 275355714Skris and Linux), "threads" is the default. 275455714Skris [Bodo Moeller] 275555714Skris 275655714Skris *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 275755714Skris [Bodo Moeller] 275855714Skris 275955714Skris *) Install various scripts to $(OPENSSLDIR)/misc, not to 276055714Skris $(INSTALLTOP)/bin -- they shouldn't clutter directories 276155714Skris such as /usr/local/bin. 276255714Skris [Bodo Moeller] 276355714Skris 276455714Skris *) "make linux-shared" to build shared libraries. 276555714Skris [Niels Poppe <niels@netbox.org>] 276655714Skris 276755714Skris *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 276855714Skris [Ulf M�ller] 276955714Skris 277055714Skris *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 277155714Skris extension adding in x509 utility. 277255714Skris [Steve Henson] 277355714Skris 277455714Skris *) Remove NOPROTO sections and error code comments. 277555714Skris [Ulf M�ller] 277655714Skris 277755714Skris *) Partial rewrite of the DEF file generator to now parse the ANSI 277855714Skris prototypes. 277955714Skris [Steve Henson] 278055714Skris 278155714Skris *) New Configure options --prefix=DIR and --openssldir=DIR. 278255714Skris [Ulf M�ller] 278355714Skris 278455714Skris *) Complete rewrite of the error code script(s). It is all now handled 278555714Skris by one script at the top level which handles error code gathering, 278655714Skris header rewriting and C source file generation. It should be much better 278755714Skris than the old method: it now uses a modified version of Ulf's parser to 278855714Skris read the ANSI prototypes in all header files (thus the old K&R definitions 278955714Skris aren't needed for error creation any more) and do a better job of 279055714Skris translating function codes into names. The old 'ASN1 error code imbedded 279155714Skris in a comment' is no longer necessary and it doesn't use .err files which 279255714Skris have now been deleted. Also the error code call doesn't have to appear all 279355714Skris on one line (which resulted in some large lines...). 279455714Skris [Steve Henson] 279555714Skris 279655714Skris *) Change #include filenames from <foo.h> to <openssl/foo.h>. 279755714Skris [Bodo Moeller] 279855714Skris 279955714Skris *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 280055714Skris 0 (which usually indicates a closed connection), but continue reading. 280155714Skris [Bodo Moeller] 280255714Skris 280355714Skris *) Fix some race conditions. 280455714Skris [Bodo Moeller] 280555714Skris 280655714Skris *) Add support for CRL distribution points extension. Add Certificate 280755714Skris Policies and CRL distribution points documentation. 280855714Skris [Steve Henson] 280955714Skris 281055714Skris *) Move the autogenerated header file parts to crypto/opensslconf.h. 281155714Skris [Ulf M�ller] 281255714Skris 281355714Skris *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 281455714Skris 8 of keying material. Merlin has also confirmed interop with this fix 281555714Skris between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 281655714Skris [Merlin Hughes <merlin@baltimore.ie>] 281755714Skris 281855714Skris *) Fix lots of warnings. 281955714Skris [Richard Levitte <levitte@stacken.kth.se>] 282055714Skris 282155714Skris *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 282255714Skris the directory spec didn't end with a LIST_SEPARATOR_CHAR. 282355714Skris [Richard Levitte <levitte@stacken.kth.se>] 282455714Skris 282555714Skris *) Fix problems with sizeof(long) == 8. 282655714Skris [Andy Polyakov <appro@fy.chalmers.se>] 282755714Skris 282855714Skris *) Change functions to ANSI C. 282955714Skris [Ulf M�ller] 283055714Skris 283155714Skris *) Fix typos in error codes. 283255714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 283355714Skris 283455714Skris *) Remove defunct assembler files from Configure. 283555714Skris [Ulf M�ller] 283655714Skris 283755714Skris *) SPARC v8 assembler BIGNUM implementation. 283855714Skris [Andy Polyakov <appro@fy.chalmers.se>] 283955714Skris 284055714Skris *) Support for Certificate Policies extension: both print and set. 284155714Skris Various additions to support the r2i method this uses. 284255714Skris [Steve Henson] 284355714Skris 284455714Skris *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 284555714Skris return a const string when you are expecting an allocated buffer. 284655714Skris [Ben Laurie] 284755714Skris 284855714Skris *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 284955714Skris types DirectoryString and DisplayText. 285055714Skris [Steve Henson] 285155714Skris 285255714Skris *) Add code to allow r2i extensions to access the configuration database, 285355714Skris add an LHASH database driver and add several ctx helper functions. 285455714Skris [Steve Henson] 285555714Skris 285655714Skris *) Fix an evil bug in bn_expand2() which caused various BN functions to 285755714Skris fail when they extended the size of a BIGNUM. 285855714Skris [Steve Henson] 285955714Skris 286055714Skris *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 286155714Skris support typesafe stack. 286255714Skris [Steve Henson] 286355714Skris 286455714Skris *) Fix typo in SSL_[gs]et_options(). 286555714Skris [Nils Frostberg <nils@medcom.se>] 286655714Skris 286755714Skris *) Delete various functions and files that belonged to the (now obsolete) 286855714Skris old X509V3 handling code. 286955714Skris [Steve Henson] 287055714Skris 287155714Skris *) New Configure option "rsaref". 287255714Skris [Ulf M�ller] 287355714Skris 287455714Skris *) Don't auto-generate pem.h. 287555714Skris [Bodo Moeller] 287655714Skris 287755714Skris *) Introduce type-safe ASN.1 SETs. 287855714Skris [Ben Laurie] 287955714Skris 288055714Skris *) Convert various additional casted stacks to type-safe STACK_OF() variants. 288155714Skris [Ben Laurie, Ralf S. Engelschall, Steve Henson] 288255714Skris 288355714Skris *) Introduce type-safe STACKs. This will almost certainly break lots of code 288455714Skris that links with OpenSSL (well at least cause lots of warnings), but fear 288555714Skris not: the conversion is trivial, and it eliminates loads of evil casts. A 288655714Skris few STACKed things have been converted already. Feel free to convert more. 288755714Skris In the fullness of time, I'll do away with the STACK type altogether. 288855714Skris [Ben Laurie] 288955714Skris 289055714Skris *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 289155714Skris specified in <certfile> by updating the entry in the index.txt file. 289255714Skris This way one no longer has to edit the index.txt file manually for 289355714Skris revoking a certificate. The -revoke option does the gory details now. 289455714Skris [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 289555714Skris 289655714Skris *) Fix `openssl crl -noout -text' combination where `-noout' killed the 289755714Skris `-text' option at all and this way the `-noout -text' combination was 289855714Skris inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 289955714Skris [Ralf S. Engelschall] 290055714Skris 290155714Skris *) Make sure a corresponding plain text error message exists for the 290255714Skris X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 290355714Skris verify callback function determined that a certificate was revoked. 290455714Skris [Ralf S. Engelschall] 290555714Skris 290655714Skris *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 290755714Skris ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 290855714Skris all available cipers including rc5, which was forgotten until now. 290955714Skris In order to let the testing shell script know which algorithms 291055714Skris are available, a new (up to now undocumented) command 291155714Skris "openssl list-cipher-commands" is used. 291255714Skris [Bodo Moeller] 291355714Skris 291455714Skris *) Bugfix: s_client occasionally would sleep in select() when 291555714Skris it should have checked SSL_pending() first. 291655714Skris [Bodo Moeller] 291755714Skris 291855714Skris *) New functions DSA_do_sign and DSA_do_verify to provide access to 291955714Skris the raw DSA values prior to ASN.1 encoding. 292055714Skris [Ulf M�ller] 292155714Skris 292255714Skris *) Tweaks to Configure 292355714Skris [Niels Poppe <niels@netbox.org>] 292455714Skris 292555714Skris *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 292655714Skris yet... 292755714Skris [Steve Henson] 292855714Skris 292955714Skris *) New variables $(RANLIB) and $(PERL) in the Makefiles. 293055714Skris [Ulf M�ller] 293155714Skris 293255714Skris *) New config option to avoid instructions that are illegal on the 80386. 293355714Skris The default code is faster, but requires at least a 486. 293455714Skris [Ulf M�ller] 293555714Skris 293655714Skris *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 293755714Skris SSL2_SERVER_VERSION (not used at all) macros, which are now the 293855714Skris same as SSL2_VERSION anyway. 293955714Skris [Bodo Moeller] 294055714Skris 294155714Skris *) New "-showcerts" option for s_client. 294255714Skris [Bodo Moeller] 294355714Skris 294455714Skris *) Still more PKCS#12 integration. Add pkcs12 application to openssl 294555714Skris application. Various cleanups and fixes. 294655714Skris [Steve Henson] 294755714Skris 294855714Skris *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 294955714Skris modify error routines to work internally. Add error codes and PBE init 295055714Skris to library startup routines. 295155714Skris [Steve Henson] 295255714Skris 295355714Skris *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 295455714Skris packing functions to asn1 and evp. Changed function names and error 295555714Skris codes along the way. 295655714Skris [Steve Henson] 295755714Skris 295855714Skris *) PKCS12 integration: and so it begins... First of several patches to 295955714Skris slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 296055714Skris objects to objects.h 296155714Skris [Steve Henson] 296255714Skris 296355714Skris *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 296455714Skris and display support for Thawte strong extranet extension. 296555714Skris [Steve Henson] 296655714Skris 296755714Skris *) Add LinuxPPC support. 296855714Skris [Jeff Dubrule <igor@pobox.org>] 296955714Skris 297055714Skris *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 297155714Skris bn_div_words in alpha.s. 297255714Skris [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 297355714Skris 297455714Skris *) Make sure the RSA OAEP test is skipped under -DRSAref because 297555714Skris OAEP isn't supported when OpenSSL is built with RSAref. 297655714Skris [Ulf Moeller <ulf@fitug.de>] 297755714Skris 297855714Skris *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 297955714Skris so they no longer are missing under -DNOPROTO. 298055714Skris [Soren S. Jorvang <soren@t.dk>] 298155714Skris 298255714Skris 298355714Skris Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 298455714Skris 298555714Skris *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 298655714Skris doesn't work when the session is reused. Coming soon! 298755714Skris [Ben Laurie] 298855714Skris 298955714Skris *) Fix a security hole, that allows sessions to be reused in the wrong 299055714Skris context thus bypassing client cert protection! All software that uses 299155714Skris client certs and session caches in multiple contexts NEEDS PATCHING to 299255714Skris allow session reuse! A fuller solution is in the works. 299355714Skris [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 299455714Skris 299555714Skris *) Some more source tree cleanups (removed obsolete files 299655714Skris crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 299755714Skris permission on "config" script to be executable) and a fix for the INSTALL 299855714Skris document. 299955714Skris [Ulf Moeller <ulf@fitug.de>] 300055714Skris 300155714Skris *) Remove some legacy and erroneous uses of malloc, free instead of 300255714Skris Malloc, Free. 300355714Skris [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 300455714Skris 300555714Skris *) Make rsa_oaep_test return non-zero on error. 300655714Skris [Ulf Moeller <ulf@fitug.de>] 300755714Skris 300855714Skris *) Add support for native Solaris shared libraries. Configure 300955714Skris solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 301055714Skris if someone would make that last step automatic. 301155714Skris [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 301255714Skris 301355714Skris *) ctx_size was not built with the right compiler during "make links". Fixed. 301455714Skris [Ben Laurie] 301555714Skris 301655714Skris *) Change the meaning of 'ALL' in the cipher list. It now means "everything 301755714Skris except NULL ciphers". This means the default cipher list will no longer 301855714Skris enable NULL ciphers. They need to be specifically enabled e.g. with 301955714Skris the string "DEFAULT:eNULL". 302055714Skris [Steve Henson] 302155714Skris 302255714Skris *) Fix to RSA private encryption routines: if p < q then it would 302355714Skris occasionally produce an invalid result. This will only happen with 302455714Skris externally generated keys because OpenSSL (and SSLeay) ensure p > q. 302555714Skris [Steve Henson] 302655714Skris 302755714Skris *) Be less restrictive and allow also `perl util/perlpath.pl 302855714Skris /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 302955714Skris because this way one can also use an interpreter named `perl5' (which is 303055714Skris usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 303155714Skris installed as `perl'). 303255714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 303355714Skris 303455714Skris *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 303555714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 303655714Skris 303755714Skris *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 303855714Skris advapi32.lib to Win32 build and change the pem test comparision 303955714Skris to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 304055714Skris suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 304155714Skris and crypto/des/ede_cbcm_enc.c. 304255714Skris [Steve Henson] 304355714Skris 304455714Skris *) DES quad checksum was broken on big-endian architectures. Fixed. 304555714Skris [Ben Laurie] 304655714Skris 304755714Skris *) Comment out two functions in bio.h that aren't implemented. Fix up the 304855714Skris Win32 test batch file so it (might) work again. The Win32 test batch file 304955714Skris is horrible: I feel ill.... 305055714Skris [Steve Henson] 305155714Skris 305255714Skris *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 305355714Skris in e_os.h. Audit of header files to check ANSI and non ANSI 305455714Skris sections: 10 functions were absent from non ANSI section and not exported 305555714Skris from Windows DLLs. Fixed up libeay.num for new functions. 305655714Skris [Steve Henson] 305755714Skris 305855714Skris *) Make `openssl version' output lines consistent. 305955714Skris [Ralf S. Engelschall] 306055714Skris 306155714Skris *) Fix Win32 symbol export lists for BIO functions: Added 306255714Skris BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 306355714Skris to ms/libeay{16,32}.def. 306455714Skris [Ralf S. Engelschall] 306555714Skris 306655714Skris *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 306755714Skris fine under Unix and passes some trivial tests I've now added. But the 306855714Skris whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 306955714Skris added to make sure no one expects that this stuff really works in the 307055714Skris OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 307155714Skris up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 307255714Skris openssl_bio.xs. 307355714Skris [Ralf S. Engelschall] 307455714Skris 307555714Skris *) Fix the generation of two part addresses in perl. 307655714Skris [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 307755714Skris 307855714Skris *) Add config entry for Linux on MIPS. 307955714Skris [John Tobey <jtobey@channel1.com>] 308055714Skris 308155714Skris *) Make links whenever Configure is run, unless we are on Windoze. 308255714Skris [Ben Laurie] 308355714Skris 308455714Skris *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 308555714Skris Currently only issuerAltName and AuthorityKeyIdentifier make any sense 308655714Skris in CRLs. 308755714Skris [Steve Henson] 308855714Skris 308955714Skris *) Add a useful kludge to allow package maintainers to specify compiler and 309055714Skris other platforms details on the command line without having to patch the 309155714Skris Configure script everytime: One now can use ``perl Configure 309255714Skris <id>:<details>'', i.e. platform ids are allowed to have details appended 309355714Skris to them (seperated by colons). This is treated as there would be a static 309455714Skris pre-configured entry in Configure's %table under key <id> with value 309555714Skris <details> and ``perl Configure <id>'' is called. So, when you want to 309655714Skris perform a quick test-compile under FreeBSD 3.1 with pgcc and without 309755714Skris assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 309855714Skris now, which overrides the FreeBSD-elf entry on-the-fly. 309955714Skris [Ralf S. Engelschall] 310055714Skris 310155714Skris *) Disable new TLS1 ciphersuites by default: they aren't official yet. 310255714Skris [Ben Laurie] 310355714Skris 310455714Skris *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 310555714Skris on the `perl Configure ...' command line. This way one can compile 310655714Skris OpenSSL libraries with Position Independent Code (PIC) which is needed 310755714Skris for linking it into DSOs. 310855714Skris [Ralf S. Engelschall] 310955714Skris 311055714Skris *) Remarkably, export ciphers were totally broken and no-one had noticed! 311155714Skris Fixed. 311255714Skris [Ben Laurie] 311355714Skris 311455714Skris *) Cleaned up the LICENSE document: The official contact for any license 311555714Skris questions now is the OpenSSL core team under openssl-core@openssl.org. 311655714Skris And add a paragraph about the dual-license situation to make sure people 311755714Skris recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 311855714Skris to the OpenSSL toolkit. 311955714Skris [Ralf S. Engelschall] 312055714Skris 312155714Skris *) General source tree makefile cleanups: Made `making xxx in yyy...' 312255714Skris display consistent in the source tree and replaced `/bin/rm' by `rm'. 312355714Skris Additonally cleaned up the `make links' target: Remove unnecessary 312455714Skris semicolons, subsequent redundant removes, inline point.sh into mklink.sh 312555714Skris to speed processing and no longer clutter the display with confusing 312655714Skris stuff. Instead only the actually done links are displayed. 312755714Skris [Ralf S. Engelschall] 312855714Skris 312955714Skris *) Permit null encryption ciphersuites, used for authentication only. It used 313055714Skris to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 313155714Skris It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 313255714Skris encryption. 313355714Skris [Ben Laurie] 313455714Skris 313555714Skris *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 313655714Skris signed attributes when verifying signatures (this would break them), 313755714Skris the detached data encoding was wrong and public keys obtained using 313855714Skris X509_get_pubkey() weren't freed. 313955714Skris [Steve Henson] 314055714Skris 314155714Skris *) Add text documentation for the BUFFER functions. Also added a work around 314255714Skris to a Win95 console bug. This was triggered by the password read stuff: the 314355714Skris last character typed gets carried over to the next fread(). If you were 314455714Skris generating a new cert request using 'req' for example then the last 314555714Skris character of the passphrase would be CR which would then enter the first 314655714Skris field as blank. 314755714Skris [Steve Henson] 314855714Skris 314955714Skris *) Added the new `Includes OpenSSL Cryptography Software' button as 315055714Skris doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 315155714Skris button and can be used by applications based on OpenSSL to show the 315255714Skris relationship to the OpenSSL project. 315355714Skris [Ralf S. Engelschall] 315455714Skris 315555714Skris *) Remove confusing variables in function signatures in files 315655714Skris ssl/ssl_lib.c and ssl/ssl.h. 315755714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 315855714Skris 315955714Skris *) Don't install bss_file.c under PREFIX/include/ 316055714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 316155714Skris 316255714Skris *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 316355714Skris functions that return function pointers and has support for NT specific 316455714Skris stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 316555714Skris #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 316655714Skris unsigned to signed types: this was killing the Win32 compile. 316755714Skris [Steve Henson] 316855714Skris 316955714Skris *) Add new certificate file to stack functions, 317055714Skris SSL_add_dir_cert_subjects_to_stack() and 317155714Skris SSL_add_file_cert_subjects_to_stack(). These largely supplant 317255714Skris SSL_load_client_CA_file(), and can be used to add multiple certs easily 317355714Skris to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 317455714Skris This means that Apache-SSL and similar packages don't have to mess around 317555714Skris to add as many CAs as they want to the preferred list. 317655714Skris [Ben Laurie] 317755714Skris 317855714Skris *) Experiment with doxygen documentation. Currently only partially applied to 317955714Skris ssl/ssl_lib.c. 318055714Skris See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 318155714Skris openssl.doxy as the configuration file. 318255714Skris [Ben Laurie] 318355714Skris 318455714Skris *) Get rid of remaining C++-style comments which strict C compilers hate. 318555714Skris [Ralf S. Engelschall, pointed out by Carlos Amengual] 318655714Skris 318755714Skris *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 318855714Skris compiled in by default: it has problems with large keys. 318955714Skris [Steve Henson] 319055714Skris 319155714Skris *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 319255714Skris DH private keys and/or callback functions which directly correspond to 319355714Skris their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 319455714Skris is needed for applications which have to configure certificates on a 319555714Skris per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 319655714Skris (e.g. s_server). 319755714Skris For the RSA certificate situation is makes no difference, but 319855714Skris for the DSA certificate situation this fixes the "no shared cipher" 319955714Skris problem where the OpenSSL cipher selection procedure failed because the 320055714Skris temporary keys were not overtaken from the context and the API provided 320155714Skris no way to reconfigure them. 320255714Skris The new functions now let applications reconfigure the stuff and they 320355714Skris are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 320455714Skris SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 320555714Skris non-public-API function ssl_cert_instantiate() is used as a helper 320655714Skris function and also to reduce code redundancy inside ssl_rsa.c. 320755714Skris [Ralf S. Engelschall] 320855714Skris 320955714Skris *) Move s_server -dcert and -dkey options out of the undocumented feature 321055714Skris area because they are useful for the DSA situation and should be 321155714Skris recognized by the users. 321255714Skris [Ralf S. Engelschall] 321355714Skris 321455714Skris *) Fix the cipher decision scheme for export ciphers: the export bits are 321555714Skris *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 321655714Skris SSL_EXP_MASK. So, the original variable has to be used instead of the 321755714Skris already masked variable. 321855714Skris [Richard Levitte <levitte@stacken.kth.se>] 321955714Skris 322055714Skris *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 322155714Skris [Richard Levitte <levitte@stacken.kth.se>] 322255714Skris 322355714Skris *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 322455714Skris from `int' to `unsigned int' because it's a length and initialized by 322555714Skris EVP_DigestFinal() which expects an `unsigned int *'. 322655714Skris [Richard Levitte <levitte@stacken.kth.se>] 322755714Skris 322855714Skris *) Don't hard-code path to Perl interpreter on shebang line of Configure 322955714Skris script. Instead use the usual Shell->Perl transition trick. 323055714Skris [Ralf S. Engelschall] 323155714Skris 323255714Skris *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 323355714Skris (in addition to RSA certificates) to match the behaviour of `openssl dsa 323455714Skris -noout -modulus' as it's already the case for `openssl rsa -noout 323555714Skris -modulus'. For RSA the -modulus is the real "modulus" while for DSA 323655714Skris currently the public key is printed (a decision which was already done by 323755714Skris `openssl dsa -modulus' in the past) which serves a similar purpose. 323855714Skris Additionally the NO_RSA no longer completely removes the whole -modulus 323955714Skris option; it now only avoids using the RSA stuff. Same applies to NO_DSA 324055714Skris now, too. 324155714Skris [Ralf S. Engelschall] 324255714Skris 324355714Skris *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 324455714Skris BIO. See the source (crypto/evp/bio_ok.c) for more info. 324555714Skris [Arne Ansper <arne@ats.cyber.ee>] 324655714Skris 324755714Skris *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 324855714Skris to be added. Now both 'req' and 'ca' can use new objects defined in the 324955714Skris config file. 325055714Skris [Steve Henson] 325155714Skris 325255714Skris *) Add cool BIO that does syslog (or event log on NT). 325355714Skris [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 325455714Skris 325555714Skris *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 325655714Skris TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 325755714Skris TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 325855714Skris Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 325955714Skris [Ben Laurie] 326055714Skris 326155714Skris *) Add preliminary config info for new extension code. 326255714Skris [Steve Henson] 326355714Skris 326455714Skris *) Make RSA_NO_PADDING really use no padding. 326555714Skris [Ulf Moeller <ulf@fitug.de>] 326655714Skris 326755714Skris *) Generate errors when private/public key check is done. 326855714Skris [Ben Laurie] 326955714Skris 327055714Skris *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 327155714Skris for some CRL extensions and new objects added. 327255714Skris [Steve Henson] 327355714Skris 327455714Skris *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 327555714Skris key usage extension and fuller support for authority key id. 327655714Skris [Steve Henson] 327755714Skris 327855714Skris *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 327955714Skris padding method for RSA, which is recommended for new applications in PKCS 328055714Skris #1 v2.0 (RFC 2437, October 1998). 328155714Skris OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 328255714Skris foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 328355714Skris against Bleichbacher's attack on RSA. 328455714Skris [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 328555714Skris Ben Laurie] 328655714Skris 328755714Skris *) Updates to the new SSL compression code 328855714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 328955714Skris 329055714Skris *) Fix so that the version number in the master secret, when passed 329155714Skris via RSA, checks that if TLS was proposed, but we roll back to SSLv3 329255714Skris (because the server will not accept higher), that the version number 329355714Skris is 0x03,0x01, not 0x03,0x00 329455714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 329555714Skris 329655714Skris *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 329755714Skris leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 329855714Skris in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 329955714Skris [Steve Henson] 330055714Skris 330155714Skris *) Support for RAW extensions where an arbitrary extension can be 330255714Skris created by including its DER encoding. See apps/openssl.cnf for 330355714Skris an example. 330455714Skris [Steve Henson] 330555714Skris 330655714Skris *) Make sure latest Perl versions don't interpret some generated C array 330755714Skris code as Perl array code in the crypto/err/err_genc.pl script. 330855714Skris [Lars Weber <3weber@informatik.uni-hamburg.de>] 330955714Skris 331055714Skris *) Modify ms/do_ms.bat to not generate assembly language makefiles since 331155714Skris not many people have the assembler. Various Win32 compilation fixes and 331255714Skris update to the INSTALL.W32 file with (hopefully) more accurate Win32 331355714Skris build instructions. 331455714Skris [Steve Henson] 331555714Skris 331655714Skris *) Modify configure script 'Configure' to automatically create crypto/date.h 331755714Skris file under Win32 and also build pem.h from pem.org. New script 331855714Skris util/mkfiles.pl to create the MINFO file on environments that can't do a 331955714Skris 'make files': perl util/mkfiles.pl >MINFO should work. 332055714Skris [Steve Henson] 332155714Skris 332255714Skris *) Major rework of DES function declarations, in the pursuit of correctness 332355714Skris and purity. As a result, many evil casts evaporated, and some weirdness, 332455714Skris too. You may find this causes warnings in your code. Zapping your evil 332555714Skris casts will probably fix them. Mostly. 332655714Skris [Ben Laurie] 332755714Skris 332855714Skris *) Fix for a typo in asn1.h. Bug fix to object creation script 332955714Skris obj_dat.pl. It considered a zero in an object definition to mean 333055714Skris "end of object": none of the objects in objects.h have any zeros 333155714Skris so it wasn't spotted. 333255714Skris [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 333355714Skris 333455714Skris *) Add support for Triple DES Cipher Block Chaining with Output Feedback 333555714Skris Masking (CBCM). In the absence of test vectors, the best I have been able 333655714Skris to do is check that the decrypt undoes the encrypt, so far. Send me test 333755714Skris vectors if you have them. 333855714Skris [Ben Laurie] 333955714Skris 334055714Skris *) Correct calculation of key length for export ciphers (too much space was 334155714Skris allocated for null ciphers). This has not been tested! 334255714Skris [Ben Laurie] 334355714Skris 334455714Skris *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 334555714Skris message is now correct (it understands "crypto" and "ssl" on its 334655714Skris command line). There is also now an "update" option. This will update 334755714Skris the util/ssleay.num and util/libeay.num files with any new functions. 334855714Skris If you do a: 334955714Skris perl util/mkdef.pl crypto ssl update 335055714Skris it will update them. 335155714Skris [Steve Henson] 335255714Skris 335355714Skris *) Overhauled the Perl interface (perl/*): 335455714Skris - ported BN stuff to OpenSSL's different BN library 335555714Skris - made the perl/ source tree CVS-aware 335655714Skris - renamed the package from SSLeay to OpenSSL (the files still contain 335755714Skris their history because I've copied them in the repository) 335855714Skris - removed obsolete files (the test scripts will be replaced 335955714Skris by better Test::Harness variants in the future) 336055714Skris [Ralf S. Engelschall] 336155714Skris 336255714Skris *) First cut for a very conservative source tree cleanup: 336355714Skris 1. merge various obsolete readme texts into doc/ssleay.txt 336455714Skris where we collect the old documents and readme texts. 336555714Skris 2. remove the first part of files where I'm already sure that we no 336655714Skris longer need them because of three reasons: either they are just temporary 336755714Skris files which were left by Eric or they are preserved original files where 336855714Skris I've verified that the diff is also available in the CVS via "cvs diff 336955714Skris -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 337055714Skris the crypto/md/ stuff). 337155714Skris [Ralf S. Engelschall] 337255714Skris 337355714Skris *) More extension code. Incomplete support for subject and issuer alt 337455714Skris name, issuer and authority key id. Change the i2v function parameters 337555714Skris and add an extra 'crl' parameter in the X509V3_CTX structure: guess 337655714Skris what that's for :-) Fix to ASN1 macro which messed up 337755714Skris IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 337855714Skris [Steve Henson] 337955714Skris 338055714Skris *) Preliminary support for ENUMERATED type. This is largely copied from the 338155714Skris INTEGER code. 338255714Skris [Steve Henson] 338355714Skris 338455714Skris *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 338555714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 338655714Skris 338755714Skris *) Make sure `make rehash' target really finds the `openssl' program. 338855714Skris [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 338955714Skris 339055714Skris *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 339155714Skris like to hear about it if this slows down other processors. 339255714Skris [Ben Laurie] 339355714Skris 339455714Skris *) Add CygWin32 platform information to Configure script. 339555714Skris [Alan Batie <batie@aahz.jf.intel.com>] 339655714Skris 339755714Skris *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 339855714Skris [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 339955714Skris 340055714Skris *) New program nseq to manipulate netscape certificate sequences 340155714Skris [Steve Henson] 340255714Skris 340355714Skris *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 340455714Skris few typos. 340555714Skris [Steve Henson] 340655714Skris 340755714Skris *) Fixes to BN code. Previously the default was to define BN_RECURSION 340855714Skris but the BN code had some problems that would cause failures when 340955714Skris doing certificate verification and some other functions. 341055714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 341155714Skris 341255714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 341355714Skris [Steve Henson] 341455714Skris 341555714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 341655714Skris [Steve Henson] 341755714Skris 341855714Skris *) Add several PKIX and private extended key usage OIDs. 341955714Skris [Steve Henson] 342055714Skris 342155714Skris *) Modify the 'ca' program to handle the new extension code. Modify 342255714Skris openssl.cnf for new extension format, add comments. 342355714Skris [Steve Henson] 342455714Skris 342555714Skris *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 342655714Skris and add a sample to openssl.cnf so req -x509 now adds appropriate 342755714Skris CA extensions. 342855714Skris [Steve Henson] 342955714Skris 343055714Skris *) Continued X509 V3 changes. Add to other makefiles, integrate with the 343155714Skris error code, add initial support to X509_print() and x509 application. 343255714Skris [Steve Henson] 343355714Skris 343455714Skris *) Takes a deep breath and start addding X509 V3 extension support code. Add 343555714Skris files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 343655714Skris stuff is currently isolated and isn't even compiled yet. 343755714Skris [Steve Henson] 343855714Skris 343955714Skris *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 344055714Skris ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 344155714Skris Removed the versions check from X509 routines when loading extensions: 344255714Skris this allows certain broken certificates that don't set the version 344355714Skris properly to be processed. 344455714Skris [Steve Henson] 344555714Skris 344655714Skris *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 344755714Skris Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 344855714Skris can still be regenerated with "make depend". 344955714Skris [Ben Laurie] 345055714Skris 345155714Skris *) Spelling mistake in C version of CAST-128. 345255714Skris [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 345355714Skris 345455714Skris *) Changes to the error generation code. The perl script err-code.pl 345555714Skris now reads in the old error codes and retains the old numbers, only 345655714Skris adding new ones if necessary. It also only changes the .err files if new 345755714Skris codes are added. The makefiles have been modified to only insert errors 345855714Skris when needed (to avoid needlessly modifying header files). This is done 345955714Skris by only inserting errors if the .err file is newer than the auto generated 346055714Skris C file. To rebuild all the error codes from scratch (the old behaviour) 346155714Skris either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 346255714Skris or delete all the .err files. 346355714Skris [Steve Henson] 346455714Skris 346555714Skris *) CAST-128 was incorrectly implemented for short keys. The C version has 346655714Skris been fixed, but is untested. The assembler versions are also fixed, but 346755714Skris new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 346855714Skris to regenerate it if needed. 346955714Skris [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 347055714Skris Hagino <itojun@kame.net>] 347155714Skris 347255714Skris *) File was opened incorrectly in randfile.c. 347355714Skris [Ulf M�ller <ulf@fitug.de>] 347455714Skris 347555714Skris *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 347655714Skris functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 347755714Skris GeneralizedTime. ASN1_TIME is the proper type used in certificates et 347855714Skris al: it's just almost always a UTCTime. Note this patch adds new error 347955714Skris codes so do a "make errors" if there are problems. 348055714Skris [Steve Henson] 348155714Skris 348255714Skris *) Correct Linux 1 recognition in config. 348355714Skris [Ulf M�ller <ulf@fitug.de>] 348455714Skris 348555714Skris *) Remove pointless MD5 hash when using DSA keys in ca. 348655714Skris [Anonymous <nobody@replay.com>] 348755714Skris 348855714Skris *) Generate an error if given an empty string as a cert directory. Also 348955714Skris generate an error if handed NULL (previously returned 0 to indicate an 349055714Skris error, but didn't set one). 349155714Skris [Ben Laurie, reported by Anonymous <nobody@replay.com>] 349255714Skris 349355714Skris *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 349455714Skris [Ben Laurie] 349555714Skris 349655714Skris *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 349755714Skris parameters. This was causing a warning which killed off the Win32 compile. 349855714Skris [Steve Henson] 349955714Skris 350055714Skris *) Remove C++ style comments from crypto/bn/bn_local.h. 350155714Skris [Neil Costigan <neil.costigan@celocom.com>] 350255714Skris 350355714Skris *) The function OBJ_txt2nid was broken. It was supposed to return a nid 350455714Skris based on a text string, looking up short and long names and finally 350555714Skris "dot" format. The "dot" format stuff didn't work. Added new function 350655714Skris OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 350755714Skris OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 350855714Skris OID is not part of the table. 350955714Skris [Steve Henson] 351055714Skris 351155714Skris *) Add prototypes to X509 lookup/verify methods, fixing a bug in 351255714Skris X509_LOOKUP_by_alias(). 351355714Skris [Ben Laurie] 351455714Skris 351555714Skris *) Sort openssl functions by name. 351655714Skris [Ben Laurie] 351755714Skris 351855714Skris *) Get the gendsa program working (hopefully) and add it to app list. Remove 351955714Skris encryption from sample DSA keys (in case anyone is interested the password 352055714Skris was "1234"). 352155714Skris [Steve Henson] 352255714Skris 352355714Skris *) Make _all_ *_free functions accept a NULL pointer. 352455714Skris [Frans Heymans <fheymans@isaserver.be>] 352555714Skris 352655714Skris *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 352755714Skris NULL pointers. 352855714Skris [Anonymous <nobody@replay.com>] 352955714Skris 353055714Skris *) s_server should send the CAfile as acceptable CAs, not its own cert. 353155714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 353255714Skris 353355714Skris *) Don't blow it for numeric -newkey arguments to apps/req. 353455714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 353555714Skris 353655714Skris *) Temp key "for export" tests were wrong in s3_srvr.c. 353755714Skris [Anonymous <nobody@replay.com>] 353855714Skris 353955714Skris *) Add prototype for temp key callback functions 354055714Skris SSL_CTX_set_tmp_{rsa,dh}_callback(). 354155714Skris [Ben Laurie] 354255714Skris 354355714Skris *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 354455714Skris DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 354555714Skris [Steve Henson] 354655714Skris 354755714Skris *) X509_name_add_entry() freed the wrong thing after an error. 354855714Skris [Arne Ansper <arne@ats.cyber.ee>] 354955714Skris 355055714Skris *) rsa_eay.c would attempt to free a NULL context. 355155714Skris [Arne Ansper <arne@ats.cyber.ee>] 355255714Skris 355355714Skris *) BIO_s_socket() had a broken should_retry() on Windoze. 355455714Skris [Arne Ansper <arne@ats.cyber.ee>] 355555714Skris 355655714Skris *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 355755714Skris [Arne Ansper <arne@ats.cyber.ee>] 355855714Skris 355955714Skris *) Make sure the already existing X509_STORE->depth variable is initialized 356055714Skris in X509_STORE_new(), but document the fact that this variable is still 356155714Skris unused in the certificate verification process. 356255714Skris [Ralf S. Engelschall] 356355714Skris 356455714Skris *) Fix the various library and apps files to free up pkeys obtained from 356555714Skris X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 356655714Skris [Steve Henson] 356755714Skris 356855714Skris *) Fix reference counting in X509_PUBKEY_get(). This makes 356955714Skris demos/maurice/example2.c work, amongst others, probably. 357055714Skris [Steve Henson and Ben Laurie] 357155714Skris 357255714Skris *) First cut of a cleanup for apps/. First the `ssleay' program is now named 357355714Skris `openssl' and second, the shortcut symlinks for the `openssl <command>' 357455714Skris are no longer created. This way we have a single and consistent command 357555714Skris line interface `openssl <command>', similar to `cvs <command>'. 357655714Skris [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 357755714Skris 357855714Skris *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 357955714Skris BIT STRING wrapper always have zero unused bits. 358055714Skris [Steve Henson] 358155714Skris 358255714Skris *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 358355714Skris [Steve Henson] 358455714Skris 358555714Skris *) Make the top-level INSTALL documentation easier to understand. 358655714Skris [Paul Sutton] 358755714Skris 358855714Skris *) Makefiles updated to exit if an error occurs in a sub-directory 358955714Skris make (including if user presses ^C) [Paul Sutton] 359055714Skris 359155714Skris *) Make Montgomery context stuff explicit in RSA data structure. 359255714Skris [Ben Laurie] 359355714Skris 359455714Skris *) Fix build order of pem and err to allow for generated pem.h. 359555714Skris [Ben Laurie] 359655714Skris 359755714Skris *) Fix renumbering bug in X509_NAME_delete_entry(). 359855714Skris [Ben Laurie] 359955714Skris 360055714Skris *) Enhanced the err-ins.pl script so it makes the error library number 360155714Skris global and can add a library name. This is needed for external ASN1 and 360255714Skris other error libraries. 360355714Skris [Steve Henson] 360455714Skris 360555714Skris *) Fixed sk_insert which never worked properly. 360655714Skris [Steve Henson] 360755714Skris 360855714Skris *) Fix ASN1 macros so they can handle indefinite length construted 360955714Skris EXPLICIT tags. Some non standard certificates use these: they can now 361055714Skris be read in. 361155714Skris [Steve Henson] 361255714Skris 361355714Skris *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 361455714Skris into a single doc/ssleay.txt bundle. This way the information is still 361555714Skris preserved but no longer messes up this directory. Now it's new room for 361655714Skris the new set of documenation files. 361755714Skris [Ralf S. Engelschall] 361855714Skris 361955714Skris *) SETs were incorrectly DER encoded. This was a major pain, because they 362055714Skris shared code with SEQUENCEs, which aren't coded the same. This means that 362155714Skris almost everything to do with SETs or SEQUENCEs has either changed name or 362255714Skris number of arguments. 362355714Skris [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 362455714Skris 362555714Skris *) Fix test data to work with the above. 362655714Skris [Ben Laurie] 362755714Skris 362855714Skris *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 362955714Skris was already fixed by Eric for 0.9.1 it seems. 363055714Skris [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 363155714Skris 363255714Skris *) Autodetect FreeBSD3. 363355714Skris [Ben Laurie] 363455714Skris 363555714Skris *) Fix various bugs in Configure. This affects the following platforms: 363655714Skris nextstep 363755714Skris ncr-scde 363855714Skris unixware-2.0 363955714Skris unixware-2.0-pentium 364055714Skris sco5-cc. 364155714Skris [Ben Laurie] 364255714Skris 364355714Skris *) Eliminate generated files from CVS. Reorder tests to regenerate files 364455714Skris before they are needed. 364555714Skris [Ben Laurie] 364655714Skris 364755714Skris *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 364855714Skris [Ben Laurie] 364955714Skris 365055714Skris 365155714Skris Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 365255714Skris 365355714Skris *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 365455714Skris changed SSLeay to OpenSSL in version strings. 365555714Skris [Ralf S. Engelschall] 365655714Skris 365755714Skris *) Some fixups to the top-level documents. 365855714Skris [Paul Sutton] 365955714Skris 366055714Skris *) Fixed the nasty bug where rsaref.h was not found under compile-time 366155714Skris because the symlink to include/ was missing. 366255714Skris [Ralf S. Engelschall] 366355714Skris 366455714Skris *) Incorporated the popular no-RSA/DSA-only patches 366555714Skris which allow to compile a RSA-free SSLeay. 366655714Skris [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 366755714Skris 366855714Skris *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 366955714Skris when "ssleay" is still not found. 367055714Skris [Ralf S. Engelschall] 367155714Skris 367255714Skris *) Added more platforms to Configure: Cray T3E, HPUX 11, 367355714Skris [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 367455714Skris 367555714Skris *) Updated the README file. 367655714Skris [Ralf S. Engelschall] 367755714Skris 367855714Skris *) Added various .cvsignore files in the CVS repository subdirs 367955714Skris to make a "cvs update" really silent. 368055714Skris [Ralf S. Engelschall] 368155714Skris 368255714Skris *) Recompiled the error-definition header files and added 368355714Skris missing symbols to the Win32 linker tables. 368455714Skris [Ralf S. Engelschall] 368555714Skris 368655714Skris *) Cleaned up the top-level documents; 368755714Skris o new files: CHANGES and LICENSE 368855714Skris o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 368955714Skris o merged COPYRIGHT into LICENSE 369055714Skris o removed obsolete TODO file 369155714Skris o renamed MICROSOFT to INSTALL.W32 369255714Skris [Ralf S. Engelschall] 369355714Skris 369455714Skris *) Removed dummy files from the 0.9.1b source tree: 369555714Skris crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 369655714Skris crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 369755714Skris crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 369855714Skris crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 369955714Skris util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 370055714Skris [Ralf S. Engelschall] 370155714Skris 370255714Skris *) Added various platform portability fixes. 370355714Skris [Mark J. Cox] 370455714Skris 370555714Skris *) The Genesis of the OpenSSL rpject: 370655714Skris We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 370755714Skris Young and Tim J. Hudson created while they were working for C2Net until 370855714Skris summer 1998. 370955714Skris [The OpenSSL Project] 371055714Skris 371155714Skris 371255714Skris Changes between 0.9.0b and 0.9.1b [not released] 371355714Skris 371455714Skris *) Updated a few CA certificates under certs/ 371555714Skris [Eric A. Young] 371655714Skris 371755714Skris *) Changed some BIGNUM api stuff. 371855714Skris [Eric A. Young] 371955714Skris 372055714Skris *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 372155714Skris DGUX x86, Linux Alpha, etc. 372255714Skris [Eric A. Young] 372355714Skris 372455714Skris *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 372555714Skris RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 372655714Skris available). 372755714Skris [Eric A. Young] 372855714Skris 372955714Skris *) Add -strparse option to asn1pars program which parses nested 373055714Skris binary structures 373155714Skris [Dr Stephen Henson <shenson@bigfoot.com>] 373255714Skris 373355714Skris *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 373455714Skris [Eric A. Young] 373555714Skris 373655714Skris *) DSA fix for "ca" program. 373755714Skris [Eric A. Young] 373855714Skris 373955714Skris *) Added "-genkey" option to "dsaparam" program. 374055714Skris [Eric A. Young] 374155714Skris 374255714Skris *) Added RIPE MD160 (rmd160) message digest. 374355714Skris [Eric A. Young] 374455714Skris 374555714Skris *) Added -a (all) option to "ssleay version" command. 374655714Skris [Eric A. Young] 374755714Skris 374855714Skris *) Added PLATFORM define which is the id given to Configure. 374955714Skris [Eric A. Young] 375055714Skris 375155714Skris *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 375255714Skris [Eric A. Young] 375355714Skris 375455714Skris *) Extended the ASN.1 parser routines. 375555714Skris [Eric A. Young] 375655714Skris 375755714Skris *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 375855714Skris [Eric A. Young] 375955714Skris 376055714Skris *) Added a BN_CTX to the BN library. 376155714Skris [Eric A. Young] 376255714Skris 376355714Skris *) Fixed the weak key values in DES library 376455714Skris [Eric A. Young] 376555714Skris 376655714Skris *) Changed API in EVP library for cipher aliases. 376755714Skris [Eric A. Young] 376855714Skris 376955714Skris *) Added support for RC2/64bit cipher. 377055714Skris [Eric A. Young] 377155714Skris 377255714Skris *) Converted the lhash library to the crypto/mem.c functions. 377355714Skris [Eric A. Young] 377455714Skris 377555714Skris *) Added more recognized ASN.1 object ids. 377655714Skris [Eric A. Young] 377755714Skris 377855714Skris *) Added more RSA padding checks for SSL/TLS. 377955714Skris [Eric A. Young] 378055714Skris 378155714Skris *) Added BIO proxy/filter functionality. 378255714Skris [Eric A. Young] 378355714Skris 378455714Skris *) Added extra_certs to SSL_CTX which can be used 378555714Skris send extra CA certificates to the client in the CA cert chain sending 378655714Skris process. It can be configured with SSL_CTX_add_extra_chain_cert(). 378755714Skris [Eric A. Young] 378855714Skris 378955714Skris *) Now Fortezza is denied in the authentication phase because 379055714Skris this is key exchange mechanism is not supported by SSLeay at all. 379155714Skris [Eric A. Young] 379255714Skris 379355714Skris *) Additional PKCS1 checks. 379455714Skris [Eric A. Young] 379555714Skris 379655714Skris *) Support the string "TLSv1" for all TLS v1 ciphers. 379755714Skris [Eric A. Young] 379855714Skris 379955714Skris *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 380055714Skris ex_data index of the SSL context in the X509_STORE_CTX ex_data. 380155714Skris [Eric A. Young] 380255714Skris 380355714Skris *) Fixed a few memory leaks. 380455714Skris [Eric A. Young] 380555714Skris 380655714Skris *) Fixed various code and comment typos. 380755714Skris [Eric A. Young] 380855714Skris 380955714Skris *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 381055714Skris bytes sent in the client random. 381155714Skris [Edward Bishop <ebishop@spyglass.com>] 381255714Skris 3813