CHANGES revision 89837
155714Skris 255714Skris OpenSSL CHANGES 355714Skris _______________ 455714Skris 589837Skris Changes between 0.9.6b and 0.9.6c [21 dec 2001] 689837Skris 789837Skris *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 889837Skris <Dominikus.Scherkl@biodata.com>. (The previous implementation 989837Skris worked incorrectly for those cases where range = 10..._2 and 1089837Skris 3*range is two bits longer than range.) 1189837Skris [Bodo Moeller] 1289837Skris 1389837Skris *) Only add signing time to PKCS7 structures if it is not already 1489837Skris present. 1589837Skris [Steve Henson] 1689837Skris 1789837Skris *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 1889837Skris OBJ_ld_ce should be OBJ_id_ce. 1989837Skris Also some ip-pda OIDs in crypto/objects/objects.txt were 2089837Skris incorrect (cf. RFC 3039). 2189837Skris [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 2289837Skris 2389837Skris *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 2489837Skris returns early because it has nothing to do. 2589837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 2689837Skris 2789837Skris *) [In 0.9.6c-engine release:] 2889837Skris Fix mutex callback return values in crypto/engine/hw_ncipher.c. 2989837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 3089837Skris 3189837Skris *) [In 0.9.6c-engine release:] 3289837Skris Add support for Cryptographic Appliance's keyserver technology. 3389837Skris (Use engine 'keyclient') 3489837Skris [Cryptographic Appliances and Geoff Thorpe] 3589837Skris 3689837Skris *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 3789837Skris is called via tools/c89.sh because arguments have to be 3889837Skris rearranged (all '-L' options must appear before the first object 3989837Skris modules). 4089837Skris [Richard Shapiro <rshapiro@abinitio.com>] 4189837Skris 4289837Skris *) [In 0.9.6c-engine release:] 4389837Skris Add support for Broadcom crypto accelerator cards, backported 4489837Skris from 0.9.7. 4589837Skris [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 4689837Skris 4789837Skris *) [In 0.9.6c-engine release:] 4889837Skris Add support for SureWare crypto accelerator cards from 4989837Skris Baltimore Technologies. (Use engine 'sureware') 5089837Skris [Baltimore Technologies and Mark Cox] 5189837Skris 5289837Skris *) [In 0.9.6c-engine release:] 5389837Skris Add support for crypto accelerator cards from Accelerated 5489837Skris Encryption Processing, www.aep.ie. (Use engine 'aep') 5589837Skris [AEP Inc. and Mark Cox] 5689837Skris 5789837Skris *) Add a configuration entry for gcc on UnixWare. 5889837Skris [Gary Benson <gbenson@redhat.com>] 5989837Skris 6089837Skris *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 6189837Skris messages are stored in a single piece (fixed-length part and 6289837Skris variable-length part combined) and fix various bugs found on the way. 6389837Skris [Bodo Moeller] 6489837Skris 6589837Skris *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 6689837Skris instead. BIO_gethostbyname() does not know what timeouts are 6789837Skris appropriate, so entries would stay in cache even when they have 6889837Skris become invalid. 6989837Skris [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 7089837Skris 7189837Skris *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 7289837Skris faced with a pathologically small ClientHello fragment that does 7389837Skris not contain client_version: Instead of aborting with an error, 7489837Skris simply choose the highest available protocol version (i.e., 7589837Skris TLS 1.0 unless it is disabled). In practice, ClientHello 7689837Skris messages are never sent like this, but this change gives us 7789837Skris strictly correct behaviour at least for TLS. 7889837Skris [Bodo Moeller] 7989837Skris 8089837Skris *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 8189837Skris never resets s->method to s->ctx->method when called from within 8289837Skris one of the SSL handshake functions. 8389837Skris [Bodo Moeller; problem pointed out by Niko Baric] 8489837Skris 8589837Skris *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 8689837Skris (sent using the client's version number) if client_version is 8789837Skris smaller than the protocol version in use. Also change 8889837Skris ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 8989837Skris the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 9089837Skris the client will at least see that alert. 9189837Skris [Bodo Moeller] 9289837Skris 9389837Skris *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 9489837Skris correctly. 9589837Skris [Bodo Moeller] 9689837Skris 9789837Skris *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 9889837Skris client receives HelloRequest while in a handshake. 9989837Skris [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 10089837Skris 10189837Skris *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 10289837Skris should end in 'break', not 'goto end' which circuments various 10389837Skris cleanups done in state SSL_ST_OK. But session related stuff 10489837Skris must be disabled for SSL_ST_OK in the case that we just sent a 10589837Skris HelloRequest. 10689837Skris 10789837Skris Also avoid some overhead by not calling ssl_init_wbio_buffer() 10889837Skris before just sending a HelloRequest. 10989837Skris [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 11089837Skris 11189837Skris *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 11289837Skris reveal whether illegal block cipher padding was found or a MAC 11389837Skris verification error occured. (Neither SSLerr() codes nor alerts 11489837Skris are directly visible to potential attackers, but the information 11589837Skris may leak via logfiles.) 11689837Skris 11789837Skris Similar changes are not required for the SSL 2.0 implementation 11889837Skris because the number of padding bytes is sent in clear for SSL 2.0, 11989837Skris and the extra bytes are just ignored. However ssl/s2_pkt.c 12089837Skris failed to verify that the purported number of padding bytes is in 12189837Skris the legal range. 12289837Skris [Bodo Moeller] 12389837Skris 12489837Skris *) Add OpenUNIX-8 support including shared libraries 12589837Skris (Boyd Lynn Gerber <gerberb@zenez.com>). 12689837Skris [Lutz Jaenicke] 12789837Skris 12889837Skris *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 12989837Skris 'wristwatch attack' using huge encoding parameters (cf. 13089837Skris James H. Manger's CRYPTO 2001 paper). Note that the 13189837Skris RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 13289837Skris encoding parameters and hence was not vulnerable. 13389837Skris [Bodo Moeller] 13489837Skris 13589837Skris *) BN_sqr() bug fix. 13689837Skris [Ulf M�ller, reported by Jim Ellis <jim.ellis@cavium.com>] 13789837Skris 13889837Skris *) Rabin-Miller test analyses assume uniformly distributed witnesses, 13989837Skris so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 14089837Skris followed by modular reduction. 14189837Skris [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 14289837Skris 14389837Skris *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 14489837Skris equivalent based on BN_pseudo_rand() instead of BN_rand(). 14589837Skris [Bodo Moeller] 14689837Skris 14789837Skris *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 14889837Skris This function was broken, as the check for a new client hello message 14989837Skris to handle SGC did not allow these large messages. 15089837Skris (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 15189837Skris [Lutz Jaenicke] 15289837Skris 15389837Skris *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 15489837Skris [Lutz Jaenicke] 15589837Skris 15689837Skris *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 15789837Skris for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 15889837Skris [Lutz Jaenicke] 15989837Skris 16089837Skris *) Rework the configuration and shared library support for Tru64 Unix. 16189837Skris The configuration part makes use of modern compiler features and 16289837Skris still retains old compiler behavior for those that run older versions 16389837Skris of the OS. The shared library support part includes a variant that 16489837Skris uses the RPATH feature, and is available through the special 16589837Skris configuration target "alpha-cc-rpath", which will never be selected 16689837Skris automatically. 16789837Skris [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 16889837Skris 16989837Skris *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 17089837Skris with the same message size as in ssl3_get_certificate_request(). 17189837Skris Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 17289837Skris messages might inadvertently be reject as too long. 17389837Skris [Petr Lampa <lampa@fee.vutbr.cz>] 17489837Skris 17589837Skris *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 17689837Skris [Andy Polyakov] 17789837Skris 17889837Skris *) Modified SSL library such that the verify_callback that has been set 17989837Skris specificly for an SSL object with SSL_set_verify() is actually being 18089837Skris used. Before the change, a verify_callback set with this function was 18189837Skris ignored and the verify_callback() set in the SSL_CTX at the time of 18289837Skris the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 18389837Skris to allow the necessary settings. 18489837Skris [Lutz Jaenicke] 18589837Skris 18689837Skris *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 18789837Skris explicitly to NULL, as at least on Solaris 8 this seems not always to be 18889837Skris done automatically (in contradiction to the requirements of the C 18989837Skris standard). This made problems when used from OpenSSH. 19089837Skris [Lutz Jaenicke] 19189837Skris 19289837Skris *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 19389837Skris dh->length and always used 19489837Skris 19589837Skris BN_rand_range(priv_key, dh->p). 19689837Skris 19789837Skris BN_rand_range() is not necessary for Diffie-Hellman, and this 19889837Skris specific range makes Diffie-Hellman unnecessarily inefficient if 19989837Skris dh->length (recommended exponent length) is much smaller than the 20089837Skris length of dh->p. We could use BN_rand_range() if the order of 20189837Skris the subgroup was stored in the DH structure, but we only have 20289837Skris dh->length. 20389837Skris 20489837Skris So switch back to 20589837Skris 20689837Skris BN_rand(priv_key, l, ...) 20789837Skris 20889837Skris where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 20989837Skris otherwise. 21089837Skris [Bodo Moeller] 21189837Skris 21289837Skris *) In 21389837Skris 21489837Skris RSA_eay_public_encrypt 21589837Skris RSA_eay_private_decrypt 21689837Skris RSA_eay_private_encrypt (signing) 21789837Skris RSA_eay_public_decrypt (signature verification) 21889837Skris 21989837Skris (default implementations for RSA_public_encrypt, 22089837Skris RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 22189837Skris always reject numbers >= n. 22289837Skris [Bodo Moeller] 22389837Skris 22489837Skris *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 22589837Skris to synchronize access to 'locking_thread'. This is necessary on 22689837Skris systems where access to 'locking_thread' (an 'unsigned long' 22789837Skris variable) is not atomic. 22889837Skris [Bodo Moeller] 22989837Skris 23089837Skris *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 23189837Skris *before* setting the 'crypto_lock_rand' flag. The previous code had 23289837Skris a race condition if 0 is a valid thread ID. 23389837Skris [Travis Vitek <vitek@roguewave.com>] 23489837Skris 23589837Skris *) Add support for shared libraries under Irix. 23689837Skris [Albert Chin-A-Young <china@thewrittenword.com>] 23789837Skris 23889837Skris *) Add configuration option to build on Linux on both big-endian and 23989837Skris little-endian MIPS. 24089837Skris [Ralf Baechle <ralf@uni-koblenz.de>] 24189837Skris 24289837Skris *) Add the possibility to create shared libraries on HP-UX. 24389837Skris [Richard Levitte] 24489837Skris 24579998Skris Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 24679998Skris 24779998Skris *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 24879998Skris to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 24979998Skris Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 25079998Skris PRNG state recovery was possible based on the output of 25179998Skris one PRNG request appropriately sized to gain knowledge on 25279998Skris 'md' followed by enough consecutive 1-byte PRNG requests 25379998Skris to traverse all of 'state'. 25479998Skris 25579998Skris 1. When updating 'md_local' (the current thread's copy of 'md') 25679998Skris during PRNG output generation, hash all of the previous 25779998Skris 'md_local' value, not just the half used for PRNG output. 25879998Skris 25979998Skris 2. Make the number of bytes from 'state' included into the hash 26079998Skris independent from the number of PRNG bytes requested. 26179998Skris 26279998Skris The first measure alone would be sufficient to avoid 26379998Skris Markku-Juhani's attack. (Actually it had never occurred 26479998Skris to me that the half of 'md_local' used for chaining was the 26579998Skris half from which PRNG output bytes were taken -- I had always 26679998Skris assumed that the secret half would be used.) The second 26779998Skris measure makes sure that additional data from 'state' is never 26879998Skris mixed into 'md_local' in small portions; this heuristically 26979998Skris further strengthens the PRNG. 27079998Skris [Bodo Moeller] 27179998Skris 27279998Skris *) Fix crypto/bn/asm/mips3.s. 27379998Skris [Andy Polyakov] 27479998Skris 27579998Skris *) When only the key is given to "enc", the IV is undefined. Print out 27679998Skris an error message in this case. 27779998Skris [Lutz Jaenicke] 27879998Skris 27979998Skris *) Handle special case when X509_NAME is empty in X509 printing routines. 28079998Skris [Steve Henson] 28179998Skris 28279998Skris *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 28379998Skris positive and less than q. 28479998Skris [Bodo Moeller] 28579998Skris 28679998Skris *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 28779998Skris used: it isn't thread safe and the add_lock_callback should handle 28879998Skris that itself. 28979998Skris [Paul Rose <Paul.Rose@bridge.com>] 29079998Skris 29179998Skris *) Verify that incoming data obeys the block size in 29279998Skris ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 29379998Skris [Bodo Moeller] 29479998Skris 29579998Skris *) Fix OAEP check. 29679998Skris [Ulf M�ller, Bodo M�ller] 29779998Skris 29879998Skris *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 29989837Skris RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 30079998Skris when fixing the server behaviour for backwards-compatible 'client 30179998Skris hello' messages. (Note that the attack is impractical against 30279998Skris SSL 3.0 and TLS 1.0 anyway because length and version checking 30379998Skris means that the probability of guessing a valid ciphertext is 30479998Skris around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 30579998Skris paper.) 30679998Skris 30779998Skris Before 0.9.5, the countermeasure (hide the error by generating a 30879998Skris random 'decryption result') did not work properly because 30979998Skris ERR_clear_error() was missing, meaning that SSL_get_error() would 31079998Skris detect the supposedly ignored error. 31179998Skris 31279998Skris Both problems are now fixed. 31379998Skris [Bodo Moeller] 31479998Skris 31579998Skris *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 31679998Skris (previously it was 1024). 31779998Skris [Bodo Moeller] 31879998Skris 31979998Skris *) Fix for compatibility mode trust settings: ignore trust settings 32079998Skris unless some valid trust or reject settings are present. 32179998Skris [Steve Henson] 32279998Skris 32379998Skris *) Fix for blowfish EVP: its a variable length cipher. 32479998Skris [Steve Henson] 32579998Skris 32679998Skris *) Fix various bugs related to DSA S/MIME verification. Handle missing 32779998Skris parameters in DSA public key structures and return an error in the 32879998Skris DSA routines if parameters are absent. 32979998Skris [Steve Henson] 33079998Skris 33179998Skris *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 33279998Skris in the current directory if neither $RANDFILE nor $HOME was set. 33379998Skris RAND_file_name() in 0.9.6a returned NULL in this case. This has 33479998Skris caused some confusion to Windows users who haven't defined $HOME. 33579998Skris Thus RAND_file_name() is changed again: e_os.h can define a 33679998Skris DEFAULT_HOME, which will be used if $HOME is not set. 33779998Skris For Windows, we use "C:"; on other platforms, we still require 33879998Skris environment variables. 33979998Skris 34079998Skris *) Move 'if (!initialized) RAND_poll()' into regions protected by 34179998Skris CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 34279998Skris having multiple threads call RAND_poll() concurrently. 34379998Skris [Bodo Moeller] 34479998Skris 34579998Skris *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 34679998Skris combination of a flag and a thread ID variable. 34779998Skris Otherwise while one thread is in ssleay_rand_bytes (which sets the 34879998Skris flag), *other* threads can enter ssleay_add_bytes without obeying 34989837Skris the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 35079998Skris that they do not hold after the first thread unsets add_do_not_lock). 35179998Skris [Bodo Moeller] 35279998Skris 35379998Skris *) Change bctest again: '-x' expressions are not available in all 35479998Skris versions of 'test'. 35579998Skris [Bodo Moeller] 35679998Skris 35776866Skris Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 35872613Skris 35976866Skris *) Fix a couple of memory leaks in PKCS7_dataDecode() 36076866Skris [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 36176866Skris 36276866Skris *) Change Configure and Makefiles to provide EXE_EXT, which will contain 36376866Skris the default extension for executables, if any. Also, make the perl 36476866Skris scripts that use symlink() to test if it really exists and use "cp" 36576866Skris if it doesn't. All this made OpenSSL compilable and installable in 36676866Skris CygWin. 36776866Skris [Richard Levitte] 36876866Skris 36976866Skris *) Fix for asn1_GetSequence() for indefinite length constructed data. 37076866Skris If SEQUENCE is length is indefinite just set c->slen to the total 37176866Skris amount of data available. 37276866Skris [Steve Henson, reported by shige@FreeBSD.org] 37376866Skris [This change does not apply to 0.9.7.] 37476866Skris 37576866Skris *) Change bctest to avoid here-documents inside command substitution 37676866Skris (workaround for FreeBSD /bin/sh bug). 37776866Skris For compatibility with Ultrix, avoid shell functions (introduced 37876866Skris in the bctest version that searches along $PATH). 37976866Skris [Bodo Moeller] 38076866Skris 38176866Skris *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 38276866Skris with des_encrypt() defined on some operating systems, like Solaris 38376866Skris and UnixWare. 38476866Skris [Richard Levitte] 38576866Skris 38676866Skris *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 38776866Skris On the Importance of Eliminating Errors in Cryptographic 38876866Skris Computations, J. Cryptology 14 (2001) 2, 101-119, 38976866Skris http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 39076866Skris [Ulf Moeller] 39176866Skris 39276866Skris *) MIPS assembler BIGNUM division bug fix. 39376866Skris [Andy Polyakov] 39476866Skris 39576866Skris *) Disabled incorrect Alpha assembler code. 39676866Skris [Richard Levitte] 39776866Skris 39876866Skris *) Fix PKCS#7 decode routines so they correctly update the length 39976866Skris after reading an EOC for the EXPLICIT tag. 40076866Skris [Steve Henson] 40176866Skris [This change does not apply to 0.9.7.] 40276866Skris 40376866Skris *) Fix bug in PKCS#12 key generation routines. This was triggered 40476866Skris if a 3DES key was generated with a 0 initial byte. Include 40576866Skris PKCS12_BROKEN_KEYGEN compilation option to retain the old 40676866Skris (but broken) behaviour. 40776866Skris [Steve Henson] 40876866Skris 40976866Skris *) Enhance bctest to search for a working bc along $PATH and print 41076866Skris it when found. 41176866Skris [Tim Rice <tim@multitalents.net> via Richard Levitte] 41276866Skris 41376866Skris *) Fix memory leaks in err.c: free err_data string if necessary; 41476866Skris don't write to the wrong index in ERR_set_error_data. 41576866Skris [Bodo Moeller] 41676866Skris 41776866Skris *) Implement ssl23_peek (analogous to ssl23_read), which previously 41876866Skris did not exist. 41976866Skris [Bodo Moeller] 42076866Skris 42176866Skris *) Replace rdtsc with _emit statements for VC++ version 5. 42276866Skris [Jeremy Cooper <jeremy@baymoo.org>] 42376866Skris 42476866Skris *) Make it possible to reuse SSLv2 sessions. 42576866Skris [Richard Levitte] 42676866Skris 42776866Skris *) In copy_email() check for >= 0 as a return value for 42876866Skris X509_NAME_get_index_by_NID() since 0 is a valid index. 42976866Skris [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 43076866Skris 43176866Skris *) Avoid coredump with unsupported or invalid public keys by checking if 43276866Skris X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 43376866Skris PKCS7_verify() fails with non detached data. 43476866Skris [Steve Henson] 43576866Skris 43676866Skris *) Don't use getenv in library functions when run as setuid/setgid. 43776866Skris New function OPENSSL_issetugid(). 43876866Skris [Ulf Moeller] 43976866Skris 44076866Skris *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 44176866Skris due to incorrect handling of multi-threading: 44276866Skris 44376866Skris 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 44476866Skris 44576866Skris 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 44676866Skris 44776866Skris 3. Count how many times MemCheck_off() has been called so that 44876866Skris nested use can be treated correctly. This also avoids 44976866Skris inband-signalling in the previous code (which relied on the 45076866Skris assumption that thread ID 0 is impossible). 45176866Skris [Bodo Moeller] 45276866Skris 45376866Skris *) Add "-rand" option also to s_client and s_server. 45476866Skris [Lutz Jaenicke] 45576866Skris 45676866Skris *) Fix CPU detection on Irix 6.x. 45776866Skris [Kurt Hockenbury <khockenb@stevens-tech.edu> and 45876866Skris "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 45976866Skris 46076866Skris *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 46176866Skris was empty. 46276866Skris [Steve Henson] 46376866Skris [This change does not apply to 0.9.7.] 46476866Skris 46576866Skris *) Use the cached encoding of an X509_NAME structure rather than 46676866Skris copying it. This is apparently the reason for the libsafe "errors" 46776866Skris but the code is actually correct. 46876866Skris [Steve Henson] 46976866Skris 47072613Skris *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 47172613Skris Bleichenbacher's DSA attack. 47276866Skris Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 47376866Skris to be set and top=0 forces the highest bit to be set; top=-1 is new 47476866Skris and leaves the highest bit random. 47576866Skris [Ulf Moeller, Bodo Moeller] 47672613Skris 47772613Skris *) In the NCONF_...-based implementations for CONF_... queries 47872613Skris (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 47972613Skris a temporary CONF structure with the data component set to NULL 48072613Skris (which gives segmentation faults in lh_retrieve). 48172613Skris Instead, use NULL for the CONF pointer in CONF_get_string and 48272613Skris CONF_get_number (which may use environment variables) and directly 48372613Skris return NULL from CONF_get_section. 48472613Skris [Bodo Moeller] 48572613Skris 48672613Skris *) Fix potential buffer overrun for EBCDIC. 48772613Skris [Ulf Moeller] 48872613Skris 48972613Skris *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 49072613Skris keyUsage if basicConstraints absent for a CA. 49172613Skris [Steve Henson] 49272613Skris 49372613Skris *) Make SMIME_write_PKCS7() write mail header values with a format that 49472613Skris is more generally accepted (no spaces before the semicolon), since 49572613Skris some programs can't parse those values properly otherwise. Also make 49672613Skris sure BIO's that break lines after each write do not create invalid 49772613Skris headers. 49872613Skris [Richard Levitte] 49972613Skris 50072613Skris *) Make the CRL encoding routines work with empty SEQUENCE OF. The 50172613Skris macros previously used would not encode an empty SEQUENCE OF 50272613Skris and break the signature. 50372613Skris [Steve Henson] 50476866Skris [This change does not apply to 0.9.7.] 50572613Skris 50672613Skris *) Zero the premaster secret after deriving the master secret in 50772613Skris DH ciphersuites. 50872613Skris [Steve Henson] 50972613Skris 51072613Skris *) Add some EVP_add_digest_alias registrations (as found in 51172613Skris OpenSSL_add_all_digests()) to SSL_library_init() 51272613Skris aka OpenSSL_add_ssl_algorithms(). This provides improved 51372613Skris compatibility with peers using X.509 certificates 51472613Skris with unconventional AlgorithmIdentifier OIDs. 51572613Skris [Bodo Moeller] 51672613Skris 51772613Skris *) Fix for Irix with NO_ASM. 51872613Skris ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 51972613Skris 52072613Skris *) ./config script fixes. 52172613Skris [Ulf Moeller, Richard Levitte] 52272613Skris 52372613Skris *) Fix 'openssl passwd -1'. 52472613Skris [Bodo Moeller] 52572613Skris 52672613Skris *) Change PKCS12_key_gen_asc() so it can cope with non null 52772613Skris terminated strings whose length is passed in the passlen 52872613Skris parameter, for example from PEM callbacks. This was done 52972613Skris by adding an extra length parameter to asc2uni(). 53072613Skris [Steve Henson, reported by <oddissey@samsung.co.kr>] 53172613Skris 53272613Skris *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 53372613Skris call failed, free the DSA structure. 53472613Skris [Bodo Moeller] 53572613Skris 53672613Skris *) Fix to uni2asc() to cope with zero length Unicode strings. 53772613Skris These are present in some PKCS#12 files. 53872613Skris [Steve Henson] 53972613Skris 54072613Skris *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 54172613Skris Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 54272613Skris when writing a 32767 byte record. 54372613Skris [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 54472613Skris 54572613Skris *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 54672613Skris obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 54772613Skris 54872613Skris (RSA objects have a reference count access to which is protected 54972613Skris by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 55072613Skris so they are meant to be shared between threads.) 55172613Skris [Bodo Moeller, Geoff Thorpe; original patch submitted by 55272613Skris "Reddie, Steven" <Steven.Reddie@ca.com>] 55372613Skris 55472613Skris *) Fix a deadlock in CRYPTO_mem_leaks(). 55572613Skris [Bodo Moeller] 55672613Skris 55776866Skris *) Use better test patterns in bntest. 55876866Skris [Ulf M�ller] 55976866Skris 56072613Skris *) rand_win.c fix for Borland C. 56172613Skris [Ulf M�ller] 56272613Skris 56372613Skris *) BN_rshift bugfix for n == 0. 56472613Skris [Bodo Moeller] 56572613Skris 56676866Skris *) Add a 'bctest' script that checks for some known 'bc' bugs 56776866Skris so that 'make test' does not abort just because 'bc' is broken. 56876866Skris [Bodo Moeller] 56976866Skris 57072613Skris *) Store verify_result within SSL_SESSION also for client side to 57172613Skris avoid potential security hole. (Re-used sessions on the client side 57272613Skris always resulted in verify_result==X509_V_OK, not using the original 57372613Skris result of the server certificate verification.) 57472613Skris [Lutz Jaenicke] 57572613Skris 57672613Skris *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 57772613Skris SSL3_RT_APPLICATION_DATA, return 0. 57872613Skris Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 57972613Skris [Bodo Moeller] 58072613Skris 58172613Skris *) Fix SSL_peek: 58272613Skris Both ssl2_peek and ssl3_peek, which were totally broken in earlier 58372613Skris releases, have been re-implemented by renaming the previous 58472613Skris implementations of ssl2_read and ssl3_read to ssl2_read_internal 58572613Skris and ssl3_read_internal, respectively, and adding 'peek' parameters 58672613Skris to them. The new ssl[23]_{read,peek} functions are calls to 58772613Skris ssl[23]_read_internal with the 'peek' flag set appropriately. 58872613Skris A 'peek' parameter has also been added to ssl3_read_bytes, which 58972613Skris does the actual work for ssl3_read_internal. 59072613Skris [Bodo Moeller] 59172613Skris 59276866Skris *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 59376866Skris the method-specific "init()" handler. Also clean up ex_data after 59476866Skris calling the method-specific "finish()" handler. Previously, this was 59576866Skris happening the other way round. 59676866Skris [Geoff Thorpe] 59776866Skris 59872613Skris *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 59972613Skris The previous value, 12, was not always sufficient for BN_mod_exp(). 60072613Skris [Bodo Moeller] 60172613Skris 60276866Skris *) Make sure that shared libraries get the internal name engine with 60376866Skris the full version number and not just 0. This should mark the 60476866Skris shared libraries as not backward compatible. Of course, this should 60576866Skris be changed again when we can guarantee backward binary compatibility. 60676866Skris [Richard Levitte] 60776866Skris 60872613Skris *) Fix typo in get_cert_by_subject() in by_dir.c 60972613Skris [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 61072613Skris 61176866Skris *) Rework the system to generate shared libraries: 61276866Skris 61376866Skris - Make note of the expected extension for the shared libraries and 61476866Skris if there is a need for symbolic links from for example libcrypto.so.0 61576866Skris to libcrypto.so.0.9.7. There is extended info in Configure for 61676866Skris that. 61776866Skris 61876866Skris - Make as few rebuilds of the shared libraries as possible. 61976866Skris 62076866Skris - Still avoid linking the OpenSSL programs with the shared libraries. 62176866Skris 62276866Skris - When installing, install the shared libraries separately from the 62376866Skris static ones. 62476866Skris [Richard Levitte] 62576866Skris 62672613Skris *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 62772613Skris 62872613Skris Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 62972613Skris and not in SSL_clear because the latter is also used by the 63072613Skris accept/connect functions; previously, the settings made by 63172613Skris SSL_set_read_ahead would be lost during the handshake. 63272613Skris [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 63372613Skris 63472613Skris *) Correct util/mkdef.pl to be selective about disabled algorithms. 63572613Skris Previously, it would create entries for disableed algorithms no 63672613Skris matter what. 63772613Skris [Richard Levitte] 63872613Skris 63972613Skris *) Added several new manual pages for SSL_* function. 64072613Skris [Lutz Jaenicke] 64172613Skris 64268651Skris Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 64368651Skris 64468651Skris *) In ssl23_get_client_hello, generate an error message when faced 64568651Skris with an initial SSL 3.0/TLS record that is too small to contain the 64668651Skris first two bytes of the ClientHello message, i.e. client_version. 64768651Skris (Note that this is a pathologic case that probably has never happened 64868651Skris in real life.) The previous approach was to use the version number 64968651Skris from the record header as a substitute; but our protocol choice 65068651Skris should not depend on that one because it is not authenticated 65168651Skris by the Finished messages. 65268651Skris [Bodo Moeller] 65368651Skris 65468651Skris *) More robust randomness gathering functions for Windows. 65568651Skris [Jeffrey Altman <jaltman@columbia.edu>] 65668651Skris 65768651Skris *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 65868651Skris not set then we don't setup the error code for issuer check errors 65968651Skris to avoid possibly overwriting other errors which the callback does 66068651Skris handle. If an application does set the flag then we assume it knows 66168651Skris what it is doing and can handle the new informational codes 66268651Skris appropriately. 66368651Skris [Steve Henson] 66468651Skris 66568651Skris *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 66668651Skris a general "ANY" type, as such it should be able to decode anything 66768651Skris including tagged types. However it didn't check the class so it would 66868651Skris wrongly interpret tagged types in the same way as their universal 66968651Skris counterpart and unknown types were just rejected. Changed so that the 67068651Skris tagged and unknown types are handled in the same way as a SEQUENCE: 67168651Skris that is the encoding is stored intact. There is also a new type 67268651Skris "V_ASN1_OTHER" which is used when the class is not universal, in this 67368651Skris case we have no idea what the actual type is so we just lump them all 67468651Skris together. 67568651Skris [Steve Henson] 67668651Skris 67768651Skris *) On VMS, stdout may very well lead to a file that is written to 67868651Skris in a record-oriented fashion. That means that every write() will 67968651Skris write a separate record, which will be read separately by the 68068651Skris programs trying to read from it. This can be very confusing. 68168651Skris 68268651Skris The solution is to put a BIO filter in the way that will buffer 68368651Skris text until a linefeed is reached, and then write everything a 68468651Skris line at a time, so every record written will be an actual line, 68568651Skris not chunks of lines and not (usually doesn't happen, but I've 68668651Skris seen it once) several lines in one record. BIO_f_linebuffer() is 68768651Skris the answer. 68868651Skris 68968651Skris Currently, it's a VMS-only method, because that's where it has 69068651Skris been tested well enough. 69168651Skris [Richard Levitte] 69268651Skris 69368651Skris *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 69468651Skris it can return incorrect results. 69568651Skris (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 69668651Skris but it was in 0.9.6-beta[12].) 69768651Skris [Bodo Moeller] 69868651Skris 69968651Skris *) Disable the check for content being present when verifying detached 70068651Skris signatures in pk7_smime.c. Some versions of Netscape (wrongly) 70168651Skris include zero length content when signing messages. 70268651Skris [Steve Henson] 70368651Skris 70468651Skris *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 70568651Skris BIO_ctrl (for BIO pairs). 70668651Skris [Bodo M�ller] 70768651Skris 70868651Skris *) Add DSO method for VMS. 70968651Skris [Richard Levitte] 71068651Skris 71168651Skris *) Bug fix: Montgomery multiplication could produce results with the 71268651Skris wrong sign. 71368651Skris [Ulf M�ller] 71468651Skris 71568651Skris *) Add RPM specification openssl.spec and modify it to build three 71668651Skris packages. The default package contains applications, application 71768651Skris documentation and run-time libraries. The devel package contains 71868651Skris include files, static libraries and function documentation. The 71968651Skris doc package contains the contents of the doc directory. The original 72068651Skris openssl.spec was provided by Damien Miller <djm@mindrot.org>. 72168651Skris [Richard Levitte] 72268651Skris 72368651Skris *) Add a large number of documentation files for many SSL routines. 72468651Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 72568651Skris 72668651Skris *) Add a configuration entry for Sony News 4. 72768651Skris [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 72868651Skris 72968651Skris *) Don't set the two most significant bits to one when generating a 73068651Skris random number < q in the DSA library. 73168651Skris [Ulf M�ller] 73268651Skris 73368651Skris *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 73468651Skris behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 73568651Skris the underlying transport is blocking) if a handshake took place. 73668651Skris (The default behaviour is needed by applications such as s_client 73768651Skris and s_server that use select() to determine when to use SSL_read; 73868651Skris but for applications that know in advance when to expect data, it 73968651Skris just makes things more complicated.) 74068651Skris [Bodo Moeller] 74168651Skris 74268651Skris *) Add RAND_egd_bytes(), which gives control over the number of bytes read 74368651Skris from EGD. 74468651Skris [Ben Laurie] 74568651Skris 74668651Skris *) Add a few more EBCDIC conditionals that make `req' and `x509' 74768651Skris work better on such systems. 74868651Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 74968651Skris 75068651Skris *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 75168651Skris Update PKCS12_parse() so it copies the friendlyName and the 75268651Skris keyid to the certificates aux info. 75368651Skris [Steve Henson] 75468651Skris 75568651Skris *) Fix bug in PKCS7_verify() which caused an infinite loop 75668651Skris if there was more than one signature. 75768651Skris [Sven Uszpelkat <su@celocom.de>] 75868651Skris 75968651Skris *) Major change in util/mkdef.pl to include extra information 76068651Skris about each symbol, as well as presentig variables as well 76168651Skris as functions. This change means that there's n more need 76268651Skris to rebuild the .num files when some algorithms are excluded. 76368651Skris [Richard Levitte] 76468651Skris 76568651Skris *) Allow the verify time to be set by an application, 76668651Skris rather than always using the current time. 76768651Skris [Steve Henson] 76868651Skris 76968651Skris *) Phase 2 verify code reorganisation. The certificate 77068651Skris verify code now looks up an issuer certificate by a 77168651Skris number of criteria: subject name, authority key id 77268651Skris and key usage. It also verifies self signed certificates 77368651Skris by the same criteria. The main comparison function is 77468651Skris X509_check_issued() which performs these checks. 77568651Skris 77668651Skris Lot of changes were necessary in order to support this 77768651Skris without completely rewriting the lookup code. 77868651Skris 77968651Skris Authority and subject key identifier are now cached. 78068651Skris 78168651Skris The LHASH 'certs' is X509_STORE has now been replaced 78268651Skris by a STACK_OF(X509_OBJECT). This is mainly because an 78368651Skris LHASH can't store or retrieve multiple objects with 78468651Skris the same hash value. 78568651Skris 78668651Skris As a result various functions (which were all internal 78768651Skris use only) have changed to handle the new X509_STORE 78868651Skris structure. This will break anything that messed round 78968651Skris with X509_STORE internally. 79068651Skris 79168651Skris The functions X509_STORE_add_cert() now checks for an 79268651Skris exact match, rather than just subject name. 79368651Skris 79468651Skris The X509_STORE API doesn't directly support the retrieval 79568651Skris of multiple certificates matching a given criteria, however 79668651Skris this can be worked round by performing a lookup first 79768651Skris (which will fill the cache with candidate certificates) 79868651Skris and then examining the cache for matches. This is probably 79968651Skris the best we can do without throwing out X509_LOOKUP 80068651Skris entirely (maybe later...). 80168651Skris 80268651Skris The X509_VERIFY_CTX structure has been enhanced considerably. 80368651Skris 80468651Skris All certificate lookup operations now go via a get_issuer() 80568651Skris callback. Although this currently uses an X509_STORE it 80668651Skris can be replaced by custom lookups. This is a simple way 80768651Skris to bypass the X509_STORE hackery necessary to make this 80868651Skris work and makes it possible to use more efficient techniques 80968651Skris in future. A very simple version which uses a simple 81068651Skris STACK for its trusted certificate store is also provided 81168651Skris using X509_STORE_CTX_trusted_stack(). 81268651Skris 81368651Skris The verify_cb() and verify() callbacks now have equivalents 81468651Skris in the X509_STORE_CTX structure. 81568651Skris 81668651Skris X509_STORE_CTX also has a 'flags' field which can be used 81768651Skris to customise the verify behaviour. 81868651Skris [Steve Henson] 81968651Skris 82068651Skris *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 82168651Skris excludes S/MIME capabilities. 82268651Skris [Steve Henson] 82368651Skris 82468651Skris *) When a certificate request is read in keep a copy of the 82568651Skris original encoding of the signed data and use it when outputing 82668651Skris again. Signatures then use the original encoding rather than 82768651Skris a decoded, encoded version which may cause problems if the 82868651Skris request is improperly encoded. 82968651Skris [Steve Henson] 83068651Skris 83168651Skris *) For consistency with other BIO_puts implementations, call 83268651Skris buffer_write(b, ...) directly in buffer_puts instead of calling 83368651Skris BIO_write(b, ...). 83468651Skris 83568651Skris In BIO_puts, increment b->num_write as in BIO_write. 83668651Skris [Peter.Sylvester@EdelWeb.fr] 83768651Skris 83868651Skris *) Fix BN_mul_word for the case where the word is 0. (We have to use 83968651Skris BN_zero, we may not return a BIGNUM with an array consisting of 84068651Skris words set to zero.) 84168651Skris [Bodo Moeller] 84268651Skris 84368651Skris *) Avoid calling abort() from within the library when problems are 84468651Skris detected, except if preprocessor symbols have been defined 84568651Skris (such as REF_CHECK, BN_DEBUG etc.). 84668651Skris [Bodo Moeller] 84768651Skris 84868651Skris *) New openssl application 'rsautl'. This utility can be 84968651Skris used for low level RSA operations. DER public key 85068651Skris BIO/fp routines also added. 85168651Skris [Steve Henson] 85268651Skris 85368651Skris *) New Configure entry and patches for compiling on QNX 4. 85468651Skris [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 85568651Skris 85668651Skris *) A demo state-machine implementation was sponsored by 85768651Skris Nuron (http://www.nuron.com/) and is now available in 85868651Skris demos/state_machine. 85968651Skris [Ben Laurie] 86068651Skris 86168651Skris *) New options added to the 'dgst' utility for signature 86268651Skris generation and verification. 86368651Skris [Steve Henson] 86468651Skris 86568651Skris *) Unrecognized PKCS#7 content types are now handled via a 86668651Skris catch all ASN1_TYPE structure. This allows unsupported 86768651Skris types to be stored as a "blob" and an application can 86868651Skris encode and decode it manually. 86968651Skris [Steve Henson] 87068651Skris 87168651Skris *) Fix various signed/unsigned issues to make a_strex.c 87268651Skris compile under VC++. 87368651Skris [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 87468651Skris 87568651Skris *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 87668651Skris length if passed a buffer. ASN1_INTEGER_to_BN failed 87768651Skris if passed a NULL BN and its argument was negative. 87868651Skris [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 87968651Skris 88068651Skris *) Modification to PKCS#7 encoding routines to output definite 88168651Skris length encoding. Since currently the whole structures are in 88268651Skris memory there's not real point in using indefinite length 88368651Skris constructed encoding. However if OpenSSL is compiled with 88468651Skris the flag PKCS7_INDEFINITE_ENCODING the old form is used. 88568651Skris [Steve Henson] 88668651Skris 88768651Skris *) Added BIO_vprintf() and BIO_vsnprintf(). 88868651Skris [Richard Levitte] 88968651Skris 89068651Skris *) Added more prefixes to parse for in the the strings written 89168651Skris through a logging bio, to cover all the levels that are available 89268651Skris through syslog. The prefixes are now: 89368651Skris 89468651Skris PANIC, EMERG, EMR => LOG_EMERG 89568651Skris ALERT, ALR => LOG_ALERT 89668651Skris CRIT, CRI => LOG_CRIT 89768651Skris ERROR, ERR => LOG_ERR 89868651Skris WARNING, WARN, WAR => LOG_WARNING 89968651Skris NOTICE, NOTE, NOT => LOG_NOTICE 90068651Skris INFO, INF => LOG_INFO 90168651Skris DEBUG, DBG => LOG_DEBUG 90268651Skris 90368651Skris and as before, if none of those prefixes are present at the 90468651Skris beginning of the string, LOG_ERR is chosen. 90568651Skris 90668651Skris On Win32, the LOG_* levels are mapped according to this: 90768651Skris 90868651Skris LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 90968651Skris LOG_WARNING => EVENTLOG_WARNING_TYPE 91068651Skris LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 91168651Skris 91268651Skris [Richard Levitte] 91368651Skris 91468651Skris *) Made it possible to reconfigure with just the configuration 91568651Skris argument "reconf" or "reconfigure". The command line arguments 91668651Skris are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 91768651Skris and are retrieved from there when reconfiguring. 91868651Skris [Richard Levitte] 91968651Skris 92068651Skris *) MD4 implemented. 92168651Skris [Assar Westerlund <assar@sics.se>, Richard Levitte] 92268651Skris 92368651Skris *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 92468651Skris [Richard Levitte] 92568651Skris 92668651Skris *) The obj_dat.pl script was messing up the sorting of object 92768651Skris names. The reason was that it compared the quoted version 92868651Skris of strings as a result "OCSP" > "OCSP Signing" because 92968651Skris " > SPACE. Changed script to store unquoted versions of 93068651Skris names and add quotes on output. It was also omitting some 93168651Skris names from the lookup table if they were given a default 93268651Skris value (that is if SN is missing it is given the same 93368651Skris value as LN and vice versa), these are now added on the 93468651Skris grounds that if an object has a name we should be able to 93568651Skris look it up. Finally added warning output when duplicate 93668651Skris short or long names are found. 93768651Skris [Steve Henson] 93868651Skris 93968651Skris *) Changes needed for Tandem NSK. 94068651Skris [Scott Uroff <scott@xypro.com>] 94168651Skris 94268651Skris *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 94368651Skris RSA_padding_check_SSLv23(), special padding was never detected 94468651Skris and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 94568651Skris version rollback attacks was not effective. 94668651Skris 94768651Skris In s23_clnt.c, don't use special rollback-attack detection padding 94868651Skris (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 94968651Skris client; similarly, in s23_srvr.c, don't do the rollback check if 95068651Skris SSL 2.0 is the only protocol enabled in the server. 95168651Skris [Bodo Moeller] 95268651Skris 95368651Skris *) Make it possible to get hexdumps of unprintable data with 'openssl 95468651Skris asn1parse'. By implication, the functions ASN1_parse_dump() and 95568651Skris BIO_dump_indent() are added. 95668651Skris [Richard Levitte] 95768651Skris 95868651Skris *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 95968651Skris these print out strings and name structures based on various 96068651Skris flags including RFC2253 support and proper handling of 96168651Skris multibyte characters. Added options to the 'x509' utility 96268651Skris to allow the various flags to be set. 96368651Skris [Steve Henson] 96468651Skris 96568651Skris *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 96668651Skris Also change the functions X509_cmp_current_time() and 96768651Skris X509_gmtime_adj() work with an ASN1_TIME structure, 96868651Skris this will enable certificates using GeneralizedTime in validity 96968651Skris dates to be checked. 97068651Skris [Steve Henson] 97168651Skris 97268651Skris *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 97368651Skris negative public key encodings) on by default, 97468651Skris NO_NEG_PUBKEY_BUG can be set to disable it. 97568651Skris [Steve Henson] 97668651Skris 97768651Skris *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 97868651Skris content octets. An i2c_ASN1_OBJECT is unnecessary because 97968651Skris the encoding can be trivially obtained from the structure. 98068651Skris [Steve Henson] 98168651Skris 98268651Skris *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 98368651Skris not read locks (CRYPTO_r_[un]lock). 98468651Skris [Bodo Moeller] 98568651Skris 98668651Skris *) A first attempt at creating official support for shared 98768651Skris libraries through configuration. I've kept it so the 98868651Skris default is static libraries only, and the OpenSSL programs 98968651Skris are always statically linked for now, but there are 99068651Skris preparations for dynamic linking in place. 99189837Skris This has been tested on Linux and Tru64. 99268651Skris [Richard Levitte] 99368651Skris 99468651Skris *) Randomness polling function for Win9x, as described in: 99568651Skris Peter Gutmann, Software Generation of Practically Strong 99668651Skris Random Numbers. 99768651Skris [Ulf M�ller] 99868651Skris 99968651Skris *) Fix so PRNG is seeded in req if using an already existing 100068651Skris DSA key. 100168651Skris [Steve Henson] 100268651Skris 100368651Skris *) New options to smime application. -inform and -outform 100468651Skris allow alternative formats for the S/MIME message including 100568651Skris PEM and DER. The -content option allows the content to be 100668651Skris specified separately. This should allow things like Netscape 100768651Skris form signing output easier to verify. 100868651Skris [Steve Henson] 100968651Skris 101068651Skris *) Fix the ASN1 encoding of tags using the 'long form'. 101168651Skris [Steve Henson] 101268651Skris 101368651Skris *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 101468651Skris STRING types. These convert content octets to and from the 101568651Skris underlying type. The actual tag and length octets are 101668651Skris already assumed to have been read in and checked. These 101768651Skris are needed because all other string types have virtually 101868651Skris identical handling apart from the tag. By having versions 101968651Skris of the ASN1 functions that just operate on content octets 102068651Skris IMPLICIT tagging can be handled properly. It also allows 102168651Skris the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 102268651Skris and ASN1_INTEGER are identical apart from the tag. 102368651Skris [Steve Henson] 102468651Skris 102568651Skris *) Change the handling of OID objects as follows: 102668651Skris 102768651Skris - New object identifiers are inserted in objects.txt, following 102868651Skris the syntax given in objects.README. 102968651Skris - objects.pl is used to process obj_mac.num and create a new 103068651Skris obj_mac.h. 103168651Skris - obj_dat.pl is used to create a new obj_dat.h, using the data in 103268651Skris obj_mac.h. 103368651Skris 103468651Skris This is currently kind of a hack, and the perl code in objects.pl 103568651Skris isn't very elegant, but it works as I intended. The simplest way 103668651Skris to check that it worked correctly is to look in obj_dat.h and 103768651Skris check the array nid_objs and make sure the objects haven't moved 103868651Skris around (this is important!). Additions are OK, as well as 103968651Skris consistent name changes. 104068651Skris [Richard Levitte] 104168651Skris 104268651Skris *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 104368651Skris [Bodo Moeller] 104468651Skris 104568651Skris *) Addition of the command line parameter '-rand file' to 'openssl req'. 104668651Skris The given file adds to whatever has already been seeded into the 104768651Skris random pool through the RANDFILE configuration file option or 104868651Skris environment variable, or the default random state file. 104968651Skris [Richard Levitte] 105068651Skris 105168651Skris *) mkstack.pl now sorts each macro group into lexical order. 105268651Skris Previously the output order depended on the order the files 105368651Skris appeared in the directory, resulting in needless rewriting 105468651Skris of safestack.h . 105568651Skris [Steve Henson] 105668651Skris 105768651Skris *) Patches to make OpenSSL compile under Win32 again. Mostly 105868651Skris work arounds for the VC++ problem that it treats func() as 105968651Skris func(void). Also stripped out the parts of mkdef.pl that 106068651Skris added extra typesafe functions: these no longer exist. 106168651Skris [Steve Henson] 106268651Skris 106368651Skris *) Reorganisation of the stack code. The macros are now all 106468651Skris collected in safestack.h . Each macro is defined in terms of 106568651Skris a "stack macro" of the form SKM_<name>(type, a, b). The 106668651Skris DEBUG_SAFESTACK is now handled in terms of function casts, 106768651Skris this has the advantage of retaining type safety without the 106868651Skris use of additional functions. If DEBUG_SAFESTACK is not defined 106968651Skris then the non typesafe macros are used instead. Also modified the 107068651Skris mkstack.pl script to handle the new form. Needs testing to see 107168651Skris if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 107268651Skris the default if no major problems. Similar behaviour for ASN1_SET_OF 107368651Skris and PKCS12_STACK_OF. 107468651Skris [Steve Henson] 107568651Skris 107668651Skris *) When some versions of IIS use the 'NET' form of private key the 107768651Skris key derivation algorithm is different. Normally MD5(password) is 107868651Skris used as a 128 bit RC4 key. In the modified case 107968651Skris MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 108068651Skris new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 108168651Skris as the old Netscape_RSA functions except they have an additional 108268651Skris 'sgckey' parameter which uses the modified algorithm. Also added 108368651Skris an -sgckey command line option to the rsa utility. Thanks to 108468651Skris Adrian Peck <bertie@ncipher.com> for posting details of the modified 108568651Skris algorithm to openssl-dev. 108668651Skris [Steve Henson] 108768651Skris 108868651Skris *) The evp_local.h macros were using 'c.##kname' which resulted in 108968651Skris invalid expansion on some systems (SCO 5.0.5 for example). 109068651Skris Corrected to 'c.kname'. 109168651Skris [Phillip Porch <root@theporch.com>] 109268651Skris 109368651Skris *) New X509_get1_email() and X509_REQ_get1_email() functions that return 109468651Skris a STACK of email addresses from a certificate or request, these look 109568651Skris in the subject name and the subject alternative name extensions and 109668651Skris omit any duplicate addresses. 109768651Skris [Steve Henson] 109868651Skris 109968651Skris *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 110068651Skris This makes DSA verification about 2 % faster. 110168651Skris [Bodo Moeller] 110268651Skris 110368651Skris *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 110468651Skris (meaning that now 2^5 values will be precomputed, which is only 4 KB 110568651Skris plus overhead for 1024 bit moduli). 110668651Skris This makes exponentiations about 0.5 % faster for 1024 bit 110768651Skris exponents (as measured by "openssl speed rsa2048"). 110868651Skris [Bodo Moeller] 110968651Skris 111068651Skris *) Rename memory handling macros to avoid conflicts with other 111168651Skris software: 111268651Skris Malloc => OPENSSL_malloc 111368651Skris Malloc_locked => OPENSSL_malloc_locked 111468651Skris Realloc => OPENSSL_realloc 111568651Skris Free => OPENSSL_free 111668651Skris [Richard Levitte] 111768651Skris 111868651Skris *) New function BN_mod_exp_mont_word for small bases (roughly 15% 111968651Skris faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 112068651Skris [Bodo Moeller] 112168651Skris 112268651Skris *) CygWin32 support. 112368651Skris [John Jarvie <jjarvie@newsguy.com>] 112468651Skris 112568651Skris *) The type-safe stack code has been rejigged. It is now only compiled 112668651Skris in when OpenSSL is configured with the DEBUG_SAFESTACK option and 112768651Skris by default all type-specific stack functions are "#define"d back to 112868651Skris standard stack functions. This results in more streamlined output 112968651Skris but retains the type-safety checking possibilities of the original 113068651Skris approach. 113168651Skris [Geoff Thorpe] 113268651Skris 113368651Skris *) The STACK code has been cleaned up, and certain type declarations 113468651Skris that didn't make a lot of sense have been brought in line. This has 113568651Skris also involved a cleanup of sorts in safestack.h to more correctly 113668651Skris map type-safe stack functions onto their plain stack counterparts. 113768651Skris This work has also resulted in a variety of "const"ifications of 113868651Skris lots of the code, especially "_cmp" operations which should normally 113968651Skris be prototyped with "const" parameters anyway. 114068651Skris [Geoff Thorpe] 114168651Skris 114268651Skris *) When generating bytes for the first time in md_rand.c, 'stir the pool' 114368651Skris by seeding with STATE_SIZE dummy bytes (with zero entropy count). 114468651Skris (The PRNG state consists of two parts, the large pool 'state' and 'md', 114568651Skris where all of 'md' is used each time the PRNG is used, but 'state' 114668651Skris is used only indexed by a cyclic counter. As entropy may not be 114768651Skris well distributed from the beginning, 'md' is important as a 114868651Skris chaining variable. However, the output function chains only half 114968651Skris of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 115068651Skris all of 'md', and seeding with STATE_SIZE dummy bytes will result 115168651Skris in all of 'state' being rewritten, with the new values depending 115268651Skris on virtually all of 'md'. This overcomes the 80 bit limitation.) 115368651Skris [Bodo Moeller] 115468651Skris 115568651Skris *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 115668651Skris the handshake is continued after ssl_verify_cert_chain(); 115768651Skris otherwise, if SSL_VERIFY_NONE is set, remaining error codes 115868651Skris can lead to 'unexplainable' connection aborts later. 115968651Skris [Bodo Moeller; problem tracked down by Lutz Jaenicke] 116068651Skris 116168651Skris *) Major EVP API cipher revision. 116268651Skris Add hooks for extra EVP features. This allows various cipher 116368651Skris parameters to be set in the EVP interface. Support added for variable 116468651Skris key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 116568651Skris setting of RC2 and RC5 parameters. 116668651Skris 116768651Skris Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 116868651Skris ciphers. 116968651Skris 117068651Skris Remove lots of duplicated code from the EVP library. For example *every* 117168651Skris cipher init() function handles the 'iv' in the same way according to the 117268651Skris cipher mode. They also all do nothing if the 'key' parameter is NULL and 117368651Skris for CFB and OFB modes they zero ctx->num. 117468651Skris 117568651Skris New functionality allows removal of S/MIME code RC2 hack. 117668651Skris 117768651Skris Most of the routines have the same form and so can be declared in terms 117868651Skris of macros. 117968651Skris 118068651Skris By shifting this to the top level EVP_CipherInit() it can be removed from 118168651Skris all individual ciphers. If the cipher wants to handle IVs or keys 118268651Skris differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 118368651Skris flags. 118468651Skris 118568651Skris Change lots of functions like EVP_EncryptUpdate() to now return a 118668651Skris value: although software versions of the algorithms cannot fail 118768651Skris any installed hardware versions can. 118868651Skris [Steve Henson] 118968651Skris 119068651Skris *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 119168651Skris this option is set, tolerate broken clients that send the negotiated 119268651Skris protocol version number instead of the requested protocol version 119368651Skris number. 119468651Skris [Bodo Moeller] 119568651Skris 119668651Skris *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 119768651Skris i.e. non-zero for export ciphersuites, zero otherwise. 119868651Skris Previous versions had this flag inverted, inconsistent with 119968651Skris rsa_tmp_cb (..._TMP_RSA_CB). 120068651Skris [Bodo Moeller; problem reported by Amit Chopra] 120168651Skris 120268651Skris *) Add missing DSA library text string. Work around for some IIS 120368651Skris key files with invalid SEQUENCE encoding. 120468651Skris [Steve Henson] 120568651Skris 120668651Skris *) Add a document (doc/standards.txt) that list all kinds of standards 120768651Skris and so on that are implemented in OpenSSL. 120868651Skris [Richard Levitte] 120968651Skris 121068651Skris *) Enhance c_rehash script. Old version would mishandle certificates 121168651Skris with the same subject name hash and wouldn't handle CRLs at all. 121268651Skris Added -fingerprint option to crl utility, to support new c_rehash 121368651Skris features. 121468651Skris [Steve Henson] 121568651Skris 121668651Skris *) Eliminate non-ANSI declarations in crypto.h and stack.h. 121768651Skris [Ulf M�ller] 121868651Skris 121968651Skris *) Fix for SSL server purpose checking. Server checking was 122068651Skris rejecting certificates which had extended key usage present 122168651Skris but no ssl client purpose. 122268651Skris [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 122368651Skris 122468651Skris *) Make PKCS#12 code work with no password. The PKCS#12 spec 122568651Skris is a little unclear about how a blank password is handled. 122668651Skris Since the password in encoded as a BMPString with terminating 122768651Skris double NULL a zero length password would end up as just the 122868651Skris double NULL. However no password at all is different and is 122968651Skris handled differently in the PKCS#12 key generation code. NS 123068651Skris treats a blank password as zero length. MSIE treats it as no 123168651Skris password on export: but it will try both on import. We now do 123268651Skris the same: PKCS12_parse() tries zero length and no password if 123368651Skris the password is set to "" or NULL (NULL is now a valid password: 123468651Skris it wasn't before) as does the pkcs12 application. 123568651Skris [Steve Henson] 123668651Skris 123768651Skris *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 123868651Skris perror when PEM_read_bio_X509_REQ fails, the error message must 123968651Skris be obtained from the error queue. 124068651Skris [Bodo Moeller] 124168651Skris 124268651Skris *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 124368651Skris it in ERR_remove_state if appropriate, and change ERR_get_state 124468651Skris accordingly to avoid race conditions (this is necessary because 124568651Skris thread_hash is no longer constant once set). 124668651Skris [Bodo Moeller] 124768651Skris 124868651Skris *) Bugfix for linux-elf makefile.one. 124968651Skris [Ulf M�ller] 125068651Skris 125168651Skris *) RSA_get_default_method() will now cause a default 125268651Skris RSA_METHOD to be chosen if one doesn't exist already. 125368651Skris Previously this was only set during a call to RSA_new() 125468651Skris or RSA_new_method(NULL) meaning it was possible for 125568651Skris RSA_get_default_method() to return NULL. 125668651Skris [Geoff Thorpe] 125768651Skris 125868651Skris *) Added native name translation to the existing DSO code 125968651Skris that will convert (if the flag to do so is set) filenames 126068651Skris that are sufficiently small and have no path information 126168651Skris into a canonical native form. Eg. "blah" converted to 126268651Skris "libblah.so" or "blah.dll" etc. 126368651Skris [Geoff Thorpe] 126468651Skris 126568651Skris *) New function ERR_error_string_n(e, buf, len) which is like 126668651Skris ERR_error_string(e, buf), but writes at most 'len' bytes 126768651Skris including the 0 terminator. For ERR_error_string_n, 'buf' 126868651Skris may not be NULL. 126968651Skris [Damien Miller <djm@mindrot.org>, Bodo Moeller] 127068651Skris 127168651Skris *) CONF library reworked to become more general. A new CONF 127268651Skris configuration file reader "class" is implemented as well as a 127368651Skris new functions (NCONF_*, for "New CONF") to handle it. The now 127468651Skris old CONF_* functions are still there, but are reimplemented to 127568651Skris work in terms of the new functions. Also, a set of functions 127668651Skris to handle the internal storage of the configuration data is 127768651Skris provided to make it easier to write new configuration file 127868651Skris reader "classes" (I can definitely see something reading a 127968651Skris configuration file in XML format, for example), called _CONF_*, 128068651Skris or "the configuration storage API"... 128168651Skris 128268651Skris The new configuration file reading functions are: 128368651Skris 128468651Skris NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 128568651Skris NCONF_get_section, NCONF_get_string, NCONF_get_numbre 128668651Skris 128768651Skris NCONF_default, NCONF_WIN32 128868651Skris 128968651Skris NCONF_dump_fp, NCONF_dump_bio 129068651Skris 129168651Skris NCONF_default and NCONF_WIN32 are method (or "class") choosers, 129268651Skris NCONF_new creates a new CONF object. This works in the same way 129368651Skris as other interfaces in OpenSSL, like the BIO interface. 129468651Skris NCONF_dump_* dump the internal storage of the configuration file, 129568651Skris which is useful for debugging. All other functions take the same 129668651Skris arguments as the old CONF_* functions wth the exception of the 129768651Skris first that must be a `CONF *' instead of a `LHASH *'. 129868651Skris 129968651Skris To make it easer to use the new classes with the old CONF_* functions, 130068651Skris the function CONF_set_default_method is provided. 130168651Skris [Richard Levitte] 130268651Skris 130368651Skris *) Add '-tls1' option to 'openssl ciphers', which was already 130468651Skris mentioned in the documentation but had not been implemented. 130568651Skris (This option is not yet really useful because even the additional 130668651Skris experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 130768651Skris [Bodo Moeller] 130868651Skris 130968651Skris *) Initial DSO code added into libcrypto for letting OpenSSL (and 131068651Skris OpenSSL-based applications) load shared libraries and bind to 131168651Skris them in a portable way. 131268651Skris [Geoff Thorpe, with contributions from Richard Levitte] 131368651Skris 131459191Skris Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 131559191Skris 131659191Skris *) Make sure _lrotl and _lrotr are only used with MSVC. 131759191Skris 131859191Skris *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 131959191Skris (the default implementation of RAND_status). 132059191Skris 132159191Skris *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 132259191Skris to '-clrext' (= clear extensions), as intended and documented. 132359191Skris [Bodo Moeller; inconsistency pointed out by Michael Attili 132459191Skris <attili@amaxo.com>] 132559191Skris 132659191Skris *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 132759191Skris was larger than the MD block size. 132859191Skris [Steve Henson, pointed out by Yost William <YostW@tce.com>] 132959191Skris 133059191Skris *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 133159191Skris fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 133259191Skris using the passed key: if the passed key was a private key the result 133359191Skris of X509_print(), for example, would be to print out all the private key 133459191Skris components. 133559191Skris [Steve Henson] 133659191Skris 133759191Skris *) des_quad_cksum() byte order bug fix. 133859191Skris [Ulf M�ller, using the problem description in krb4-0.9.7, where 133959191Skris the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 134059191Skris 134159191Skris *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 134259191Skris discouraged. 134359191Skris [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 134459191Skris 134559191Skris *) For easily testing in shell scripts whether some command 134659191Skris 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 134759191Skris returns with exit code 0 iff no command of the given name is available. 134859191Skris 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 134959191Skris the output goes to stdout and nothing is printed to stderr. 135059191Skris Additional arguments are always ignored. 135159191Skris 135259191Skris Since for each cipher there is a command of the same name, 135359191Skris the 'no-cipher' compilation switches can be tested this way. 135459191Skris 135559191Skris ('openssl no-XXX' is not able to detect pseudo-commands such 135659191Skris as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 135759191Skris [Bodo Moeller] 135859191Skris 135959191Skris *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 136059191Skris [Bodo Moeller] 136159191Skris 136259191Skris *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 136359191Skris is set; it will be thrown away anyway because each handshake creates 136459191Skris its own key. 136559191Skris ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 136659191Skris to parameters -- in previous versions (since OpenSSL 0.9.3) the 136759191Skris 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 136859191Skris you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 136959191Skris [Bodo Moeller] 137059191Skris 137159191Skris *) New s_client option -ign_eof: EOF at stdin is ignored, and 137259191Skris 'Q' and 'R' lose their special meanings (quit/renegotiate). 137359191Skris This is part of what -quiet does; unlike -quiet, -ign_eof 137459191Skris does not suppress any output. 137559191Skris [Richard Levitte] 137659191Skris 137759191Skris *) Add compatibility options to the purpose and trust code. The 137859191Skris purpose X509_PURPOSE_ANY is "any purpose" which automatically 137959191Skris accepts a certificate or CA, this was the previous behaviour, 138059191Skris with all the associated security issues. 138159191Skris 138259191Skris X509_TRUST_COMPAT is the old trust behaviour: only and 138359191Skris automatically trust self signed roots in certificate store. A 138459191Skris new trust setting X509_TRUST_DEFAULT is used to specify that 138559191Skris a purpose has no associated trust setting and it should instead 138659191Skris use the value in the default purpose. 138759191Skris [Steve Henson] 138859191Skris 138959191Skris *) Fix the PKCS#8 DSA private key code so it decodes keys again 139059191Skris and fix a memory leak. 139159191Skris [Steve Henson] 139259191Skris 139359191Skris *) In util/mkerr.pl (which implements 'make errors'), preserve 139459191Skris reason strings from the previous version of the .c file, as 139559191Skris the default to have only downcase letters (and digits) in 139659191Skris automatically generated reasons codes is not always appropriate. 139759191Skris [Bodo Moeller] 139859191Skris 139959191Skris *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 140059191Skris using strerror. Previously, ERR_reason_error_string() returned 140159191Skris library names as reason strings for SYSerr; but SYSerr is a special 140259191Skris case where small numbers are errno values, not library numbers. 140359191Skris [Bodo Moeller] 140459191Skris 140559191Skris *) Add '-dsaparam' option to 'openssl dhparam' application. This 140659191Skris converts DSA parameters into DH parameters. (When creating parameters, 140759191Skris DSA_generate_parameters is used.) 140859191Skris [Bodo Moeller] 140959191Skris 141059191Skris *) Include 'length' (recommended exponent length) in C code generated 141159191Skris by 'openssl dhparam -C'. 141259191Skris [Bodo Moeller] 141359191Skris 141459191Skris *) The second argument to set_label in perlasm was already being used 141559191Skris so couldn't be used as a "file scope" flag. Moved to third argument 141659191Skris which was free. 141759191Skris [Steve Henson] 141859191Skris 141959191Skris *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 142059191Skris instead of RAND_bytes for encryption IVs and salts. 142159191Skris [Bodo Moeller] 142259191Skris 142359191Skris *) Include RAND_status() into RAND_METHOD instead of implementing 142459191Skris it only for md_rand.c Otherwise replacing the PRNG by calling 142559191Skris RAND_set_rand_method would be impossible. 142659191Skris [Bodo Moeller] 142759191Skris 142859191Skris *) Don't let DSA_generate_key() enter an infinite loop if the random 142959191Skris number generation fails. 143059191Skris [Bodo Moeller] 143159191Skris 143259191Skris *) New 'rand' application for creating pseudo-random output. 143359191Skris [Bodo Moeller] 143459191Skris 143559191Skris *) Added configuration support for Linux/IA64 143659191Skris [Rolf Haberrecker <rolf@suse.de>] 143759191Skris 143859191Skris *) Assembler module support for Mingw32. 143959191Skris [Ulf M�ller] 144059191Skris 144159191Skris *) Shared library support for HPUX (in shlib/). 144259191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 144359191Skris 144459191Skris *) Shared library support for Solaris gcc. 144559191Skris [Lutz Behnke <behnke@trustcenter.de>] 144659191Skris 144759191Skris Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 144859191Skris 144959191Skris *) PKCS7_encrypt() was adding text MIME headers twice because they 145059191Skris were added manually and by SMIME_crlf_copy(). 145159191Skris [Steve Henson] 145259191Skris 145359191Skris *) In bntest.c don't call BN_rand with zero bits argument. 145459191Skris [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 145559191Skris 145659191Skris *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 145759191Skris case was implemented. This caused BN_div_recp() to fail occasionally. 145859191Skris [Ulf M�ller] 145959191Skris 146059191Skris *) Add an optional second argument to the set_label() in the perl 146159191Skris assembly language builder. If this argument exists and is set 146259191Skris to 1 it signals that the assembler should use a symbol whose 146359191Skris scope is the entire file, not just the current function. This 146459191Skris is needed with MASM which uses the format label:: for this scope. 146559191Skris [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 146659191Skris 146759191Skris *) Change the ASN1 types so they are typedefs by default. Before 146859191Skris almost all types were #define'd to ASN1_STRING which was causing 146959191Skris STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 147059191Skris for example. 147159191Skris [Steve Henson] 147259191Skris 147359191Skris *) Change names of new functions to the new get1/get0 naming 147459191Skris convention: After 'get1', the caller owns a reference count 147559191Skris and has to call ..._free; 'get0' returns a pointer to some 147659191Skris data structure without incrementing reference counters. 147759191Skris (Some of the existing 'get' functions increment a reference 147859191Skris counter, some don't.) 147959191Skris Similarly, 'set1' and 'add1' functions increase reference 148059191Skris counters or duplicate objects. 148159191Skris [Steve Henson] 148259191Skris 148359191Skris *) Allow for the possibility of temp RSA key generation failure: 148459191Skris the code used to assume it always worked and crashed on failure. 148559191Skris [Steve Henson] 148659191Skris 148759191Skris *) Fix potential buffer overrun problem in BIO_printf(). 148859191Skris [Ulf M�ller, using public domain code by Patrick Powell; problem 148959191Skris pointed out by David Sacerdote <das33@cornell.edu>] 149059191Skris 149159191Skris *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 149259191Skris RAND_egd() and RAND_status(). In the command line application, 149359191Skris the EGD socket can be specified like a seed file using RANDFILE 149459191Skris or -rand. 149559191Skris [Ulf M�ller] 149659191Skris 149759191Skris *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 149859191Skris Some CAs (e.g. Verisign) distribute certificates in this form. 149959191Skris [Steve Henson] 150059191Skris 150159191Skris *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 150259191Skris list to exclude them. This means that no special compilation option 150359191Skris is needed to use anonymous DH: it just needs to be included in the 150459191Skris cipher list. 150559191Skris [Steve Henson] 150659191Skris 150759191Skris *) Change the EVP_MD_CTX_type macro so its meaning consistent with 150859191Skris EVP_MD_type. The old functionality is available in a new macro called 150959191Skris EVP_MD_md(). Change code that uses it and update docs. 151059191Skris [Steve Henson] 151159191Skris 151259191Skris *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 151359191Skris where the 'void *' argument is replaced by a function pointer argument. 151459191Skris Previously 'void *' was abused to point to functions, which works on 151559191Skris many platforms, but is not correct. As these functions are usually 151659191Skris called by macros defined in OpenSSL header files, most source code 151759191Skris should work without changes. 151859191Skris [Richard Levitte] 151959191Skris 152059191Skris *) <openssl/opensslconf.h> (which is created by Configure) now contains 152159191Skris sections with information on -D... compiler switches used for 152259191Skris compiling the library so that applications can see them. To enable 152359191Skris one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 152459191Skris must be defined. E.g., 152559191Skris #define OPENSSL_ALGORITHM_DEFINES 152659191Skris #include <openssl/opensslconf.h> 152759191Skris defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 152859191Skris [Richard Levitte, Ulf and Bodo M�ller] 152959191Skris 153059191Skris *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 153159191Skris record layer. 153259191Skris [Bodo Moeller] 153359191Skris 153459191Skris *) Change the 'other' type in certificate aux info to a STACK_OF 153559191Skris X509_ALGOR. Although not an AlgorithmIdentifier as such it has 153659191Skris the required ASN1 format: arbitrary types determined by an OID. 153759191Skris [Steve Henson] 153859191Skris 153959191Skris *) Add some PEM_write_X509_REQ_NEW() functions and a command line 154059191Skris argument to 'req'. This is not because the function is newer or 154159191Skris better than others it just uses the work 'NEW' in the certificate 154259191Skris request header lines. Some software needs this. 154359191Skris [Steve Henson] 154459191Skris 154559191Skris *) Reorganise password command line arguments: now passwords can be 154659191Skris obtained from various sources. Delete the PEM_cb function and make 154759191Skris it the default behaviour: i.e. if the callback is NULL and the 154859191Skris usrdata argument is not NULL interpret it as a null terminated pass 154959191Skris phrase. If usrdata and the callback are NULL then the pass phrase 155059191Skris is prompted for as usual. 155159191Skris [Steve Henson] 155259191Skris 155359191Skris *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 155459191Skris the support is automatically enabled. The resulting binaries will 155559191Skris autodetect the card and use it if present. 155659191Skris [Ben Laurie and Compaq Inc.] 155759191Skris 155859191Skris *) Work around for Netscape hang bug. This sends certificate request 155959191Skris and server done in one record. Since this is perfectly legal in the 156059191Skris SSL/TLS protocol it isn't a "bug" option and is on by default. See 156159191Skris the bugs/SSLv3 entry for more info. 156259191Skris [Steve Henson] 156359191Skris 156459191Skris *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 156559191Skris [Andy Polyakov] 156659191Skris 156759191Skris *) Add -rand argument to smime and pkcs12 applications and read/write 156859191Skris of seed file. 156959191Skris [Steve Henson] 157059191Skris 157159191Skris *) New 'passwd' tool for crypt(3) and apr1 password hashes. 157259191Skris [Bodo Moeller] 157359191Skris 157459191Skris *) Add command line password options to the remaining applications. 157559191Skris [Steve Henson] 157659191Skris 157759191Skris *) Bug fix for BN_div_recp() for numerators with an even number of 157859191Skris bits. 157959191Skris [Ulf M�ller] 158059191Skris 158159191Skris *) More tests in bntest.c, and changed test_bn output. 158259191Skris [Ulf M�ller] 158359191Skris 158459191Skris *) ./config recognizes MacOS X now. 158559191Skris [Andy Polyakov] 158659191Skris 158759191Skris *) Bug fix for BN_div() when the first words of num and divsor are 158859191Skris equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 158959191Skris [Ulf M�ller] 159059191Skris 159159191Skris *) Add support for various broken PKCS#8 formats, and command line 159259191Skris options to produce them. 159359191Skris [Steve Henson] 159459191Skris 159559191Skris *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 159659191Skris get temporary BIGNUMs from a BN_CTX. 159759191Skris [Ulf M�ller] 159859191Skris 159959191Skris *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 160059191Skris for p == 0. 160159191Skris [Ulf M�ller] 160259191Skris 160359191Skris *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 160459191Skris include a #define from the old name to the new. The original intent 160559191Skris was that statically linked binaries could for example just call 160659191Skris SSLeay_add_all_ciphers() to just add ciphers to the table and not 160759191Skris link with digests. This never worked becayse SSLeay_add_all_digests() 160859191Skris and SSLeay_add_all_ciphers() were in the same source file so calling 160959191Skris one would link with the other. They are now in separate source files. 161059191Skris [Steve Henson] 161159191Skris 161259191Skris *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 161359191Skris [Steve Henson] 161459191Skris 161559191Skris *) Use a less unusual form of the Miller-Rabin primality test (it used 161659191Skris a binary algorithm for exponentiation integrated into the Miller-Rabin 161759191Skris loop, our standard modexp algorithms are faster). 161859191Skris [Bodo Moeller] 161959191Skris 162059191Skris *) Support for the EBCDIC character set completed. 162159191Skris [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 162259191Skris 162359191Skris *) Source code cleanups: use const where appropriate, eliminate casts, 162459191Skris use void * instead of char * in lhash. 162559191Skris [Ulf M�ller] 162659191Skris 162759191Skris *) Bugfix: ssl3_send_server_key_exchange was not restartable 162859191Skris (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 162959191Skris this the server could overwrite ephemeral keys that the client 163059191Skris has already seen). 163159191Skris [Bodo Moeller] 163259191Skris 163359191Skris *) Turn DSA_is_prime into a macro that calls BN_is_prime, 163459191Skris using 50 iterations of the Rabin-Miller test. 163559191Skris 163659191Skris DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 163759191Skris iterations of the Rabin-Miller test as required by the appendix 163859191Skris to FIPS PUB 186[-1]) instead of DSA_is_prime. 163959191Skris As BN_is_prime_fasttest includes trial division, DSA parameter 164059191Skris generation becomes much faster. 164159191Skris 164259191Skris This implies a change for the callback functions in DSA_is_prime 164359191Skris and DSA_generate_parameters: The callback function is called once 164459191Skris for each positive witness in the Rabin-Miller test, not just 164559191Skris occasionally in the inner loop; and the parameters to the 164659191Skris callback function now provide an iteration count for the outer 164759191Skris loop rather than for the current invocation of the inner loop. 164859191Skris DSA_generate_parameters additionally can call the callback 164959191Skris function with an 'iteration count' of -1, meaning that a 165059191Skris candidate has passed the trial division test (when q is generated 165159191Skris from an application-provided seed, trial division is skipped). 165259191Skris [Bodo Moeller] 165359191Skris 165459191Skris *) New function BN_is_prime_fasttest that optionally does trial 165559191Skris division before starting the Rabin-Miller test and has 165659191Skris an additional BN_CTX * argument (whereas BN_is_prime always 165759191Skris has to allocate at least one BN_CTX). 165859191Skris 'callback(1, -1, cb_arg)' is called when a number has passed the 165959191Skris trial division stage. 166059191Skris [Bodo Moeller] 166159191Skris 166259191Skris *) Fix for bug in CRL encoding. The validity dates weren't being handled 166359191Skris as ASN1_TIME. 166459191Skris [Steve Henson] 166559191Skris 166659191Skris *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 166759191Skris [Steve Henson] 166859191Skris 166959191Skris *) New function BN_pseudo_rand(). 167059191Skris [Ulf M�ller] 167159191Skris 167259191Skris *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 167359191Skris bignum version of BN_from_montgomery() with the working code from 167459191Skris SSLeay 0.9.0 (the word based version is faster anyway), and clean up 167559191Skris the comments. 167659191Skris [Ulf M�ller] 167759191Skris 167859191Skris *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 167959191Skris made it impossible to use the same SSL_SESSION data structure in 168059191Skris SSL2 clients in multiple threads. 168159191Skris [Bodo Moeller] 168259191Skris 168359191Skris *) The return value of RAND_load_file() no longer counts bytes obtained 168459191Skris by stat(). RAND_load_file(..., -1) is new and uses the complete file 168559191Skris to seed the PRNG (previously an explicit byte count was required). 168659191Skris [Ulf M�ller, Bodo M�ller] 168759191Skris 168859191Skris *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 168959191Skris used (char *) instead of (void *) and had casts all over the place. 169059191Skris [Steve Henson] 169159191Skris 169259191Skris *) Make BN_generate_prime() return NULL on error if ret!=NULL. 169359191Skris [Ulf M�ller] 169459191Skris 169559191Skris *) Retain source code compatibility for BN_prime_checks macro: 169659191Skris BN_is_prime(..., BN_prime_checks, ...) now uses 169759191Skris BN_prime_checks_for_size to determine the appropriate number of 169859191Skris Rabin-Miller iterations. 169959191Skris [Ulf M�ller] 170059191Skris 170159191Skris *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 170259191Skris DH_CHECK_P_NOT_SAFE_PRIME. 170359191Skris (Check if this is true? OpenPGP calls them "strong".) 170459191Skris [Ulf M�ller] 170559191Skris 170659191Skris *) Merge the functionality of "dh" and "gendh" programs into a new program 170759191Skris "dhparam". The old programs are retained for now but will handle DH keys 170859191Skris (instead of parameters) in future. 170959191Skris [Steve Henson] 171059191Skris 171159191Skris *) Make the ciphers, s_server and s_client programs check the return values 171259191Skris when a new cipher list is set. 171359191Skris [Steve Henson] 171459191Skris 171559191Skris *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 171659191Skris ciphers. Before when the 56bit ciphers were enabled the sorting was 171759191Skris wrong. 171859191Skris 171959191Skris The syntax for the cipher sorting has been extended to support sorting by 172059191Skris cipher-strength (using the strength_bits hard coded in the tables). 172159191Skris The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 172259191Skris 172359191Skris Fix a bug in the cipher-command parser: when supplying a cipher command 172459191Skris string with an "undefined" symbol (neither command nor alphanumeric 172559191Skris [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 172659191Skris an error is flagged. 172759191Skris 172859191Skris Due to the strength-sorting extension, the code of the 172959191Skris ssl_create_cipher_list() function was completely rearranged. I hope that 173059191Skris the readability was also increased :-) 173159191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 173259191Skris 173359191Skris *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 173459191Skris for the first serial number and places 2 in the serial number file. This 173559191Skris avoids problems when the root CA is created with serial number zero and 173659191Skris the first user certificate has the same issuer name and serial number 173759191Skris as the root CA. 173859191Skris [Steve Henson] 173959191Skris 174059191Skris *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 174159191Skris the new code. Add documentation for this stuff. 174259191Skris [Steve Henson] 174359191Skris 174459191Skris *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 174559191Skris X509_*() to X509at_*() on the grounds that they don't handle X509 174659191Skris structures and behave in an analagous way to the X509v3 functions: 174759191Skris they shouldn't be called directly but wrapper functions should be used 174859191Skris instead. 174959191Skris 175059191Skris So we also now have some wrapper functions that call the X509at functions 175159191Skris when passed certificate requests. (TO DO: similar things can be done with 175259191Skris PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 175359191Skris things. Some of these need some d2i or i2d and print functionality 175459191Skris because they handle more complex structures.) 175559191Skris [Steve Henson] 175659191Skris 175759191Skris *) Add missing #ifndefs that caused missing symbols when building libssl 175859191Skris as a shared library without RSA. Use #ifndef NO_SSL2 instead of 175959191Skris NO_RSA in ssl/s2*.c. 176059191Skris [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 176159191Skris 176259191Skris *) Precautions against using the PRNG uninitialized: RAND_bytes() now 176359191Skris has a return value which indicates the quality of the random data 176459191Skris (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 176559191Skris error queue. New function RAND_pseudo_bytes() generates output that is 176659191Skris guaranteed to be unique but not unpredictable. RAND_add is like 176759191Skris RAND_seed, but takes an extra argument for an entropy estimate 176859191Skris (RAND_seed always assumes full entropy). 176959191Skris [Ulf M�ller] 177059191Skris 177159191Skris *) Do more iterations of Rabin-Miller probable prime test (specifically, 177259191Skris 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 177359191Skris instead of only 2 for all lengths; see BN_prime_checks_for_size definition 177459191Skris in crypto/bn/bn_prime.c for the complete table). This guarantees a 177559191Skris false-positive rate of at most 2^-80 for random input. 177659191Skris [Bodo Moeller] 177759191Skris 177859191Skris *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 177959191Skris [Bodo Moeller] 178059191Skris 178159191Skris *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 178259191Skris in the 0.9.5 release), this returns the chain 178359191Skris from an X509_CTX structure with a dup of the stack and all 178459191Skris the X509 reference counts upped: so the stack will exist 178559191Skris after X509_CTX_cleanup() has been called. Modify pkcs12.c 178659191Skris to use this. 178759191Skris 178859191Skris Also make SSL_SESSION_print() print out the verify return 178959191Skris code. 179059191Skris [Steve Henson] 179159191Skris 179259191Skris *) Add manpage for the pkcs12 command. Also change the default 179359191Skris behaviour so MAC iteration counts are used unless the new 179459191Skris -nomaciter option is used. This improves file security and 179559191Skris only older versions of MSIE (4.0 for example) need it. 179659191Skris [Steve Henson] 179759191Skris 179859191Skris *) Honor the no-xxx Configure options when creating .DEF files. 179959191Skris [Ulf M�ller] 180059191Skris 180159191Skris *) Add PKCS#10 attributes to field table: challengePassword, 180259191Skris unstructuredName and unstructuredAddress. These are taken from 180359191Skris draft PKCS#9 v2.0 but are compatible with v1.2 provided no 180459191Skris international characters are used. 180559191Skris 180659191Skris More changes to X509_ATTRIBUTE code: allow the setting of types 180759191Skris based on strings. Remove the 'loc' parameter when adding 180859191Skris attributes because these will be a SET OF encoding which is sorted 180959191Skris in ASN1 order. 181059191Skris [Steve Henson] 181159191Skris 181259191Skris *) Initial changes to the 'req' utility to allow request generation 181359191Skris automation. This will allow an application to just generate a template 181459191Skris file containing all the field values and have req construct the 181559191Skris request. 181659191Skris 181759191Skris Initial support for X509_ATTRIBUTE handling. Stacks of these are 181859191Skris used all over the place including certificate requests and PKCS#7 181959191Skris structures. They are currently handled manually where necessary with 182059191Skris some primitive wrappers for PKCS#7. The new functions behave in a 182159191Skris manner analogous to the X509 extension functions: they allow 182259191Skris attributes to be looked up by NID and added. 182359191Skris 182459191Skris Later something similar to the X509V3 code would be desirable to 182559191Skris automatically handle the encoding, decoding and printing of the 182659191Skris more complex types. The string types like challengePassword can 182759191Skris be handled by the string table functions. 182859191Skris 182959191Skris Also modified the multi byte string table handling. Now there is 183059191Skris a 'global mask' which masks out certain types. The table itself 183159191Skris can use the flag STABLE_NO_MASK to ignore the mask setting: this 183259191Skris is useful when for example there is only one permissible type 183359191Skris (as in countryName) and using the mask might result in no valid 183459191Skris types at all. 183559191Skris [Steve Henson] 183659191Skris 183759191Skris *) Clean up 'Finished' handling, and add functions SSL_get_finished and 183859191Skris SSL_get_peer_finished to allow applications to obtain the latest 183959191Skris Finished messages sent to the peer or expected from the peer, 184059191Skris respectively. (SSL_get_peer_finished is usually the Finished message 184159191Skris actually received from the peer, otherwise the protocol will be aborted.) 184259191Skris 184359191Skris As the Finished message are message digests of the complete handshake 184459191Skris (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 184559191Skris be used for external authentication procedures when the authentication 184659191Skris provided by SSL/TLS is not desired or is not enough. 184759191Skris [Bodo Moeller] 184859191Skris 184959191Skris *) Enhanced support for Alpha Linux is added. Now ./config checks if 185059191Skris the host supports BWX extension and if Compaq C is present on the 185159191Skris $PATH. Just exploiting of the BWX extension results in 20-30% 185259191Skris performance kick for some algorithms, e.g. DES and RC4 to mention 185359191Skris a couple. Compaq C in turn generates ~20% faster code for MD5 and 185459191Skris SHA1. 185559191Skris [Andy Polyakov] 185659191Skris 185759191Skris *) Add support for MS "fast SGC". This is arguably a violation of the 185859191Skris SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 185959191Skris weak crypto and after checking the certificate is SGC a second one 186059191Skris with strong crypto. MS SGC stops the first handshake after receiving 186159191Skris the server certificate message and sends a second client hello. Since 186259191Skris a server will typically do all the time consuming operations before 186359191Skris expecting any further messages from the client (server key exchange 186459191Skris is the most expensive) there is little difference between the two. 186559191Skris 186659191Skris To get OpenSSL to support MS SGC we have to permit a second client 186759191Skris hello message after we have sent server done. In addition we have to 186859191Skris reset the MAC if we do get this second client hello. 186959191Skris [Steve Henson] 187059191Skris 187159191Skris *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 187259191Skris if a DER encoded private key is RSA or DSA traditional format. Changed 187359191Skris d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 187459191Skris format DER encoded private key. Newer code should use PKCS#8 format which 187559191Skris has the key type encoded in the ASN1 structure. Added DER private key 187659191Skris support to pkcs8 application. 187759191Skris [Steve Henson] 187859191Skris 187959191Skris *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 188059191Skris ciphersuites has been selected (as required by the SSL 3/TLS 1 188159191Skris specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 188259191Skris is set, we interpret this as a request to violate the specification 188359191Skris (the worst that can happen is a handshake failure, and 'correct' 188459191Skris behaviour would result in a handshake failure anyway). 188559191Skris [Bodo Moeller] 188659191Skris 188759191Skris *) In SSL_CTX_add_session, take into account that there might be multiple 188859191Skris SSL_SESSION structures with the same session ID (e.g. when two threads 188959191Skris concurrently obtain them from an external cache). 189059191Skris The internal cache can handle only one SSL_SESSION with a given ID, 189159191Skris so if there's a conflict, we now throw out the old one to achieve 189259191Skris consistency. 189359191Skris [Bodo Moeller] 189459191Skris 189559191Skris *) Add OIDs for idea and blowfish in CBC mode. This will allow both 189659191Skris to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 189759191Skris some routines that use cipher OIDs: some ciphers do not have OIDs 189859191Skris defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 189959191Skris example. 190059191Skris [Steve Henson] 190159191Skris 190259191Skris *) Simplify the trust setting structure and code. Now we just have 190359191Skris two sequences of OIDs for trusted and rejected settings. These will 190459191Skris typically have values the same as the extended key usage extension 190559191Skris and any application specific purposes. 190659191Skris 190759191Skris The trust checking code now has a default behaviour: it will just 190859191Skris check for an object with the same NID as the passed id. Functions can 190959191Skris be provided to override either the default behaviour or the behaviour 191059191Skris for a given id. SSL client, server and email already have functions 191159191Skris in place for compatibility: they check the NID and also return "trusted" 191259191Skris if the certificate is self signed. 191359191Skris [Steve Henson] 191459191Skris 191559191Skris *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 191659191Skris traditional format into an EVP_PKEY structure. 191759191Skris [Steve Henson] 191859191Skris 191959191Skris *) Add a password callback function PEM_cb() which either prompts for 192059191Skris a password if usr_data is NULL or otherwise assumes it is a null 192159191Skris terminated password. Allow passwords to be passed on command line 192259191Skris environment or config files in a few more utilities. 192359191Skris [Steve Henson] 192459191Skris 192559191Skris *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 192659191Skris keys. Add some short names for PKCS#8 PBE algorithms and allow them 192759191Skris to be specified on the command line for the pkcs8 and pkcs12 utilities. 192859191Skris Update documentation. 192959191Skris [Steve Henson] 193059191Skris 193159191Skris *) Support for ASN1 "NULL" type. This could be handled before by using 193259191Skris ASN1_TYPE but there wasn't any function that would try to read a NULL 193359191Skris and produce an error if it couldn't. For compatibility we also have 193459191Skris ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 193559191Skris don't allocate anything because they don't need to. 193659191Skris [Steve Henson] 193759191Skris 193859191Skris *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 193959191Skris for details. 194059191Skris [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 194159191Skris 194259191Skris *) Rebuild of the memory allocation routines used by OpenSSL code and 194359191Skris possibly others as well. The purpose is to make an interface that 194459191Skris provide hooks so anyone can build a separate set of allocation and 194559191Skris deallocation routines to be used by OpenSSL, for example memory 194659191Skris pool implementations, or something else, which was previously hard 194759191Skris since Malloc(), Realloc() and Free() were defined as macros having 194859191Skris the values malloc, realloc and free, respectively (except for Win32 194959191Skris compilations). The same is provided for memory debugging code. 195059191Skris OpenSSL already comes with functionality to find memory leaks, but 195159191Skris this gives people a chance to debug other memory problems. 195259191Skris 195359191Skris With these changes, a new set of functions and macros have appeared: 195459191Skris 195568651Skris CRYPTO_set_mem_debug_functions() [F] 195668651Skris CRYPTO_get_mem_debug_functions() [F] 195768651Skris CRYPTO_dbg_set_options() [F] 195868651Skris CRYPTO_dbg_get_options() [F] 195968651Skris CRYPTO_malloc_debug_init() [M] 196059191Skris 196159191Skris The memory debug functions are NULL by default, unless the library 196259191Skris is compiled with CRYPTO_MDEBUG or friends is defined. If someone 196359191Skris wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 196459191Skris gives the standard debugging functions that come with OpenSSL) or 196559191Skris CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 196659191Skris provided by the library user) must be used. When the standard 196759191Skris debugging functions are used, CRYPTO_dbg_set_options can be used to 196859191Skris request additional information: 196959191Skris CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 197059191Skris the CRYPTO_MDEBUG_xxx macro when compiling the library. 197159191Skris 197259191Skris Also, things like CRYPTO_set_mem_functions will always give the 197359191Skris expected result (the new set of functions is used for allocation 197459191Skris and deallocation) at all times, regardless of platform and compiler 197559191Skris options. 197659191Skris 197759191Skris To finish it up, some functions that were never use in any other 197859191Skris way than through macros have a new API and new semantic: 197959191Skris 198059191Skris CRYPTO_dbg_malloc() 198159191Skris CRYPTO_dbg_realloc() 198259191Skris CRYPTO_dbg_free() 198359191Skris 198459191Skris All macros of value have retained their old syntax. 198559191Skris [Richard Levitte and Bodo Moeller] 198659191Skris 198759191Skris *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 198859191Skris ordering of SMIMECapabilities wasn't in "strength order" and there 198959191Skris was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 199059191Skris algorithm. 199159191Skris [Steve Henson] 199259191Skris 199359191Skris *) Some ASN1 types with illegal zero length encoding (INTEGER, 199459191Skris ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 199559191Skris [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 199659191Skris 199759191Skris *) Merge in my S/MIME library for OpenSSL. This provides a simple 199859191Skris S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 199959191Skris functionality to handle multipart/signed properly) and a utility 200059191Skris called 'smime' to call all this stuff. This is based on code I 200159191Skris originally wrote for Celo who have kindly allowed it to be 200259191Skris included in OpenSSL. 200359191Skris [Steve Henson] 200459191Skris 200559191Skris *) Add variants des_set_key_checked and des_set_key_unchecked of 200659191Skris des_set_key (aka des_key_sched). Global variable des_check_key 200759191Skris decides which of these is called by des_set_key; this way 200859191Skris des_check_key behaves as it always did, but applications and 200959191Skris the library itself, which was buggy for des_check_key == 1, 201059191Skris have a cleaner way to pick the version they need. 201159191Skris [Bodo Moeller] 201259191Skris 201359191Skris *) New function PKCS12_newpass() which changes the password of a 201459191Skris PKCS12 structure. 201559191Skris [Steve Henson] 201659191Skris 201759191Skris *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 201859191Skris dynamic mix. In both cases the ids can be used as an index into the 201959191Skris table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 202059191Skris functions so they accept a list of the field values and the 202159191Skris application doesn't need to directly manipulate the X509_TRUST 202259191Skris structure. 202359191Skris [Steve Henson] 202459191Skris 202559191Skris *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 202659191Skris need initialising. 202759191Skris [Steve Henson] 202859191Skris 202959191Skris *) Modify the way the V3 extension code looks up extensions. This now 203059191Skris works in a similar way to the object code: we have some "standard" 203159191Skris extensions in a static table which is searched with OBJ_bsearch() 203259191Skris and the application can add dynamic ones if needed. The file 203359191Skris crypto/x509v3/ext_dat.h now has the info: this file needs to be 203459191Skris updated whenever a new extension is added to the core code and kept 203559191Skris in ext_nid order. There is a simple program 'tabtest.c' which checks 203659191Skris this. New extensions are not added too often so this file can readily 203759191Skris be maintained manually. 203859191Skris 203959191Skris There are two big advantages in doing things this way. The extensions 204059191Skris can be looked up immediately and no longer need to be "added" using 204159191Skris X509V3_add_standard_extensions(): this function now does nothing. 204259191Skris [Side note: I get *lots* of email saying the extension code doesn't 204359191Skris work because people forget to call this function] 204459191Skris Also no dynamic allocation is done unless new extensions are added: 204559191Skris so if we don't add custom extensions there is no need to call 204659191Skris X509V3_EXT_cleanup(). 204759191Skris [Steve Henson] 204859191Skris 204959191Skris *) Modify enc utility's salting as follows: make salting the default. Add a 205059191Skris magic header, so unsalted files fail gracefully instead of just decrypting 205159191Skris to garbage. This is because not salting is a big security hole, so people 205259191Skris should be discouraged from doing it. 205359191Skris [Ben Laurie] 205459191Skris 205559191Skris *) Fixes and enhancements to the 'x509' utility. It allowed a message 205659191Skris digest to be passed on the command line but it only used this 205759191Skris parameter when signing a certificate. Modified so all relevant 205859191Skris operations are affected by the digest parameter including the 205959191Skris -fingerprint and -x509toreq options. Also -x509toreq choked if a 206059191Skris DSA key was used because it didn't fix the digest. 206159191Skris [Steve Henson] 206259191Skris 206359191Skris *) Initial certificate chain verify code. Currently tests the untrusted 206459191Skris certificates for consistency with the verify purpose (which is set 206559191Skris when the X509_STORE_CTX structure is set up) and checks the pathlength. 206659191Skris 206759191Skris There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 206859191Skris this is because it will reject chains with invalid extensions whereas 206959191Skris every previous version of OpenSSL and SSLeay made no checks at all. 207059191Skris 207159191Skris Trust code: checks the root CA for the relevant trust settings. Trust 207259191Skris settings have an initial value consistent with the verify purpose: e.g. 207359191Skris if the verify purpose is for SSL client use it expects the CA to be 207459191Skris trusted for SSL client use. However the default value can be changed to 207559191Skris permit custom trust settings: one example of this would be to only trust 207659191Skris certificates from a specific "secure" set of CAs. 207759191Skris 207859191Skris Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 207959191Skris which should be used for version portability: especially since the 208059191Skris verify structure is likely to change more often now. 208159191Skris 208259191Skris SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 208359191Skris to set them. If not set then assume SSL clients will verify SSL servers 208459191Skris and vice versa. 208559191Skris 208659191Skris Two new options to the verify program: -untrusted allows a set of 208759191Skris untrusted certificates to be passed in and -purpose which sets the 208859191Skris intended purpose of the certificate. If a purpose is set then the 208959191Skris new chain verify code is used to check extension consistency. 209059191Skris [Steve Henson] 209159191Skris 209259191Skris *) Support for the authority information access extension. 209359191Skris [Steve Henson] 209459191Skris 209559191Skris *) Modify RSA and DSA PEM read routines to transparently handle 209659191Skris PKCS#8 format private keys. New *_PUBKEY_* functions that handle 209759191Skris public keys in a format compatible with certificate 209859191Skris SubjectPublicKeyInfo structures. Unfortunately there were already 209959191Skris functions called *_PublicKey_* which used various odd formats so 210059191Skris these are retained for compatibility: however the DSA variants were 210159191Skris never in a public release so they have been deleted. Changed dsa/rsa 210259191Skris utilities to handle the new format: note no releases ever handled public 210359191Skris keys so we should be OK. 210459191Skris 210559191Skris The primary motivation for this change is to avoid the same fiasco 210659191Skris that dogs private keys: there are several incompatible private key 210759191Skris formats some of which are standard and some OpenSSL specific and 210859191Skris require various evil hacks to allow partial transparent handling and 210959191Skris even then it doesn't work with DER formats. Given the option anything 211059191Skris other than PKCS#8 should be dumped: but the other formats have to 211159191Skris stay in the name of compatibility. 211259191Skris 211359191Skris With public keys and the benefit of hindsight one standard format 211459191Skris is used which works with EVP_PKEY, RSA or DSA structures: though 211559191Skris it clearly returns an error if you try to read the wrong kind of key. 211659191Skris 211759191Skris Added a -pubkey option to the 'x509' utility to output the public key. 211859191Skris Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 211959191Skris (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 212059191Skris EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 212159191Skris that do the same as the EVP_PKEY_assign_*() except they up the 212259191Skris reference count of the added key (they don't "swallow" the 212359191Skris supplied key). 212459191Skris [Steve Henson] 212559191Skris 212659191Skris *) Fixes to crypto/x509/by_file.c the code to read in certificates and 212759191Skris CRLs would fail if the file contained no certificates or no CRLs: 212859191Skris added a new function to read in both types and return the number 212959191Skris read: this means that if none are read it will be an error. The 213059191Skris DER versions of the certificate and CRL reader would always fail 213159191Skris because it isn't possible to mix certificates and CRLs in DER format 213259191Skris without choking one or the other routine. Changed this to just read 213359191Skris a certificate: this is the best we can do. Also modified the code 213459191Skris in apps/verify.c to take notice of return codes: it was previously 213559191Skris attempting to read in certificates from NULL pointers and ignoring 213659191Skris any errors: this is one reason why the cert and CRL reader seemed 213759191Skris to work. It doesn't check return codes from the default certificate 213859191Skris routines: these may well fail if the certificates aren't installed. 213959191Skris [Steve Henson] 214059191Skris 214159191Skris *) Code to support otherName option in GeneralName. 214259191Skris [Steve Henson] 214359191Skris 214459191Skris *) First update to verify code. Change the verify utility 214559191Skris so it warns if it is passed a self signed certificate: 214659191Skris for consistency with the normal behaviour. X509_verify 214759191Skris has been modified to it will now verify a self signed 214859191Skris certificate if *exactly* the same certificate appears 214959191Skris in the store: it was previously impossible to trust a 215059191Skris single self signed certificate. This means that: 215159191Skris openssl verify ss.pem 215259191Skris now gives a warning about a self signed certificate but 215359191Skris openssl verify -CAfile ss.pem ss.pem 215459191Skris is OK. 215559191Skris [Steve Henson] 215659191Skris 215759191Skris *) For servers, store verify_result in SSL_SESSION data structure 215859191Skris (and add it to external session representation). 215959191Skris This is needed when client certificate verifications fails, 216059191Skris but an application-provided verification callback (set by 216159191Skris SSL_CTX_set_cert_verify_callback) allows accepting the session 216259191Skris anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 216359191Skris but returns 1): When the session is reused, we have to set 216459191Skris ssl->verify_result to the appropriate error code to avoid 216559191Skris security holes. 216659191Skris [Bodo Moeller, problem pointed out by Lutz Jaenicke] 216759191Skris 216859191Skris *) Fix a bug in the new PKCS#7 code: it didn't consider the 216959191Skris case in PKCS7_dataInit() where the signed PKCS7 structure 217059191Skris didn't contain any existing data because it was being created. 217159191Skris [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 217259191Skris 217359191Skris *) Add a salt to the key derivation routines in enc.c. This 217459191Skris forms the first 8 bytes of the encrypted file. Also add a 217559191Skris -S option to allow a salt to be input on the command line. 217659191Skris [Steve Henson] 217759191Skris 217859191Skris *) New function X509_cmp(). Oddly enough there wasn't a function 217959191Skris to compare two certificates. We do this by working out the SHA1 218059191Skris hash and comparing that. X509_cmp() will be needed by the trust 218159191Skris code. 218259191Skris [Steve Henson] 218359191Skris 218459191Skris *) SSL_get1_session() is like SSL_get_session(), but increments 218559191Skris the reference count in the SSL_SESSION returned. 218659191Skris [Geoff Thorpe <geoff@eu.c2.net>] 218759191Skris 218859191Skris *) Fix for 'req': it was adding a null to request attributes. 218959191Skris Also change the X509_LOOKUP and X509_INFO code to handle 219059191Skris certificate auxiliary information. 219159191Skris [Steve Henson] 219259191Skris 219359191Skris *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 219459191Skris the 'enc' command. 219559191Skris [Steve Henson] 219659191Skris 219759191Skris *) Add the possibility to add extra information to the memory leak 219859191Skris detecting output, to form tracebacks, showing from where each 219959191Skris allocation was originated: CRYPTO_push_info("constant string") adds 220059191Skris the string plus current file name and line number to a per-thread 220159191Skris stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 220259191Skris is like calling CYRPTO_pop_info() until the stack is empty. 220359191Skris Also updated memory leak detection code to be multi-thread-safe. 220459191Skris [Richard Levitte] 220559191Skris 220659191Skris *) Add options -text and -noout to pkcs7 utility and delete the 220759191Skris encryption options which never did anything. Update docs. 220859191Skris [Steve Henson] 220959191Skris 221059191Skris *) Add options to some of the utilities to allow the pass phrase 221159191Skris to be included on either the command line (not recommended on 221259191Skris OSes like Unix) or read from the environment. Update the 221359191Skris manpages and fix a few bugs. 221459191Skris [Steve Henson] 221559191Skris 221659191Skris *) Add a few manpages for some of the openssl commands. 221759191Skris [Steve Henson] 221859191Skris 221959191Skris *) Fix the -revoke option in ca. It was freeing up memory twice, 222059191Skris leaking and not finding already revoked certificates. 222159191Skris [Steve Henson] 222259191Skris 222359191Skris *) Extensive changes to support certificate auxiliary information. 222459191Skris This involves the use of X509_CERT_AUX structure and X509_AUX 222559191Skris functions. An X509_AUX function such as PEM_read_X509_AUX() 222659191Skris can still read in a certificate file in the usual way but it 222759191Skris will also read in any additional "auxiliary information". By 222859191Skris doing things this way a fair degree of compatibility can be 222959191Skris retained: existing certificates can have this information added 223059191Skris using the new 'x509' options. 223159191Skris 223259191Skris Current auxiliary information includes an "alias" and some trust 223359191Skris settings. The trust settings will ultimately be used in enhanced 223459191Skris certificate chain verification routines: currently a certificate 223559191Skris can only be trusted if it is self signed and then it is trusted 223659191Skris for all purposes. 223759191Skris [Steve Henson] 223859191Skris 223959191Skris *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 224059191Skris The problem was that one of the replacement routines had not been working 224159191Skris since SSLeay releases. For now the offending routine has been replaced 224259191Skris with non-optimised assembler. Even so, this now gives around 95% 224359191Skris performance improvement for 1024 bit RSA signs. 224459191Skris [Mark Cox] 224559191Skris 224659191Skris *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 224759191Skris handling. Most clients have the effective key size in bits equal to 224859191Skris the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 224959191Skris A few however don't do this and instead use the size of the decrypted key 225059191Skris to determine the RC2 key length and the AlgorithmIdentifier to determine 225159191Skris the effective key length. In this case the effective key length can still 225259191Skris be 40 bits but the key length can be 168 bits for example. This is fixed 225359191Skris by manually forcing an RC2 key into the EVP_PKEY structure because the 225459191Skris EVP code can't currently handle unusual RC2 key sizes: it always assumes 225559191Skris the key length and effective key length are equal. 225659191Skris [Steve Henson] 225759191Skris 225859191Skris *) Add a bunch of functions that should simplify the creation of 225959191Skris X509_NAME structures. Now you should be able to do: 226059191Skris X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 226159191Skris and have it automatically work out the correct field type and fill in 226259191Skris the structures. The more adventurous can try: 226359191Skris X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 226459191Skris and it will (hopefully) work out the correct multibyte encoding. 226559191Skris [Steve Henson] 226659191Skris 226759191Skris *) Change the 'req' utility to use the new field handling and multibyte 226859191Skris copy routines. Before the DN field creation was handled in an ad hoc 226959191Skris way in req, ca, and x509 which was rather broken and didn't support 227059191Skris BMPStrings or UTF8Strings. Since some software doesn't implement 227159191Skris BMPStrings or UTF8Strings yet, they can be enabled using the config file 227259191Skris using the dirstring_type option. See the new comment in the default 227359191Skris openssl.cnf for more info. 227459191Skris [Steve Henson] 227559191Skris 227659191Skris *) Make crypto/rand/md_rand.c more robust: 227759191Skris - Assure unique random numbers after fork(). 227859191Skris - Make sure that concurrent threads access the global counter and 227959191Skris md serializably so that we never lose entropy in them 228059191Skris or use exactly the same state in multiple threads. 228159191Skris Access to the large state is not always serializable because 228259191Skris the additional locking could be a performance killer, and 228359191Skris md should be large enough anyway. 228459191Skris [Bodo Moeller] 228559191Skris 228659191Skris *) New file apps/app_rand.c with commonly needed functionality 228759191Skris for handling the random seed file. 228859191Skris 228959191Skris Use the random seed file in some applications that previously did not: 229059191Skris ca, 229159191Skris dsaparam -genkey (which also ignored its '-rand' option), 229259191Skris s_client, 229359191Skris s_server, 229459191Skris x509 (when signing). 229559191Skris Except on systems with /dev/urandom, it is crucial to have a random 229659191Skris seed file at least for key creation, DSA signing, and for DH exchanges; 229759191Skris for RSA signatures we could do without one. 229859191Skris 229959191Skris gendh and gendsa (unlike genrsa) used to read only the first byte 230059191Skris of each file listed in the '-rand' option. The function as previously 230159191Skris found in genrsa is now in app_rand.c and is used by all programs 230259191Skris that support '-rand'. 230359191Skris [Bodo Moeller] 230459191Skris 230559191Skris *) In RAND_write_file, use mode 0600 for creating files; 230659191Skris don't just chmod when it may be too late. 230759191Skris [Bodo Moeller] 230859191Skris 230959191Skris *) Report an error from X509_STORE_load_locations 231059191Skris when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 231159191Skris [Bill Perry] 231259191Skris 231359191Skris *) New function ASN1_mbstring_copy() this copies a string in either 231459191Skris ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 231559191Skris into an ASN1_STRING type. A mask of permissible types is passed 231659191Skris and it chooses the "minimal" type to use or an error if not type 231759191Skris is suitable. 231859191Skris [Steve Henson] 231959191Skris 232059191Skris *) Add function equivalents to the various macros in asn1.h. The old 232159191Skris macros are retained with an M_ prefix. Code inside the library can 232259191Skris use the M_ macros. External code (including the openssl utility) 232359191Skris should *NOT* in order to be "shared library friendly". 232459191Skris [Steve Henson] 232559191Skris 232659191Skris *) Add various functions that can check a certificate's extensions 232759191Skris to see if it usable for various purposes such as SSL client, 232859191Skris server or S/MIME and CAs of these types. This is currently 232959191Skris VERY EXPERIMENTAL but will ultimately be used for certificate chain 233059191Skris verification. Also added a -purpose flag to x509 utility to 233159191Skris print out all the purposes. 233259191Skris [Steve Henson] 233359191Skris 233459191Skris *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 233559191Skris functions. 233659191Skris [Steve Henson] 233759191Skris 233859191Skris *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 233959191Skris for, obtain and decode and extension and obtain its critical flag. 234059191Skris This allows all the necessary extension code to be handled in a 234159191Skris single function call. 234259191Skris [Steve Henson] 234359191Skris 234459191Skris *) RC4 tune-up featuring 30-40% performance improvement on most RISC 234559191Skris platforms. See crypto/rc4/rc4_enc.c for further details. 234659191Skris [Andy Polyakov] 234759191Skris 234859191Skris *) New -noout option to asn1parse. This causes no output to be produced 234959191Skris its main use is when combined with -strparse and -out to extract data 235059191Skris from a file (which may not be in ASN.1 format). 235159191Skris [Steve Henson] 235259191Skris 235359191Skris *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 235459191Skris when producing the local key id. 235559191Skris [Richard Levitte <levitte@stacken.kth.se>] 235659191Skris 235759191Skris *) New option -dhparam in s_server. This allows a DH parameter file to be 235859191Skris stated explicitly. If it is not stated then it tries the first server 235959191Skris certificate file. The previous behaviour hard coded the filename 236059191Skris "server.pem". 236159191Skris [Steve Henson] 236259191Skris 236359191Skris *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 236459191Skris a public key to be input or output. For example: 236559191Skris openssl rsa -in key.pem -pubout -out pubkey.pem 236659191Skris Also added necessary DSA public key functions to handle this. 236759191Skris [Steve Henson] 236859191Skris 236959191Skris *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 237059191Skris in the message. This was handled by allowing 237159191Skris X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 237259191Skris [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 237359191Skris 237459191Skris *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 237559191Skris to the end of the strings whereas this didn't. This would cause problems 237659191Skris if strings read with d2i_ASN1_bytes() were later modified. 237759191Skris [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 237859191Skris 237959191Skris *) Fix for base64 decode bug. When a base64 bio reads only one line of 238059191Skris data and it contains EOF it will end up returning an error. This is 238159191Skris caused by input 46 bytes long. The cause is due to the way base64 238259191Skris BIOs find the start of base64 encoded data. They do this by trying a 238359191Skris trial decode on each line until they find one that works. When they 238459191Skris do a flag is set and it starts again knowing it can pass all the 238559191Skris data directly through the decoder. Unfortunately it doesn't reset 238659191Skris the context it uses. This means that if EOF is reached an attempt 238759191Skris is made to pass two EOFs through the context and this causes the 238859191Skris resulting error. This can also cause other problems as well. As is 238959191Skris usual with these problems it takes *ages* to find and the fix is 239059191Skris trivial: move one line. 239159191Skris [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 239259191Skris 239359191Skris *) Ugly workaround to get s_client and s_server working under Windows. The 239459191Skris old code wouldn't work because it needed to select() on sockets and the 239559191Skris tty (for keypresses and to see if data could be written). Win32 only 239659191Skris supports select() on sockets so we select() with a 1s timeout on the 239759191Skris sockets and then see if any characters are waiting to be read, if none 239859191Skris are present then we retry, we also assume we can always write data to 239959191Skris the tty. This isn't nice because the code then blocks until we've 240059191Skris received a complete line of data and it is effectively polling the 240159191Skris keyboard at 1s intervals: however it's quite a bit better than not 240259191Skris working at all :-) A dedicated Windows application might handle this 240359191Skris with an event loop for example. 240459191Skris [Steve Henson] 240559191Skris 240659191Skris *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 240759191Skris and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 240859191Skris will be called when RSA_sign() and RSA_verify() are used. This is useful 240959191Skris if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 241059191Skris For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 241159191Skris should *not* be used: RSA_sign() and RSA_verify() must be used instead. 241259191Skris This necessitated the support of an extra signature type NID_md5_sha1 241359191Skris for SSL signatures and modifications to the SSL library to use it instead 241459191Skris of calling RSA_public_decrypt() and RSA_private_encrypt(). 241559191Skris [Steve Henson] 241659191Skris 241759191Skris *) Add new -verify -CAfile and -CApath options to the crl program, these 241859191Skris will lookup a CRL issuers certificate and verify the signature in a 241959191Skris similar way to the verify program. Tidy up the crl program so it 242059191Skris no longer accesses structures directly. Make the ASN1 CRL parsing a bit 242159191Skris less strict. It will now permit CRL extensions even if it is not 242259191Skris a V2 CRL: this will allow it to tolerate some broken CRLs. 242359191Skris [Steve Henson] 242459191Skris 242559191Skris *) Initialize all non-automatic variables each time one of the openssl 242659191Skris sub-programs is started (this is necessary as they may be started 242759191Skris multiple times from the "OpenSSL>" prompt). 242859191Skris [Lennart Bang, Bodo Moeller] 242959191Skris 243059191Skris *) Preliminary compilation option RSA_NULL which disables RSA crypto without 243159191Skris removing all other RSA functionality (this is what NO_RSA does). This 243259191Skris is so (for example) those in the US can disable those operations covered 243359191Skris by the RSA patent while allowing storage and parsing of RSA keys and RSA 243459191Skris key generation. 243559191Skris [Steve Henson] 243659191Skris 243759191Skris *) Non-copying interface to BIO pairs. 243859191Skris (still largely untested) 243959191Skris [Bodo Moeller] 244059191Skris 244159191Skris *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 244259191Skris ASCII string. This was handled independently in various places before. 244359191Skris [Steve Henson] 244459191Skris 244559191Skris *) New functions UTF8_getc() and UTF8_putc() that parse and generate 244659191Skris UTF8 strings a character at a time. 244759191Skris [Steve Henson] 244859191Skris 244959191Skris *) Use client_version from client hello to select the protocol 245059191Skris (s23_srvr.c) and for RSA client key exchange verification 245159191Skris (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 245259191Skris [Bodo Moeller] 245359191Skris 245459191Skris *) Add various utility functions to handle SPKACs, these were previously 245559191Skris handled by poking round in the structure internals. Added new function 245659191Skris NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 245759191Skris print, verify and generate SPKACs. Based on an original idea from 245859191Skris Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 245959191Skris [Steve Henson] 246059191Skris 246159191Skris *) RIPEMD160 is operational on all platforms and is back in 'make test'. 246259191Skris [Andy Polyakov] 246359191Skris 246459191Skris *) Allow the config file extension section to be overwritten on the 246559191Skris command line. Based on an original idea from Massimiliano Pala 246659191Skris <madwolf@comune.modena.it>. The new option is called -extensions 246759191Skris and can be applied to ca, req and x509. Also -reqexts to override 246859191Skris the request extensions in req and -crlexts to override the crl extensions 246959191Skris in ca. 247059191Skris [Steve Henson] 247159191Skris 247259191Skris *) Add new feature to the SPKAC handling in ca. Now you can include 247359191Skris the same field multiple times by preceding it by "XXXX." for example: 247459191Skris 1.OU="Unit name 1" 247559191Skris 2.OU="Unit name 2" 247659191Skris this is the same syntax as used in the req config file. 247759191Skris [Steve Henson] 247859191Skris 247959191Skris *) Allow certificate extensions to be added to certificate requests. These 248059191Skris are specified in a 'req_extensions' option of the req section of the 248159191Skris config file. They can be printed out with the -text option to req but 248259191Skris are otherwise ignored at present. 248359191Skris [Steve Henson] 248459191Skris 248559191Skris *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 248659191Skris data read consists of only the final block it would not decrypted because 248759191Skris EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 248859191Skris A misplaced 'break' also meant the decrypted final block might not be 248959191Skris copied until the next read. 249059191Skris [Steve Henson] 249159191Skris 249259191Skris *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 249359191Skris a few extra parameters to the DH structure: these will be useful if 249459191Skris for example we want the value of 'q' or implement X9.42 DH. 249559191Skris [Steve Henson] 249659191Skris 249759191Skris *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 249859191Skris provides hooks that allow the default DSA functions or functions on a 249959191Skris "per key" basis to be replaced. This allows hardware acceleration and 250059191Skris hardware key storage to be handled without major modification to the 250159191Skris library. Also added low level modexp hooks and CRYPTO_EX structure and 250259191Skris associated functions. 250359191Skris [Steve Henson] 250459191Skris 250559191Skris *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 250659191Skris as "read only": it can't be written to and the buffer it points to will 250759191Skris not be freed. Reading from a read only BIO is much more efficient than 250859191Skris a normal memory BIO. This was added because there are several times when 250959191Skris an area of memory needs to be read from a BIO. The previous method was 251059191Skris to create a memory BIO and write the data to it, this results in two 251159191Skris copies of the data and an O(n^2) reading algorithm. There is a new 251259191Skris function BIO_new_mem_buf() which creates a read only memory BIO from 251359191Skris an area of memory. Also modified the PKCS#7 routines to use read only 251459191Skris memory BIOs. 251559191Skris [Steve Henson] 251659191Skris 251759191Skris *) Bugfix: ssl23_get_client_hello did not work properly when called in 251859191Skris state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 251959191Skris a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 252059191Skris but a retry condition occured while trying to read the rest. 252159191Skris [Bodo Moeller] 252259191Skris 252359191Skris *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 252459191Skris NID_pkcs7_encrypted by default: this was wrong since this should almost 252559191Skris always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 252659191Skris the encrypted data type: this is a more sensible place to put it and it 252759191Skris allows the PKCS#12 code to be tidied up that duplicated this 252859191Skris functionality. 252959191Skris [Steve Henson] 253059191Skris 253159191Skris *) Changed obj_dat.pl script so it takes its input and output files on 253259191Skris the command line. This should avoid shell escape redirection problems 253359191Skris under Win32. 253459191Skris [Steve Henson] 253559191Skris 253659191Skris *) Initial support for certificate extension requests, these are included 253759191Skris in things like Xenroll certificate requests. Included functions to allow 253859191Skris extensions to be obtained and added. 253959191Skris [Steve Henson] 254059191Skris 254159191Skris *) -crlf option to s_client and s_server for sending newlines as 254259191Skris CRLF (as required by many protocols). 254359191Skris [Bodo Moeller] 254459191Skris 254555714Skris Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 254655714Skris 254755714Skris *) Install libRSAglue.a when OpenSSL is built with RSAref. 254855714Skris [Ralf S. Engelschall] 254955714Skris 255055714Skris *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 255155714Skris [Andrija Antonijevic <TheAntony2@bigfoot.com>] 255255714Skris 255355714Skris *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 255455714Skris program. 255555714Skris [Steve Henson] 255655714Skris 255755714Skris *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 255855714Skris DH parameters/keys (q is lost during that conversion, but the resulting 255955714Skris DH parameters contain its length). 256055714Skris 256155714Skris For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 256255714Skris much faster than DH_generate_parameters (which creates parameters 256355714Skris where p = 2*q + 1), and also the smaller q makes DH computations 256455714Skris much more efficient (160-bit exponentiation instead of 1024-bit 256555714Skris exponentiation); so this provides a convenient way to support DHE 256655714Skris ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 256755714Skris utter importance to use 256855714Skris SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 256955714Skris or 257055714Skris SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 257155714Skris when such DH parameters are used, because otherwise small subgroup 257255714Skris attacks may become possible! 257355714Skris [Bodo Moeller] 257455714Skris 257555714Skris *) Avoid memory leak in i2d_DHparams. 257655714Skris [Bodo Moeller] 257755714Skris 257855714Skris *) Allow the -k option to be used more than once in the enc program: 257955714Skris this allows the same encrypted message to be read by multiple recipients. 258055714Skris [Steve Henson] 258155714Skris 258255714Skris *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 258355714Skris an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 258455714Skris it will always use the numerical form of the OID, even if it has a short 258555714Skris or long name. 258655714Skris [Steve Henson] 258755714Skris 258855714Skris *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 258955714Skris method only got called if p,q,dmp1,dmq1,iqmp components were present, 259055714Skris otherwise bn_mod_exp was called. In the case of hardware keys for example 259155714Skris no private key components need be present and it might store extra data 259259191Skris in the RSA structure, which cannot be accessed from bn_mod_exp. 259359191Skris By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 259459191Skris private key operations. 259555714Skris [Steve Henson] 259655714Skris 259755714Skris *) Added support for SPARC Linux. 259855714Skris [Andy Polyakov] 259955714Skris 260055714Skris *) pem_password_cb function type incompatibly changed from 260155714Skris typedef int pem_password_cb(char *buf, int size, int rwflag); 260255714Skris to 260355714Skris ....(char *buf, int size, int rwflag, void *userdata); 260455714Skris so that applications can pass data to their callbacks: 260555714Skris The PEM[_ASN1]_{read,write}... functions and macros now take an 260655714Skris additional void * argument, which is just handed through whenever 260755714Skris the password callback is called. 260859191Skris [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 260955714Skris 261055714Skris New function SSL_CTX_set_default_passwd_cb_userdata. 261155714Skris 261255714Skris Compatibility note: As many C implementations push function arguments 261355714Skris onto the stack in reverse order, the new library version is likely to 261455714Skris interoperate with programs that have been compiled with the old 261555714Skris pem_password_cb definition (PEM_whatever takes some data that 261655714Skris happens to be on the stack as its last argument, and the callback 261755714Skris just ignores this garbage); but there is no guarantee whatsoever that 261855714Skris this will work. 261955714Skris 262055714Skris *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 262155714Skris (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 262255714Skris problems not only on Windows, but also on some Unix platforms. 262355714Skris To avoid problematic command lines, these definitions are now in an 262455714Skris auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 262555714Skris for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 262655714Skris [Bodo Moeller] 262755714Skris 262855714Skris *) MIPS III/IV assembler module is reimplemented. 262955714Skris [Andy Polyakov] 263055714Skris 263155714Skris *) More DES library cleanups: remove references to srand/rand and 263255714Skris delete an unused file. 263355714Skris [Ulf M�ller] 263455714Skris 263555714Skris *) Add support for the the free Netwide assembler (NASM) under Win32, 263655714Skris since not many people have MASM (ml) and it can be hard to obtain. 263755714Skris This is currently experimental but it seems to work OK and pass all 263855714Skris the tests. Check out INSTALL.W32 for info. 263955714Skris [Steve Henson] 264055714Skris 264155714Skris *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 264255714Skris without temporary keys kept an extra copy of the server key, 264355714Skris and connections with temporary keys did not free everything in case 264455714Skris of an error. 264555714Skris [Bodo Moeller] 264655714Skris 264755714Skris *) New function RSA_check_key and new openssl rsa option -check 264855714Skris for verifying the consistency of RSA keys. 264955714Skris [Ulf Moeller, Bodo Moeller] 265055714Skris 265155714Skris *) Various changes to make Win32 compile work: 265255714Skris 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 265355714Skris 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 265455714Skris comparison" warnings. 265555714Skris 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 265655714Skris [Steve Henson] 265755714Skris 265855714Skris *) Add a debugging option to PKCS#5 v2 key generation function: when 265955714Skris you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 266055714Skris derived keys are printed to stderr. 266155714Skris [Steve Henson] 266255714Skris 266355714Skris *) Copy the flags in ASN1_STRING_dup(). 266455714Skris [Roman E. Pavlov <pre@mo.msk.ru>] 266555714Skris 266655714Skris *) The x509 application mishandled signing requests containing DSA 266755714Skris keys when the signing key was also DSA and the parameters didn't match. 266855714Skris 266955714Skris It was supposed to omit the parameters when they matched the signing key: 267055714Skris the verifying software was then supposed to automatically use the CA's 267155714Skris parameters if they were absent from the end user certificate. 267255714Skris 267355714Skris Omitting parameters is no longer recommended. The test was also 267455714Skris the wrong way round! This was probably due to unusual behaviour in 267555714Skris EVP_cmp_parameters() which returns 1 if the parameters match. 267655714Skris This meant that parameters were omitted when they *didn't* match and 267755714Skris the certificate was useless. Certificates signed with 'ca' didn't have 267855714Skris this bug. 267955714Skris [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 268055714Skris 268155714Skris *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 268255714Skris The interface is as follows: 268355714Skris Applications can use 268455714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 268555714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 268655714Skris "off" is now the default. 268755714Skris The library internally uses 268855714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 268955714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 269055714Skris to disable memory-checking temporarily. 269155714Skris 269255714Skris Some inconsistent states that previously were possible (and were 269355714Skris even the default) are now avoided. 269455714Skris 269555714Skris -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 269655714Skris with each memory chunk allocated; this is occasionally more helpful 269755714Skris than just having a counter. 269855714Skris 269955714Skris -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 270055714Skris 270155714Skris -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 270255714Skris extensions. 270355714Skris [Bodo Moeller] 270455714Skris 270555714Skris *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 270655714Skris which largely parallels "options", but is for changing API behaviour, 270755714Skris whereas "options" are about protocol behaviour. 270855714Skris Initial "mode" flags are: 270955714Skris 271055714Skris SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 271155714Skris a single record has been written. 271255714Skris SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 271355714Skris retries use the same buffer location. 271455714Skris (But all of the contents must be 271555714Skris copied!) 271655714Skris [Bodo Moeller] 271755714Skris 271879998Skris *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 271955714Skris worked. 272055714Skris 272155714Skris *) Fix problems with no-hmac etc. 272255714Skris [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 272355714Skris 272455714Skris *) New functions RSA_get_default_method(), RSA_set_method() and 272555714Skris RSA_get_method(). These allows replacement of RSA_METHODs without having 272655714Skris to mess around with the internals of an RSA structure. 272755714Skris [Steve Henson] 272855714Skris 272955714Skris *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 273055714Skris Also really enable memory leak checks in openssl.c and in some 273155714Skris test programs. 273255714Skris [Chad C. Mulligan, Bodo Moeller] 273355714Skris 273455714Skris *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 273555714Skris up the length of negative integers. This has now been simplified to just 273655714Skris store the length when it is first determined and use it later, rather 273755714Skris than trying to keep track of where data is copied and updating it to 273855714Skris point to the end. 273955714Skris [Steve Henson, reported by Brien Wheeler 274055714Skris <bwheeler@authentica-security.com>] 274155714Skris 274255714Skris *) Add a new function PKCS7_signatureVerify. This allows the verification 274355714Skris of a PKCS#7 signature but with the signing certificate passed to the 274455714Skris function itself. This contrasts with PKCS7_dataVerify which assumes the 274555714Skris certificate is present in the PKCS#7 structure. This isn't always the 274655714Skris case: certificates can be omitted from a PKCS#7 structure and be 274755714Skris distributed by "out of band" means (such as a certificate database). 274855714Skris [Steve Henson] 274955714Skris 275055714Skris *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 275155714Skris function prototypes in pem.h, also change util/mkdef.pl to add the 275255714Skris necessary function names. 275355714Skris [Steve Henson] 275455714Skris 275555714Skris *) mk1mf.pl (used by Windows builds) did not properly read the 275655714Skris options set by Configure in the top level Makefile, and Configure 275755714Skris was not even able to write more than one option correctly. 275855714Skris Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 275955714Skris [Bodo Moeller] 276055714Skris 276155714Skris *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 276255714Skris file to be loaded from a BIO or FILE pointer. The BIO version will 276355714Skris for example allow memory BIOs to contain config info. 276455714Skris [Steve Henson] 276555714Skris 276655714Skris *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 276755714Skris Whoever hopes to achieve shared-library compatibility across versions 276855714Skris must use this, not the compile-time macro. 276955714Skris (Exercise 0.9.4: Which is the minimum library version required by 277055714Skris such programs?) 277155714Skris Note: All this applies only to multi-threaded programs, others don't 277255714Skris need locks. 277355714Skris [Bodo Moeller] 277455714Skris 277555714Skris *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 277655714Skris through a BIO pair triggered the default case, i.e. 277755714Skris SSLerr(...,SSL_R_UNKNOWN_STATE). 277855714Skris [Bodo Moeller] 277955714Skris 278055714Skris *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 278155714Skris can use the SSL library even if none of the specific BIOs is 278255714Skris appropriate. 278355714Skris [Bodo Moeller] 278455714Skris 278555714Skris *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 278655714Skris for the encoded length. 278755714Skris [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 278855714Skris 278955714Skris *) Add initial documentation of the X509V3 functions. 279055714Skris [Steve Henson] 279155714Skris 279255714Skris *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 279355714Skris PEM_write_bio_PKCS8PrivateKey() that are equivalent to 279455714Skris PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 279555714Skris secure PKCS#8 private key format with a high iteration count. 279655714Skris [Steve Henson] 279755714Skris 279855714Skris *) Fix determination of Perl interpreter: A perl or perl5 279955714Skris _directory_ in $PATH was also accepted as the interpreter. 280055714Skris [Ralf S. Engelschall] 280155714Skris 280255714Skris *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 280355714Skris wrong with it but it was very old and did things like calling 280455714Skris PEM_ASN1_read() directly and used MD5 for the hash not to mention some 280555714Skris unusual formatting. 280655714Skris [Steve Henson] 280755714Skris 280855714Skris *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 280955714Skris to use the new extension code. 281055714Skris [Steve Henson] 281155714Skris 281255714Skris *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 281355714Skris with macros. This should make it easier to change their form, add extra 281455714Skris arguments etc. Fix a few PEM prototypes which didn't have cipher as a 281555714Skris constant. 281655714Skris [Steve Henson] 281755714Skris 281855714Skris *) Add to configuration table a new entry that can specify an alternative 281955714Skris name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 282055714Skris according to Mark Crispin <MRC@Panda.COM>. 282155714Skris [Bodo Moeller] 282255714Skris 282355714Skris#if 0 282455714Skris *) DES CBC did not update the IV. Weird. 282555714Skris [Ben Laurie] 282655714Skris#else 282755714Skris des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 282855714Skris Changing the behaviour of the former might break existing programs -- 282955714Skris where IV updating is needed, des_ncbc_encrypt can be used. 283055714Skris#endif 283155714Skris 283255714Skris *) When bntest is run from "make test" it drives bc to check its 283355714Skris calculations, as well as internally checking them. If an internal check 283455714Skris fails, it needs to cause bc to give a non-zero result or make test carries 283555714Skris on without noticing the failure. Fixed. 283655714Skris [Ben Laurie] 283755714Skris 283855714Skris *) DES library cleanups. 283955714Skris [Ulf M�ller] 284055714Skris 284155714Skris *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 284255714Skris used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 284355714Skris ciphers. NOTE: although the key derivation function has been verified 284455714Skris against some published test vectors it has not been extensively tested 284555714Skris yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 284655714Skris of v2.0. 284755714Skris [Steve Henson] 284855714Skris 284955714Skris *) Instead of "mkdir -p", which is not fully portable, use new 285055714Skris Perl script "util/mkdir-p.pl". 285155714Skris [Bodo Moeller] 285255714Skris 285355714Skris *) Rewrite the way password based encryption (PBE) is handled. It used to 285455714Skris assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 285555714Skris structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 285655714Skris but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 285755714Skris the 'parameter' field of the AlgorithmIdentifier is passed to the 285855714Skris underlying key generation function so it must do its own ASN1 parsing. 285955714Skris This has also changed the EVP_PBE_CipherInit() function which now has a 286055714Skris 'parameter' argument instead of literal salt and iteration count values 286155714Skris and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 286255714Skris [Steve Henson] 286355714Skris 286455714Skris *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 286555714Skris and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 286655714Skris Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 286755714Skris KEY" because this clashed with PKCS#8 unencrypted string. Since this 286855714Skris value was just used as a "magic string" and not used directly its 286955714Skris value doesn't matter. 287055714Skris [Steve Henson] 287155714Skris 287255714Skris *) Introduce some semblance of const correctness to BN. Shame C doesn't 287355714Skris support mutable. 287455714Skris [Ben Laurie] 287555714Skris 287655714Skris *) "linux-sparc64" configuration (ultrapenguin). 287755714Skris [Ray Miller <ray.miller@oucs.ox.ac.uk>] 287855714Skris "linux-sparc" configuration. 287955714Skris [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 288055714Skris 288155714Skris *) config now generates no-xxx options for missing ciphers. 288255714Skris [Ulf M�ller] 288355714Skris 288455714Skris *) Support the EBCDIC character set (work in progress). 288555714Skris File ebcdic.c not yet included because it has a different license. 288655714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 288755714Skris 288855714Skris *) Support BS2000/OSD-POSIX. 288955714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 289055714Skris 289155714Skris *) Make callbacks for key generation use void * instead of char *. 289255714Skris [Ben Laurie] 289355714Skris 289455714Skris *) Make S/MIME samples compile (not yet tested). 289555714Skris [Ben Laurie] 289655714Skris 289755714Skris *) Additional typesafe stacks. 289855714Skris [Ben Laurie] 289955714Skris 290055714Skris *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 290155714Skris [Bodo Moeller] 290255714Skris 290355714Skris 290455714Skris Changes between 0.9.3 and 0.9.3a [29 May 1999] 290555714Skris 290655714Skris *) New configuration variant "sco5-gcc". 290755714Skris 290855714Skris *) Updated some demos. 290955714Skris [Sean O Riordain, Wade Scholine] 291055714Skris 291155714Skris *) Add missing BIO_free at exit of pkcs12 application. 291255714Skris [Wu Zhigang] 291355714Skris 291455714Skris *) Fix memory leak in conf.c. 291555714Skris [Steve Henson] 291655714Skris 291755714Skris *) Updates for Win32 to assembler version of MD5. 291855714Skris [Steve Henson] 291955714Skris 292055714Skris *) Set #! path to perl in apps/der_chop to where we found it 292155714Skris instead of using a fixed path. 292255714Skris [Bodo Moeller] 292355714Skris 292455714Skris *) SHA library changes for irix64-mips4-cc. 292555714Skris [Andy Polyakov] 292655714Skris 292755714Skris *) Improvements for VMS support. 292855714Skris [Richard Levitte] 292955714Skris 293055714Skris 293155714Skris Changes between 0.9.2b and 0.9.3 [24 May 1999] 293255714Skris 293355714Skris *) Bignum library bug fix. IRIX 6 passes "make test" now! 293455714Skris This also avoids the problems with SC4.2 and unpatched SC5. 293555714Skris [Andy Polyakov <appro@fy.chalmers.se>] 293655714Skris 293755714Skris *) New functions sk_num, sk_value and sk_set to replace the previous macros. 293855714Skris These are required because of the typesafe stack would otherwise break 293955714Skris existing code. If old code used a structure member which used to be STACK 294055714Skris and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 294155714Skris sk_num or sk_value it would produce an error because the num, data members 294255714Skris are not present in STACK_OF. Now it just produces a warning. sk_set 294355714Skris replaces the old method of assigning a value to sk_value 294455714Skris (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 294555714Skris that does this will no longer work (and should use sk_set instead) but 294655714Skris this could be regarded as a "questionable" behaviour anyway. 294755714Skris [Steve Henson] 294855714Skris 294955714Skris *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 295055714Skris correctly handle encrypted S/MIME data. 295155714Skris [Steve Henson] 295255714Skris 295355714Skris *) Change type of various DES function arguments from des_cblock 295455714Skris (which means, in function argument declarations, pointer to char) 295555714Skris to des_cblock * (meaning pointer to array with 8 char elements), 295655714Skris which allows the compiler to do more typechecking; it was like 295755714Skris that back in SSLeay, but with lots of ugly casts. 295855714Skris 295955714Skris Introduce new type const_des_cblock. 296055714Skris [Bodo Moeller] 296155714Skris 296255714Skris *) Reorganise the PKCS#7 library and get rid of some of the more obvious 296355714Skris problems: find RecipientInfo structure that matches recipient certificate 296455714Skris and initialise the ASN1 structures properly based on passed cipher. 296555714Skris [Steve Henson] 296655714Skris 296755714Skris *) Belatedly make the BN tests actually check the results. 296855714Skris [Ben Laurie] 296955714Skris 297055714Skris *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 297155714Skris to and from BNs: it was completely broken. New compilation option 297255714Skris NEG_PUBKEY_BUG to allow for some broken certificates that encode public 297355714Skris key elements as negative integers. 297455714Skris [Steve Henson] 297555714Skris 297655714Skris *) Reorganize and speed up MD5. 297755714Skris [Andy Polyakov <appro@fy.chalmers.se>] 297855714Skris 297955714Skris *) VMS support. 298055714Skris [Richard Levitte <richard@levitte.org>] 298155714Skris 298255714Skris *) New option -out to asn1parse to allow the parsed structure to be 298355714Skris output to a file. This is most useful when combined with the -strparse 298455714Skris option to examine the output of things like OCTET STRINGS. 298555714Skris [Steve Henson] 298655714Skris 298755714Skris *) Make SSL library a little more fool-proof by not requiring any longer 298855714Skris that SSL_set_{accept,connect}_state be called before 298955714Skris SSL_{accept,connect} may be used (SSL_set_..._state is omitted 299055714Skris in many applications because usually everything *appeared* to work as 299155714Skris intended anyway -- now it really works as intended). 299255714Skris [Bodo Moeller] 299355714Skris 299455714Skris *) Move openssl.cnf out of lib/. 299555714Skris [Ulf M�ller] 299655714Skris 299755714Skris *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 299855714Skris -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 299955714Skris -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 300055714Skris [Ralf S. Engelschall] 300155714Skris 300255714Skris *) Various fixes to the EVP and PKCS#7 code. It may now be able to 300355714Skris handle PKCS#7 enveloped data properly. 300455714Skris [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 300555714Skris 300655714Skris *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 300755714Skris copying pointers. The cert_st handling is changed by this in 300855714Skris various ways (and thus what used to be known as ctx->default_cert 300955714Skris is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 301055714Skris any longer when s->cert does not give us what we need). 301155714Skris ssl_cert_instantiate becomes obsolete by this change. 301255714Skris As soon as we've got the new code right (possibly it already is?), 301355714Skris we have solved a couple of bugs of the earlier code where s->cert 301455714Skris was used as if it could not have been shared with other SSL structures. 301555714Skris 301655714Skris Note that using the SSL API in certain dirty ways now will result 301755714Skris in different behaviour than observed with earlier library versions: 301855714Skris Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 301955714Skris does not influence s as it used to. 302055714Skris 302155714Skris In order to clean up things more thoroughly, inside SSL_SESSION 302255714Skris we don't use CERT any longer, but a new structure SESS_CERT 302355714Skris that holds per-session data (if available); currently, this is 302455714Skris the peer's certificate chain and, for clients, the server's certificate 302555714Skris and temporary key. CERT holds only those values that can have 302655714Skris meaningful defaults in an SSL_CTX. 302755714Skris [Bodo Moeller] 302855714Skris 302955714Skris *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 303055714Skris from the internal representation. Various PKCS#7 fixes: remove some 303155714Skris evil casts and set the enc_dig_alg field properly based on the signing 303255714Skris key type. 303355714Skris [Steve Henson] 303455714Skris 303555714Skris *) Allow PKCS#12 password to be set from the command line or the 303655714Skris environment. Let 'ca' get its config file name from the environment 303755714Skris variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 303855714Skris and 'x509'). 303955714Skris [Steve Henson] 304055714Skris 304155714Skris *) Allow certificate policies extension to use an IA5STRING for the 304255714Skris organization field. This is contrary to the PKIX definition but 304355714Skris VeriSign uses it and IE5 only recognises this form. Document 'x509' 304455714Skris extension option. 304555714Skris [Steve Henson] 304655714Skris 304755714Skris *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 304855714Skris without disallowing inline assembler and the like for non-pedantic builds. 304955714Skris [Ben Laurie] 305055714Skris 305155714Skris *) Support Borland C++ builder. 305255714Skris [Janez Jere <jj@void.si>, modified by Ulf M�ller] 305355714Skris 305455714Skris *) Support Mingw32. 305555714Skris [Ulf M�ller] 305655714Skris 305755714Skris *) SHA-1 cleanups and performance enhancements. 305855714Skris [Andy Polyakov <appro@fy.chalmers.se>] 305955714Skris 306055714Skris *) Sparc v8plus assembler for the bignum library. 306155714Skris [Andy Polyakov <appro@fy.chalmers.se>] 306255714Skris 306355714Skris *) Accept any -xxx and +xxx compiler options in Configure. 306455714Skris [Ulf M�ller] 306555714Skris 306655714Skris *) Update HPUX configuration. 306755714Skris [Anonymous] 306855714Skris 306955714Skris *) Add missing sk_<type>_unshift() function to safestack.h 307055714Skris [Ralf S. Engelschall] 307155714Skris 307255714Skris *) New function SSL_CTX_use_certificate_chain_file that sets the 307355714Skris "extra_cert"s in addition to the certificate. (This makes sense 307455714Skris only for "PEM" format files, as chains as a whole are not 307555714Skris DER-encoded.) 307655714Skris [Bodo Moeller] 307755714Skris 307855714Skris *) Support verify_depth from the SSL API. 307955714Skris x509_vfy.c had what can be considered an off-by-one-error: 308055714Skris Its depth (which was not part of the external interface) 308155714Skris was actually counting the number of certificates in a chain; 308255714Skris now it really counts the depth. 308355714Skris [Bodo Moeller] 308455714Skris 308555714Skris *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 308655714Skris instead of X509err, which often resulted in confusing error 308755714Skris messages since the error codes are not globally unique 308855714Skris (e.g. an alleged error in ssl3_accept when a certificate 308955714Skris didn't match the private key). 309055714Skris 309155714Skris *) New function SSL_CTX_set_session_id_context that allows to set a default 309255714Skris value (so that you don't need SSL_set_session_id_context for each 309355714Skris connection using the SSL_CTX). 309455714Skris [Bodo Moeller] 309555714Skris 309655714Skris *) OAEP decoding bug fix. 309755714Skris [Ulf M�ller] 309855714Skris 309955714Skris *) Support INSTALL_PREFIX for package builders, as proposed by 310055714Skris David Harris. 310155714Skris [Bodo Moeller] 310255714Skris 310355714Skris *) New Configure options "threads" and "no-threads". For systems 310455714Skris where the proper compiler options are known (currently Solaris 310555714Skris and Linux), "threads" is the default. 310655714Skris [Bodo Moeller] 310755714Skris 310855714Skris *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 310955714Skris [Bodo Moeller] 311055714Skris 311155714Skris *) Install various scripts to $(OPENSSLDIR)/misc, not to 311255714Skris $(INSTALLTOP)/bin -- they shouldn't clutter directories 311355714Skris such as /usr/local/bin. 311455714Skris [Bodo Moeller] 311555714Skris 311655714Skris *) "make linux-shared" to build shared libraries. 311755714Skris [Niels Poppe <niels@netbox.org>] 311855714Skris 311955714Skris *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 312055714Skris [Ulf M�ller] 312155714Skris 312255714Skris *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 312355714Skris extension adding in x509 utility. 312455714Skris [Steve Henson] 312555714Skris 312655714Skris *) Remove NOPROTO sections and error code comments. 312755714Skris [Ulf M�ller] 312855714Skris 312955714Skris *) Partial rewrite of the DEF file generator to now parse the ANSI 313055714Skris prototypes. 313155714Skris [Steve Henson] 313255714Skris 313355714Skris *) New Configure options --prefix=DIR and --openssldir=DIR. 313455714Skris [Ulf M�ller] 313555714Skris 313655714Skris *) Complete rewrite of the error code script(s). It is all now handled 313755714Skris by one script at the top level which handles error code gathering, 313855714Skris header rewriting and C source file generation. It should be much better 313955714Skris than the old method: it now uses a modified version of Ulf's parser to 314055714Skris read the ANSI prototypes in all header files (thus the old K&R definitions 314155714Skris aren't needed for error creation any more) and do a better job of 314255714Skris translating function codes into names. The old 'ASN1 error code imbedded 314355714Skris in a comment' is no longer necessary and it doesn't use .err files which 314455714Skris have now been deleted. Also the error code call doesn't have to appear all 314555714Skris on one line (which resulted in some large lines...). 314655714Skris [Steve Henson] 314755714Skris 314855714Skris *) Change #include filenames from <foo.h> to <openssl/foo.h>. 314955714Skris [Bodo Moeller] 315055714Skris 315155714Skris *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 315255714Skris 0 (which usually indicates a closed connection), but continue reading. 315355714Skris [Bodo Moeller] 315455714Skris 315555714Skris *) Fix some race conditions. 315655714Skris [Bodo Moeller] 315755714Skris 315855714Skris *) Add support for CRL distribution points extension. Add Certificate 315955714Skris Policies and CRL distribution points documentation. 316055714Skris [Steve Henson] 316155714Skris 316255714Skris *) Move the autogenerated header file parts to crypto/opensslconf.h. 316355714Skris [Ulf M�ller] 316455714Skris 316555714Skris *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 316655714Skris 8 of keying material. Merlin has also confirmed interop with this fix 316755714Skris between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 316855714Skris [Merlin Hughes <merlin@baltimore.ie>] 316955714Skris 317055714Skris *) Fix lots of warnings. 317155714Skris [Richard Levitte <levitte@stacken.kth.se>] 317255714Skris 317355714Skris *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 317455714Skris the directory spec didn't end with a LIST_SEPARATOR_CHAR. 317555714Skris [Richard Levitte <levitte@stacken.kth.se>] 317655714Skris 317755714Skris *) Fix problems with sizeof(long) == 8. 317855714Skris [Andy Polyakov <appro@fy.chalmers.se>] 317955714Skris 318055714Skris *) Change functions to ANSI C. 318155714Skris [Ulf M�ller] 318255714Skris 318355714Skris *) Fix typos in error codes. 318455714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 318555714Skris 318655714Skris *) Remove defunct assembler files from Configure. 318755714Skris [Ulf M�ller] 318855714Skris 318955714Skris *) SPARC v8 assembler BIGNUM implementation. 319055714Skris [Andy Polyakov <appro@fy.chalmers.se>] 319155714Skris 319255714Skris *) Support for Certificate Policies extension: both print and set. 319355714Skris Various additions to support the r2i method this uses. 319455714Skris [Steve Henson] 319555714Skris 319655714Skris *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 319755714Skris return a const string when you are expecting an allocated buffer. 319855714Skris [Ben Laurie] 319955714Skris 320055714Skris *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 320155714Skris types DirectoryString and DisplayText. 320255714Skris [Steve Henson] 320355714Skris 320455714Skris *) Add code to allow r2i extensions to access the configuration database, 320555714Skris add an LHASH database driver and add several ctx helper functions. 320655714Skris [Steve Henson] 320755714Skris 320855714Skris *) Fix an evil bug in bn_expand2() which caused various BN functions to 320955714Skris fail when they extended the size of a BIGNUM. 321055714Skris [Steve Henson] 321155714Skris 321255714Skris *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 321355714Skris support typesafe stack. 321455714Skris [Steve Henson] 321555714Skris 321655714Skris *) Fix typo in SSL_[gs]et_options(). 321755714Skris [Nils Frostberg <nils@medcom.se>] 321855714Skris 321955714Skris *) Delete various functions and files that belonged to the (now obsolete) 322055714Skris old X509V3 handling code. 322155714Skris [Steve Henson] 322255714Skris 322355714Skris *) New Configure option "rsaref". 322455714Skris [Ulf M�ller] 322555714Skris 322655714Skris *) Don't auto-generate pem.h. 322755714Skris [Bodo Moeller] 322855714Skris 322955714Skris *) Introduce type-safe ASN.1 SETs. 323055714Skris [Ben Laurie] 323155714Skris 323255714Skris *) Convert various additional casted stacks to type-safe STACK_OF() variants. 323355714Skris [Ben Laurie, Ralf S. Engelschall, Steve Henson] 323455714Skris 323555714Skris *) Introduce type-safe STACKs. This will almost certainly break lots of code 323655714Skris that links with OpenSSL (well at least cause lots of warnings), but fear 323755714Skris not: the conversion is trivial, and it eliminates loads of evil casts. A 323855714Skris few STACKed things have been converted already. Feel free to convert more. 323955714Skris In the fullness of time, I'll do away with the STACK type altogether. 324055714Skris [Ben Laurie] 324155714Skris 324255714Skris *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 324355714Skris specified in <certfile> by updating the entry in the index.txt file. 324455714Skris This way one no longer has to edit the index.txt file manually for 324555714Skris revoking a certificate. The -revoke option does the gory details now. 324655714Skris [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 324755714Skris 324855714Skris *) Fix `openssl crl -noout -text' combination where `-noout' killed the 324955714Skris `-text' option at all and this way the `-noout -text' combination was 325055714Skris inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 325155714Skris [Ralf S. Engelschall] 325255714Skris 325355714Skris *) Make sure a corresponding plain text error message exists for the 325455714Skris X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 325555714Skris verify callback function determined that a certificate was revoked. 325655714Skris [Ralf S. Engelschall] 325755714Skris 325855714Skris *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 325955714Skris ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 326055714Skris all available cipers including rc5, which was forgotten until now. 326155714Skris In order to let the testing shell script know which algorithms 326255714Skris are available, a new (up to now undocumented) command 326355714Skris "openssl list-cipher-commands" is used. 326455714Skris [Bodo Moeller] 326555714Skris 326655714Skris *) Bugfix: s_client occasionally would sleep in select() when 326755714Skris it should have checked SSL_pending() first. 326855714Skris [Bodo Moeller] 326955714Skris 327055714Skris *) New functions DSA_do_sign and DSA_do_verify to provide access to 327155714Skris the raw DSA values prior to ASN.1 encoding. 327255714Skris [Ulf M�ller] 327355714Skris 327455714Skris *) Tweaks to Configure 327555714Skris [Niels Poppe <niels@netbox.org>] 327655714Skris 327755714Skris *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 327855714Skris yet... 327955714Skris [Steve Henson] 328055714Skris 328155714Skris *) New variables $(RANLIB) and $(PERL) in the Makefiles. 328255714Skris [Ulf M�ller] 328355714Skris 328455714Skris *) New config option to avoid instructions that are illegal on the 80386. 328555714Skris The default code is faster, but requires at least a 486. 328655714Skris [Ulf M�ller] 328755714Skris 328855714Skris *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 328955714Skris SSL2_SERVER_VERSION (not used at all) macros, which are now the 329055714Skris same as SSL2_VERSION anyway. 329155714Skris [Bodo Moeller] 329255714Skris 329355714Skris *) New "-showcerts" option for s_client. 329455714Skris [Bodo Moeller] 329555714Skris 329655714Skris *) Still more PKCS#12 integration. Add pkcs12 application to openssl 329755714Skris application. Various cleanups and fixes. 329855714Skris [Steve Henson] 329955714Skris 330055714Skris *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 330155714Skris modify error routines to work internally. Add error codes and PBE init 330255714Skris to library startup routines. 330355714Skris [Steve Henson] 330455714Skris 330555714Skris *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 330655714Skris packing functions to asn1 and evp. Changed function names and error 330755714Skris codes along the way. 330855714Skris [Steve Henson] 330955714Skris 331055714Skris *) PKCS12 integration: and so it begins... First of several patches to 331155714Skris slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 331255714Skris objects to objects.h 331355714Skris [Steve Henson] 331455714Skris 331555714Skris *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 331655714Skris and display support for Thawte strong extranet extension. 331755714Skris [Steve Henson] 331855714Skris 331955714Skris *) Add LinuxPPC support. 332055714Skris [Jeff Dubrule <igor@pobox.org>] 332155714Skris 332255714Skris *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 332355714Skris bn_div_words in alpha.s. 332455714Skris [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 332555714Skris 332655714Skris *) Make sure the RSA OAEP test is skipped under -DRSAref because 332755714Skris OAEP isn't supported when OpenSSL is built with RSAref. 332855714Skris [Ulf Moeller <ulf@fitug.de>] 332955714Skris 333055714Skris *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 333155714Skris so they no longer are missing under -DNOPROTO. 333255714Skris [Soren S. Jorvang <soren@t.dk>] 333355714Skris 333455714Skris 333555714Skris Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 333655714Skris 333755714Skris *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 333855714Skris doesn't work when the session is reused. Coming soon! 333955714Skris [Ben Laurie] 334055714Skris 334155714Skris *) Fix a security hole, that allows sessions to be reused in the wrong 334255714Skris context thus bypassing client cert protection! All software that uses 334355714Skris client certs and session caches in multiple contexts NEEDS PATCHING to 334455714Skris allow session reuse! A fuller solution is in the works. 334555714Skris [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 334655714Skris 334755714Skris *) Some more source tree cleanups (removed obsolete files 334855714Skris crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 334955714Skris permission on "config" script to be executable) and a fix for the INSTALL 335055714Skris document. 335155714Skris [Ulf Moeller <ulf@fitug.de>] 335255714Skris 335355714Skris *) Remove some legacy and erroneous uses of malloc, free instead of 335455714Skris Malloc, Free. 335555714Skris [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 335655714Skris 335755714Skris *) Make rsa_oaep_test return non-zero on error. 335855714Skris [Ulf Moeller <ulf@fitug.de>] 335955714Skris 336055714Skris *) Add support for native Solaris shared libraries. Configure 336155714Skris solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 336255714Skris if someone would make that last step automatic. 336355714Skris [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 336455714Skris 336555714Skris *) ctx_size was not built with the right compiler during "make links". Fixed. 336655714Skris [Ben Laurie] 336755714Skris 336855714Skris *) Change the meaning of 'ALL' in the cipher list. It now means "everything 336955714Skris except NULL ciphers". This means the default cipher list will no longer 337055714Skris enable NULL ciphers. They need to be specifically enabled e.g. with 337155714Skris the string "DEFAULT:eNULL". 337255714Skris [Steve Henson] 337355714Skris 337455714Skris *) Fix to RSA private encryption routines: if p < q then it would 337555714Skris occasionally produce an invalid result. This will only happen with 337655714Skris externally generated keys because OpenSSL (and SSLeay) ensure p > q. 337755714Skris [Steve Henson] 337855714Skris 337955714Skris *) Be less restrictive and allow also `perl util/perlpath.pl 338055714Skris /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 338155714Skris because this way one can also use an interpreter named `perl5' (which is 338255714Skris usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 338355714Skris installed as `perl'). 338455714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 338555714Skris 338655714Skris *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 338755714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 338855714Skris 338955714Skris *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 339055714Skris advapi32.lib to Win32 build and change the pem test comparision 339155714Skris to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 339255714Skris suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 339355714Skris and crypto/des/ede_cbcm_enc.c. 339455714Skris [Steve Henson] 339555714Skris 339655714Skris *) DES quad checksum was broken on big-endian architectures. Fixed. 339755714Skris [Ben Laurie] 339855714Skris 339955714Skris *) Comment out two functions in bio.h that aren't implemented. Fix up the 340055714Skris Win32 test batch file so it (might) work again. The Win32 test batch file 340155714Skris is horrible: I feel ill.... 340255714Skris [Steve Henson] 340355714Skris 340455714Skris *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 340555714Skris in e_os.h. Audit of header files to check ANSI and non ANSI 340655714Skris sections: 10 functions were absent from non ANSI section and not exported 340755714Skris from Windows DLLs. Fixed up libeay.num for new functions. 340855714Skris [Steve Henson] 340955714Skris 341055714Skris *) Make `openssl version' output lines consistent. 341155714Skris [Ralf S. Engelschall] 341255714Skris 341355714Skris *) Fix Win32 symbol export lists for BIO functions: Added 341455714Skris BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 341555714Skris to ms/libeay{16,32}.def. 341655714Skris [Ralf S. Engelschall] 341755714Skris 341855714Skris *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 341955714Skris fine under Unix and passes some trivial tests I've now added. But the 342055714Skris whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 342155714Skris added to make sure no one expects that this stuff really works in the 342255714Skris OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 342355714Skris up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 342455714Skris openssl_bio.xs. 342555714Skris [Ralf S. Engelschall] 342655714Skris 342755714Skris *) Fix the generation of two part addresses in perl. 342855714Skris [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 342955714Skris 343055714Skris *) Add config entry for Linux on MIPS. 343155714Skris [John Tobey <jtobey@channel1.com>] 343255714Skris 343355714Skris *) Make links whenever Configure is run, unless we are on Windoze. 343455714Skris [Ben Laurie] 343555714Skris 343655714Skris *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 343755714Skris Currently only issuerAltName and AuthorityKeyIdentifier make any sense 343855714Skris in CRLs. 343955714Skris [Steve Henson] 344055714Skris 344155714Skris *) Add a useful kludge to allow package maintainers to specify compiler and 344255714Skris other platforms details on the command line without having to patch the 344355714Skris Configure script everytime: One now can use ``perl Configure 344455714Skris <id>:<details>'', i.e. platform ids are allowed to have details appended 344555714Skris to them (seperated by colons). This is treated as there would be a static 344655714Skris pre-configured entry in Configure's %table under key <id> with value 344755714Skris <details> and ``perl Configure <id>'' is called. So, when you want to 344855714Skris perform a quick test-compile under FreeBSD 3.1 with pgcc and without 344955714Skris assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 345055714Skris now, which overrides the FreeBSD-elf entry on-the-fly. 345155714Skris [Ralf S. Engelschall] 345255714Skris 345355714Skris *) Disable new TLS1 ciphersuites by default: they aren't official yet. 345455714Skris [Ben Laurie] 345555714Skris 345655714Skris *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 345755714Skris on the `perl Configure ...' command line. This way one can compile 345855714Skris OpenSSL libraries with Position Independent Code (PIC) which is needed 345955714Skris for linking it into DSOs. 346055714Skris [Ralf S. Engelschall] 346155714Skris 346255714Skris *) Remarkably, export ciphers were totally broken and no-one had noticed! 346355714Skris Fixed. 346455714Skris [Ben Laurie] 346555714Skris 346655714Skris *) Cleaned up the LICENSE document: The official contact for any license 346755714Skris questions now is the OpenSSL core team under openssl-core@openssl.org. 346855714Skris And add a paragraph about the dual-license situation to make sure people 346955714Skris recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 347055714Skris to the OpenSSL toolkit. 347155714Skris [Ralf S. Engelschall] 347255714Skris 347355714Skris *) General source tree makefile cleanups: Made `making xxx in yyy...' 347455714Skris display consistent in the source tree and replaced `/bin/rm' by `rm'. 347555714Skris Additonally cleaned up the `make links' target: Remove unnecessary 347655714Skris semicolons, subsequent redundant removes, inline point.sh into mklink.sh 347755714Skris to speed processing and no longer clutter the display with confusing 347855714Skris stuff. Instead only the actually done links are displayed. 347955714Skris [Ralf S. Engelschall] 348055714Skris 348155714Skris *) Permit null encryption ciphersuites, used for authentication only. It used 348255714Skris to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 348355714Skris It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 348455714Skris encryption. 348555714Skris [Ben Laurie] 348655714Skris 348755714Skris *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 348855714Skris signed attributes when verifying signatures (this would break them), 348955714Skris the detached data encoding was wrong and public keys obtained using 349055714Skris X509_get_pubkey() weren't freed. 349155714Skris [Steve Henson] 349255714Skris 349355714Skris *) Add text documentation for the BUFFER functions. Also added a work around 349455714Skris to a Win95 console bug. This was triggered by the password read stuff: the 349555714Skris last character typed gets carried over to the next fread(). If you were 349655714Skris generating a new cert request using 'req' for example then the last 349755714Skris character of the passphrase would be CR which would then enter the first 349855714Skris field as blank. 349955714Skris [Steve Henson] 350055714Skris 350155714Skris *) Added the new `Includes OpenSSL Cryptography Software' button as 350255714Skris doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 350355714Skris button and can be used by applications based on OpenSSL to show the 350455714Skris relationship to the OpenSSL project. 350555714Skris [Ralf S. Engelschall] 350655714Skris 350755714Skris *) Remove confusing variables in function signatures in files 350855714Skris ssl/ssl_lib.c and ssl/ssl.h. 350955714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 351055714Skris 351155714Skris *) Don't install bss_file.c under PREFIX/include/ 351255714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 351355714Skris 351455714Skris *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 351555714Skris functions that return function pointers and has support for NT specific 351655714Skris stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 351755714Skris #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 351855714Skris unsigned to signed types: this was killing the Win32 compile. 351955714Skris [Steve Henson] 352055714Skris 352155714Skris *) Add new certificate file to stack functions, 352255714Skris SSL_add_dir_cert_subjects_to_stack() and 352355714Skris SSL_add_file_cert_subjects_to_stack(). These largely supplant 352455714Skris SSL_load_client_CA_file(), and can be used to add multiple certs easily 352555714Skris to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 352655714Skris This means that Apache-SSL and similar packages don't have to mess around 352755714Skris to add as many CAs as they want to the preferred list. 352855714Skris [Ben Laurie] 352955714Skris 353055714Skris *) Experiment with doxygen documentation. Currently only partially applied to 353155714Skris ssl/ssl_lib.c. 353255714Skris See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 353355714Skris openssl.doxy as the configuration file. 353455714Skris [Ben Laurie] 353555714Skris 353655714Skris *) Get rid of remaining C++-style comments which strict C compilers hate. 353755714Skris [Ralf S. Engelschall, pointed out by Carlos Amengual] 353855714Skris 353955714Skris *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 354055714Skris compiled in by default: it has problems with large keys. 354155714Skris [Steve Henson] 354255714Skris 354355714Skris *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 354455714Skris DH private keys and/or callback functions which directly correspond to 354555714Skris their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 354655714Skris is needed for applications which have to configure certificates on a 354755714Skris per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 354855714Skris (e.g. s_server). 354955714Skris For the RSA certificate situation is makes no difference, but 355055714Skris for the DSA certificate situation this fixes the "no shared cipher" 355155714Skris problem where the OpenSSL cipher selection procedure failed because the 355255714Skris temporary keys were not overtaken from the context and the API provided 355355714Skris no way to reconfigure them. 355455714Skris The new functions now let applications reconfigure the stuff and they 355555714Skris are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 355655714Skris SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 355755714Skris non-public-API function ssl_cert_instantiate() is used as a helper 355855714Skris function and also to reduce code redundancy inside ssl_rsa.c. 355955714Skris [Ralf S. Engelschall] 356055714Skris 356155714Skris *) Move s_server -dcert and -dkey options out of the undocumented feature 356255714Skris area because they are useful for the DSA situation and should be 356355714Skris recognized by the users. 356455714Skris [Ralf S. Engelschall] 356555714Skris 356655714Skris *) Fix the cipher decision scheme for export ciphers: the export bits are 356755714Skris *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 356855714Skris SSL_EXP_MASK. So, the original variable has to be used instead of the 356955714Skris already masked variable. 357055714Skris [Richard Levitte <levitte@stacken.kth.se>] 357155714Skris 357255714Skris *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 357355714Skris [Richard Levitte <levitte@stacken.kth.se>] 357455714Skris 357555714Skris *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 357655714Skris from `int' to `unsigned int' because it's a length and initialized by 357755714Skris EVP_DigestFinal() which expects an `unsigned int *'. 357855714Skris [Richard Levitte <levitte@stacken.kth.se>] 357955714Skris 358055714Skris *) Don't hard-code path to Perl interpreter on shebang line of Configure 358155714Skris script. Instead use the usual Shell->Perl transition trick. 358255714Skris [Ralf S. Engelschall] 358355714Skris 358455714Skris *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 358555714Skris (in addition to RSA certificates) to match the behaviour of `openssl dsa 358655714Skris -noout -modulus' as it's already the case for `openssl rsa -noout 358755714Skris -modulus'. For RSA the -modulus is the real "modulus" while for DSA 358855714Skris currently the public key is printed (a decision which was already done by 358955714Skris `openssl dsa -modulus' in the past) which serves a similar purpose. 359055714Skris Additionally the NO_RSA no longer completely removes the whole -modulus 359155714Skris option; it now only avoids using the RSA stuff. Same applies to NO_DSA 359255714Skris now, too. 359355714Skris [Ralf S. Engelschall] 359455714Skris 359555714Skris *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 359655714Skris BIO. See the source (crypto/evp/bio_ok.c) for more info. 359755714Skris [Arne Ansper <arne@ats.cyber.ee>] 359855714Skris 359955714Skris *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 360055714Skris to be added. Now both 'req' and 'ca' can use new objects defined in the 360155714Skris config file. 360255714Skris [Steve Henson] 360355714Skris 360455714Skris *) Add cool BIO that does syslog (or event log on NT). 360555714Skris [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 360655714Skris 360755714Skris *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 360855714Skris TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 360955714Skris TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 361055714Skris Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 361155714Skris [Ben Laurie] 361255714Skris 361355714Skris *) Add preliminary config info for new extension code. 361455714Skris [Steve Henson] 361555714Skris 361655714Skris *) Make RSA_NO_PADDING really use no padding. 361755714Skris [Ulf Moeller <ulf@fitug.de>] 361855714Skris 361955714Skris *) Generate errors when private/public key check is done. 362055714Skris [Ben Laurie] 362155714Skris 362255714Skris *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 362355714Skris for some CRL extensions and new objects added. 362455714Skris [Steve Henson] 362555714Skris 362655714Skris *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 362755714Skris key usage extension and fuller support for authority key id. 362855714Skris [Steve Henson] 362955714Skris 363055714Skris *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 363155714Skris padding method for RSA, which is recommended for new applications in PKCS 363255714Skris #1 v2.0 (RFC 2437, October 1998). 363355714Skris OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 363455714Skris foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 363555714Skris against Bleichbacher's attack on RSA. 363655714Skris [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 363755714Skris Ben Laurie] 363855714Skris 363955714Skris *) Updates to the new SSL compression code 364055714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 364155714Skris 364255714Skris *) Fix so that the version number in the master secret, when passed 364355714Skris via RSA, checks that if TLS was proposed, but we roll back to SSLv3 364455714Skris (because the server will not accept higher), that the version number 364555714Skris is 0x03,0x01, not 0x03,0x00 364655714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 364755714Skris 364855714Skris *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 364955714Skris leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 365055714Skris in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 365155714Skris [Steve Henson] 365255714Skris 365355714Skris *) Support for RAW extensions where an arbitrary extension can be 365455714Skris created by including its DER encoding. See apps/openssl.cnf for 365555714Skris an example. 365655714Skris [Steve Henson] 365755714Skris 365855714Skris *) Make sure latest Perl versions don't interpret some generated C array 365955714Skris code as Perl array code in the crypto/err/err_genc.pl script. 366055714Skris [Lars Weber <3weber@informatik.uni-hamburg.de>] 366155714Skris 366255714Skris *) Modify ms/do_ms.bat to not generate assembly language makefiles since 366355714Skris not many people have the assembler. Various Win32 compilation fixes and 366455714Skris update to the INSTALL.W32 file with (hopefully) more accurate Win32 366555714Skris build instructions. 366655714Skris [Steve Henson] 366755714Skris 366855714Skris *) Modify configure script 'Configure' to automatically create crypto/date.h 366955714Skris file under Win32 and also build pem.h from pem.org. New script 367055714Skris util/mkfiles.pl to create the MINFO file on environments that can't do a 367155714Skris 'make files': perl util/mkfiles.pl >MINFO should work. 367255714Skris [Steve Henson] 367355714Skris 367455714Skris *) Major rework of DES function declarations, in the pursuit of correctness 367555714Skris and purity. As a result, many evil casts evaporated, and some weirdness, 367655714Skris too. You may find this causes warnings in your code. Zapping your evil 367755714Skris casts will probably fix them. Mostly. 367855714Skris [Ben Laurie] 367955714Skris 368055714Skris *) Fix for a typo in asn1.h. Bug fix to object creation script 368155714Skris obj_dat.pl. It considered a zero in an object definition to mean 368255714Skris "end of object": none of the objects in objects.h have any zeros 368355714Skris so it wasn't spotted. 368455714Skris [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 368555714Skris 368655714Skris *) Add support for Triple DES Cipher Block Chaining with Output Feedback 368755714Skris Masking (CBCM). In the absence of test vectors, the best I have been able 368855714Skris to do is check that the decrypt undoes the encrypt, so far. Send me test 368955714Skris vectors if you have them. 369055714Skris [Ben Laurie] 369155714Skris 369255714Skris *) Correct calculation of key length for export ciphers (too much space was 369355714Skris allocated for null ciphers). This has not been tested! 369455714Skris [Ben Laurie] 369555714Skris 369655714Skris *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 369755714Skris message is now correct (it understands "crypto" and "ssl" on its 369855714Skris command line). There is also now an "update" option. This will update 369955714Skris the util/ssleay.num and util/libeay.num files with any new functions. 370055714Skris If you do a: 370155714Skris perl util/mkdef.pl crypto ssl update 370255714Skris it will update them. 370355714Skris [Steve Henson] 370455714Skris 370555714Skris *) Overhauled the Perl interface (perl/*): 370655714Skris - ported BN stuff to OpenSSL's different BN library 370755714Skris - made the perl/ source tree CVS-aware 370855714Skris - renamed the package from SSLeay to OpenSSL (the files still contain 370955714Skris their history because I've copied them in the repository) 371055714Skris - removed obsolete files (the test scripts will be replaced 371155714Skris by better Test::Harness variants in the future) 371255714Skris [Ralf S. Engelschall] 371355714Skris 371455714Skris *) First cut for a very conservative source tree cleanup: 371555714Skris 1. merge various obsolete readme texts into doc/ssleay.txt 371655714Skris where we collect the old documents and readme texts. 371755714Skris 2. remove the first part of files where I'm already sure that we no 371855714Skris longer need them because of three reasons: either they are just temporary 371955714Skris files which were left by Eric or they are preserved original files where 372055714Skris I've verified that the diff is also available in the CVS via "cvs diff 372155714Skris -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 372255714Skris the crypto/md/ stuff). 372355714Skris [Ralf S. Engelschall] 372455714Skris 372555714Skris *) More extension code. Incomplete support for subject and issuer alt 372655714Skris name, issuer and authority key id. Change the i2v function parameters 372755714Skris and add an extra 'crl' parameter in the X509V3_CTX structure: guess 372855714Skris what that's for :-) Fix to ASN1 macro which messed up 372955714Skris IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 373055714Skris [Steve Henson] 373155714Skris 373255714Skris *) Preliminary support for ENUMERATED type. This is largely copied from the 373355714Skris INTEGER code. 373455714Skris [Steve Henson] 373555714Skris 373655714Skris *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 373755714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 373855714Skris 373955714Skris *) Make sure `make rehash' target really finds the `openssl' program. 374055714Skris [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 374155714Skris 374255714Skris *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 374355714Skris like to hear about it if this slows down other processors. 374455714Skris [Ben Laurie] 374555714Skris 374655714Skris *) Add CygWin32 platform information to Configure script. 374755714Skris [Alan Batie <batie@aahz.jf.intel.com>] 374855714Skris 374955714Skris *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 375055714Skris [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 375155714Skris 375255714Skris *) New program nseq to manipulate netscape certificate sequences 375355714Skris [Steve Henson] 375455714Skris 375555714Skris *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 375655714Skris few typos. 375755714Skris [Steve Henson] 375855714Skris 375955714Skris *) Fixes to BN code. Previously the default was to define BN_RECURSION 376055714Skris but the BN code had some problems that would cause failures when 376155714Skris doing certificate verification and some other functions. 376255714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 376355714Skris 376455714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 376555714Skris [Steve Henson] 376655714Skris 376755714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 376855714Skris [Steve Henson] 376955714Skris 377055714Skris *) Add several PKIX and private extended key usage OIDs. 377155714Skris [Steve Henson] 377255714Skris 377355714Skris *) Modify the 'ca' program to handle the new extension code. Modify 377455714Skris openssl.cnf for new extension format, add comments. 377555714Skris [Steve Henson] 377655714Skris 377755714Skris *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 377855714Skris and add a sample to openssl.cnf so req -x509 now adds appropriate 377955714Skris CA extensions. 378055714Skris [Steve Henson] 378155714Skris 378255714Skris *) Continued X509 V3 changes. Add to other makefiles, integrate with the 378355714Skris error code, add initial support to X509_print() and x509 application. 378455714Skris [Steve Henson] 378555714Skris 378655714Skris *) Takes a deep breath and start addding X509 V3 extension support code. Add 378755714Skris files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 378855714Skris stuff is currently isolated and isn't even compiled yet. 378955714Skris [Steve Henson] 379055714Skris 379155714Skris *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 379255714Skris ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 379355714Skris Removed the versions check from X509 routines when loading extensions: 379455714Skris this allows certain broken certificates that don't set the version 379555714Skris properly to be processed. 379655714Skris [Steve Henson] 379755714Skris 379855714Skris *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 379955714Skris Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 380055714Skris can still be regenerated with "make depend". 380155714Skris [Ben Laurie] 380255714Skris 380355714Skris *) Spelling mistake in C version of CAST-128. 380455714Skris [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 380555714Skris 380655714Skris *) Changes to the error generation code. The perl script err-code.pl 380755714Skris now reads in the old error codes and retains the old numbers, only 380855714Skris adding new ones if necessary. It also only changes the .err files if new 380955714Skris codes are added. The makefiles have been modified to only insert errors 381055714Skris when needed (to avoid needlessly modifying header files). This is done 381155714Skris by only inserting errors if the .err file is newer than the auto generated 381255714Skris C file. To rebuild all the error codes from scratch (the old behaviour) 381355714Skris either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 381455714Skris or delete all the .err files. 381555714Skris [Steve Henson] 381655714Skris 381755714Skris *) CAST-128 was incorrectly implemented for short keys. The C version has 381855714Skris been fixed, but is untested. The assembler versions are also fixed, but 381955714Skris new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 382055714Skris to regenerate it if needed. 382155714Skris [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 382255714Skris Hagino <itojun@kame.net>] 382355714Skris 382455714Skris *) File was opened incorrectly in randfile.c. 382555714Skris [Ulf M�ller <ulf@fitug.de>] 382655714Skris 382755714Skris *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 382855714Skris functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 382955714Skris GeneralizedTime. ASN1_TIME is the proper type used in certificates et 383055714Skris al: it's just almost always a UTCTime. Note this patch adds new error 383155714Skris codes so do a "make errors" if there are problems. 383255714Skris [Steve Henson] 383355714Skris 383455714Skris *) Correct Linux 1 recognition in config. 383555714Skris [Ulf M�ller <ulf@fitug.de>] 383655714Skris 383755714Skris *) Remove pointless MD5 hash when using DSA keys in ca. 383855714Skris [Anonymous <nobody@replay.com>] 383955714Skris 384055714Skris *) Generate an error if given an empty string as a cert directory. Also 384155714Skris generate an error if handed NULL (previously returned 0 to indicate an 384255714Skris error, but didn't set one). 384355714Skris [Ben Laurie, reported by Anonymous <nobody@replay.com>] 384455714Skris 384555714Skris *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 384655714Skris [Ben Laurie] 384755714Skris 384855714Skris *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 384955714Skris parameters. This was causing a warning which killed off the Win32 compile. 385055714Skris [Steve Henson] 385155714Skris 385255714Skris *) Remove C++ style comments from crypto/bn/bn_local.h. 385355714Skris [Neil Costigan <neil.costigan@celocom.com>] 385455714Skris 385555714Skris *) The function OBJ_txt2nid was broken. It was supposed to return a nid 385655714Skris based on a text string, looking up short and long names and finally 385755714Skris "dot" format. The "dot" format stuff didn't work. Added new function 385855714Skris OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 385955714Skris OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 386055714Skris OID is not part of the table. 386155714Skris [Steve Henson] 386255714Skris 386355714Skris *) Add prototypes to X509 lookup/verify methods, fixing a bug in 386455714Skris X509_LOOKUP_by_alias(). 386555714Skris [Ben Laurie] 386655714Skris 386755714Skris *) Sort openssl functions by name. 386855714Skris [Ben Laurie] 386955714Skris 387055714Skris *) Get the gendsa program working (hopefully) and add it to app list. Remove 387155714Skris encryption from sample DSA keys (in case anyone is interested the password 387255714Skris was "1234"). 387355714Skris [Steve Henson] 387455714Skris 387555714Skris *) Make _all_ *_free functions accept a NULL pointer. 387655714Skris [Frans Heymans <fheymans@isaserver.be>] 387755714Skris 387855714Skris *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 387955714Skris NULL pointers. 388055714Skris [Anonymous <nobody@replay.com>] 388155714Skris 388255714Skris *) s_server should send the CAfile as acceptable CAs, not its own cert. 388355714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 388455714Skris 388555714Skris *) Don't blow it for numeric -newkey arguments to apps/req. 388655714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 388755714Skris 388855714Skris *) Temp key "for export" tests were wrong in s3_srvr.c. 388955714Skris [Anonymous <nobody@replay.com>] 389055714Skris 389155714Skris *) Add prototype for temp key callback functions 389255714Skris SSL_CTX_set_tmp_{rsa,dh}_callback(). 389355714Skris [Ben Laurie] 389455714Skris 389555714Skris *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 389655714Skris DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 389755714Skris [Steve Henson] 389855714Skris 389955714Skris *) X509_name_add_entry() freed the wrong thing after an error. 390055714Skris [Arne Ansper <arne@ats.cyber.ee>] 390155714Skris 390255714Skris *) rsa_eay.c would attempt to free a NULL context. 390355714Skris [Arne Ansper <arne@ats.cyber.ee>] 390455714Skris 390555714Skris *) BIO_s_socket() had a broken should_retry() on Windoze. 390655714Skris [Arne Ansper <arne@ats.cyber.ee>] 390755714Skris 390855714Skris *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 390955714Skris [Arne Ansper <arne@ats.cyber.ee>] 391055714Skris 391155714Skris *) Make sure the already existing X509_STORE->depth variable is initialized 391255714Skris in X509_STORE_new(), but document the fact that this variable is still 391355714Skris unused in the certificate verification process. 391455714Skris [Ralf S. Engelschall] 391555714Skris 391655714Skris *) Fix the various library and apps files to free up pkeys obtained from 391755714Skris X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 391855714Skris [Steve Henson] 391955714Skris 392055714Skris *) Fix reference counting in X509_PUBKEY_get(). This makes 392155714Skris demos/maurice/example2.c work, amongst others, probably. 392255714Skris [Steve Henson and Ben Laurie] 392355714Skris 392455714Skris *) First cut of a cleanup for apps/. First the `ssleay' program is now named 392555714Skris `openssl' and second, the shortcut symlinks for the `openssl <command>' 392655714Skris are no longer created. This way we have a single and consistent command 392755714Skris line interface `openssl <command>', similar to `cvs <command>'. 392855714Skris [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 392955714Skris 393055714Skris *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 393155714Skris BIT STRING wrapper always have zero unused bits. 393255714Skris [Steve Henson] 393355714Skris 393455714Skris *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 393555714Skris [Steve Henson] 393655714Skris 393755714Skris *) Make the top-level INSTALL documentation easier to understand. 393855714Skris [Paul Sutton] 393955714Skris 394055714Skris *) Makefiles updated to exit if an error occurs in a sub-directory 394155714Skris make (including if user presses ^C) [Paul Sutton] 394255714Skris 394355714Skris *) Make Montgomery context stuff explicit in RSA data structure. 394455714Skris [Ben Laurie] 394555714Skris 394655714Skris *) Fix build order of pem and err to allow for generated pem.h. 394755714Skris [Ben Laurie] 394855714Skris 394955714Skris *) Fix renumbering bug in X509_NAME_delete_entry(). 395055714Skris [Ben Laurie] 395155714Skris 395255714Skris *) Enhanced the err-ins.pl script so it makes the error library number 395355714Skris global and can add a library name. This is needed for external ASN1 and 395455714Skris other error libraries. 395555714Skris [Steve Henson] 395655714Skris 395755714Skris *) Fixed sk_insert which never worked properly. 395855714Skris [Steve Henson] 395955714Skris 396055714Skris *) Fix ASN1 macros so they can handle indefinite length construted 396155714Skris EXPLICIT tags. Some non standard certificates use these: they can now 396255714Skris be read in. 396355714Skris [Steve Henson] 396455714Skris 396555714Skris *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 396655714Skris into a single doc/ssleay.txt bundle. This way the information is still 396755714Skris preserved but no longer messes up this directory. Now it's new room for 396855714Skris the new set of documenation files. 396955714Skris [Ralf S. Engelschall] 397055714Skris 397155714Skris *) SETs were incorrectly DER encoded. This was a major pain, because they 397255714Skris shared code with SEQUENCEs, which aren't coded the same. This means that 397355714Skris almost everything to do with SETs or SEQUENCEs has either changed name or 397455714Skris number of arguments. 397555714Skris [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 397655714Skris 397755714Skris *) Fix test data to work with the above. 397855714Skris [Ben Laurie] 397955714Skris 398055714Skris *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 398155714Skris was already fixed by Eric for 0.9.1 it seems. 398255714Skris [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 398355714Skris 398455714Skris *) Autodetect FreeBSD3. 398555714Skris [Ben Laurie] 398655714Skris 398755714Skris *) Fix various bugs in Configure. This affects the following platforms: 398855714Skris nextstep 398955714Skris ncr-scde 399055714Skris unixware-2.0 399155714Skris unixware-2.0-pentium 399255714Skris sco5-cc. 399355714Skris [Ben Laurie] 399455714Skris 399555714Skris *) Eliminate generated files from CVS. Reorder tests to regenerate files 399655714Skris before they are needed. 399755714Skris [Ben Laurie] 399855714Skris 399955714Skris *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 400055714Skris [Ben Laurie] 400155714Skris 400255714Skris 400355714Skris Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 400455714Skris 400555714Skris *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 400655714Skris changed SSLeay to OpenSSL in version strings. 400755714Skris [Ralf S. Engelschall] 400855714Skris 400955714Skris *) Some fixups to the top-level documents. 401055714Skris [Paul Sutton] 401155714Skris 401255714Skris *) Fixed the nasty bug where rsaref.h was not found under compile-time 401355714Skris because the symlink to include/ was missing. 401455714Skris [Ralf S. Engelschall] 401555714Skris 401655714Skris *) Incorporated the popular no-RSA/DSA-only patches 401755714Skris which allow to compile a RSA-free SSLeay. 401855714Skris [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 401955714Skris 402055714Skris *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 402155714Skris when "ssleay" is still not found. 402255714Skris [Ralf S. Engelschall] 402355714Skris 402455714Skris *) Added more platforms to Configure: Cray T3E, HPUX 11, 402555714Skris [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 402655714Skris 402755714Skris *) Updated the README file. 402855714Skris [Ralf S. Engelschall] 402955714Skris 403055714Skris *) Added various .cvsignore files in the CVS repository subdirs 403155714Skris to make a "cvs update" really silent. 403255714Skris [Ralf S. Engelschall] 403355714Skris 403455714Skris *) Recompiled the error-definition header files and added 403555714Skris missing symbols to the Win32 linker tables. 403655714Skris [Ralf S. Engelschall] 403755714Skris 403855714Skris *) Cleaned up the top-level documents; 403955714Skris o new files: CHANGES and LICENSE 404055714Skris o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 404155714Skris o merged COPYRIGHT into LICENSE 404255714Skris o removed obsolete TODO file 404355714Skris o renamed MICROSOFT to INSTALL.W32 404455714Skris [Ralf S. Engelschall] 404555714Skris 404655714Skris *) Removed dummy files from the 0.9.1b source tree: 404755714Skris crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 404855714Skris crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 404955714Skris crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 405055714Skris crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 405155714Skris util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 405255714Skris [Ralf S. Engelschall] 405355714Skris 405455714Skris *) Added various platform portability fixes. 405555714Skris [Mark J. Cox] 405655714Skris 405755714Skris *) The Genesis of the OpenSSL rpject: 405855714Skris We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 405955714Skris Young and Tim J. Hudson created while they were working for C2Net until 406055714Skris summer 1998. 406155714Skris [The OpenSSL Project] 406255714Skris 406355714Skris 406455714Skris Changes between 0.9.0b and 0.9.1b [not released] 406555714Skris 406655714Skris *) Updated a few CA certificates under certs/ 406755714Skris [Eric A. Young] 406855714Skris 406955714Skris *) Changed some BIGNUM api stuff. 407055714Skris [Eric A. Young] 407155714Skris 407255714Skris *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 407355714Skris DGUX x86, Linux Alpha, etc. 407455714Skris [Eric A. Young] 407555714Skris 407655714Skris *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 407755714Skris RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 407855714Skris available). 407955714Skris [Eric A. Young] 408055714Skris 408155714Skris *) Add -strparse option to asn1pars program which parses nested 408255714Skris binary structures 408355714Skris [Dr Stephen Henson <shenson@bigfoot.com>] 408455714Skris 408555714Skris *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 408655714Skris [Eric A. Young] 408755714Skris 408855714Skris *) DSA fix for "ca" program. 408955714Skris [Eric A. Young] 409055714Skris 409155714Skris *) Added "-genkey" option to "dsaparam" program. 409255714Skris [Eric A. Young] 409355714Skris 409455714Skris *) Added RIPE MD160 (rmd160) message digest. 409555714Skris [Eric A. Young] 409655714Skris 409755714Skris *) Added -a (all) option to "ssleay version" command. 409855714Skris [Eric A. Young] 409955714Skris 410055714Skris *) Added PLATFORM define which is the id given to Configure. 410155714Skris [Eric A. Young] 410255714Skris 410355714Skris *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 410455714Skris [Eric A. Young] 410555714Skris 410655714Skris *) Extended the ASN.1 parser routines. 410755714Skris [Eric A. Young] 410855714Skris 410955714Skris *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 411055714Skris [Eric A. Young] 411155714Skris 411255714Skris *) Added a BN_CTX to the BN library. 411355714Skris [Eric A. Young] 411455714Skris 411555714Skris *) Fixed the weak key values in DES library 411655714Skris [Eric A. Young] 411755714Skris 411855714Skris *) Changed API in EVP library for cipher aliases. 411955714Skris [Eric A. Young] 412055714Skris 412155714Skris *) Added support for RC2/64bit cipher. 412255714Skris [Eric A. Young] 412355714Skris 412455714Skris *) Converted the lhash library to the crypto/mem.c functions. 412555714Skris [Eric A. Young] 412655714Skris 412755714Skris *) Added more recognized ASN.1 object ids. 412855714Skris [Eric A. Young] 412955714Skris 413055714Skris *) Added more RSA padding checks for SSL/TLS. 413155714Skris [Eric A. Young] 413255714Skris 413355714Skris *) Added BIO proxy/filter functionality. 413455714Skris [Eric A. Young] 413555714Skris 413655714Skris *) Added extra_certs to SSL_CTX which can be used 413755714Skris send extra CA certificates to the client in the CA cert chain sending 413855714Skris process. It can be configured with SSL_CTX_add_extra_chain_cert(). 413955714Skris [Eric A. Young] 414055714Skris 414155714Skris *) Now Fortezza is denied in the authentication phase because 414255714Skris this is key exchange mechanism is not supported by SSLeay at all. 414355714Skris [Eric A. Young] 414455714Skris 414555714Skris *) Additional PKCS1 checks. 414655714Skris [Eric A. Young] 414755714Skris 414855714Skris *) Support the string "TLSv1" for all TLS v1 ciphers. 414955714Skris [Eric A. Young] 415055714Skris 415155714Skris *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 415255714Skris ex_data index of the SSL context in the X509_STORE_CTX ex_data. 415355714Skris [Eric A. Young] 415455714Skris 415555714Skris *) Fixed a few memory leaks. 415655714Skris [Eric A. Young] 415755714Skris 415855714Skris *) Fixed various code and comment typos. 415955714Skris [Eric A. Young] 416055714Skris 416155714Skris *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 416255714Skris bytes sent in the client random. 416355714Skris [Edward Bishop <ebishop@spyglass.com>] 416455714Skris 4165