CHANGES revision 100936
155714Skris 255714Skris OpenSSL CHANGES 355714Skris _______________ 455714Skris 5100936Snectar Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 6100936Snectar 7100936Snectar *) Fix cipher selection routines: ciphers without encryption had no flags 8100936Snectar for the cipher strength set and where therefore not handled correctly 9100936Snectar by the selection routines (PR #130). 10100936Snectar [Lutz Jaenicke] 11100936Snectar 12100936Snectar *) Fix EVP_dsa_sha macro. 13100936Snectar [Nils Larsch] 14100936Snectar 15100936Snectar *) New option 16100936Snectar SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 17100936Snectar for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 18100936Snectar that was added in OpenSSL 0.9.6d. 19100936Snectar 20100936Snectar As the countermeasure turned out to be incompatible with some 21100936Snectar broken SSL implementations, the new option is part of SSL_OP_ALL. 22100936Snectar SSL_OP_ALL is usually employed when compatibility with weird SSL 23100936Snectar implementations is desired (e.g. '-bugs' option to 's_client' and 24100936Snectar 's_server'), so the new option is automatically set in many 25100936Snectar applications. 26100936Snectar [Bodo Moeller] 27100936Snectar 28100936Snectar *) Changes in security patch: 29100936Snectar 30100936Snectar Changes marked "(CHATS)" were sponsored by the Defense Advanced 31100936Snectar Research Projects Agency (DARPA) and Air Force Research Laboratory, 32100936Snectar Air Force Materiel Command, USAF, under agreement number 33100936Snectar F30602-01-2-0537. 34100936Snectar 35100936Snectar *) Add various sanity checks to asn1_get_length() to reject 36100936Snectar the ASN1 length bytes if they exceed sizeof(long), will appear 37100936Snectar negative or the content length exceeds the length of the 38100936Snectar supplied buffer. 39100936Snectar [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 40100936Snectar 41100936Snectar *) Assertions for various potential buffer overflows, not known to 42100936Snectar happen in practice. 43100936Snectar [Ben Laurie (CHATS)] 44100936Snectar 45100936Snectar *) Various temporary buffers to hold ASCII versions of integers were 46100936Snectar too small for 64 bit platforms. (CAN-2002-0655) 47100936Snectar [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 48100936Snectar 49100936Snectar *) Remote buffer overflow in SSL3 protocol - an attacker could 50100936Snectar supply an oversized session ID to a client. (CAN-2002-0656) 51100936Snectar [Ben Laurie (CHATS)] 52100936Snectar 53100936Snectar *) Remote buffer overflow in SSL2 protocol - an attacker could 54100936Snectar supply an oversized client master key. (CAN-2002-0656) 55100936Snectar [Ben Laurie (CHATS)] 56100936Snectar 57100928Snectar Changes between 0.9.6c and 0.9.6d [9 May 2002] 58100928Snectar 59100928Snectar *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 60100928Snectar encoded as NULL) with id-dsa-with-sha1. 61100928Snectar [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 62100928Snectar 63100928Snectar *) Check various X509_...() return values in apps/req.c. 64100928Snectar [Nils Larsch <nla@trustcenter.de>] 65100928Snectar 66100928Snectar *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 67100928Snectar an end-of-file condition would erronously be flagged, when the CRLF 68100928Snectar was just at the end of a processed block. The bug was discovered when 69100928Snectar processing data through a buffering memory BIO handing the data to a 70100928Snectar BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 71100928Snectar <ptsekov@syntrex.com> and Nedelcho Stanev. 72100928Snectar [Lutz Jaenicke] 73100928Snectar 74100928Snectar *) Implement a countermeasure against a vulnerability recently found 75100928Snectar in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 76100928Snectar before application data chunks to avoid the use of known IVs 77100928Snectar with data potentially chosen by the attacker. 78100928Snectar [Bodo Moeller] 79100928Snectar 80100928Snectar *) Fix length checks in ssl3_get_client_hello(). 81100928Snectar [Bodo Moeller] 82100928Snectar 83100928Snectar *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 84100928Snectar to prevent ssl3_read_internal() from incorrectly assuming that 85100928Snectar ssl3_read_bytes() found application data while handshake 86100928Snectar processing was enabled when in fact s->s3->in_read_app_data was 87100928Snectar merely automatically cleared during the initial handshake. 88100928Snectar [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 89100928Snectar 90100928Snectar *) Fix object definitions for Private and Enterprise: they were not 91100928Snectar recognized in their shortname (=lowercase) representation. Extend 92100928Snectar obj_dat.pl to issue an error when using undefined keywords instead 93100928Snectar of silently ignoring the problem (Svenning Sorensen 94100928Snectar <sss@sss.dnsalias.net>). 95100928Snectar [Lutz Jaenicke] 96100928Snectar 97100928Snectar *) Fix DH_generate_parameters() so that it works for 'non-standard' 98100928Snectar generators, i.e. generators other than 2 and 5. (Previously, the 99100928Snectar code did not properly initialise the 'add' and 'rem' values to 100100928Snectar BN_generate_prime().) 101100928Snectar 102100928Snectar In the new general case, we do not insist that 'generator' is 103100928Snectar actually a primitive root: This requirement is rather pointless; 104100928Snectar a generator of the order-q subgroup is just as good, if not 105100928Snectar better. 106100928Snectar [Bodo Moeller] 107100928Snectar 108100928Snectar *) Map new X509 verification errors to alerts. Discovered and submitted by 109100928Snectar Tom Wu <tom@arcot.com>. 110100928Snectar [Lutz Jaenicke] 111100928Snectar 112100928Snectar *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 113100928Snectar returning non-zero before the data has been completely received 114100928Snectar when using non-blocking I/O. 115100928Snectar [Bodo Moeller; problem pointed out by John Hughes] 116100928Snectar 117100928Snectar *) Some of the ciphers missed the strength entry (SSL_LOW etc). 118100928Snectar [Ben Laurie, Lutz Jaenicke] 119100928Snectar 120100928Snectar *) Fix bug in SSL_clear(): bad sessions were not removed (found by 121100928Snectar Yoram Zahavi <YoramZ@gilian.com>). 122100928Snectar [Lutz Jaenicke] 123100928Snectar 124100928Snectar *) Add information about CygWin 1.3 and on, and preserve proper 125100928Snectar configuration for the versions before that. 126100928Snectar [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 127100928Snectar 128100928Snectar *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 129100928Snectar check whether we deal with a copy of a session and do not delete from 130100928Snectar the cache in this case. Problem reported by "Izhar Shoshani Levi" 131100928Snectar <izhar@checkpoint.com>. 132100928Snectar [Lutz Jaenicke] 133100928Snectar 134100928Snectar *) Do not store session data into the internal session cache, if it 135100928Snectar is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 136100928Snectar flag is set). Proposed by Aslam <aslam@funk.com>. 137100928Snectar [Lutz Jaenicke] 138100928Snectar 139100928Snectar *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 140100928Snectar value is 0. 141100928Snectar [Richard Levitte] 142100928Snectar 143100928Snectar *) [In 0.9.6c-engine release:] 144100928Snectar Fix a crashbug and a logic bug in hwcrhk_load_pubkey() 145100928Snectar [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 146100928Snectar 147100928Snectar *) Add the configuration target linux-s390x. 148100928Snectar [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 149100928Snectar 150100928Snectar *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 151100928Snectar ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 152100928Snectar variable as an indication that a ClientHello message has been 153100928Snectar received. As the flag value will be lost between multiple 154100928Snectar invocations of ssl3_accept when using non-blocking I/O, the 155100928Snectar function may not be aware that a handshake has actually taken 156100928Snectar place, thus preventing a new session from being added to the 157100928Snectar session cache. 158100928Snectar 159100928Snectar To avoid this problem, we now set s->new_session to 2 instead of 160100928Snectar using a local variable. 161100928Snectar [Lutz Jaenicke, Bodo Moeller] 162100928Snectar 163100928Snectar *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 164100928Snectar if the SSL_R_LENGTH_MISMATCH error is detected. 165100928Snectar [Geoff Thorpe, Bodo Moeller] 166100928Snectar 167100928Snectar *) New 'shared_ldflag' column in Configure platform table. 168100928Snectar [Richard Levitte] 169100928Snectar 170100928Snectar *) Fix EVP_CIPHER_mode macro. 171100928Snectar ["Dan S. Camper" <dan@bti.net>] 172100928Snectar 173100928Snectar *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 174100928Snectar type, we must throw them away by setting rr->length to 0. 175100928Snectar [D P Chang <dpc@qualys.com>] 176100928Snectar 17789837Skris Changes between 0.9.6b and 0.9.6c [21 dec 2001] 17889837Skris 17989837Skris *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 18089837Skris <Dominikus.Scherkl@biodata.com>. (The previous implementation 18189837Skris worked incorrectly for those cases where range = 10..._2 and 18289837Skris 3*range is two bits longer than range.) 18389837Skris [Bodo Moeller] 18489837Skris 18589837Skris *) Only add signing time to PKCS7 structures if it is not already 18689837Skris present. 18789837Skris [Steve Henson] 18889837Skris 18989837Skris *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 19089837Skris OBJ_ld_ce should be OBJ_id_ce. 19189837Skris Also some ip-pda OIDs in crypto/objects/objects.txt were 19289837Skris incorrect (cf. RFC 3039). 19389837Skris [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 19489837Skris 19589837Skris *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 19689837Skris returns early because it has nothing to do. 19789837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 19889837Skris 19989837Skris *) [In 0.9.6c-engine release:] 20089837Skris Fix mutex callback return values in crypto/engine/hw_ncipher.c. 20189837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 20289837Skris 20389837Skris *) [In 0.9.6c-engine release:] 20489837Skris Add support for Cryptographic Appliance's keyserver technology. 20589837Skris (Use engine 'keyclient') 20689837Skris [Cryptographic Appliances and Geoff Thorpe] 20789837Skris 20889837Skris *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 20989837Skris is called via tools/c89.sh because arguments have to be 21089837Skris rearranged (all '-L' options must appear before the first object 21189837Skris modules). 21289837Skris [Richard Shapiro <rshapiro@abinitio.com>] 21389837Skris 21489837Skris *) [In 0.9.6c-engine release:] 21589837Skris Add support for Broadcom crypto accelerator cards, backported 21689837Skris from 0.9.7. 21789837Skris [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 21889837Skris 21989837Skris *) [In 0.9.6c-engine release:] 22089837Skris Add support for SureWare crypto accelerator cards from 22189837Skris Baltimore Technologies. (Use engine 'sureware') 22289837Skris [Baltimore Technologies and Mark Cox] 22389837Skris 22489837Skris *) [In 0.9.6c-engine release:] 22589837Skris Add support for crypto accelerator cards from Accelerated 22689837Skris Encryption Processing, www.aep.ie. (Use engine 'aep') 22789837Skris [AEP Inc. and Mark Cox] 22889837Skris 22989837Skris *) Add a configuration entry for gcc on UnixWare. 23089837Skris [Gary Benson <gbenson@redhat.com>] 23189837Skris 23289837Skris *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 23389837Skris messages are stored in a single piece (fixed-length part and 23489837Skris variable-length part combined) and fix various bugs found on the way. 23589837Skris [Bodo Moeller] 23689837Skris 23789837Skris *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 23889837Skris instead. BIO_gethostbyname() does not know what timeouts are 23989837Skris appropriate, so entries would stay in cache even when they have 24089837Skris become invalid. 24189837Skris [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 24289837Skris 24389837Skris *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 24489837Skris faced with a pathologically small ClientHello fragment that does 24589837Skris not contain client_version: Instead of aborting with an error, 24689837Skris simply choose the highest available protocol version (i.e., 24789837Skris TLS 1.0 unless it is disabled). In practice, ClientHello 24889837Skris messages are never sent like this, but this change gives us 24989837Skris strictly correct behaviour at least for TLS. 25089837Skris [Bodo Moeller] 25189837Skris 25289837Skris *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 25389837Skris never resets s->method to s->ctx->method when called from within 25489837Skris one of the SSL handshake functions. 25589837Skris [Bodo Moeller; problem pointed out by Niko Baric] 25689837Skris 25789837Skris *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 25889837Skris (sent using the client's version number) if client_version is 25989837Skris smaller than the protocol version in use. Also change 26089837Skris ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 26189837Skris the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 26289837Skris the client will at least see that alert. 26389837Skris [Bodo Moeller] 26489837Skris 26589837Skris *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 26689837Skris correctly. 26789837Skris [Bodo Moeller] 26889837Skris 26989837Skris *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 27089837Skris client receives HelloRequest while in a handshake. 27189837Skris [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 27289837Skris 27389837Skris *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 27489837Skris should end in 'break', not 'goto end' which circuments various 27589837Skris cleanups done in state SSL_ST_OK. But session related stuff 27689837Skris must be disabled for SSL_ST_OK in the case that we just sent a 27789837Skris HelloRequest. 27889837Skris 27989837Skris Also avoid some overhead by not calling ssl_init_wbio_buffer() 28089837Skris before just sending a HelloRequest. 28189837Skris [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 28289837Skris 28389837Skris *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 28489837Skris reveal whether illegal block cipher padding was found or a MAC 28589837Skris verification error occured. (Neither SSLerr() codes nor alerts 28689837Skris are directly visible to potential attackers, but the information 28789837Skris may leak via logfiles.) 28889837Skris 28989837Skris Similar changes are not required for the SSL 2.0 implementation 29089837Skris because the number of padding bytes is sent in clear for SSL 2.0, 29189837Skris and the extra bytes are just ignored. However ssl/s2_pkt.c 29289837Skris failed to verify that the purported number of padding bytes is in 29389837Skris the legal range. 29489837Skris [Bodo Moeller] 29589837Skris 29689837Skris *) Add OpenUNIX-8 support including shared libraries 29789837Skris (Boyd Lynn Gerber <gerberb@zenez.com>). 29889837Skris [Lutz Jaenicke] 29989837Skris 30089837Skris *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 30189837Skris 'wristwatch attack' using huge encoding parameters (cf. 30289837Skris James H. Manger's CRYPTO 2001 paper). Note that the 30389837Skris RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 30489837Skris encoding parameters and hence was not vulnerable. 30589837Skris [Bodo Moeller] 30689837Skris 30789837Skris *) BN_sqr() bug fix. 30889837Skris [Ulf M�ller, reported by Jim Ellis <jim.ellis@cavium.com>] 30989837Skris 31089837Skris *) Rabin-Miller test analyses assume uniformly distributed witnesses, 31189837Skris so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 31289837Skris followed by modular reduction. 31389837Skris [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 31489837Skris 31589837Skris *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 31689837Skris equivalent based on BN_pseudo_rand() instead of BN_rand(). 31789837Skris [Bodo Moeller] 31889837Skris 31989837Skris *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 32089837Skris This function was broken, as the check for a new client hello message 32189837Skris to handle SGC did not allow these large messages. 32289837Skris (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 32389837Skris [Lutz Jaenicke] 32489837Skris 32589837Skris *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 32689837Skris [Lutz Jaenicke] 32789837Skris 32889837Skris *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 32989837Skris for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 33089837Skris [Lutz Jaenicke] 33189837Skris 33289837Skris *) Rework the configuration and shared library support for Tru64 Unix. 33389837Skris The configuration part makes use of modern compiler features and 33489837Skris still retains old compiler behavior for those that run older versions 33589837Skris of the OS. The shared library support part includes a variant that 33689837Skris uses the RPATH feature, and is available through the special 33789837Skris configuration target "alpha-cc-rpath", which will never be selected 33889837Skris automatically. 33989837Skris [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 34089837Skris 34189837Skris *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 34289837Skris with the same message size as in ssl3_get_certificate_request(). 34389837Skris Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 34489837Skris messages might inadvertently be reject as too long. 34589837Skris [Petr Lampa <lampa@fee.vutbr.cz>] 34689837Skris 34789837Skris *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 34889837Skris [Andy Polyakov] 34989837Skris 35089837Skris *) Modified SSL library such that the verify_callback that has been set 35189837Skris specificly for an SSL object with SSL_set_verify() is actually being 35289837Skris used. Before the change, a verify_callback set with this function was 35389837Skris ignored and the verify_callback() set in the SSL_CTX at the time of 35489837Skris the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 35589837Skris to allow the necessary settings. 35689837Skris [Lutz Jaenicke] 35789837Skris 35889837Skris *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 35989837Skris explicitly to NULL, as at least on Solaris 8 this seems not always to be 36089837Skris done automatically (in contradiction to the requirements of the C 36189837Skris standard). This made problems when used from OpenSSH. 36289837Skris [Lutz Jaenicke] 36389837Skris 36489837Skris *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 36589837Skris dh->length and always used 36689837Skris 36789837Skris BN_rand_range(priv_key, dh->p). 36889837Skris 36989837Skris BN_rand_range() is not necessary for Diffie-Hellman, and this 37089837Skris specific range makes Diffie-Hellman unnecessarily inefficient if 37189837Skris dh->length (recommended exponent length) is much smaller than the 37289837Skris length of dh->p. We could use BN_rand_range() if the order of 37389837Skris the subgroup was stored in the DH structure, but we only have 37489837Skris dh->length. 37589837Skris 37689837Skris So switch back to 37789837Skris 37889837Skris BN_rand(priv_key, l, ...) 37989837Skris 38089837Skris where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 38189837Skris otherwise. 38289837Skris [Bodo Moeller] 38389837Skris 38489837Skris *) In 38589837Skris 38689837Skris RSA_eay_public_encrypt 38789837Skris RSA_eay_private_decrypt 38889837Skris RSA_eay_private_encrypt (signing) 38989837Skris RSA_eay_public_decrypt (signature verification) 39089837Skris 39189837Skris (default implementations for RSA_public_encrypt, 39289837Skris RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 39389837Skris always reject numbers >= n. 39489837Skris [Bodo Moeller] 39589837Skris 39689837Skris *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 39789837Skris to synchronize access to 'locking_thread'. This is necessary on 39889837Skris systems where access to 'locking_thread' (an 'unsigned long' 39989837Skris variable) is not atomic. 40089837Skris [Bodo Moeller] 40189837Skris 40289837Skris *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 40389837Skris *before* setting the 'crypto_lock_rand' flag. The previous code had 40489837Skris a race condition if 0 is a valid thread ID. 40589837Skris [Travis Vitek <vitek@roguewave.com>] 40689837Skris 40789837Skris *) Add support for shared libraries under Irix. 40889837Skris [Albert Chin-A-Young <china@thewrittenword.com>] 40989837Skris 41089837Skris *) Add configuration option to build on Linux on both big-endian and 41189837Skris little-endian MIPS. 41289837Skris [Ralf Baechle <ralf@uni-koblenz.de>] 41389837Skris 41489837Skris *) Add the possibility to create shared libraries on HP-UX. 41589837Skris [Richard Levitte] 41689837Skris 41779998Skris Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 41879998Skris 41979998Skris *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 42079998Skris to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 42179998Skris Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 42279998Skris PRNG state recovery was possible based on the output of 42379998Skris one PRNG request appropriately sized to gain knowledge on 42479998Skris 'md' followed by enough consecutive 1-byte PRNG requests 42579998Skris to traverse all of 'state'. 42679998Skris 42779998Skris 1. When updating 'md_local' (the current thread's copy of 'md') 42879998Skris during PRNG output generation, hash all of the previous 42979998Skris 'md_local' value, not just the half used for PRNG output. 43079998Skris 43179998Skris 2. Make the number of bytes from 'state' included into the hash 43279998Skris independent from the number of PRNG bytes requested. 43379998Skris 43479998Skris The first measure alone would be sufficient to avoid 43579998Skris Markku-Juhani's attack. (Actually it had never occurred 43679998Skris to me that the half of 'md_local' used for chaining was the 43779998Skris half from which PRNG output bytes were taken -- I had always 43879998Skris assumed that the secret half would be used.) The second 43979998Skris measure makes sure that additional data from 'state' is never 44079998Skris mixed into 'md_local' in small portions; this heuristically 44179998Skris further strengthens the PRNG. 44279998Skris [Bodo Moeller] 44379998Skris 44479998Skris *) Fix crypto/bn/asm/mips3.s. 44579998Skris [Andy Polyakov] 44679998Skris 44779998Skris *) When only the key is given to "enc", the IV is undefined. Print out 44879998Skris an error message in this case. 44979998Skris [Lutz Jaenicke] 45079998Skris 45179998Skris *) Handle special case when X509_NAME is empty in X509 printing routines. 45279998Skris [Steve Henson] 45379998Skris 45479998Skris *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 45579998Skris positive and less than q. 45679998Skris [Bodo Moeller] 45779998Skris 45879998Skris *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 45979998Skris used: it isn't thread safe and the add_lock_callback should handle 46079998Skris that itself. 46179998Skris [Paul Rose <Paul.Rose@bridge.com>] 46279998Skris 46379998Skris *) Verify that incoming data obeys the block size in 46479998Skris ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 46579998Skris [Bodo Moeller] 46679998Skris 46779998Skris *) Fix OAEP check. 46879998Skris [Ulf M�ller, Bodo M�ller] 46979998Skris 47079998Skris *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 47189837Skris RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 47279998Skris when fixing the server behaviour for backwards-compatible 'client 47379998Skris hello' messages. (Note that the attack is impractical against 47479998Skris SSL 3.0 and TLS 1.0 anyway because length and version checking 47579998Skris means that the probability of guessing a valid ciphertext is 47679998Skris around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 47779998Skris paper.) 47879998Skris 47979998Skris Before 0.9.5, the countermeasure (hide the error by generating a 48079998Skris random 'decryption result') did not work properly because 48179998Skris ERR_clear_error() was missing, meaning that SSL_get_error() would 48279998Skris detect the supposedly ignored error. 48379998Skris 48479998Skris Both problems are now fixed. 48579998Skris [Bodo Moeller] 48679998Skris 48779998Skris *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 48879998Skris (previously it was 1024). 48979998Skris [Bodo Moeller] 49079998Skris 49179998Skris *) Fix for compatibility mode trust settings: ignore trust settings 49279998Skris unless some valid trust or reject settings are present. 49379998Skris [Steve Henson] 49479998Skris 49579998Skris *) Fix for blowfish EVP: its a variable length cipher. 49679998Skris [Steve Henson] 49779998Skris 49879998Skris *) Fix various bugs related to DSA S/MIME verification. Handle missing 49979998Skris parameters in DSA public key structures and return an error in the 50079998Skris DSA routines if parameters are absent. 50179998Skris [Steve Henson] 50279998Skris 50379998Skris *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 50479998Skris in the current directory if neither $RANDFILE nor $HOME was set. 50579998Skris RAND_file_name() in 0.9.6a returned NULL in this case. This has 50679998Skris caused some confusion to Windows users who haven't defined $HOME. 50779998Skris Thus RAND_file_name() is changed again: e_os.h can define a 50879998Skris DEFAULT_HOME, which will be used if $HOME is not set. 50979998Skris For Windows, we use "C:"; on other platforms, we still require 51079998Skris environment variables. 51179998Skris 51279998Skris *) Move 'if (!initialized) RAND_poll()' into regions protected by 51379998Skris CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 51479998Skris having multiple threads call RAND_poll() concurrently. 51579998Skris [Bodo Moeller] 51679998Skris 51779998Skris *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 51879998Skris combination of a flag and a thread ID variable. 51979998Skris Otherwise while one thread is in ssleay_rand_bytes (which sets the 52079998Skris flag), *other* threads can enter ssleay_add_bytes without obeying 52189837Skris the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 52279998Skris that they do not hold after the first thread unsets add_do_not_lock). 52379998Skris [Bodo Moeller] 52479998Skris 52579998Skris *) Change bctest again: '-x' expressions are not available in all 52679998Skris versions of 'test'. 52779998Skris [Bodo Moeller] 52879998Skris 52976866Skris Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 53072613Skris 53176866Skris *) Fix a couple of memory leaks in PKCS7_dataDecode() 53276866Skris [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 53376866Skris 53476866Skris *) Change Configure and Makefiles to provide EXE_EXT, which will contain 53576866Skris the default extension for executables, if any. Also, make the perl 53676866Skris scripts that use symlink() to test if it really exists and use "cp" 53776866Skris if it doesn't. All this made OpenSSL compilable and installable in 53876866Skris CygWin. 53976866Skris [Richard Levitte] 54076866Skris 54176866Skris *) Fix for asn1_GetSequence() for indefinite length constructed data. 54276866Skris If SEQUENCE is length is indefinite just set c->slen to the total 54376866Skris amount of data available. 54476866Skris [Steve Henson, reported by shige@FreeBSD.org] 54576866Skris [This change does not apply to 0.9.7.] 54676866Skris 54776866Skris *) Change bctest to avoid here-documents inside command substitution 54876866Skris (workaround for FreeBSD /bin/sh bug). 54976866Skris For compatibility with Ultrix, avoid shell functions (introduced 55076866Skris in the bctest version that searches along $PATH). 55176866Skris [Bodo Moeller] 55276866Skris 55376866Skris *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 55476866Skris with des_encrypt() defined on some operating systems, like Solaris 55576866Skris and UnixWare. 55676866Skris [Richard Levitte] 55776866Skris 55876866Skris *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 55976866Skris On the Importance of Eliminating Errors in Cryptographic 56076866Skris Computations, J. Cryptology 14 (2001) 2, 101-119, 56176866Skris http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 56276866Skris [Ulf Moeller] 56376866Skris 56476866Skris *) MIPS assembler BIGNUM division bug fix. 56576866Skris [Andy Polyakov] 56676866Skris 56776866Skris *) Disabled incorrect Alpha assembler code. 56876866Skris [Richard Levitte] 56976866Skris 57076866Skris *) Fix PKCS#7 decode routines so they correctly update the length 57176866Skris after reading an EOC for the EXPLICIT tag. 57276866Skris [Steve Henson] 57376866Skris [This change does not apply to 0.9.7.] 57476866Skris 57576866Skris *) Fix bug in PKCS#12 key generation routines. This was triggered 57676866Skris if a 3DES key was generated with a 0 initial byte. Include 57776866Skris PKCS12_BROKEN_KEYGEN compilation option to retain the old 57876866Skris (but broken) behaviour. 57976866Skris [Steve Henson] 58076866Skris 58176866Skris *) Enhance bctest to search for a working bc along $PATH and print 58276866Skris it when found. 58376866Skris [Tim Rice <tim@multitalents.net> via Richard Levitte] 58476866Skris 58576866Skris *) Fix memory leaks in err.c: free err_data string if necessary; 58676866Skris don't write to the wrong index in ERR_set_error_data. 58776866Skris [Bodo Moeller] 58876866Skris 58976866Skris *) Implement ssl23_peek (analogous to ssl23_read), which previously 59076866Skris did not exist. 59176866Skris [Bodo Moeller] 59276866Skris 59376866Skris *) Replace rdtsc with _emit statements for VC++ version 5. 59476866Skris [Jeremy Cooper <jeremy@baymoo.org>] 59576866Skris 59676866Skris *) Make it possible to reuse SSLv2 sessions. 59776866Skris [Richard Levitte] 59876866Skris 59976866Skris *) In copy_email() check for >= 0 as a return value for 60076866Skris X509_NAME_get_index_by_NID() since 0 is a valid index. 60176866Skris [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 60276866Skris 60376866Skris *) Avoid coredump with unsupported or invalid public keys by checking if 60476866Skris X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 60576866Skris PKCS7_verify() fails with non detached data. 60676866Skris [Steve Henson] 60776866Skris 60876866Skris *) Don't use getenv in library functions when run as setuid/setgid. 60976866Skris New function OPENSSL_issetugid(). 61076866Skris [Ulf Moeller] 61176866Skris 61276866Skris *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 61376866Skris due to incorrect handling of multi-threading: 61476866Skris 61576866Skris 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 61676866Skris 61776866Skris 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 61876866Skris 61976866Skris 3. Count how many times MemCheck_off() has been called so that 62076866Skris nested use can be treated correctly. This also avoids 62176866Skris inband-signalling in the previous code (which relied on the 62276866Skris assumption that thread ID 0 is impossible). 62376866Skris [Bodo Moeller] 62476866Skris 62576866Skris *) Add "-rand" option also to s_client and s_server. 62676866Skris [Lutz Jaenicke] 62776866Skris 62876866Skris *) Fix CPU detection on Irix 6.x. 62976866Skris [Kurt Hockenbury <khockenb@stevens-tech.edu> and 63076866Skris "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 63176866Skris 63276866Skris *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 63376866Skris was empty. 63476866Skris [Steve Henson] 63576866Skris [This change does not apply to 0.9.7.] 63676866Skris 63776866Skris *) Use the cached encoding of an X509_NAME structure rather than 63876866Skris copying it. This is apparently the reason for the libsafe "errors" 63976866Skris but the code is actually correct. 64076866Skris [Steve Henson] 64176866Skris 64272613Skris *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 64372613Skris Bleichenbacher's DSA attack. 64476866Skris Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 64576866Skris to be set and top=0 forces the highest bit to be set; top=-1 is new 64676866Skris and leaves the highest bit random. 64776866Skris [Ulf Moeller, Bodo Moeller] 64872613Skris 64972613Skris *) In the NCONF_...-based implementations for CONF_... queries 65072613Skris (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 65172613Skris a temporary CONF structure with the data component set to NULL 65272613Skris (which gives segmentation faults in lh_retrieve). 65372613Skris Instead, use NULL for the CONF pointer in CONF_get_string and 65472613Skris CONF_get_number (which may use environment variables) and directly 65572613Skris return NULL from CONF_get_section. 65672613Skris [Bodo Moeller] 65772613Skris 65872613Skris *) Fix potential buffer overrun for EBCDIC. 65972613Skris [Ulf Moeller] 66072613Skris 66172613Skris *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 66272613Skris keyUsage if basicConstraints absent for a CA. 66372613Skris [Steve Henson] 66472613Skris 66572613Skris *) Make SMIME_write_PKCS7() write mail header values with a format that 66672613Skris is more generally accepted (no spaces before the semicolon), since 66772613Skris some programs can't parse those values properly otherwise. Also make 66872613Skris sure BIO's that break lines after each write do not create invalid 66972613Skris headers. 67072613Skris [Richard Levitte] 67172613Skris 67272613Skris *) Make the CRL encoding routines work with empty SEQUENCE OF. The 67372613Skris macros previously used would not encode an empty SEQUENCE OF 67472613Skris and break the signature. 67572613Skris [Steve Henson] 67676866Skris [This change does not apply to 0.9.7.] 67772613Skris 67872613Skris *) Zero the premaster secret after deriving the master secret in 67972613Skris DH ciphersuites. 68072613Skris [Steve Henson] 68172613Skris 68272613Skris *) Add some EVP_add_digest_alias registrations (as found in 68372613Skris OpenSSL_add_all_digests()) to SSL_library_init() 68472613Skris aka OpenSSL_add_ssl_algorithms(). This provides improved 68572613Skris compatibility with peers using X.509 certificates 68672613Skris with unconventional AlgorithmIdentifier OIDs. 68772613Skris [Bodo Moeller] 68872613Skris 68972613Skris *) Fix for Irix with NO_ASM. 69072613Skris ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 69172613Skris 69272613Skris *) ./config script fixes. 69372613Skris [Ulf Moeller, Richard Levitte] 69472613Skris 69572613Skris *) Fix 'openssl passwd -1'. 69672613Skris [Bodo Moeller] 69772613Skris 69872613Skris *) Change PKCS12_key_gen_asc() so it can cope with non null 69972613Skris terminated strings whose length is passed in the passlen 70072613Skris parameter, for example from PEM callbacks. This was done 70172613Skris by adding an extra length parameter to asc2uni(). 70272613Skris [Steve Henson, reported by <oddissey@samsung.co.kr>] 70372613Skris 70472613Skris *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 70572613Skris call failed, free the DSA structure. 70672613Skris [Bodo Moeller] 70772613Skris 70872613Skris *) Fix to uni2asc() to cope with zero length Unicode strings. 70972613Skris These are present in some PKCS#12 files. 71072613Skris [Steve Henson] 71172613Skris 71272613Skris *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 71372613Skris Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 71472613Skris when writing a 32767 byte record. 71572613Skris [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 71672613Skris 71772613Skris *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 71872613Skris obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 71972613Skris 72072613Skris (RSA objects have a reference count access to which is protected 72172613Skris by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 72272613Skris so they are meant to be shared between threads.) 72372613Skris [Bodo Moeller, Geoff Thorpe; original patch submitted by 72472613Skris "Reddie, Steven" <Steven.Reddie@ca.com>] 72572613Skris 72672613Skris *) Fix a deadlock in CRYPTO_mem_leaks(). 72772613Skris [Bodo Moeller] 72872613Skris 72976866Skris *) Use better test patterns in bntest. 73076866Skris [Ulf M�ller] 73176866Skris 73272613Skris *) rand_win.c fix for Borland C. 73372613Skris [Ulf M�ller] 73472613Skris 73572613Skris *) BN_rshift bugfix for n == 0. 73672613Skris [Bodo Moeller] 73772613Skris 73876866Skris *) Add a 'bctest' script that checks for some known 'bc' bugs 73976866Skris so that 'make test' does not abort just because 'bc' is broken. 74076866Skris [Bodo Moeller] 74176866Skris 74272613Skris *) Store verify_result within SSL_SESSION also for client side to 74372613Skris avoid potential security hole. (Re-used sessions on the client side 74472613Skris always resulted in verify_result==X509_V_OK, not using the original 74572613Skris result of the server certificate verification.) 74672613Skris [Lutz Jaenicke] 74772613Skris 74872613Skris *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 74972613Skris SSL3_RT_APPLICATION_DATA, return 0. 75072613Skris Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 75172613Skris [Bodo Moeller] 75272613Skris 75372613Skris *) Fix SSL_peek: 75472613Skris Both ssl2_peek and ssl3_peek, which were totally broken in earlier 75572613Skris releases, have been re-implemented by renaming the previous 75672613Skris implementations of ssl2_read and ssl3_read to ssl2_read_internal 75772613Skris and ssl3_read_internal, respectively, and adding 'peek' parameters 75872613Skris to them. The new ssl[23]_{read,peek} functions are calls to 75972613Skris ssl[23]_read_internal with the 'peek' flag set appropriately. 76072613Skris A 'peek' parameter has also been added to ssl3_read_bytes, which 76172613Skris does the actual work for ssl3_read_internal. 76272613Skris [Bodo Moeller] 76372613Skris 76476866Skris *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 76576866Skris the method-specific "init()" handler. Also clean up ex_data after 76676866Skris calling the method-specific "finish()" handler. Previously, this was 76776866Skris happening the other way round. 76876866Skris [Geoff Thorpe] 76976866Skris 77072613Skris *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 77172613Skris The previous value, 12, was not always sufficient for BN_mod_exp(). 77272613Skris [Bodo Moeller] 77372613Skris 77476866Skris *) Make sure that shared libraries get the internal name engine with 77576866Skris the full version number and not just 0. This should mark the 77676866Skris shared libraries as not backward compatible. Of course, this should 77776866Skris be changed again when we can guarantee backward binary compatibility. 77876866Skris [Richard Levitte] 77976866Skris 78072613Skris *) Fix typo in get_cert_by_subject() in by_dir.c 78172613Skris [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 78272613Skris 78376866Skris *) Rework the system to generate shared libraries: 78476866Skris 78576866Skris - Make note of the expected extension for the shared libraries and 78676866Skris if there is a need for symbolic links from for example libcrypto.so.0 78776866Skris to libcrypto.so.0.9.7. There is extended info in Configure for 78876866Skris that. 78976866Skris 79076866Skris - Make as few rebuilds of the shared libraries as possible. 79176866Skris 79276866Skris - Still avoid linking the OpenSSL programs with the shared libraries. 79376866Skris 79476866Skris - When installing, install the shared libraries separately from the 79576866Skris static ones. 79676866Skris [Richard Levitte] 79776866Skris 79872613Skris *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 79972613Skris 80072613Skris Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 80172613Skris and not in SSL_clear because the latter is also used by the 80272613Skris accept/connect functions; previously, the settings made by 80372613Skris SSL_set_read_ahead would be lost during the handshake. 80472613Skris [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 80572613Skris 80672613Skris *) Correct util/mkdef.pl to be selective about disabled algorithms. 80772613Skris Previously, it would create entries for disableed algorithms no 80872613Skris matter what. 80972613Skris [Richard Levitte] 81072613Skris 81172613Skris *) Added several new manual pages for SSL_* function. 81272613Skris [Lutz Jaenicke] 81372613Skris 81468651Skris Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 81568651Skris 81668651Skris *) In ssl23_get_client_hello, generate an error message when faced 81768651Skris with an initial SSL 3.0/TLS record that is too small to contain the 81868651Skris first two bytes of the ClientHello message, i.e. client_version. 81968651Skris (Note that this is a pathologic case that probably has never happened 82068651Skris in real life.) The previous approach was to use the version number 82168651Skris from the record header as a substitute; but our protocol choice 82268651Skris should not depend on that one because it is not authenticated 82368651Skris by the Finished messages. 82468651Skris [Bodo Moeller] 82568651Skris 82668651Skris *) More robust randomness gathering functions for Windows. 82768651Skris [Jeffrey Altman <jaltman@columbia.edu>] 82868651Skris 82968651Skris *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 83068651Skris not set then we don't setup the error code for issuer check errors 83168651Skris to avoid possibly overwriting other errors which the callback does 83268651Skris handle. If an application does set the flag then we assume it knows 83368651Skris what it is doing and can handle the new informational codes 83468651Skris appropriately. 83568651Skris [Steve Henson] 83668651Skris 83768651Skris *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 83868651Skris a general "ANY" type, as such it should be able to decode anything 83968651Skris including tagged types. However it didn't check the class so it would 84068651Skris wrongly interpret tagged types in the same way as their universal 84168651Skris counterpart and unknown types were just rejected. Changed so that the 84268651Skris tagged and unknown types are handled in the same way as a SEQUENCE: 84368651Skris that is the encoding is stored intact. There is also a new type 84468651Skris "V_ASN1_OTHER" which is used when the class is not universal, in this 84568651Skris case we have no idea what the actual type is so we just lump them all 84668651Skris together. 84768651Skris [Steve Henson] 84868651Skris 84968651Skris *) On VMS, stdout may very well lead to a file that is written to 85068651Skris in a record-oriented fashion. That means that every write() will 85168651Skris write a separate record, which will be read separately by the 85268651Skris programs trying to read from it. This can be very confusing. 85368651Skris 85468651Skris The solution is to put a BIO filter in the way that will buffer 85568651Skris text until a linefeed is reached, and then write everything a 85668651Skris line at a time, so every record written will be an actual line, 85768651Skris not chunks of lines and not (usually doesn't happen, but I've 85868651Skris seen it once) several lines in one record. BIO_f_linebuffer() is 85968651Skris the answer. 86068651Skris 86168651Skris Currently, it's a VMS-only method, because that's where it has 86268651Skris been tested well enough. 86368651Skris [Richard Levitte] 86468651Skris 86568651Skris *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 86668651Skris it can return incorrect results. 86768651Skris (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 86868651Skris but it was in 0.9.6-beta[12].) 86968651Skris [Bodo Moeller] 87068651Skris 87168651Skris *) Disable the check for content being present when verifying detached 87268651Skris signatures in pk7_smime.c. Some versions of Netscape (wrongly) 87368651Skris include zero length content when signing messages. 87468651Skris [Steve Henson] 87568651Skris 87668651Skris *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 87768651Skris BIO_ctrl (for BIO pairs). 87868651Skris [Bodo M�ller] 87968651Skris 88068651Skris *) Add DSO method for VMS. 88168651Skris [Richard Levitte] 88268651Skris 88368651Skris *) Bug fix: Montgomery multiplication could produce results with the 88468651Skris wrong sign. 88568651Skris [Ulf M�ller] 88668651Skris 88768651Skris *) Add RPM specification openssl.spec and modify it to build three 88868651Skris packages. The default package contains applications, application 88968651Skris documentation and run-time libraries. The devel package contains 89068651Skris include files, static libraries and function documentation. The 89168651Skris doc package contains the contents of the doc directory. The original 89268651Skris openssl.spec was provided by Damien Miller <djm@mindrot.org>. 89368651Skris [Richard Levitte] 89468651Skris 89568651Skris *) Add a large number of documentation files for many SSL routines. 89668651Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 89768651Skris 89868651Skris *) Add a configuration entry for Sony News 4. 89968651Skris [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 90068651Skris 90168651Skris *) Don't set the two most significant bits to one when generating a 90268651Skris random number < q in the DSA library. 90368651Skris [Ulf M�ller] 90468651Skris 90568651Skris *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 90668651Skris behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 90768651Skris the underlying transport is blocking) if a handshake took place. 90868651Skris (The default behaviour is needed by applications such as s_client 90968651Skris and s_server that use select() to determine when to use SSL_read; 91068651Skris but for applications that know in advance when to expect data, it 91168651Skris just makes things more complicated.) 91268651Skris [Bodo Moeller] 91368651Skris 91468651Skris *) Add RAND_egd_bytes(), which gives control over the number of bytes read 91568651Skris from EGD. 91668651Skris [Ben Laurie] 91768651Skris 91868651Skris *) Add a few more EBCDIC conditionals that make `req' and `x509' 91968651Skris work better on such systems. 92068651Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 92168651Skris 92268651Skris *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 92368651Skris Update PKCS12_parse() so it copies the friendlyName and the 92468651Skris keyid to the certificates aux info. 92568651Skris [Steve Henson] 92668651Skris 92768651Skris *) Fix bug in PKCS7_verify() which caused an infinite loop 92868651Skris if there was more than one signature. 92968651Skris [Sven Uszpelkat <su@celocom.de>] 93068651Skris 93168651Skris *) Major change in util/mkdef.pl to include extra information 93268651Skris about each symbol, as well as presentig variables as well 93368651Skris as functions. This change means that there's n more need 93468651Skris to rebuild the .num files when some algorithms are excluded. 93568651Skris [Richard Levitte] 93668651Skris 93768651Skris *) Allow the verify time to be set by an application, 93868651Skris rather than always using the current time. 93968651Skris [Steve Henson] 94068651Skris 94168651Skris *) Phase 2 verify code reorganisation. The certificate 94268651Skris verify code now looks up an issuer certificate by a 94368651Skris number of criteria: subject name, authority key id 94468651Skris and key usage. It also verifies self signed certificates 94568651Skris by the same criteria. The main comparison function is 94668651Skris X509_check_issued() which performs these checks. 94768651Skris 94868651Skris Lot of changes were necessary in order to support this 94968651Skris without completely rewriting the lookup code. 95068651Skris 95168651Skris Authority and subject key identifier are now cached. 95268651Skris 95368651Skris The LHASH 'certs' is X509_STORE has now been replaced 95468651Skris by a STACK_OF(X509_OBJECT). This is mainly because an 95568651Skris LHASH can't store or retrieve multiple objects with 95668651Skris the same hash value. 95768651Skris 95868651Skris As a result various functions (which were all internal 95968651Skris use only) have changed to handle the new X509_STORE 96068651Skris structure. This will break anything that messed round 96168651Skris with X509_STORE internally. 96268651Skris 96368651Skris The functions X509_STORE_add_cert() now checks for an 96468651Skris exact match, rather than just subject name. 96568651Skris 96668651Skris The X509_STORE API doesn't directly support the retrieval 96768651Skris of multiple certificates matching a given criteria, however 96868651Skris this can be worked round by performing a lookup first 96968651Skris (which will fill the cache with candidate certificates) 97068651Skris and then examining the cache for matches. This is probably 97168651Skris the best we can do without throwing out X509_LOOKUP 97268651Skris entirely (maybe later...). 97368651Skris 97468651Skris The X509_VERIFY_CTX structure has been enhanced considerably. 97568651Skris 97668651Skris All certificate lookup operations now go via a get_issuer() 97768651Skris callback. Although this currently uses an X509_STORE it 97868651Skris can be replaced by custom lookups. This is a simple way 97968651Skris to bypass the X509_STORE hackery necessary to make this 98068651Skris work and makes it possible to use more efficient techniques 98168651Skris in future. A very simple version which uses a simple 98268651Skris STACK for its trusted certificate store is also provided 98368651Skris using X509_STORE_CTX_trusted_stack(). 98468651Skris 98568651Skris The verify_cb() and verify() callbacks now have equivalents 98668651Skris in the X509_STORE_CTX structure. 98768651Skris 98868651Skris X509_STORE_CTX also has a 'flags' field which can be used 98968651Skris to customise the verify behaviour. 99068651Skris [Steve Henson] 99168651Skris 99268651Skris *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 99368651Skris excludes S/MIME capabilities. 99468651Skris [Steve Henson] 99568651Skris 99668651Skris *) When a certificate request is read in keep a copy of the 99768651Skris original encoding of the signed data and use it when outputing 99868651Skris again. Signatures then use the original encoding rather than 99968651Skris a decoded, encoded version which may cause problems if the 100068651Skris request is improperly encoded. 100168651Skris [Steve Henson] 100268651Skris 100368651Skris *) For consistency with other BIO_puts implementations, call 100468651Skris buffer_write(b, ...) directly in buffer_puts instead of calling 100568651Skris BIO_write(b, ...). 100668651Skris 100768651Skris In BIO_puts, increment b->num_write as in BIO_write. 100868651Skris [Peter.Sylvester@EdelWeb.fr] 100968651Skris 101068651Skris *) Fix BN_mul_word for the case where the word is 0. (We have to use 101168651Skris BN_zero, we may not return a BIGNUM with an array consisting of 101268651Skris words set to zero.) 101368651Skris [Bodo Moeller] 101468651Skris 101568651Skris *) Avoid calling abort() from within the library when problems are 101668651Skris detected, except if preprocessor symbols have been defined 101768651Skris (such as REF_CHECK, BN_DEBUG etc.). 101868651Skris [Bodo Moeller] 101968651Skris 102068651Skris *) New openssl application 'rsautl'. This utility can be 102168651Skris used for low level RSA operations. DER public key 102268651Skris BIO/fp routines also added. 102368651Skris [Steve Henson] 102468651Skris 102568651Skris *) New Configure entry and patches for compiling on QNX 4. 102668651Skris [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 102768651Skris 102868651Skris *) A demo state-machine implementation was sponsored by 102968651Skris Nuron (http://www.nuron.com/) and is now available in 103068651Skris demos/state_machine. 103168651Skris [Ben Laurie] 103268651Skris 103368651Skris *) New options added to the 'dgst' utility for signature 103468651Skris generation and verification. 103568651Skris [Steve Henson] 103668651Skris 103768651Skris *) Unrecognized PKCS#7 content types are now handled via a 103868651Skris catch all ASN1_TYPE structure. This allows unsupported 103968651Skris types to be stored as a "blob" and an application can 104068651Skris encode and decode it manually. 104168651Skris [Steve Henson] 104268651Skris 104368651Skris *) Fix various signed/unsigned issues to make a_strex.c 104468651Skris compile under VC++. 104568651Skris [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 104668651Skris 104768651Skris *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 104868651Skris length if passed a buffer. ASN1_INTEGER_to_BN failed 104968651Skris if passed a NULL BN and its argument was negative. 105068651Skris [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 105168651Skris 105268651Skris *) Modification to PKCS#7 encoding routines to output definite 105368651Skris length encoding. Since currently the whole structures are in 105468651Skris memory there's not real point in using indefinite length 105568651Skris constructed encoding. However if OpenSSL is compiled with 105668651Skris the flag PKCS7_INDEFINITE_ENCODING the old form is used. 105768651Skris [Steve Henson] 105868651Skris 105968651Skris *) Added BIO_vprintf() and BIO_vsnprintf(). 106068651Skris [Richard Levitte] 106168651Skris 106268651Skris *) Added more prefixes to parse for in the the strings written 106368651Skris through a logging bio, to cover all the levels that are available 106468651Skris through syslog. The prefixes are now: 106568651Skris 106668651Skris PANIC, EMERG, EMR => LOG_EMERG 106768651Skris ALERT, ALR => LOG_ALERT 106868651Skris CRIT, CRI => LOG_CRIT 106968651Skris ERROR, ERR => LOG_ERR 107068651Skris WARNING, WARN, WAR => LOG_WARNING 107168651Skris NOTICE, NOTE, NOT => LOG_NOTICE 107268651Skris INFO, INF => LOG_INFO 107368651Skris DEBUG, DBG => LOG_DEBUG 107468651Skris 107568651Skris and as before, if none of those prefixes are present at the 107668651Skris beginning of the string, LOG_ERR is chosen. 107768651Skris 107868651Skris On Win32, the LOG_* levels are mapped according to this: 107968651Skris 108068651Skris LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 108168651Skris LOG_WARNING => EVENTLOG_WARNING_TYPE 108268651Skris LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 108368651Skris 108468651Skris [Richard Levitte] 108568651Skris 108668651Skris *) Made it possible to reconfigure with just the configuration 108768651Skris argument "reconf" or "reconfigure". The command line arguments 108868651Skris are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 108968651Skris and are retrieved from there when reconfiguring. 109068651Skris [Richard Levitte] 109168651Skris 109268651Skris *) MD4 implemented. 109368651Skris [Assar Westerlund <assar@sics.se>, Richard Levitte] 109468651Skris 109568651Skris *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 109668651Skris [Richard Levitte] 109768651Skris 109868651Skris *) The obj_dat.pl script was messing up the sorting of object 109968651Skris names. The reason was that it compared the quoted version 110068651Skris of strings as a result "OCSP" > "OCSP Signing" because 110168651Skris " > SPACE. Changed script to store unquoted versions of 110268651Skris names and add quotes on output. It was also omitting some 110368651Skris names from the lookup table if they were given a default 110468651Skris value (that is if SN is missing it is given the same 110568651Skris value as LN and vice versa), these are now added on the 110668651Skris grounds that if an object has a name we should be able to 110768651Skris look it up. Finally added warning output when duplicate 110868651Skris short or long names are found. 110968651Skris [Steve Henson] 111068651Skris 111168651Skris *) Changes needed for Tandem NSK. 111268651Skris [Scott Uroff <scott@xypro.com>] 111368651Skris 111468651Skris *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 111568651Skris RSA_padding_check_SSLv23(), special padding was never detected 111668651Skris and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 111768651Skris version rollback attacks was not effective. 111868651Skris 111968651Skris In s23_clnt.c, don't use special rollback-attack detection padding 112068651Skris (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 112168651Skris client; similarly, in s23_srvr.c, don't do the rollback check if 112268651Skris SSL 2.0 is the only protocol enabled in the server. 112368651Skris [Bodo Moeller] 112468651Skris 112568651Skris *) Make it possible to get hexdumps of unprintable data with 'openssl 112668651Skris asn1parse'. By implication, the functions ASN1_parse_dump() and 112768651Skris BIO_dump_indent() are added. 112868651Skris [Richard Levitte] 112968651Skris 113068651Skris *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 113168651Skris these print out strings and name structures based on various 113268651Skris flags including RFC2253 support and proper handling of 113368651Skris multibyte characters. Added options to the 'x509' utility 113468651Skris to allow the various flags to be set. 113568651Skris [Steve Henson] 113668651Skris 113768651Skris *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 113868651Skris Also change the functions X509_cmp_current_time() and 113968651Skris X509_gmtime_adj() work with an ASN1_TIME structure, 114068651Skris this will enable certificates using GeneralizedTime in validity 114168651Skris dates to be checked. 114268651Skris [Steve Henson] 114368651Skris 114468651Skris *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 114568651Skris negative public key encodings) on by default, 114668651Skris NO_NEG_PUBKEY_BUG can be set to disable it. 114768651Skris [Steve Henson] 114868651Skris 114968651Skris *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 115068651Skris content octets. An i2c_ASN1_OBJECT is unnecessary because 115168651Skris the encoding can be trivially obtained from the structure. 115268651Skris [Steve Henson] 115368651Skris 115468651Skris *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 115568651Skris not read locks (CRYPTO_r_[un]lock). 115668651Skris [Bodo Moeller] 115768651Skris 115868651Skris *) A first attempt at creating official support for shared 115968651Skris libraries through configuration. I've kept it so the 116068651Skris default is static libraries only, and the OpenSSL programs 116168651Skris are always statically linked for now, but there are 116268651Skris preparations for dynamic linking in place. 116389837Skris This has been tested on Linux and Tru64. 116468651Skris [Richard Levitte] 116568651Skris 116668651Skris *) Randomness polling function for Win9x, as described in: 116768651Skris Peter Gutmann, Software Generation of Practically Strong 116868651Skris Random Numbers. 116968651Skris [Ulf M�ller] 117068651Skris 117168651Skris *) Fix so PRNG is seeded in req if using an already existing 117268651Skris DSA key. 117368651Skris [Steve Henson] 117468651Skris 117568651Skris *) New options to smime application. -inform and -outform 117668651Skris allow alternative formats for the S/MIME message including 117768651Skris PEM and DER. The -content option allows the content to be 117868651Skris specified separately. This should allow things like Netscape 117968651Skris form signing output easier to verify. 118068651Skris [Steve Henson] 118168651Skris 118268651Skris *) Fix the ASN1 encoding of tags using the 'long form'. 118368651Skris [Steve Henson] 118468651Skris 118568651Skris *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 118668651Skris STRING types. These convert content octets to and from the 118768651Skris underlying type. The actual tag and length octets are 118868651Skris already assumed to have been read in and checked. These 118968651Skris are needed because all other string types have virtually 119068651Skris identical handling apart from the tag. By having versions 119168651Skris of the ASN1 functions that just operate on content octets 119268651Skris IMPLICIT tagging can be handled properly. It also allows 119368651Skris the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 119468651Skris and ASN1_INTEGER are identical apart from the tag. 119568651Skris [Steve Henson] 119668651Skris 119768651Skris *) Change the handling of OID objects as follows: 119868651Skris 119968651Skris - New object identifiers are inserted in objects.txt, following 120068651Skris the syntax given in objects.README. 120168651Skris - objects.pl is used to process obj_mac.num and create a new 120268651Skris obj_mac.h. 120368651Skris - obj_dat.pl is used to create a new obj_dat.h, using the data in 120468651Skris obj_mac.h. 120568651Skris 120668651Skris This is currently kind of a hack, and the perl code in objects.pl 120768651Skris isn't very elegant, but it works as I intended. The simplest way 120868651Skris to check that it worked correctly is to look in obj_dat.h and 120968651Skris check the array nid_objs and make sure the objects haven't moved 121068651Skris around (this is important!). Additions are OK, as well as 121168651Skris consistent name changes. 121268651Skris [Richard Levitte] 121368651Skris 121468651Skris *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 121568651Skris [Bodo Moeller] 121668651Skris 121768651Skris *) Addition of the command line parameter '-rand file' to 'openssl req'. 121868651Skris The given file adds to whatever has already been seeded into the 121968651Skris random pool through the RANDFILE configuration file option or 122068651Skris environment variable, or the default random state file. 122168651Skris [Richard Levitte] 122268651Skris 122368651Skris *) mkstack.pl now sorts each macro group into lexical order. 122468651Skris Previously the output order depended on the order the files 122568651Skris appeared in the directory, resulting in needless rewriting 122668651Skris of safestack.h . 122768651Skris [Steve Henson] 122868651Skris 122968651Skris *) Patches to make OpenSSL compile under Win32 again. Mostly 123068651Skris work arounds for the VC++ problem that it treats func() as 123168651Skris func(void). Also stripped out the parts of mkdef.pl that 123268651Skris added extra typesafe functions: these no longer exist. 123368651Skris [Steve Henson] 123468651Skris 123568651Skris *) Reorganisation of the stack code. The macros are now all 123668651Skris collected in safestack.h . Each macro is defined in terms of 123768651Skris a "stack macro" of the form SKM_<name>(type, a, b). The 123868651Skris DEBUG_SAFESTACK is now handled in terms of function casts, 123968651Skris this has the advantage of retaining type safety without the 124068651Skris use of additional functions. If DEBUG_SAFESTACK is not defined 124168651Skris then the non typesafe macros are used instead. Also modified the 124268651Skris mkstack.pl script to handle the new form. Needs testing to see 124368651Skris if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 124468651Skris the default if no major problems. Similar behaviour for ASN1_SET_OF 124568651Skris and PKCS12_STACK_OF. 124668651Skris [Steve Henson] 124768651Skris 124868651Skris *) When some versions of IIS use the 'NET' form of private key the 124968651Skris key derivation algorithm is different. Normally MD5(password) is 125068651Skris used as a 128 bit RC4 key. In the modified case 125168651Skris MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 125268651Skris new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 125368651Skris as the old Netscape_RSA functions except they have an additional 125468651Skris 'sgckey' parameter which uses the modified algorithm. Also added 125568651Skris an -sgckey command line option to the rsa utility. Thanks to 125668651Skris Adrian Peck <bertie@ncipher.com> for posting details of the modified 125768651Skris algorithm to openssl-dev. 125868651Skris [Steve Henson] 125968651Skris 126068651Skris *) The evp_local.h macros were using 'c.##kname' which resulted in 126168651Skris invalid expansion on some systems (SCO 5.0.5 for example). 126268651Skris Corrected to 'c.kname'. 126368651Skris [Phillip Porch <root@theporch.com>] 126468651Skris 126568651Skris *) New X509_get1_email() and X509_REQ_get1_email() functions that return 126668651Skris a STACK of email addresses from a certificate or request, these look 126768651Skris in the subject name and the subject alternative name extensions and 126868651Skris omit any duplicate addresses. 126968651Skris [Steve Henson] 127068651Skris 127168651Skris *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 127268651Skris This makes DSA verification about 2 % faster. 127368651Skris [Bodo Moeller] 127468651Skris 127568651Skris *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 127668651Skris (meaning that now 2^5 values will be precomputed, which is only 4 KB 127768651Skris plus overhead for 1024 bit moduli). 127868651Skris This makes exponentiations about 0.5 % faster for 1024 bit 127968651Skris exponents (as measured by "openssl speed rsa2048"). 128068651Skris [Bodo Moeller] 128168651Skris 128268651Skris *) Rename memory handling macros to avoid conflicts with other 128368651Skris software: 128468651Skris Malloc => OPENSSL_malloc 128568651Skris Malloc_locked => OPENSSL_malloc_locked 128668651Skris Realloc => OPENSSL_realloc 128768651Skris Free => OPENSSL_free 128868651Skris [Richard Levitte] 128968651Skris 129068651Skris *) New function BN_mod_exp_mont_word for small bases (roughly 15% 129168651Skris faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 129268651Skris [Bodo Moeller] 129368651Skris 129468651Skris *) CygWin32 support. 129568651Skris [John Jarvie <jjarvie@newsguy.com>] 129668651Skris 129768651Skris *) The type-safe stack code has been rejigged. It is now only compiled 129868651Skris in when OpenSSL is configured with the DEBUG_SAFESTACK option and 129968651Skris by default all type-specific stack functions are "#define"d back to 130068651Skris standard stack functions. This results in more streamlined output 130168651Skris but retains the type-safety checking possibilities of the original 130268651Skris approach. 130368651Skris [Geoff Thorpe] 130468651Skris 130568651Skris *) The STACK code has been cleaned up, and certain type declarations 130668651Skris that didn't make a lot of sense have been brought in line. This has 130768651Skris also involved a cleanup of sorts in safestack.h to more correctly 130868651Skris map type-safe stack functions onto their plain stack counterparts. 130968651Skris This work has also resulted in a variety of "const"ifications of 131068651Skris lots of the code, especially "_cmp" operations which should normally 131168651Skris be prototyped with "const" parameters anyway. 131268651Skris [Geoff Thorpe] 131368651Skris 131468651Skris *) When generating bytes for the first time in md_rand.c, 'stir the pool' 131568651Skris by seeding with STATE_SIZE dummy bytes (with zero entropy count). 131668651Skris (The PRNG state consists of two parts, the large pool 'state' and 'md', 131768651Skris where all of 'md' is used each time the PRNG is used, but 'state' 131868651Skris is used only indexed by a cyclic counter. As entropy may not be 131968651Skris well distributed from the beginning, 'md' is important as a 132068651Skris chaining variable. However, the output function chains only half 132168651Skris of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 132268651Skris all of 'md', and seeding with STATE_SIZE dummy bytes will result 132368651Skris in all of 'state' being rewritten, with the new values depending 132468651Skris on virtually all of 'md'. This overcomes the 80 bit limitation.) 132568651Skris [Bodo Moeller] 132668651Skris 132768651Skris *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 132868651Skris the handshake is continued after ssl_verify_cert_chain(); 132968651Skris otherwise, if SSL_VERIFY_NONE is set, remaining error codes 133068651Skris can lead to 'unexplainable' connection aborts later. 133168651Skris [Bodo Moeller; problem tracked down by Lutz Jaenicke] 133268651Skris 133368651Skris *) Major EVP API cipher revision. 133468651Skris Add hooks for extra EVP features. This allows various cipher 133568651Skris parameters to be set in the EVP interface. Support added for variable 133668651Skris key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 133768651Skris setting of RC2 and RC5 parameters. 133868651Skris 133968651Skris Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 134068651Skris ciphers. 134168651Skris 134268651Skris Remove lots of duplicated code from the EVP library. For example *every* 134368651Skris cipher init() function handles the 'iv' in the same way according to the 134468651Skris cipher mode. They also all do nothing if the 'key' parameter is NULL and 134568651Skris for CFB and OFB modes they zero ctx->num. 134668651Skris 134768651Skris New functionality allows removal of S/MIME code RC2 hack. 134868651Skris 134968651Skris Most of the routines have the same form and so can be declared in terms 135068651Skris of macros. 135168651Skris 135268651Skris By shifting this to the top level EVP_CipherInit() it can be removed from 135368651Skris all individual ciphers. If the cipher wants to handle IVs or keys 135468651Skris differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 135568651Skris flags. 135668651Skris 135768651Skris Change lots of functions like EVP_EncryptUpdate() to now return a 135868651Skris value: although software versions of the algorithms cannot fail 135968651Skris any installed hardware versions can. 136068651Skris [Steve Henson] 136168651Skris 136268651Skris *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 136368651Skris this option is set, tolerate broken clients that send the negotiated 136468651Skris protocol version number instead of the requested protocol version 136568651Skris number. 136668651Skris [Bodo Moeller] 136768651Skris 136868651Skris *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 136968651Skris i.e. non-zero for export ciphersuites, zero otherwise. 137068651Skris Previous versions had this flag inverted, inconsistent with 137168651Skris rsa_tmp_cb (..._TMP_RSA_CB). 137268651Skris [Bodo Moeller; problem reported by Amit Chopra] 137368651Skris 137468651Skris *) Add missing DSA library text string. Work around for some IIS 137568651Skris key files with invalid SEQUENCE encoding. 137668651Skris [Steve Henson] 137768651Skris 137868651Skris *) Add a document (doc/standards.txt) that list all kinds of standards 137968651Skris and so on that are implemented in OpenSSL. 138068651Skris [Richard Levitte] 138168651Skris 138268651Skris *) Enhance c_rehash script. Old version would mishandle certificates 138368651Skris with the same subject name hash and wouldn't handle CRLs at all. 138468651Skris Added -fingerprint option to crl utility, to support new c_rehash 138568651Skris features. 138668651Skris [Steve Henson] 138768651Skris 138868651Skris *) Eliminate non-ANSI declarations in crypto.h and stack.h. 138968651Skris [Ulf M�ller] 139068651Skris 139168651Skris *) Fix for SSL server purpose checking. Server checking was 139268651Skris rejecting certificates which had extended key usage present 139368651Skris but no ssl client purpose. 139468651Skris [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 139568651Skris 139668651Skris *) Make PKCS#12 code work with no password. The PKCS#12 spec 139768651Skris is a little unclear about how a blank password is handled. 139868651Skris Since the password in encoded as a BMPString with terminating 139968651Skris double NULL a zero length password would end up as just the 140068651Skris double NULL. However no password at all is different and is 140168651Skris handled differently in the PKCS#12 key generation code. NS 140268651Skris treats a blank password as zero length. MSIE treats it as no 140368651Skris password on export: but it will try both on import. We now do 140468651Skris the same: PKCS12_parse() tries zero length and no password if 140568651Skris the password is set to "" or NULL (NULL is now a valid password: 140668651Skris it wasn't before) as does the pkcs12 application. 140768651Skris [Steve Henson] 140868651Skris 140968651Skris *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 141068651Skris perror when PEM_read_bio_X509_REQ fails, the error message must 141168651Skris be obtained from the error queue. 141268651Skris [Bodo Moeller] 141368651Skris 141468651Skris *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 141568651Skris it in ERR_remove_state if appropriate, and change ERR_get_state 141668651Skris accordingly to avoid race conditions (this is necessary because 141768651Skris thread_hash is no longer constant once set). 141868651Skris [Bodo Moeller] 141968651Skris 142068651Skris *) Bugfix for linux-elf makefile.one. 142168651Skris [Ulf M�ller] 142268651Skris 142368651Skris *) RSA_get_default_method() will now cause a default 142468651Skris RSA_METHOD to be chosen if one doesn't exist already. 142568651Skris Previously this was only set during a call to RSA_new() 142668651Skris or RSA_new_method(NULL) meaning it was possible for 142768651Skris RSA_get_default_method() to return NULL. 142868651Skris [Geoff Thorpe] 142968651Skris 143068651Skris *) Added native name translation to the existing DSO code 143168651Skris that will convert (if the flag to do so is set) filenames 143268651Skris that are sufficiently small and have no path information 143368651Skris into a canonical native form. Eg. "blah" converted to 143468651Skris "libblah.so" or "blah.dll" etc. 143568651Skris [Geoff Thorpe] 143668651Skris 143768651Skris *) New function ERR_error_string_n(e, buf, len) which is like 143868651Skris ERR_error_string(e, buf), but writes at most 'len' bytes 143968651Skris including the 0 terminator. For ERR_error_string_n, 'buf' 144068651Skris may not be NULL. 144168651Skris [Damien Miller <djm@mindrot.org>, Bodo Moeller] 144268651Skris 144368651Skris *) CONF library reworked to become more general. A new CONF 144468651Skris configuration file reader "class" is implemented as well as a 144568651Skris new functions (NCONF_*, for "New CONF") to handle it. The now 144668651Skris old CONF_* functions are still there, but are reimplemented to 144768651Skris work in terms of the new functions. Also, a set of functions 144868651Skris to handle the internal storage of the configuration data is 144968651Skris provided to make it easier to write new configuration file 145068651Skris reader "classes" (I can definitely see something reading a 145168651Skris configuration file in XML format, for example), called _CONF_*, 145268651Skris or "the configuration storage API"... 145368651Skris 145468651Skris The new configuration file reading functions are: 145568651Skris 145668651Skris NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 145768651Skris NCONF_get_section, NCONF_get_string, NCONF_get_numbre 145868651Skris 145968651Skris NCONF_default, NCONF_WIN32 146068651Skris 146168651Skris NCONF_dump_fp, NCONF_dump_bio 146268651Skris 146368651Skris NCONF_default and NCONF_WIN32 are method (or "class") choosers, 146468651Skris NCONF_new creates a new CONF object. This works in the same way 146568651Skris as other interfaces in OpenSSL, like the BIO interface. 146668651Skris NCONF_dump_* dump the internal storage of the configuration file, 146768651Skris which is useful for debugging. All other functions take the same 146868651Skris arguments as the old CONF_* functions wth the exception of the 146968651Skris first that must be a `CONF *' instead of a `LHASH *'. 147068651Skris 147168651Skris To make it easer to use the new classes with the old CONF_* functions, 147268651Skris the function CONF_set_default_method is provided. 147368651Skris [Richard Levitte] 147468651Skris 147568651Skris *) Add '-tls1' option to 'openssl ciphers', which was already 147668651Skris mentioned in the documentation but had not been implemented. 147768651Skris (This option is not yet really useful because even the additional 147868651Skris experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 147968651Skris [Bodo Moeller] 148068651Skris 148168651Skris *) Initial DSO code added into libcrypto for letting OpenSSL (and 148268651Skris OpenSSL-based applications) load shared libraries and bind to 148368651Skris them in a portable way. 148468651Skris [Geoff Thorpe, with contributions from Richard Levitte] 148568651Skris 148659191Skris Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 148759191Skris 148859191Skris *) Make sure _lrotl and _lrotr are only used with MSVC. 148959191Skris 149059191Skris *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 149159191Skris (the default implementation of RAND_status). 149259191Skris 149359191Skris *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 149459191Skris to '-clrext' (= clear extensions), as intended and documented. 149559191Skris [Bodo Moeller; inconsistency pointed out by Michael Attili 149659191Skris <attili@amaxo.com>] 149759191Skris 149859191Skris *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 149959191Skris was larger than the MD block size. 150059191Skris [Steve Henson, pointed out by Yost William <YostW@tce.com>] 150159191Skris 150259191Skris *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 150359191Skris fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 150459191Skris using the passed key: if the passed key was a private key the result 150559191Skris of X509_print(), for example, would be to print out all the private key 150659191Skris components. 150759191Skris [Steve Henson] 150859191Skris 150959191Skris *) des_quad_cksum() byte order bug fix. 151059191Skris [Ulf M�ller, using the problem description in krb4-0.9.7, where 151159191Skris the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 151259191Skris 151359191Skris *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 151459191Skris discouraged. 151559191Skris [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 151659191Skris 151759191Skris *) For easily testing in shell scripts whether some command 151859191Skris 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 151959191Skris returns with exit code 0 iff no command of the given name is available. 152059191Skris 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 152159191Skris the output goes to stdout and nothing is printed to stderr. 152259191Skris Additional arguments are always ignored. 152359191Skris 152459191Skris Since for each cipher there is a command of the same name, 152559191Skris the 'no-cipher' compilation switches can be tested this way. 152659191Skris 152759191Skris ('openssl no-XXX' is not able to detect pseudo-commands such 152859191Skris as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 152959191Skris [Bodo Moeller] 153059191Skris 153159191Skris *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 153259191Skris [Bodo Moeller] 153359191Skris 153459191Skris *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 153559191Skris is set; it will be thrown away anyway because each handshake creates 153659191Skris its own key. 153759191Skris ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 153859191Skris to parameters -- in previous versions (since OpenSSL 0.9.3) the 153959191Skris 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 154059191Skris you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 154159191Skris [Bodo Moeller] 154259191Skris 154359191Skris *) New s_client option -ign_eof: EOF at stdin is ignored, and 154459191Skris 'Q' and 'R' lose their special meanings (quit/renegotiate). 154559191Skris This is part of what -quiet does; unlike -quiet, -ign_eof 154659191Skris does not suppress any output. 154759191Skris [Richard Levitte] 154859191Skris 154959191Skris *) Add compatibility options to the purpose and trust code. The 155059191Skris purpose X509_PURPOSE_ANY is "any purpose" which automatically 155159191Skris accepts a certificate or CA, this was the previous behaviour, 155259191Skris with all the associated security issues. 155359191Skris 155459191Skris X509_TRUST_COMPAT is the old trust behaviour: only and 155559191Skris automatically trust self signed roots in certificate store. A 155659191Skris new trust setting X509_TRUST_DEFAULT is used to specify that 155759191Skris a purpose has no associated trust setting and it should instead 155859191Skris use the value in the default purpose. 155959191Skris [Steve Henson] 156059191Skris 156159191Skris *) Fix the PKCS#8 DSA private key code so it decodes keys again 156259191Skris and fix a memory leak. 156359191Skris [Steve Henson] 156459191Skris 156559191Skris *) In util/mkerr.pl (which implements 'make errors'), preserve 156659191Skris reason strings from the previous version of the .c file, as 156759191Skris the default to have only downcase letters (and digits) in 156859191Skris automatically generated reasons codes is not always appropriate. 156959191Skris [Bodo Moeller] 157059191Skris 157159191Skris *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 157259191Skris using strerror. Previously, ERR_reason_error_string() returned 157359191Skris library names as reason strings for SYSerr; but SYSerr is a special 157459191Skris case where small numbers are errno values, not library numbers. 157559191Skris [Bodo Moeller] 157659191Skris 157759191Skris *) Add '-dsaparam' option to 'openssl dhparam' application. This 157859191Skris converts DSA parameters into DH parameters. (When creating parameters, 157959191Skris DSA_generate_parameters is used.) 158059191Skris [Bodo Moeller] 158159191Skris 158259191Skris *) Include 'length' (recommended exponent length) in C code generated 158359191Skris by 'openssl dhparam -C'. 158459191Skris [Bodo Moeller] 158559191Skris 158659191Skris *) The second argument to set_label in perlasm was already being used 158759191Skris so couldn't be used as a "file scope" flag. Moved to third argument 158859191Skris which was free. 158959191Skris [Steve Henson] 159059191Skris 159159191Skris *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 159259191Skris instead of RAND_bytes for encryption IVs and salts. 159359191Skris [Bodo Moeller] 159459191Skris 159559191Skris *) Include RAND_status() into RAND_METHOD instead of implementing 159659191Skris it only for md_rand.c Otherwise replacing the PRNG by calling 159759191Skris RAND_set_rand_method would be impossible. 159859191Skris [Bodo Moeller] 159959191Skris 160059191Skris *) Don't let DSA_generate_key() enter an infinite loop if the random 160159191Skris number generation fails. 160259191Skris [Bodo Moeller] 160359191Skris 160459191Skris *) New 'rand' application for creating pseudo-random output. 160559191Skris [Bodo Moeller] 160659191Skris 160759191Skris *) Added configuration support for Linux/IA64 160859191Skris [Rolf Haberrecker <rolf@suse.de>] 160959191Skris 161059191Skris *) Assembler module support for Mingw32. 161159191Skris [Ulf M�ller] 161259191Skris 161359191Skris *) Shared library support for HPUX (in shlib/). 161459191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 161559191Skris 161659191Skris *) Shared library support for Solaris gcc. 161759191Skris [Lutz Behnke <behnke@trustcenter.de>] 161859191Skris 161959191Skris Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 162059191Skris 162159191Skris *) PKCS7_encrypt() was adding text MIME headers twice because they 162259191Skris were added manually and by SMIME_crlf_copy(). 162359191Skris [Steve Henson] 162459191Skris 162559191Skris *) In bntest.c don't call BN_rand with zero bits argument. 162659191Skris [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 162759191Skris 162859191Skris *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 162959191Skris case was implemented. This caused BN_div_recp() to fail occasionally. 163059191Skris [Ulf M�ller] 163159191Skris 163259191Skris *) Add an optional second argument to the set_label() in the perl 163359191Skris assembly language builder. If this argument exists and is set 163459191Skris to 1 it signals that the assembler should use a symbol whose 163559191Skris scope is the entire file, not just the current function. This 163659191Skris is needed with MASM which uses the format label:: for this scope. 163759191Skris [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 163859191Skris 163959191Skris *) Change the ASN1 types so they are typedefs by default. Before 164059191Skris almost all types were #define'd to ASN1_STRING which was causing 164159191Skris STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 164259191Skris for example. 164359191Skris [Steve Henson] 164459191Skris 164559191Skris *) Change names of new functions to the new get1/get0 naming 164659191Skris convention: After 'get1', the caller owns a reference count 164759191Skris and has to call ..._free; 'get0' returns a pointer to some 164859191Skris data structure without incrementing reference counters. 164959191Skris (Some of the existing 'get' functions increment a reference 165059191Skris counter, some don't.) 165159191Skris Similarly, 'set1' and 'add1' functions increase reference 165259191Skris counters or duplicate objects. 165359191Skris [Steve Henson] 165459191Skris 165559191Skris *) Allow for the possibility of temp RSA key generation failure: 165659191Skris the code used to assume it always worked and crashed on failure. 165759191Skris [Steve Henson] 165859191Skris 165959191Skris *) Fix potential buffer overrun problem in BIO_printf(). 166059191Skris [Ulf M�ller, using public domain code by Patrick Powell; problem 166159191Skris pointed out by David Sacerdote <das33@cornell.edu>] 166259191Skris 166359191Skris *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 166459191Skris RAND_egd() and RAND_status(). In the command line application, 166559191Skris the EGD socket can be specified like a seed file using RANDFILE 166659191Skris or -rand. 166759191Skris [Ulf M�ller] 166859191Skris 166959191Skris *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 167059191Skris Some CAs (e.g. Verisign) distribute certificates in this form. 167159191Skris [Steve Henson] 167259191Skris 167359191Skris *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 167459191Skris list to exclude them. This means that no special compilation option 167559191Skris is needed to use anonymous DH: it just needs to be included in the 167659191Skris cipher list. 167759191Skris [Steve Henson] 167859191Skris 167959191Skris *) Change the EVP_MD_CTX_type macro so its meaning consistent with 168059191Skris EVP_MD_type. The old functionality is available in a new macro called 168159191Skris EVP_MD_md(). Change code that uses it and update docs. 168259191Skris [Steve Henson] 168359191Skris 168459191Skris *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 168559191Skris where the 'void *' argument is replaced by a function pointer argument. 168659191Skris Previously 'void *' was abused to point to functions, which works on 168759191Skris many platforms, but is not correct. As these functions are usually 168859191Skris called by macros defined in OpenSSL header files, most source code 168959191Skris should work without changes. 169059191Skris [Richard Levitte] 169159191Skris 169259191Skris *) <openssl/opensslconf.h> (which is created by Configure) now contains 169359191Skris sections with information on -D... compiler switches used for 169459191Skris compiling the library so that applications can see them. To enable 169559191Skris one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 169659191Skris must be defined. E.g., 169759191Skris #define OPENSSL_ALGORITHM_DEFINES 169859191Skris #include <openssl/opensslconf.h> 169959191Skris defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 170059191Skris [Richard Levitte, Ulf and Bodo M�ller] 170159191Skris 170259191Skris *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 170359191Skris record layer. 170459191Skris [Bodo Moeller] 170559191Skris 170659191Skris *) Change the 'other' type in certificate aux info to a STACK_OF 170759191Skris X509_ALGOR. Although not an AlgorithmIdentifier as such it has 170859191Skris the required ASN1 format: arbitrary types determined by an OID. 170959191Skris [Steve Henson] 171059191Skris 171159191Skris *) Add some PEM_write_X509_REQ_NEW() functions and a command line 171259191Skris argument to 'req'. This is not because the function is newer or 171359191Skris better than others it just uses the work 'NEW' in the certificate 171459191Skris request header lines. Some software needs this. 171559191Skris [Steve Henson] 171659191Skris 171759191Skris *) Reorganise password command line arguments: now passwords can be 171859191Skris obtained from various sources. Delete the PEM_cb function and make 171959191Skris it the default behaviour: i.e. if the callback is NULL and the 172059191Skris usrdata argument is not NULL interpret it as a null terminated pass 172159191Skris phrase. If usrdata and the callback are NULL then the pass phrase 172259191Skris is prompted for as usual. 172359191Skris [Steve Henson] 172459191Skris 172559191Skris *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 172659191Skris the support is automatically enabled. The resulting binaries will 172759191Skris autodetect the card and use it if present. 172859191Skris [Ben Laurie and Compaq Inc.] 172959191Skris 173059191Skris *) Work around for Netscape hang bug. This sends certificate request 173159191Skris and server done in one record. Since this is perfectly legal in the 173259191Skris SSL/TLS protocol it isn't a "bug" option and is on by default. See 173359191Skris the bugs/SSLv3 entry for more info. 173459191Skris [Steve Henson] 173559191Skris 173659191Skris *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 173759191Skris [Andy Polyakov] 173859191Skris 173959191Skris *) Add -rand argument to smime and pkcs12 applications and read/write 174059191Skris of seed file. 174159191Skris [Steve Henson] 174259191Skris 174359191Skris *) New 'passwd' tool for crypt(3) and apr1 password hashes. 174459191Skris [Bodo Moeller] 174559191Skris 174659191Skris *) Add command line password options to the remaining applications. 174759191Skris [Steve Henson] 174859191Skris 174959191Skris *) Bug fix for BN_div_recp() for numerators with an even number of 175059191Skris bits. 175159191Skris [Ulf M�ller] 175259191Skris 175359191Skris *) More tests in bntest.c, and changed test_bn output. 175459191Skris [Ulf M�ller] 175559191Skris 175659191Skris *) ./config recognizes MacOS X now. 175759191Skris [Andy Polyakov] 175859191Skris 175959191Skris *) Bug fix for BN_div() when the first words of num and divsor are 176059191Skris equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 176159191Skris [Ulf M�ller] 176259191Skris 176359191Skris *) Add support for various broken PKCS#8 formats, and command line 176459191Skris options to produce them. 176559191Skris [Steve Henson] 176659191Skris 176759191Skris *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 176859191Skris get temporary BIGNUMs from a BN_CTX. 176959191Skris [Ulf M�ller] 177059191Skris 177159191Skris *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 177259191Skris for p == 0. 177359191Skris [Ulf M�ller] 177459191Skris 177559191Skris *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 177659191Skris include a #define from the old name to the new. The original intent 177759191Skris was that statically linked binaries could for example just call 177859191Skris SSLeay_add_all_ciphers() to just add ciphers to the table and not 177959191Skris link with digests. This never worked becayse SSLeay_add_all_digests() 178059191Skris and SSLeay_add_all_ciphers() were in the same source file so calling 178159191Skris one would link with the other. They are now in separate source files. 178259191Skris [Steve Henson] 178359191Skris 178459191Skris *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 178559191Skris [Steve Henson] 178659191Skris 178759191Skris *) Use a less unusual form of the Miller-Rabin primality test (it used 178859191Skris a binary algorithm for exponentiation integrated into the Miller-Rabin 178959191Skris loop, our standard modexp algorithms are faster). 179059191Skris [Bodo Moeller] 179159191Skris 179259191Skris *) Support for the EBCDIC character set completed. 179359191Skris [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 179459191Skris 179559191Skris *) Source code cleanups: use const where appropriate, eliminate casts, 179659191Skris use void * instead of char * in lhash. 179759191Skris [Ulf M�ller] 179859191Skris 179959191Skris *) Bugfix: ssl3_send_server_key_exchange was not restartable 180059191Skris (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 180159191Skris this the server could overwrite ephemeral keys that the client 180259191Skris has already seen). 180359191Skris [Bodo Moeller] 180459191Skris 180559191Skris *) Turn DSA_is_prime into a macro that calls BN_is_prime, 180659191Skris using 50 iterations of the Rabin-Miller test. 180759191Skris 180859191Skris DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 180959191Skris iterations of the Rabin-Miller test as required by the appendix 181059191Skris to FIPS PUB 186[-1]) instead of DSA_is_prime. 181159191Skris As BN_is_prime_fasttest includes trial division, DSA parameter 181259191Skris generation becomes much faster. 181359191Skris 181459191Skris This implies a change for the callback functions in DSA_is_prime 181559191Skris and DSA_generate_parameters: The callback function is called once 181659191Skris for each positive witness in the Rabin-Miller test, not just 181759191Skris occasionally in the inner loop; and the parameters to the 181859191Skris callback function now provide an iteration count for the outer 181959191Skris loop rather than for the current invocation of the inner loop. 182059191Skris DSA_generate_parameters additionally can call the callback 182159191Skris function with an 'iteration count' of -1, meaning that a 182259191Skris candidate has passed the trial division test (when q is generated 182359191Skris from an application-provided seed, trial division is skipped). 182459191Skris [Bodo Moeller] 182559191Skris 182659191Skris *) New function BN_is_prime_fasttest that optionally does trial 182759191Skris division before starting the Rabin-Miller test and has 182859191Skris an additional BN_CTX * argument (whereas BN_is_prime always 182959191Skris has to allocate at least one BN_CTX). 183059191Skris 'callback(1, -1, cb_arg)' is called when a number has passed the 183159191Skris trial division stage. 183259191Skris [Bodo Moeller] 183359191Skris 183459191Skris *) Fix for bug in CRL encoding. The validity dates weren't being handled 183559191Skris as ASN1_TIME. 183659191Skris [Steve Henson] 183759191Skris 183859191Skris *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 183959191Skris [Steve Henson] 184059191Skris 184159191Skris *) New function BN_pseudo_rand(). 184259191Skris [Ulf M�ller] 184359191Skris 184459191Skris *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 184559191Skris bignum version of BN_from_montgomery() with the working code from 184659191Skris SSLeay 0.9.0 (the word based version is faster anyway), and clean up 184759191Skris the comments. 184859191Skris [Ulf M�ller] 184959191Skris 185059191Skris *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 185159191Skris made it impossible to use the same SSL_SESSION data structure in 185259191Skris SSL2 clients in multiple threads. 185359191Skris [Bodo Moeller] 185459191Skris 185559191Skris *) The return value of RAND_load_file() no longer counts bytes obtained 185659191Skris by stat(). RAND_load_file(..., -1) is new and uses the complete file 185759191Skris to seed the PRNG (previously an explicit byte count was required). 185859191Skris [Ulf M�ller, Bodo M�ller] 185959191Skris 186059191Skris *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 186159191Skris used (char *) instead of (void *) and had casts all over the place. 186259191Skris [Steve Henson] 186359191Skris 186459191Skris *) Make BN_generate_prime() return NULL on error if ret!=NULL. 186559191Skris [Ulf M�ller] 186659191Skris 186759191Skris *) Retain source code compatibility for BN_prime_checks macro: 186859191Skris BN_is_prime(..., BN_prime_checks, ...) now uses 186959191Skris BN_prime_checks_for_size to determine the appropriate number of 187059191Skris Rabin-Miller iterations. 187159191Skris [Ulf M�ller] 187259191Skris 187359191Skris *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 187459191Skris DH_CHECK_P_NOT_SAFE_PRIME. 187559191Skris (Check if this is true? OpenPGP calls them "strong".) 187659191Skris [Ulf M�ller] 187759191Skris 187859191Skris *) Merge the functionality of "dh" and "gendh" programs into a new program 187959191Skris "dhparam". The old programs are retained for now but will handle DH keys 188059191Skris (instead of parameters) in future. 188159191Skris [Steve Henson] 188259191Skris 188359191Skris *) Make the ciphers, s_server and s_client programs check the return values 188459191Skris when a new cipher list is set. 188559191Skris [Steve Henson] 188659191Skris 188759191Skris *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 188859191Skris ciphers. Before when the 56bit ciphers were enabled the sorting was 188959191Skris wrong. 189059191Skris 189159191Skris The syntax for the cipher sorting has been extended to support sorting by 189259191Skris cipher-strength (using the strength_bits hard coded in the tables). 189359191Skris The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 189459191Skris 189559191Skris Fix a bug in the cipher-command parser: when supplying a cipher command 189659191Skris string with an "undefined" symbol (neither command nor alphanumeric 189759191Skris [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 189859191Skris an error is flagged. 189959191Skris 190059191Skris Due to the strength-sorting extension, the code of the 190159191Skris ssl_create_cipher_list() function was completely rearranged. I hope that 190259191Skris the readability was also increased :-) 190359191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 190459191Skris 190559191Skris *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 190659191Skris for the first serial number and places 2 in the serial number file. This 190759191Skris avoids problems when the root CA is created with serial number zero and 190859191Skris the first user certificate has the same issuer name and serial number 190959191Skris as the root CA. 191059191Skris [Steve Henson] 191159191Skris 191259191Skris *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 191359191Skris the new code. Add documentation for this stuff. 191459191Skris [Steve Henson] 191559191Skris 191659191Skris *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 191759191Skris X509_*() to X509at_*() on the grounds that they don't handle X509 191859191Skris structures and behave in an analagous way to the X509v3 functions: 191959191Skris they shouldn't be called directly but wrapper functions should be used 192059191Skris instead. 192159191Skris 192259191Skris So we also now have some wrapper functions that call the X509at functions 192359191Skris when passed certificate requests. (TO DO: similar things can be done with 192459191Skris PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 192559191Skris things. Some of these need some d2i or i2d and print functionality 192659191Skris because they handle more complex structures.) 192759191Skris [Steve Henson] 192859191Skris 192959191Skris *) Add missing #ifndefs that caused missing symbols when building libssl 193059191Skris as a shared library without RSA. Use #ifndef NO_SSL2 instead of 193159191Skris NO_RSA in ssl/s2*.c. 193259191Skris [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 193359191Skris 193459191Skris *) Precautions against using the PRNG uninitialized: RAND_bytes() now 193559191Skris has a return value which indicates the quality of the random data 193659191Skris (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 193759191Skris error queue. New function RAND_pseudo_bytes() generates output that is 193859191Skris guaranteed to be unique but not unpredictable. RAND_add is like 193959191Skris RAND_seed, but takes an extra argument for an entropy estimate 194059191Skris (RAND_seed always assumes full entropy). 194159191Skris [Ulf M�ller] 194259191Skris 194359191Skris *) Do more iterations of Rabin-Miller probable prime test (specifically, 194459191Skris 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 194559191Skris instead of only 2 for all lengths; see BN_prime_checks_for_size definition 194659191Skris in crypto/bn/bn_prime.c for the complete table). This guarantees a 194759191Skris false-positive rate of at most 2^-80 for random input. 194859191Skris [Bodo Moeller] 194959191Skris 195059191Skris *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 195159191Skris [Bodo Moeller] 195259191Skris 195359191Skris *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 195459191Skris in the 0.9.5 release), this returns the chain 195559191Skris from an X509_CTX structure with a dup of the stack and all 195659191Skris the X509 reference counts upped: so the stack will exist 195759191Skris after X509_CTX_cleanup() has been called. Modify pkcs12.c 195859191Skris to use this. 195959191Skris 196059191Skris Also make SSL_SESSION_print() print out the verify return 196159191Skris code. 196259191Skris [Steve Henson] 196359191Skris 196459191Skris *) Add manpage for the pkcs12 command. Also change the default 196559191Skris behaviour so MAC iteration counts are used unless the new 196659191Skris -nomaciter option is used. This improves file security and 196759191Skris only older versions of MSIE (4.0 for example) need it. 196859191Skris [Steve Henson] 196959191Skris 197059191Skris *) Honor the no-xxx Configure options when creating .DEF files. 197159191Skris [Ulf M�ller] 197259191Skris 197359191Skris *) Add PKCS#10 attributes to field table: challengePassword, 197459191Skris unstructuredName and unstructuredAddress. These are taken from 197559191Skris draft PKCS#9 v2.0 but are compatible with v1.2 provided no 197659191Skris international characters are used. 197759191Skris 197859191Skris More changes to X509_ATTRIBUTE code: allow the setting of types 197959191Skris based on strings. Remove the 'loc' parameter when adding 198059191Skris attributes because these will be a SET OF encoding which is sorted 198159191Skris in ASN1 order. 198259191Skris [Steve Henson] 198359191Skris 198459191Skris *) Initial changes to the 'req' utility to allow request generation 198559191Skris automation. This will allow an application to just generate a template 198659191Skris file containing all the field values and have req construct the 198759191Skris request. 198859191Skris 198959191Skris Initial support for X509_ATTRIBUTE handling. Stacks of these are 199059191Skris used all over the place including certificate requests and PKCS#7 199159191Skris structures. They are currently handled manually where necessary with 199259191Skris some primitive wrappers for PKCS#7. The new functions behave in a 199359191Skris manner analogous to the X509 extension functions: they allow 199459191Skris attributes to be looked up by NID and added. 199559191Skris 199659191Skris Later something similar to the X509V3 code would be desirable to 199759191Skris automatically handle the encoding, decoding and printing of the 199859191Skris more complex types. The string types like challengePassword can 199959191Skris be handled by the string table functions. 200059191Skris 200159191Skris Also modified the multi byte string table handling. Now there is 200259191Skris a 'global mask' which masks out certain types. The table itself 200359191Skris can use the flag STABLE_NO_MASK to ignore the mask setting: this 200459191Skris is useful when for example there is only one permissible type 200559191Skris (as in countryName) and using the mask might result in no valid 200659191Skris types at all. 200759191Skris [Steve Henson] 200859191Skris 200959191Skris *) Clean up 'Finished' handling, and add functions SSL_get_finished and 201059191Skris SSL_get_peer_finished to allow applications to obtain the latest 201159191Skris Finished messages sent to the peer or expected from the peer, 201259191Skris respectively. (SSL_get_peer_finished is usually the Finished message 201359191Skris actually received from the peer, otherwise the protocol will be aborted.) 201459191Skris 201559191Skris As the Finished message are message digests of the complete handshake 201659191Skris (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 201759191Skris be used for external authentication procedures when the authentication 201859191Skris provided by SSL/TLS is not desired or is not enough. 201959191Skris [Bodo Moeller] 202059191Skris 202159191Skris *) Enhanced support for Alpha Linux is added. Now ./config checks if 202259191Skris the host supports BWX extension and if Compaq C is present on the 202359191Skris $PATH. Just exploiting of the BWX extension results in 20-30% 202459191Skris performance kick for some algorithms, e.g. DES and RC4 to mention 202559191Skris a couple. Compaq C in turn generates ~20% faster code for MD5 and 202659191Skris SHA1. 202759191Skris [Andy Polyakov] 202859191Skris 202959191Skris *) Add support for MS "fast SGC". This is arguably a violation of the 203059191Skris SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 203159191Skris weak crypto and after checking the certificate is SGC a second one 203259191Skris with strong crypto. MS SGC stops the first handshake after receiving 203359191Skris the server certificate message and sends a second client hello. Since 203459191Skris a server will typically do all the time consuming operations before 203559191Skris expecting any further messages from the client (server key exchange 203659191Skris is the most expensive) there is little difference between the two. 203759191Skris 203859191Skris To get OpenSSL to support MS SGC we have to permit a second client 203959191Skris hello message after we have sent server done. In addition we have to 204059191Skris reset the MAC if we do get this second client hello. 204159191Skris [Steve Henson] 204259191Skris 204359191Skris *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 204459191Skris if a DER encoded private key is RSA or DSA traditional format. Changed 204559191Skris d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 204659191Skris format DER encoded private key. Newer code should use PKCS#8 format which 204759191Skris has the key type encoded in the ASN1 structure. Added DER private key 204859191Skris support to pkcs8 application. 204959191Skris [Steve Henson] 205059191Skris 205159191Skris *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 205259191Skris ciphersuites has been selected (as required by the SSL 3/TLS 1 205359191Skris specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 205459191Skris is set, we interpret this as a request to violate the specification 205559191Skris (the worst that can happen is a handshake failure, and 'correct' 205659191Skris behaviour would result in a handshake failure anyway). 205759191Skris [Bodo Moeller] 205859191Skris 205959191Skris *) In SSL_CTX_add_session, take into account that there might be multiple 206059191Skris SSL_SESSION structures with the same session ID (e.g. when two threads 206159191Skris concurrently obtain them from an external cache). 206259191Skris The internal cache can handle only one SSL_SESSION with a given ID, 206359191Skris so if there's a conflict, we now throw out the old one to achieve 206459191Skris consistency. 206559191Skris [Bodo Moeller] 206659191Skris 206759191Skris *) Add OIDs for idea and blowfish in CBC mode. This will allow both 206859191Skris to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 206959191Skris some routines that use cipher OIDs: some ciphers do not have OIDs 207059191Skris defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 207159191Skris example. 207259191Skris [Steve Henson] 207359191Skris 207459191Skris *) Simplify the trust setting structure and code. Now we just have 207559191Skris two sequences of OIDs for trusted and rejected settings. These will 207659191Skris typically have values the same as the extended key usage extension 207759191Skris and any application specific purposes. 207859191Skris 207959191Skris The trust checking code now has a default behaviour: it will just 208059191Skris check for an object with the same NID as the passed id. Functions can 208159191Skris be provided to override either the default behaviour or the behaviour 208259191Skris for a given id. SSL client, server and email already have functions 208359191Skris in place for compatibility: they check the NID and also return "trusted" 208459191Skris if the certificate is self signed. 208559191Skris [Steve Henson] 208659191Skris 208759191Skris *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 208859191Skris traditional format into an EVP_PKEY structure. 208959191Skris [Steve Henson] 209059191Skris 209159191Skris *) Add a password callback function PEM_cb() which either prompts for 209259191Skris a password if usr_data is NULL or otherwise assumes it is a null 209359191Skris terminated password. Allow passwords to be passed on command line 209459191Skris environment or config files in a few more utilities. 209559191Skris [Steve Henson] 209659191Skris 209759191Skris *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 209859191Skris keys. Add some short names for PKCS#8 PBE algorithms and allow them 209959191Skris to be specified on the command line for the pkcs8 and pkcs12 utilities. 210059191Skris Update documentation. 210159191Skris [Steve Henson] 210259191Skris 210359191Skris *) Support for ASN1 "NULL" type. This could be handled before by using 210459191Skris ASN1_TYPE but there wasn't any function that would try to read a NULL 210559191Skris and produce an error if it couldn't. For compatibility we also have 210659191Skris ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 210759191Skris don't allocate anything because they don't need to. 210859191Skris [Steve Henson] 210959191Skris 211059191Skris *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 211159191Skris for details. 211259191Skris [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 211359191Skris 211459191Skris *) Rebuild of the memory allocation routines used by OpenSSL code and 211559191Skris possibly others as well. The purpose is to make an interface that 211659191Skris provide hooks so anyone can build a separate set of allocation and 211759191Skris deallocation routines to be used by OpenSSL, for example memory 211859191Skris pool implementations, or something else, which was previously hard 211959191Skris since Malloc(), Realloc() and Free() were defined as macros having 212059191Skris the values malloc, realloc and free, respectively (except for Win32 212159191Skris compilations). The same is provided for memory debugging code. 212259191Skris OpenSSL already comes with functionality to find memory leaks, but 212359191Skris this gives people a chance to debug other memory problems. 212459191Skris 212559191Skris With these changes, a new set of functions and macros have appeared: 212659191Skris 212768651Skris CRYPTO_set_mem_debug_functions() [F] 212868651Skris CRYPTO_get_mem_debug_functions() [F] 212968651Skris CRYPTO_dbg_set_options() [F] 213068651Skris CRYPTO_dbg_get_options() [F] 213168651Skris CRYPTO_malloc_debug_init() [M] 213259191Skris 213359191Skris The memory debug functions are NULL by default, unless the library 213459191Skris is compiled with CRYPTO_MDEBUG or friends is defined. If someone 213559191Skris wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 213659191Skris gives the standard debugging functions that come with OpenSSL) or 213759191Skris CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 213859191Skris provided by the library user) must be used. When the standard 213959191Skris debugging functions are used, CRYPTO_dbg_set_options can be used to 214059191Skris request additional information: 214159191Skris CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 214259191Skris the CRYPTO_MDEBUG_xxx macro when compiling the library. 214359191Skris 214459191Skris Also, things like CRYPTO_set_mem_functions will always give the 214559191Skris expected result (the new set of functions is used for allocation 214659191Skris and deallocation) at all times, regardless of platform and compiler 214759191Skris options. 214859191Skris 214959191Skris To finish it up, some functions that were never use in any other 215059191Skris way than through macros have a new API and new semantic: 215159191Skris 215259191Skris CRYPTO_dbg_malloc() 215359191Skris CRYPTO_dbg_realloc() 215459191Skris CRYPTO_dbg_free() 215559191Skris 215659191Skris All macros of value have retained their old syntax. 215759191Skris [Richard Levitte and Bodo Moeller] 215859191Skris 215959191Skris *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 216059191Skris ordering of SMIMECapabilities wasn't in "strength order" and there 216159191Skris was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 216259191Skris algorithm. 216359191Skris [Steve Henson] 216459191Skris 216559191Skris *) Some ASN1 types with illegal zero length encoding (INTEGER, 216659191Skris ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 216759191Skris [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 216859191Skris 216959191Skris *) Merge in my S/MIME library for OpenSSL. This provides a simple 217059191Skris S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 217159191Skris functionality to handle multipart/signed properly) and a utility 217259191Skris called 'smime' to call all this stuff. This is based on code I 217359191Skris originally wrote for Celo who have kindly allowed it to be 217459191Skris included in OpenSSL. 217559191Skris [Steve Henson] 217659191Skris 217759191Skris *) Add variants des_set_key_checked and des_set_key_unchecked of 217859191Skris des_set_key (aka des_key_sched). Global variable des_check_key 217959191Skris decides which of these is called by des_set_key; this way 218059191Skris des_check_key behaves as it always did, but applications and 218159191Skris the library itself, which was buggy for des_check_key == 1, 218259191Skris have a cleaner way to pick the version they need. 218359191Skris [Bodo Moeller] 218459191Skris 218559191Skris *) New function PKCS12_newpass() which changes the password of a 218659191Skris PKCS12 structure. 218759191Skris [Steve Henson] 218859191Skris 218959191Skris *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 219059191Skris dynamic mix. In both cases the ids can be used as an index into the 219159191Skris table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 219259191Skris functions so they accept a list of the field values and the 219359191Skris application doesn't need to directly manipulate the X509_TRUST 219459191Skris structure. 219559191Skris [Steve Henson] 219659191Skris 219759191Skris *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 219859191Skris need initialising. 219959191Skris [Steve Henson] 220059191Skris 220159191Skris *) Modify the way the V3 extension code looks up extensions. This now 220259191Skris works in a similar way to the object code: we have some "standard" 220359191Skris extensions in a static table which is searched with OBJ_bsearch() 220459191Skris and the application can add dynamic ones if needed. The file 220559191Skris crypto/x509v3/ext_dat.h now has the info: this file needs to be 220659191Skris updated whenever a new extension is added to the core code and kept 220759191Skris in ext_nid order. There is a simple program 'tabtest.c' which checks 220859191Skris this. New extensions are not added too often so this file can readily 220959191Skris be maintained manually. 221059191Skris 221159191Skris There are two big advantages in doing things this way. The extensions 221259191Skris can be looked up immediately and no longer need to be "added" using 221359191Skris X509V3_add_standard_extensions(): this function now does nothing. 221459191Skris [Side note: I get *lots* of email saying the extension code doesn't 221559191Skris work because people forget to call this function] 221659191Skris Also no dynamic allocation is done unless new extensions are added: 221759191Skris so if we don't add custom extensions there is no need to call 221859191Skris X509V3_EXT_cleanup(). 221959191Skris [Steve Henson] 222059191Skris 222159191Skris *) Modify enc utility's salting as follows: make salting the default. Add a 222259191Skris magic header, so unsalted files fail gracefully instead of just decrypting 222359191Skris to garbage. This is because not salting is a big security hole, so people 222459191Skris should be discouraged from doing it. 222559191Skris [Ben Laurie] 222659191Skris 222759191Skris *) Fixes and enhancements to the 'x509' utility. It allowed a message 222859191Skris digest to be passed on the command line but it only used this 222959191Skris parameter when signing a certificate. Modified so all relevant 223059191Skris operations are affected by the digest parameter including the 223159191Skris -fingerprint and -x509toreq options. Also -x509toreq choked if a 223259191Skris DSA key was used because it didn't fix the digest. 223359191Skris [Steve Henson] 223459191Skris 223559191Skris *) Initial certificate chain verify code. Currently tests the untrusted 223659191Skris certificates for consistency with the verify purpose (which is set 223759191Skris when the X509_STORE_CTX structure is set up) and checks the pathlength. 223859191Skris 223959191Skris There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 224059191Skris this is because it will reject chains with invalid extensions whereas 224159191Skris every previous version of OpenSSL and SSLeay made no checks at all. 224259191Skris 224359191Skris Trust code: checks the root CA for the relevant trust settings. Trust 224459191Skris settings have an initial value consistent with the verify purpose: e.g. 224559191Skris if the verify purpose is for SSL client use it expects the CA to be 224659191Skris trusted for SSL client use. However the default value can be changed to 224759191Skris permit custom trust settings: one example of this would be to only trust 224859191Skris certificates from a specific "secure" set of CAs. 224959191Skris 225059191Skris Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 225159191Skris which should be used for version portability: especially since the 225259191Skris verify structure is likely to change more often now. 225359191Skris 225459191Skris SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 225559191Skris to set them. If not set then assume SSL clients will verify SSL servers 225659191Skris and vice versa. 225759191Skris 225859191Skris Two new options to the verify program: -untrusted allows a set of 225959191Skris untrusted certificates to be passed in and -purpose which sets the 226059191Skris intended purpose of the certificate. If a purpose is set then the 226159191Skris new chain verify code is used to check extension consistency. 226259191Skris [Steve Henson] 226359191Skris 226459191Skris *) Support for the authority information access extension. 226559191Skris [Steve Henson] 226659191Skris 226759191Skris *) Modify RSA and DSA PEM read routines to transparently handle 226859191Skris PKCS#8 format private keys. New *_PUBKEY_* functions that handle 226959191Skris public keys in a format compatible with certificate 227059191Skris SubjectPublicKeyInfo structures. Unfortunately there were already 227159191Skris functions called *_PublicKey_* which used various odd formats so 227259191Skris these are retained for compatibility: however the DSA variants were 227359191Skris never in a public release so they have been deleted. Changed dsa/rsa 227459191Skris utilities to handle the new format: note no releases ever handled public 227559191Skris keys so we should be OK. 227659191Skris 227759191Skris The primary motivation for this change is to avoid the same fiasco 227859191Skris that dogs private keys: there are several incompatible private key 227959191Skris formats some of which are standard and some OpenSSL specific and 228059191Skris require various evil hacks to allow partial transparent handling and 228159191Skris even then it doesn't work with DER formats. Given the option anything 228259191Skris other than PKCS#8 should be dumped: but the other formats have to 228359191Skris stay in the name of compatibility. 228459191Skris 228559191Skris With public keys and the benefit of hindsight one standard format 228659191Skris is used which works with EVP_PKEY, RSA or DSA structures: though 228759191Skris it clearly returns an error if you try to read the wrong kind of key. 228859191Skris 228959191Skris Added a -pubkey option to the 'x509' utility to output the public key. 229059191Skris Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 229159191Skris (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 229259191Skris EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 229359191Skris that do the same as the EVP_PKEY_assign_*() except they up the 229459191Skris reference count of the added key (they don't "swallow" the 229559191Skris supplied key). 229659191Skris [Steve Henson] 229759191Skris 229859191Skris *) Fixes to crypto/x509/by_file.c the code to read in certificates and 229959191Skris CRLs would fail if the file contained no certificates or no CRLs: 230059191Skris added a new function to read in both types and return the number 230159191Skris read: this means that if none are read it will be an error. The 230259191Skris DER versions of the certificate and CRL reader would always fail 230359191Skris because it isn't possible to mix certificates and CRLs in DER format 230459191Skris without choking one or the other routine. Changed this to just read 230559191Skris a certificate: this is the best we can do. Also modified the code 230659191Skris in apps/verify.c to take notice of return codes: it was previously 230759191Skris attempting to read in certificates from NULL pointers and ignoring 230859191Skris any errors: this is one reason why the cert and CRL reader seemed 230959191Skris to work. It doesn't check return codes from the default certificate 231059191Skris routines: these may well fail if the certificates aren't installed. 231159191Skris [Steve Henson] 231259191Skris 231359191Skris *) Code to support otherName option in GeneralName. 231459191Skris [Steve Henson] 231559191Skris 231659191Skris *) First update to verify code. Change the verify utility 231759191Skris so it warns if it is passed a self signed certificate: 231859191Skris for consistency with the normal behaviour. X509_verify 231959191Skris has been modified to it will now verify a self signed 232059191Skris certificate if *exactly* the same certificate appears 232159191Skris in the store: it was previously impossible to trust a 232259191Skris single self signed certificate. This means that: 232359191Skris openssl verify ss.pem 232459191Skris now gives a warning about a self signed certificate but 232559191Skris openssl verify -CAfile ss.pem ss.pem 232659191Skris is OK. 232759191Skris [Steve Henson] 232859191Skris 232959191Skris *) For servers, store verify_result in SSL_SESSION data structure 233059191Skris (and add it to external session representation). 233159191Skris This is needed when client certificate verifications fails, 233259191Skris but an application-provided verification callback (set by 233359191Skris SSL_CTX_set_cert_verify_callback) allows accepting the session 233459191Skris anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 233559191Skris but returns 1): When the session is reused, we have to set 233659191Skris ssl->verify_result to the appropriate error code to avoid 233759191Skris security holes. 233859191Skris [Bodo Moeller, problem pointed out by Lutz Jaenicke] 233959191Skris 234059191Skris *) Fix a bug in the new PKCS#7 code: it didn't consider the 234159191Skris case in PKCS7_dataInit() where the signed PKCS7 structure 234259191Skris didn't contain any existing data because it was being created. 234359191Skris [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 234459191Skris 234559191Skris *) Add a salt to the key derivation routines in enc.c. This 234659191Skris forms the first 8 bytes of the encrypted file. Also add a 234759191Skris -S option to allow a salt to be input on the command line. 234859191Skris [Steve Henson] 234959191Skris 235059191Skris *) New function X509_cmp(). Oddly enough there wasn't a function 235159191Skris to compare two certificates. We do this by working out the SHA1 235259191Skris hash and comparing that. X509_cmp() will be needed by the trust 235359191Skris code. 235459191Skris [Steve Henson] 235559191Skris 235659191Skris *) SSL_get1_session() is like SSL_get_session(), but increments 235759191Skris the reference count in the SSL_SESSION returned. 235859191Skris [Geoff Thorpe <geoff@eu.c2.net>] 235959191Skris 236059191Skris *) Fix for 'req': it was adding a null to request attributes. 236159191Skris Also change the X509_LOOKUP and X509_INFO code to handle 236259191Skris certificate auxiliary information. 236359191Skris [Steve Henson] 236459191Skris 236559191Skris *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 236659191Skris the 'enc' command. 236759191Skris [Steve Henson] 236859191Skris 236959191Skris *) Add the possibility to add extra information to the memory leak 237059191Skris detecting output, to form tracebacks, showing from where each 237159191Skris allocation was originated: CRYPTO_push_info("constant string") adds 237259191Skris the string plus current file name and line number to a per-thread 237359191Skris stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 237459191Skris is like calling CYRPTO_pop_info() until the stack is empty. 237559191Skris Also updated memory leak detection code to be multi-thread-safe. 237659191Skris [Richard Levitte] 237759191Skris 237859191Skris *) Add options -text and -noout to pkcs7 utility and delete the 237959191Skris encryption options which never did anything. Update docs. 238059191Skris [Steve Henson] 238159191Skris 238259191Skris *) Add options to some of the utilities to allow the pass phrase 238359191Skris to be included on either the command line (not recommended on 238459191Skris OSes like Unix) or read from the environment. Update the 238559191Skris manpages and fix a few bugs. 238659191Skris [Steve Henson] 238759191Skris 238859191Skris *) Add a few manpages for some of the openssl commands. 238959191Skris [Steve Henson] 239059191Skris 239159191Skris *) Fix the -revoke option in ca. It was freeing up memory twice, 239259191Skris leaking and not finding already revoked certificates. 239359191Skris [Steve Henson] 239459191Skris 239559191Skris *) Extensive changes to support certificate auxiliary information. 239659191Skris This involves the use of X509_CERT_AUX structure and X509_AUX 239759191Skris functions. An X509_AUX function such as PEM_read_X509_AUX() 239859191Skris can still read in a certificate file in the usual way but it 239959191Skris will also read in any additional "auxiliary information". By 240059191Skris doing things this way a fair degree of compatibility can be 240159191Skris retained: existing certificates can have this information added 240259191Skris using the new 'x509' options. 240359191Skris 240459191Skris Current auxiliary information includes an "alias" and some trust 240559191Skris settings. The trust settings will ultimately be used in enhanced 240659191Skris certificate chain verification routines: currently a certificate 240759191Skris can only be trusted if it is self signed and then it is trusted 240859191Skris for all purposes. 240959191Skris [Steve Henson] 241059191Skris 241159191Skris *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 241259191Skris The problem was that one of the replacement routines had not been working 241359191Skris since SSLeay releases. For now the offending routine has been replaced 241459191Skris with non-optimised assembler. Even so, this now gives around 95% 241559191Skris performance improvement for 1024 bit RSA signs. 241659191Skris [Mark Cox] 241759191Skris 241859191Skris *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 241959191Skris handling. Most clients have the effective key size in bits equal to 242059191Skris the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 242159191Skris A few however don't do this and instead use the size of the decrypted key 242259191Skris to determine the RC2 key length and the AlgorithmIdentifier to determine 242359191Skris the effective key length. In this case the effective key length can still 242459191Skris be 40 bits but the key length can be 168 bits for example. This is fixed 242559191Skris by manually forcing an RC2 key into the EVP_PKEY structure because the 242659191Skris EVP code can't currently handle unusual RC2 key sizes: it always assumes 242759191Skris the key length and effective key length are equal. 242859191Skris [Steve Henson] 242959191Skris 243059191Skris *) Add a bunch of functions that should simplify the creation of 243159191Skris X509_NAME structures. Now you should be able to do: 243259191Skris X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 243359191Skris and have it automatically work out the correct field type and fill in 243459191Skris the structures. The more adventurous can try: 243559191Skris X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 243659191Skris and it will (hopefully) work out the correct multibyte encoding. 243759191Skris [Steve Henson] 243859191Skris 243959191Skris *) Change the 'req' utility to use the new field handling and multibyte 244059191Skris copy routines. Before the DN field creation was handled in an ad hoc 244159191Skris way in req, ca, and x509 which was rather broken and didn't support 244259191Skris BMPStrings or UTF8Strings. Since some software doesn't implement 244359191Skris BMPStrings or UTF8Strings yet, they can be enabled using the config file 244459191Skris using the dirstring_type option. See the new comment in the default 244559191Skris openssl.cnf for more info. 244659191Skris [Steve Henson] 244759191Skris 244859191Skris *) Make crypto/rand/md_rand.c more robust: 244959191Skris - Assure unique random numbers after fork(). 245059191Skris - Make sure that concurrent threads access the global counter and 245159191Skris md serializably so that we never lose entropy in them 245259191Skris or use exactly the same state in multiple threads. 245359191Skris Access to the large state is not always serializable because 245459191Skris the additional locking could be a performance killer, and 245559191Skris md should be large enough anyway. 245659191Skris [Bodo Moeller] 245759191Skris 245859191Skris *) New file apps/app_rand.c with commonly needed functionality 245959191Skris for handling the random seed file. 246059191Skris 246159191Skris Use the random seed file in some applications that previously did not: 246259191Skris ca, 246359191Skris dsaparam -genkey (which also ignored its '-rand' option), 246459191Skris s_client, 246559191Skris s_server, 246659191Skris x509 (when signing). 246759191Skris Except on systems with /dev/urandom, it is crucial to have a random 246859191Skris seed file at least for key creation, DSA signing, and for DH exchanges; 246959191Skris for RSA signatures we could do without one. 247059191Skris 247159191Skris gendh and gendsa (unlike genrsa) used to read only the first byte 247259191Skris of each file listed in the '-rand' option. The function as previously 247359191Skris found in genrsa is now in app_rand.c and is used by all programs 247459191Skris that support '-rand'. 247559191Skris [Bodo Moeller] 247659191Skris 247759191Skris *) In RAND_write_file, use mode 0600 for creating files; 247859191Skris don't just chmod when it may be too late. 247959191Skris [Bodo Moeller] 248059191Skris 248159191Skris *) Report an error from X509_STORE_load_locations 248259191Skris when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 248359191Skris [Bill Perry] 248459191Skris 248559191Skris *) New function ASN1_mbstring_copy() this copies a string in either 248659191Skris ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 248759191Skris into an ASN1_STRING type. A mask of permissible types is passed 248859191Skris and it chooses the "minimal" type to use or an error if not type 248959191Skris is suitable. 249059191Skris [Steve Henson] 249159191Skris 249259191Skris *) Add function equivalents to the various macros in asn1.h. The old 249359191Skris macros are retained with an M_ prefix. Code inside the library can 249459191Skris use the M_ macros. External code (including the openssl utility) 249559191Skris should *NOT* in order to be "shared library friendly". 249659191Skris [Steve Henson] 249759191Skris 249859191Skris *) Add various functions that can check a certificate's extensions 249959191Skris to see if it usable for various purposes such as SSL client, 250059191Skris server or S/MIME and CAs of these types. This is currently 250159191Skris VERY EXPERIMENTAL but will ultimately be used for certificate chain 250259191Skris verification. Also added a -purpose flag to x509 utility to 250359191Skris print out all the purposes. 250459191Skris [Steve Henson] 250559191Skris 250659191Skris *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 250759191Skris functions. 250859191Skris [Steve Henson] 250959191Skris 251059191Skris *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 251159191Skris for, obtain and decode and extension and obtain its critical flag. 251259191Skris This allows all the necessary extension code to be handled in a 251359191Skris single function call. 251459191Skris [Steve Henson] 251559191Skris 251659191Skris *) RC4 tune-up featuring 30-40% performance improvement on most RISC 251759191Skris platforms. See crypto/rc4/rc4_enc.c for further details. 251859191Skris [Andy Polyakov] 251959191Skris 252059191Skris *) New -noout option to asn1parse. This causes no output to be produced 252159191Skris its main use is when combined with -strparse and -out to extract data 252259191Skris from a file (which may not be in ASN.1 format). 252359191Skris [Steve Henson] 252459191Skris 252559191Skris *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 252659191Skris when producing the local key id. 252759191Skris [Richard Levitte <levitte@stacken.kth.se>] 252859191Skris 252959191Skris *) New option -dhparam in s_server. This allows a DH parameter file to be 253059191Skris stated explicitly. If it is not stated then it tries the first server 253159191Skris certificate file. The previous behaviour hard coded the filename 253259191Skris "server.pem". 253359191Skris [Steve Henson] 253459191Skris 253559191Skris *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 253659191Skris a public key to be input or output. For example: 253759191Skris openssl rsa -in key.pem -pubout -out pubkey.pem 253859191Skris Also added necessary DSA public key functions to handle this. 253959191Skris [Steve Henson] 254059191Skris 254159191Skris *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 254259191Skris in the message. This was handled by allowing 254359191Skris X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 254459191Skris [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 254559191Skris 254659191Skris *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 254759191Skris to the end of the strings whereas this didn't. This would cause problems 254859191Skris if strings read with d2i_ASN1_bytes() were later modified. 254959191Skris [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 255059191Skris 255159191Skris *) Fix for base64 decode bug. When a base64 bio reads only one line of 255259191Skris data and it contains EOF it will end up returning an error. This is 255359191Skris caused by input 46 bytes long. The cause is due to the way base64 255459191Skris BIOs find the start of base64 encoded data. They do this by trying a 255559191Skris trial decode on each line until they find one that works. When they 255659191Skris do a flag is set and it starts again knowing it can pass all the 255759191Skris data directly through the decoder. Unfortunately it doesn't reset 255859191Skris the context it uses. This means that if EOF is reached an attempt 255959191Skris is made to pass two EOFs through the context and this causes the 256059191Skris resulting error. This can also cause other problems as well. As is 256159191Skris usual with these problems it takes *ages* to find and the fix is 256259191Skris trivial: move one line. 256359191Skris [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 256459191Skris 256559191Skris *) Ugly workaround to get s_client and s_server working under Windows. The 256659191Skris old code wouldn't work because it needed to select() on sockets and the 256759191Skris tty (for keypresses and to see if data could be written). Win32 only 256859191Skris supports select() on sockets so we select() with a 1s timeout on the 256959191Skris sockets and then see if any characters are waiting to be read, if none 257059191Skris are present then we retry, we also assume we can always write data to 257159191Skris the tty. This isn't nice because the code then blocks until we've 257259191Skris received a complete line of data and it is effectively polling the 257359191Skris keyboard at 1s intervals: however it's quite a bit better than not 257459191Skris working at all :-) A dedicated Windows application might handle this 257559191Skris with an event loop for example. 257659191Skris [Steve Henson] 257759191Skris 257859191Skris *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 257959191Skris and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 258059191Skris will be called when RSA_sign() and RSA_verify() are used. This is useful 258159191Skris if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 258259191Skris For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 258359191Skris should *not* be used: RSA_sign() and RSA_verify() must be used instead. 258459191Skris This necessitated the support of an extra signature type NID_md5_sha1 258559191Skris for SSL signatures and modifications to the SSL library to use it instead 258659191Skris of calling RSA_public_decrypt() and RSA_private_encrypt(). 258759191Skris [Steve Henson] 258859191Skris 258959191Skris *) Add new -verify -CAfile and -CApath options to the crl program, these 259059191Skris will lookup a CRL issuers certificate and verify the signature in a 259159191Skris similar way to the verify program. Tidy up the crl program so it 259259191Skris no longer accesses structures directly. Make the ASN1 CRL parsing a bit 259359191Skris less strict. It will now permit CRL extensions even if it is not 259459191Skris a V2 CRL: this will allow it to tolerate some broken CRLs. 259559191Skris [Steve Henson] 259659191Skris 259759191Skris *) Initialize all non-automatic variables each time one of the openssl 259859191Skris sub-programs is started (this is necessary as they may be started 259959191Skris multiple times from the "OpenSSL>" prompt). 260059191Skris [Lennart Bang, Bodo Moeller] 260159191Skris 260259191Skris *) Preliminary compilation option RSA_NULL which disables RSA crypto without 260359191Skris removing all other RSA functionality (this is what NO_RSA does). This 260459191Skris is so (for example) those in the US can disable those operations covered 260559191Skris by the RSA patent while allowing storage and parsing of RSA keys and RSA 260659191Skris key generation. 260759191Skris [Steve Henson] 260859191Skris 260959191Skris *) Non-copying interface to BIO pairs. 261059191Skris (still largely untested) 261159191Skris [Bodo Moeller] 261259191Skris 261359191Skris *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 261459191Skris ASCII string. This was handled independently in various places before. 261559191Skris [Steve Henson] 261659191Skris 261759191Skris *) New functions UTF8_getc() and UTF8_putc() that parse and generate 261859191Skris UTF8 strings a character at a time. 261959191Skris [Steve Henson] 262059191Skris 262159191Skris *) Use client_version from client hello to select the protocol 262259191Skris (s23_srvr.c) and for RSA client key exchange verification 262359191Skris (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 262459191Skris [Bodo Moeller] 262559191Skris 262659191Skris *) Add various utility functions to handle SPKACs, these were previously 262759191Skris handled by poking round in the structure internals. Added new function 262859191Skris NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 262959191Skris print, verify and generate SPKACs. Based on an original idea from 263059191Skris Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 263159191Skris [Steve Henson] 263259191Skris 263359191Skris *) RIPEMD160 is operational on all platforms and is back in 'make test'. 263459191Skris [Andy Polyakov] 263559191Skris 263659191Skris *) Allow the config file extension section to be overwritten on the 263759191Skris command line. Based on an original idea from Massimiliano Pala 263859191Skris <madwolf@comune.modena.it>. The new option is called -extensions 263959191Skris and can be applied to ca, req and x509. Also -reqexts to override 264059191Skris the request extensions in req and -crlexts to override the crl extensions 264159191Skris in ca. 264259191Skris [Steve Henson] 264359191Skris 264459191Skris *) Add new feature to the SPKAC handling in ca. Now you can include 264559191Skris the same field multiple times by preceding it by "XXXX." for example: 264659191Skris 1.OU="Unit name 1" 264759191Skris 2.OU="Unit name 2" 264859191Skris this is the same syntax as used in the req config file. 264959191Skris [Steve Henson] 265059191Skris 265159191Skris *) Allow certificate extensions to be added to certificate requests. These 265259191Skris are specified in a 'req_extensions' option of the req section of the 265359191Skris config file. They can be printed out with the -text option to req but 265459191Skris are otherwise ignored at present. 265559191Skris [Steve Henson] 265659191Skris 265759191Skris *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 265859191Skris data read consists of only the final block it would not decrypted because 265959191Skris EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 266059191Skris A misplaced 'break' also meant the decrypted final block might not be 266159191Skris copied until the next read. 266259191Skris [Steve Henson] 266359191Skris 266459191Skris *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 266559191Skris a few extra parameters to the DH structure: these will be useful if 266659191Skris for example we want the value of 'q' or implement X9.42 DH. 266759191Skris [Steve Henson] 266859191Skris 266959191Skris *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 267059191Skris provides hooks that allow the default DSA functions or functions on a 267159191Skris "per key" basis to be replaced. This allows hardware acceleration and 267259191Skris hardware key storage to be handled without major modification to the 267359191Skris library. Also added low level modexp hooks and CRYPTO_EX structure and 267459191Skris associated functions. 267559191Skris [Steve Henson] 267659191Skris 267759191Skris *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 267859191Skris as "read only": it can't be written to and the buffer it points to will 267959191Skris not be freed. Reading from a read only BIO is much more efficient than 268059191Skris a normal memory BIO. This was added because there are several times when 268159191Skris an area of memory needs to be read from a BIO. The previous method was 268259191Skris to create a memory BIO and write the data to it, this results in two 268359191Skris copies of the data and an O(n^2) reading algorithm. There is a new 268459191Skris function BIO_new_mem_buf() which creates a read only memory BIO from 268559191Skris an area of memory. Also modified the PKCS#7 routines to use read only 268659191Skris memory BIOs. 268759191Skris [Steve Henson] 268859191Skris 268959191Skris *) Bugfix: ssl23_get_client_hello did not work properly when called in 269059191Skris state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 269159191Skris a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 269259191Skris but a retry condition occured while trying to read the rest. 269359191Skris [Bodo Moeller] 269459191Skris 269559191Skris *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 269659191Skris NID_pkcs7_encrypted by default: this was wrong since this should almost 269759191Skris always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 269859191Skris the encrypted data type: this is a more sensible place to put it and it 269959191Skris allows the PKCS#12 code to be tidied up that duplicated this 270059191Skris functionality. 270159191Skris [Steve Henson] 270259191Skris 270359191Skris *) Changed obj_dat.pl script so it takes its input and output files on 270459191Skris the command line. This should avoid shell escape redirection problems 270559191Skris under Win32. 270659191Skris [Steve Henson] 270759191Skris 270859191Skris *) Initial support for certificate extension requests, these are included 270959191Skris in things like Xenroll certificate requests. Included functions to allow 271059191Skris extensions to be obtained and added. 271159191Skris [Steve Henson] 271259191Skris 271359191Skris *) -crlf option to s_client and s_server for sending newlines as 271459191Skris CRLF (as required by many protocols). 271559191Skris [Bodo Moeller] 271659191Skris 271755714Skris Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 271855714Skris 271955714Skris *) Install libRSAglue.a when OpenSSL is built with RSAref. 272055714Skris [Ralf S. Engelschall] 272155714Skris 272255714Skris *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 272355714Skris [Andrija Antonijevic <TheAntony2@bigfoot.com>] 272455714Skris 272555714Skris *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 272655714Skris program. 272755714Skris [Steve Henson] 272855714Skris 272955714Skris *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 273055714Skris DH parameters/keys (q is lost during that conversion, but the resulting 273155714Skris DH parameters contain its length). 273255714Skris 273355714Skris For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 273455714Skris much faster than DH_generate_parameters (which creates parameters 273555714Skris where p = 2*q + 1), and also the smaller q makes DH computations 273655714Skris much more efficient (160-bit exponentiation instead of 1024-bit 273755714Skris exponentiation); so this provides a convenient way to support DHE 273855714Skris ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 273955714Skris utter importance to use 274055714Skris SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 274155714Skris or 274255714Skris SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 274355714Skris when such DH parameters are used, because otherwise small subgroup 274455714Skris attacks may become possible! 274555714Skris [Bodo Moeller] 274655714Skris 274755714Skris *) Avoid memory leak in i2d_DHparams. 274855714Skris [Bodo Moeller] 274955714Skris 275055714Skris *) Allow the -k option to be used more than once in the enc program: 275155714Skris this allows the same encrypted message to be read by multiple recipients. 275255714Skris [Steve Henson] 275355714Skris 275455714Skris *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 275555714Skris an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 275655714Skris it will always use the numerical form of the OID, even if it has a short 275755714Skris or long name. 275855714Skris [Steve Henson] 275955714Skris 276055714Skris *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 276155714Skris method only got called if p,q,dmp1,dmq1,iqmp components were present, 276255714Skris otherwise bn_mod_exp was called. In the case of hardware keys for example 276355714Skris no private key components need be present and it might store extra data 276459191Skris in the RSA structure, which cannot be accessed from bn_mod_exp. 276559191Skris By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 276659191Skris private key operations. 276755714Skris [Steve Henson] 276855714Skris 276955714Skris *) Added support for SPARC Linux. 277055714Skris [Andy Polyakov] 277155714Skris 277255714Skris *) pem_password_cb function type incompatibly changed from 277355714Skris typedef int pem_password_cb(char *buf, int size, int rwflag); 277455714Skris to 277555714Skris ....(char *buf, int size, int rwflag, void *userdata); 277655714Skris so that applications can pass data to their callbacks: 277755714Skris The PEM[_ASN1]_{read,write}... functions and macros now take an 277855714Skris additional void * argument, which is just handed through whenever 277955714Skris the password callback is called. 278059191Skris [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 278155714Skris 278255714Skris New function SSL_CTX_set_default_passwd_cb_userdata. 278355714Skris 278455714Skris Compatibility note: As many C implementations push function arguments 278555714Skris onto the stack in reverse order, the new library version is likely to 278655714Skris interoperate with programs that have been compiled with the old 278755714Skris pem_password_cb definition (PEM_whatever takes some data that 278855714Skris happens to be on the stack as its last argument, and the callback 278955714Skris just ignores this garbage); but there is no guarantee whatsoever that 279055714Skris this will work. 279155714Skris 279255714Skris *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 279355714Skris (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 279455714Skris problems not only on Windows, but also on some Unix platforms. 279555714Skris To avoid problematic command lines, these definitions are now in an 279655714Skris auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 279755714Skris for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 279855714Skris [Bodo Moeller] 279955714Skris 280055714Skris *) MIPS III/IV assembler module is reimplemented. 280155714Skris [Andy Polyakov] 280255714Skris 280355714Skris *) More DES library cleanups: remove references to srand/rand and 280455714Skris delete an unused file. 280555714Skris [Ulf M�ller] 280655714Skris 280755714Skris *) Add support for the the free Netwide assembler (NASM) under Win32, 280855714Skris since not many people have MASM (ml) and it can be hard to obtain. 280955714Skris This is currently experimental but it seems to work OK and pass all 281055714Skris the tests. Check out INSTALL.W32 for info. 281155714Skris [Steve Henson] 281255714Skris 281355714Skris *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 281455714Skris without temporary keys kept an extra copy of the server key, 281555714Skris and connections with temporary keys did not free everything in case 281655714Skris of an error. 281755714Skris [Bodo Moeller] 281855714Skris 281955714Skris *) New function RSA_check_key and new openssl rsa option -check 282055714Skris for verifying the consistency of RSA keys. 282155714Skris [Ulf Moeller, Bodo Moeller] 282255714Skris 282355714Skris *) Various changes to make Win32 compile work: 282455714Skris 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 282555714Skris 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 282655714Skris comparison" warnings. 282755714Skris 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 282855714Skris [Steve Henson] 282955714Skris 283055714Skris *) Add a debugging option to PKCS#5 v2 key generation function: when 283155714Skris you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 283255714Skris derived keys are printed to stderr. 283355714Skris [Steve Henson] 283455714Skris 283555714Skris *) Copy the flags in ASN1_STRING_dup(). 283655714Skris [Roman E. Pavlov <pre@mo.msk.ru>] 283755714Skris 283855714Skris *) The x509 application mishandled signing requests containing DSA 283955714Skris keys when the signing key was also DSA and the parameters didn't match. 284055714Skris 284155714Skris It was supposed to omit the parameters when they matched the signing key: 284255714Skris the verifying software was then supposed to automatically use the CA's 284355714Skris parameters if they were absent from the end user certificate. 284455714Skris 284555714Skris Omitting parameters is no longer recommended. The test was also 284655714Skris the wrong way round! This was probably due to unusual behaviour in 284755714Skris EVP_cmp_parameters() which returns 1 if the parameters match. 284855714Skris This meant that parameters were omitted when they *didn't* match and 284955714Skris the certificate was useless. Certificates signed with 'ca' didn't have 285055714Skris this bug. 285155714Skris [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 285255714Skris 285355714Skris *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 285455714Skris The interface is as follows: 285555714Skris Applications can use 285655714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 285755714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 285855714Skris "off" is now the default. 285955714Skris The library internally uses 286055714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 286155714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 286255714Skris to disable memory-checking temporarily. 286355714Skris 286455714Skris Some inconsistent states that previously were possible (and were 286555714Skris even the default) are now avoided. 286655714Skris 286755714Skris -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 286855714Skris with each memory chunk allocated; this is occasionally more helpful 286955714Skris than just having a counter. 287055714Skris 287155714Skris -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 287255714Skris 287355714Skris -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 287455714Skris extensions. 287555714Skris [Bodo Moeller] 287655714Skris 287755714Skris *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 287855714Skris which largely parallels "options", but is for changing API behaviour, 287955714Skris whereas "options" are about protocol behaviour. 288055714Skris Initial "mode" flags are: 288155714Skris 288255714Skris SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 288355714Skris a single record has been written. 288455714Skris SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 288555714Skris retries use the same buffer location. 288655714Skris (But all of the contents must be 288755714Skris copied!) 288855714Skris [Bodo Moeller] 288955714Skris 289079998Skris *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 289155714Skris worked. 289255714Skris 289355714Skris *) Fix problems with no-hmac etc. 289455714Skris [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 289555714Skris 289655714Skris *) New functions RSA_get_default_method(), RSA_set_method() and 289755714Skris RSA_get_method(). These allows replacement of RSA_METHODs without having 289855714Skris to mess around with the internals of an RSA structure. 289955714Skris [Steve Henson] 290055714Skris 290155714Skris *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 290255714Skris Also really enable memory leak checks in openssl.c and in some 290355714Skris test programs. 290455714Skris [Chad C. Mulligan, Bodo Moeller] 290555714Skris 290655714Skris *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 290755714Skris up the length of negative integers. This has now been simplified to just 290855714Skris store the length when it is first determined and use it later, rather 290955714Skris than trying to keep track of where data is copied and updating it to 291055714Skris point to the end. 291155714Skris [Steve Henson, reported by Brien Wheeler 291255714Skris <bwheeler@authentica-security.com>] 291355714Skris 291455714Skris *) Add a new function PKCS7_signatureVerify. This allows the verification 291555714Skris of a PKCS#7 signature but with the signing certificate passed to the 291655714Skris function itself. This contrasts with PKCS7_dataVerify which assumes the 291755714Skris certificate is present in the PKCS#7 structure. This isn't always the 291855714Skris case: certificates can be omitted from a PKCS#7 structure and be 291955714Skris distributed by "out of band" means (such as a certificate database). 292055714Skris [Steve Henson] 292155714Skris 292255714Skris *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 292355714Skris function prototypes in pem.h, also change util/mkdef.pl to add the 292455714Skris necessary function names. 292555714Skris [Steve Henson] 292655714Skris 292755714Skris *) mk1mf.pl (used by Windows builds) did not properly read the 292855714Skris options set by Configure in the top level Makefile, and Configure 292955714Skris was not even able to write more than one option correctly. 293055714Skris Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 293155714Skris [Bodo Moeller] 293255714Skris 293355714Skris *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 293455714Skris file to be loaded from a BIO or FILE pointer. The BIO version will 293555714Skris for example allow memory BIOs to contain config info. 293655714Skris [Steve Henson] 293755714Skris 293855714Skris *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 293955714Skris Whoever hopes to achieve shared-library compatibility across versions 294055714Skris must use this, not the compile-time macro. 294155714Skris (Exercise 0.9.4: Which is the minimum library version required by 294255714Skris such programs?) 294355714Skris Note: All this applies only to multi-threaded programs, others don't 294455714Skris need locks. 294555714Skris [Bodo Moeller] 294655714Skris 294755714Skris *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 294855714Skris through a BIO pair triggered the default case, i.e. 294955714Skris SSLerr(...,SSL_R_UNKNOWN_STATE). 295055714Skris [Bodo Moeller] 295155714Skris 295255714Skris *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 295355714Skris can use the SSL library even if none of the specific BIOs is 295455714Skris appropriate. 295555714Skris [Bodo Moeller] 295655714Skris 295755714Skris *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 295855714Skris for the encoded length. 295955714Skris [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 296055714Skris 296155714Skris *) Add initial documentation of the X509V3 functions. 296255714Skris [Steve Henson] 296355714Skris 296455714Skris *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 296555714Skris PEM_write_bio_PKCS8PrivateKey() that are equivalent to 296655714Skris PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 296755714Skris secure PKCS#8 private key format with a high iteration count. 296855714Skris [Steve Henson] 296955714Skris 297055714Skris *) Fix determination of Perl interpreter: A perl or perl5 297155714Skris _directory_ in $PATH was also accepted as the interpreter. 297255714Skris [Ralf S. Engelschall] 297355714Skris 297455714Skris *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 297555714Skris wrong with it but it was very old and did things like calling 297655714Skris PEM_ASN1_read() directly and used MD5 for the hash not to mention some 297755714Skris unusual formatting. 297855714Skris [Steve Henson] 297955714Skris 298055714Skris *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 298155714Skris to use the new extension code. 298255714Skris [Steve Henson] 298355714Skris 298455714Skris *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 298555714Skris with macros. This should make it easier to change their form, add extra 298655714Skris arguments etc. Fix a few PEM prototypes which didn't have cipher as a 298755714Skris constant. 298855714Skris [Steve Henson] 298955714Skris 299055714Skris *) Add to configuration table a new entry that can specify an alternative 299155714Skris name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 299255714Skris according to Mark Crispin <MRC@Panda.COM>. 299355714Skris [Bodo Moeller] 299455714Skris 299555714Skris#if 0 299655714Skris *) DES CBC did not update the IV. Weird. 299755714Skris [Ben Laurie] 299855714Skris#else 299955714Skris des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 300055714Skris Changing the behaviour of the former might break existing programs -- 300155714Skris where IV updating is needed, des_ncbc_encrypt can be used. 300255714Skris#endif 300355714Skris 300455714Skris *) When bntest is run from "make test" it drives bc to check its 300555714Skris calculations, as well as internally checking them. If an internal check 300655714Skris fails, it needs to cause bc to give a non-zero result or make test carries 300755714Skris on without noticing the failure. Fixed. 300855714Skris [Ben Laurie] 300955714Skris 301055714Skris *) DES library cleanups. 301155714Skris [Ulf M�ller] 301255714Skris 301355714Skris *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 301455714Skris used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 301555714Skris ciphers. NOTE: although the key derivation function has been verified 301655714Skris against some published test vectors it has not been extensively tested 301755714Skris yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 301855714Skris of v2.0. 301955714Skris [Steve Henson] 302055714Skris 302155714Skris *) Instead of "mkdir -p", which is not fully portable, use new 302255714Skris Perl script "util/mkdir-p.pl". 302355714Skris [Bodo Moeller] 302455714Skris 302555714Skris *) Rewrite the way password based encryption (PBE) is handled. It used to 302655714Skris assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 302755714Skris structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 302855714Skris but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 302955714Skris the 'parameter' field of the AlgorithmIdentifier is passed to the 303055714Skris underlying key generation function so it must do its own ASN1 parsing. 303155714Skris This has also changed the EVP_PBE_CipherInit() function which now has a 303255714Skris 'parameter' argument instead of literal salt and iteration count values 303355714Skris and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 303455714Skris [Steve Henson] 303555714Skris 303655714Skris *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 303755714Skris and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 303855714Skris Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 303955714Skris KEY" because this clashed with PKCS#8 unencrypted string. Since this 304055714Skris value was just used as a "magic string" and not used directly its 304155714Skris value doesn't matter. 304255714Skris [Steve Henson] 304355714Skris 304455714Skris *) Introduce some semblance of const correctness to BN. Shame C doesn't 304555714Skris support mutable. 304655714Skris [Ben Laurie] 304755714Skris 304855714Skris *) "linux-sparc64" configuration (ultrapenguin). 304955714Skris [Ray Miller <ray.miller@oucs.ox.ac.uk>] 305055714Skris "linux-sparc" configuration. 305155714Skris [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 305255714Skris 305355714Skris *) config now generates no-xxx options for missing ciphers. 305455714Skris [Ulf M�ller] 305555714Skris 305655714Skris *) Support the EBCDIC character set (work in progress). 305755714Skris File ebcdic.c not yet included because it has a different license. 305855714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 305955714Skris 306055714Skris *) Support BS2000/OSD-POSIX. 306155714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 306255714Skris 306355714Skris *) Make callbacks for key generation use void * instead of char *. 306455714Skris [Ben Laurie] 306555714Skris 306655714Skris *) Make S/MIME samples compile (not yet tested). 306755714Skris [Ben Laurie] 306855714Skris 306955714Skris *) Additional typesafe stacks. 307055714Skris [Ben Laurie] 307155714Skris 307255714Skris *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 307355714Skris [Bodo Moeller] 307455714Skris 307555714Skris 307655714Skris Changes between 0.9.3 and 0.9.3a [29 May 1999] 307755714Skris 307855714Skris *) New configuration variant "sco5-gcc". 307955714Skris 308055714Skris *) Updated some demos. 308155714Skris [Sean O Riordain, Wade Scholine] 308255714Skris 308355714Skris *) Add missing BIO_free at exit of pkcs12 application. 308455714Skris [Wu Zhigang] 308555714Skris 308655714Skris *) Fix memory leak in conf.c. 308755714Skris [Steve Henson] 308855714Skris 308955714Skris *) Updates for Win32 to assembler version of MD5. 309055714Skris [Steve Henson] 309155714Skris 309255714Skris *) Set #! path to perl in apps/der_chop to where we found it 309355714Skris instead of using a fixed path. 309455714Skris [Bodo Moeller] 309555714Skris 309655714Skris *) SHA library changes for irix64-mips4-cc. 309755714Skris [Andy Polyakov] 309855714Skris 309955714Skris *) Improvements for VMS support. 310055714Skris [Richard Levitte] 310155714Skris 310255714Skris 310355714Skris Changes between 0.9.2b and 0.9.3 [24 May 1999] 310455714Skris 310555714Skris *) Bignum library bug fix. IRIX 6 passes "make test" now! 310655714Skris This also avoids the problems with SC4.2 and unpatched SC5. 310755714Skris [Andy Polyakov <appro@fy.chalmers.se>] 310855714Skris 310955714Skris *) New functions sk_num, sk_value and sk_set to replace the previous macros. 311055714Skris These are required because of the typesafe stack would otherwise break 311155714Skris existing code. If old code used a structure member which used to be STACK 311255714Skris and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 311355714Skris sk_num or sk_value it would produce an error because the num, data members 311455714Skris are not present in STACK_OF. Now it just produces a warning. sk_set 311555714Skris replaces the old method of assigning a value to sk_value 311655714Skris (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 311755714Skris that does this will no longer work (and should use sk_set instead) but 311855714Skris this could be regarded as a "questionable" behaviour anyway. 311955714Skris [Steve Henson] 312055714Skris 312155714Skris *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 312255714Skris correctly handle encrypted S/MIME data. 312355714Skris [Steve Henson] 312455714Skris 312555714Skris *) Change type of various DES function arguments from des_cblock 312655714Skris (which means, in function argument declarations, pointer to char) 312755714Skris to des_cblock * (meaning pointer to array with 8 char elements), 312855714Skris which allows the compiler to do more typechecking; it was like 312955714Skris that back in SSLeay, but with lots of ugly casts. 313055714Skris 313155714Skris Introduce new type const_des_cblock. 313255714Skris [Bodo Moeller] 313355714Skris 313455714Skris *) Reorganise the PKCS#7 library and get rid of some of the more obvious 313555714Skris problems: find RecipientInfo structure that matches recipient certificate 313655714Skris and initialise the ASN1 structures properly based on passed cipher. 313755714Skris [Steve Henson] 313855714Skris 313955714Skris *) Belatedly make the BN tests actually check the results. 314055714Skris [Ben Laurie] 314155714Skris 314255714Skris *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 314355714Skris to and from BNs: it was completely broken. New compilation option 314455714Skris NEG_PUBKEY_BUG to allow for some broken certificates that encode public 314555714Skris key elements as negative integers. 314655714Skris [Steve Henson] 314755714Skris 314855714Skris *) Reorganize and speed up MD5. 314955714Skris [Andy Polyakov <appro@fy.chalmers.se>] 315055714Skris 315155714Skris *) VMS support. 315255714Skris [Richard Levitte <richard@levitte.org>] 315355714Skris 315455714Skris *) New option -out to asn1parse to allow the parsed structure to be 315555714Skris output to a file. This is most useful when combined with the -strparse 315655714Skris option to examine the output of things like OCTET STRINGS. 315755714Skris [Steve Henson] 315855714Skris 315955714Skris *) Make SSL library a little more fool-proof by not requiring any longer 316055714Skris that SSL_set_{accept,connect}_state be called before 316155714Skris SSL_{accept,connect} may be used (SSL_set_..._state is omitted 316255714Skris in many applications because usually everything *appeared* to work as 316355714Skris intended anyway -- now it really works as intended). 316455714Skris [Bodo Moeller] 316555714Skris 316655714Skris *) Move openssl.cnf out of lib/. 316755714Skris [Ulf M�ller] 316855714Skris 316955714Skris *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 317055714Skris -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 317155714Skris -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 317255714Skris [Ralf S. Engelschall] 317355714Skris 317455714Skris *) Various fixes to the EVP and PKCS#7 code. It may now be able to 317555714Skris handle PKCS#7 enveloped data properly. 317655714Skris [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 317755714Skris 317855714Skris *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 317955714Skris copying pointers. The cert_st handling is changed by this in 318055714Skris various ways (and thus what used to be known as ctx->default_cert 318155714Skris is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 318255714Skris any longer when s->cert does not give us what we need). 318355714Skris ssl_cert_instantiate becomes obsolete by this change. 318455714Skris As soon as we've got the new code right (possibly it already is?), 318555714Skris we have solved a couple of bugs of the earlier code where s->cert 318655714Skris was used as if it could not have been shared with other SSL structures. 318755714Skris 318855714Skris Note that using the SSL API in certain dirty ways now will result 318955714Skris in different behaviour than observed with earlier library versions: 319055714Skris Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 319155714Skris does not influence s as it used to. 319255714Skris 319355714Skris In order to clean up things more thoroughly, inside SSL_SESSION 319455714Skris we don't use CERT any longer, but a new structure SESS_CERT 319555714Skris that holds per-session data (if available); currently, this is 319655714Skris the peer's certificate chain and, for clients, the server's certificate 319755714Skris and temporary key. CERT holds only those values that can have 319855714Skris meaningful defaults in an SSL_CTX. 319955714Skris [Bodo Moeller] 320055714Skris 320155714Skris *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 320255714Skris from the internal representation. Various PKCS#7 fixes: remove some 320355714Skris evil casts and set the enc_dig_alg field properly based on the signing 320455714Skris key type. 320555714Skris [Steve Henson] 320655714Skris 320755714Skris *) Allow PKCS#12 password to be set from the command line or the 320855714Skris environment. Let 'ca' get its config file name from the environment 320955714Skris variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 321055714Skris and 'x509'). 321155714Skris [Steve Henson] 321255714Skris 321355714Skris *) Allow certificate policies extension to use an IA5STRING for the 321455714Skris organization field. This is contrary to the PKIX definition but 321555714Skris VeriSign uses it and IE5 only recognises this form. Document 'x509' 321655714Skris extension option. 321755714Skris [Steve Henson] 321855714Skris 321955714Skris *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 322055714Skris without disallowing inline assembler and the like for non-pedantic builds. 322155714Skris [Ben Laurie] 322255714Skris 322355714Skris *) Support Borland C++ builder. 322455714Skris [Janez Jere <jj@void.si>, modified by Ulf M�ller] 322555714Skris 322655714Skris *) Support Mingw32. 322755714Skris [Ulf M�ller] 322855714Skris 322955714Skris *) SHA-1 cleanups and performance enhancements. 323055714Skris [Andy Polyakov <appro@fy.chalmers.se>] 323155714Skris 323255714Skris *) Sparc v8plus assembler for the bignum library. 323355714Skris [Andy Polyakov <appro@fy.chalmers.se>] 323455714Skris 323555714Skris *) Accept any -xxx and +xxx compiler options in Configure. 323655714Skris [Ulf M�ller] 323755714Skris 323855714Skris *) Update HPUX configuration. 323955714Skris [Anonymous] 324055714Skris 324155714Skris *) Add missing sk_<type>_unshift() function to safestack.h 324255714Skris [Ralf S. Engelschall] 324355714Skris 324455714Skris *) New function SSL_CTX_use_certificate_chain_file that sets the 324555714Skris "extra_cert"s in addition to the certificate. (This makes sense 324655714Skris only for "PEM" format files, as chains as a whole are not 324755714Skris DER-encoded.) 324855714Skris [Bodo Moeller] 324955714Skris 325055714Skris *) Support verify_depth from the SSL API. 325155714Skris x509_vfy.c had what can be considered an off-by-one-error: 325255714Skris Its depth (which was not part of the external interface) 325355714Skris was actually counting the number of certificates in a chain; 325455714Skris now it really counts the depth. 325555714Skris [Bodo Moeller] 325655714Skris 325755714Skris *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 325855714Skris instead of X509err, which often resulted in confusing error 325955714Skris messages since the error codes are not globally unique 326055714Skris (e.g. an alleged error in ssl3_accept when a certificate 326155714Skris didn't match the private key). 326255714Skris 326355714Skris *) New function SSL_CTX_set_session_id_context that allows to set a default 326455714Skris value (so that you don't need SSL_set_session_id_context for each 326555714Skris connection using the SSL_CTX). 326655714Skris [Bodo Moeller] 326755714Skris 326855714Skris *) OAEP decoding bug fix. 326955714Skris [Ulf M�ller] 327055714Skris 327155714Skris *) Support INSTALL_PREFIX for package builders, as proposed by 327255714Skris David Harris. 327355714Skris [Bodo Moeller] 327455714Skris 327555714Skris *) New Configure options "threads" and "no-threads". For systems 327655714Skris where the proper compiler options are known (currently Solaris 327755714Skris and Linux), "threads" is the default. 327855714Skris [Bodo Moeller] 327955714Skris 328055714Skris *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 328155714Skris [Bodo Moeller] 328255714Skris 328355714Skris *) Install various scripts to $(OPENSSLDIR)/misc, not to 328455714Skris $(INSTALLTOP)/bin -- they shouldn't clutter directories 328555714Skris such as /usr/local/bin. 328655714Skris [Bodo Moeller] 328755714Skris 328855714Skris *) "make linux-shared" to build shared libraries. 328955714Skris [Niels Poppe <niels@netbox.org>] 329055714Skris 329155714Skris *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 329255714Skris [Ulf M�ller] 329355714Skris 329455714Skris *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 329555714Skris extension adding in x509 utility. 329655714Skris [Steve Henson] 329755714Skris 329855714Skris *) Remove NOPROTO sections and error code comments. 329955714Skris [Ulf M�ller] 330055714Skris 330155714Skris *) Partial rewrite of the DEF file generator to now parse the ANSI 330255714Skris prototypes. 330355714Skris [Steve Henson] 330455714Skris 330555714Skris *) New Configure options --prefix=DIR and --openssldir=DIR. 330655714Skris [Ulf M�ller] 330755714Skris 330855714Skris *) Complete rewrite of the error code script(s). It is all now handled 330955714Skris by one script at the top level which handles error code gathering, 331055714Skris header rewriting and C source file generation. It should be much better 331155714Skris than the old method: it now uses a modified version of Ulf's parser to 331255714Skris read the ANSI prototypes in all header files (thus the old K&R definitions 331355714Skris aren't needed for error creation any more) and do a better job of 331455714Skris translating function codes into names. The old 'ASN1 error code imbedded 331555714Skris in a comment' is no longer necessary and it doesn't use .err files which 331655714Skris have now been deleted. Also the error code call doesn't have to appear all 331755714Skris on one line (which resulted in some large lines...). 331855714Skris [Steve Henson] 331955714Skris 332055714Skris *) Change #include filenames from <foo.h> to <openssl/foo.h>. 332155714Skris [Bodo Moeller] 332255714Skris 332355714Skris *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 332455714Skris 0 (which usually indicates a closed connection), but continue reading. 332555714Skris [Bodo Moeller] 332655714Skris 332755714Skris *) Fix some race conditions. 332855714Skris [Bodo Moeller] 332955714Skris 333055714Skris *) Add support for CRL distribution points extension. Add Certificate 333155714Skris Policies and CRL distribution points documentation. 333255714Skris [Steve Henson] 333355714Skris 333455714Skris *) Move the autogenerated header file parts to crypto/opensslconf.h. 333555714Skris [Ulf M�ller] 333655714Skris 333755714Skris *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 333855714Skris 8 of keying material. Merlin has also confirmed interop with this fix 333955714Skris between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 334055714Skris [Merlin Hughes <merlin@baltimore.ie>] 334155714Skris 334255714Skris *) Fix lots of warnings. 334355714Skris [Richard Levitte <levitte@stacken.kth.se>] 334455714Skris 334555714Skris *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 334655714Skris the directory spec didn't end with a LIST_SEPARATOR_CHAR. 334755714Skris [Richard Levitte <levitte@stacken.kth.se>] 334855714Skris 334955714Skris *) Fix problems with sizeof(long) == 8. 335055714Skris [Andy Polyakov <appro@fy.chalmers.se>] 335155714Skris 335255714Skris *) Change functions to ANSI C. 335355714Skris [Ulf M�ller] 335455714Skris 335555714Skris *) Fix typos in error codes. 335655714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 335755714Skris 335855714Skris *) Remove defunct assembler files from Configure. 335955714Skris [Ulf M�ller] 336055714Skris 336155714Skris *) SPARC v8 assembler BIGNUM implementation. 336255714Skris [Andy Polyakov <appro@fy.chalmers.se>] 336355714Skris 336455714Skris *) Support for Certificate Policies extension: both print and set. 336555714Skris Various additions to support the r2i method this uses. 336655714Skris [Steve Henson] 336755714Skris 336855714Skris *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 336955714Skris return a const string when you are expecting an allocated buffer. 337055714Skris [Ben Laurie] 337155714Skris 337255714Skris *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 337355714Skris types DirectoryString and DisplayText. 337455714Skris [Steve Henson] 337555714Skris 337655714Skris *) Add code to allow r2i extensions to access the configuration database, 337755714Skris add an LHASH database driver and add several ctx helper functions. 337855714Skris [Steve Henson] 337955714Skris 338055714Skris *) Fix an evil bug in bn_expand2() which caused various BN functions to 338155714Skris fail when they extended the size of a BIGNUM. 338255714Skris [Steve Henson] 338355714Skris 338455714Skris *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 338555714Skris support typesafe stack. 338655714Skris [Steve Henson] 338755714Skris 338855714Skris *) Fix typo in SSL_[gs]et_options(). 338955714Skris [Nils Frostberg <nils@medcom.se>] 339055714Skris 339155714Skris *) Delete various functions and files that belonged to the (now obsolete) 339255714Skris old X509V3 handling code. 339355714Skris [Steve Henson] 339455714Skris 339555714Skris *) New Configure option "rsaref". 339655714Skris [Ulf M�ller] 339755714Skris 339855714Skris *) Don't auto-generate pem.h. 339955714Skris [Bodo Moeller] 340055714Skris 340155714Skris *) Introduce type-safe ASN.1 SETs. 340255714Skris [Ben Laurie] 340355714Skris 340455714Skris *) Convert various additional casted stacks to type-safe STACK_OF() variants. 340555714Skris [Ben Laurie, Ralf S. Engelschall, Steve Henson] 340655714Skris 340755714Skris *) Introduce type-safe STACKs. This will almost certainly break lots of code 340855714Skris that links with OpenSSL (well at least cause lots of warnings), but fear 340955714Skris not: the conversion is trivial, and it eliminates loads of evil casts. A 341055714Skris few STACKed things have been converted already. Feel free to convert more. 341155714Skris In the fullness of time, I'll do away with the STACK type altogether. 341255714Skris [Ben Laurie] 341355714Skris 341455714Skris *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 341555714Skris specified in <certfile> by updating the entry in the index.txt file. 341655714Skris This way one no longer has to edit the index.txt file manually for 341755714Skris revoking a certificate. The -revoke option does the gory details now. 341855714Skris [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 341955714Skris 342055714Skris *) Fix `openssl crl -noout -text' combination where `-noout' killed the 342155714Skris `-text' option at all and this way the `-noout -text' combination was 342255714Skris inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 342355714Skris [Ralf S. Engelschall] 342455714Skris 342555714Skris *) Make sure a corresponding plain text error message exists for the 342655714Skris X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 342755714Skris verify callback function determined that a certificate was revoked. 342855714Skris [Ralf S. Engelschall] 342955714Skris 343055714Skris *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 343155714Skris ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 343255714Skris all available cipers including rc5, which was forgotten until now. 343355714Skris In order to let the testing shell script know which algorithms 343455714Skris are available, a new (up to now undocumented) command 343555714Skris "openssl list-cipher-commands" is used. 343655714Skris [Bodo Moeller] 343755714Skris 343855714Skris *) Bugfix: s_client occasionally would sleep in select() when 343955714Skris it should have checked SSL_pending() first. 344055714Skris [Bodo Moeller] 344155714Skris 344255714Skris *) New functions DSA_do_sign and DSA_do_verify to provide access to 344355714Skris the raw DSA values prior to ASN.1 encoding. 344455714Skris [Ulf M�ller] 344555714Skris 344655714Skris *) Tweaks to Configure 344755714Skris [Niels Poppe <niels@netbox.org>] 344855714Skris 344955714Skris *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 345055714Skris yet... 345155714Skris [Steve Henson] 345255714Skris 345355714Skris *) New variables $(RANLIB) and $(PERL) in the Makefiles. 345455714Skris [Ulf M�ller] 345555714Skris 345655714Skris *) New config option to avoid instructions that are illegal on the 80386. 345755714Skris The default code is faster, but requires at least a 486. 345855714Skris [Ulf M�ller] 345955714Skris 346055714Skris *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 346155714Skris SSL2_SERVER_VERSION (not used at all) macros, which are now the 346255714Skris same as SSL2_VERSION anyway. 346355714Skris [Bodo Moeller] 346455714Skris 346555714Skris *) New "-showcerts" option for s_client. 346655714Skris [Bodo Moeller] 346755714Skris 346855714Skris *) Still more PKCS#12 integration. Add pkcs12 application to openssl 346955714Skris application. Various cleanups and fixes. 347055714Skris [Steve Henson] 347155714Skris 347255714Skris *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 347355714Skris modify error routines to work internally. Add error codes and PBE init 347455714Skris to library startup routines. 347555714Skris [Steve Henson] 347655714Skris 347755714Skris *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 347855714Skris packing functions to asn1 and evp. Changed function names and error 347955714Skris codes along the way. 348055714Skris [Steve Henson] 348155714Skris 348255714Skris *) PKCS12 integration: and so it begins... First of several patches to 348355714Skris slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 348455714Skris objects to objects.h 348555714Skris [Steve Henson] 348655714Skris 348755714Skris *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 348855714Skris and display support for Thawte strong extranet extension. 348955714Skris [Steve Henson] 349055714Skris 349155714Skris *) Add LinuxPPC support. 349255714Skris [Jeff Dubrule <igor@pobox.org>] 349355714Skris 349455714Skris *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 349555714Skris bn_div_words in alpha.s. 349655714Skris [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 349755714Skris 349855714Skris *) Make sure the RSA OAEP test is skipped under -DRSAref because 349955714Skris OAEP isn't supported when OpenSSL is built with RSAref. 350055714Skris [Ulf Moeller <ulf@fitug.de>] 350155714Skris 350255714Skris *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 350355714Skris so they no longer are missing under -DNOPROTO. 350455714Skris [Soren S. Jorvang <soren@t.dk>] 350555714Skris 350655714Skris 350755714Skris Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 350855714Skris 350955714Skris *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 351055714Skris doesn't work when the session is reused. Coming soon! 351155714Skris [Ben Laurie] 351255714Skris 351355714Skris *) Fix a security hole, that allows sessions to be reused in the wrong 351455714Skris context thus bypassing client cert protection! All software that uses 351555714Skris client certs and session caches in multiple contexts NEEDS PATCHING to 351655714Skris allow session reuse! A fuller solution is in the works. 351755714Skris [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 351855714Skris 351955714Skris *) Some more source tree cleanups (removed obsolete files 352055714Skris crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 352155714Skris permission on "config" script to be executable) and a fix for the INSTALL 352255714Skris document. 352355714Skris [Ulf Moeller <ulf@fitug.de>] 352455714Skris 352555714Skris *) Remove some legacy and erroneous uses of malloc, free instead of 352655714Skris Malloc, Free. 352755714Skris [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 352855714Skris 352955714Skris *) Make rsa_oaep_test return non-zero on error. 353055714Skris [Ulf Moeller <ulf@fitug.de>] 353155714Skris 353255714Skris *) Add support for native Solaris shared libraries. Configure 353355714Skris solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 353455714Skris if someone would make that last step automatic. 353555714Skris [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 353655714Skris 353755714Skris *) ctx_size was not built with the right compiler during "make links". Fixed. 353855714Skris [Ben Laurie] 353955714Skris 354055714Skris *) Change the meaning of 'ALL' in the cipher list. It now means "everything 354155714Skris except NULL ciphers". This means the default cipher list will no longer 354255714Skris enable NULL ciphers. They need to be specifically enabled e.g. with 354355714Skris the string "DEFAULT:eNULL". 354455714Skris [Steve Henson] 354555714Skris 354655714Skris *) Fix to RSA private encryption routines: if p < q then it would 354755714Skris occasionally produce an invalid result. This will only happen with 354855714Skris externally generated keys because OpenSSL (and SSLeay) ensure p > q. 354955714Skris [Steve Henson] 355055714Skris 355155714Skris *) Be less restrictive and allow also `perl util/perlpath.pl 355255714Skris /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 355355714Skris because this way one can also use an interpreter named `perl5' (which is 355455714Skris usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 355555714Skris installed as `perl'). 355655714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 355755714Skris 355855714Skris *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 355955714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 356055714Skris 356155714Skris *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 356255714Skris advapi32.lib to Win32 build and change the pem test comparision 356355714Skris to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 356455714Skris suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 356555714Skris and crypto/des/ede_cbcm_enc.c. 356655714Skris [Steve Henson] 356755714Skris 356855714Skris *) DES quad checksum was broken on big-endian architectures. Fixed. 356955714Skris [Ben Laurie] 357055714Skris 357155714Skris *) Comment out two functions in bio.h that aren't implemented. Fix up the 357255714Skris Win32 test batch file so it (might) work again. The Win32 test batch file 357355714Skris is horrible: I feel ill.... 357455714Skris [Steve Henson] 357555714Skris 357655714Skris *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 357755714Skris in e_os.h. Audit of header files to check ANSI and non ANSI 357855714Skris sections: 10 functions were absent from non ANSI section and not exported 357955714Skris from Windows DLLs. Fixed up libeay.num for new functions. 358055714Skris [Steve Henson] 358155714Skris 358255714Skris *) Make `openssl version' output lines consistent. 358355714Skris [Ralf S. Engelschall] 358455714Skris 358555714Skris *) Fix Win32 symbol export lists for BIO functions: Added 358655714Skris BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 358755714Skris to ms/libeay{16,32}.def. 358855714Skris [Ralf S. Engelschall] 358955714Skris 359055714Skris *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 359155714Skris fine under Unix and passes some trivial tests I've now added. But the 359255714Skris whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 359355714Skris added to make sure no one expects that this stuff really works in the 359455714Skris OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 359555714Skris up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 359655714Skris openssl_bio.xs. 359755714Skris [Ralf S. Engelschall] 359855714Skris 359955714Skris *) Fix the generation of two part addresses in perl. 360055714Skris [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 360155714Skris 360255714Skris *) Add config entry for Linux on MIPS. 360355714Skris [John Tobey <jtobey@channel1.com>] 360455714Skris 360555714Skris *) Make links whenever Configure is run, unless we are on Windoze. 360655714Skris [Ben Laurie] 360755714Skris 360855714Skris *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 360955714Skris Currently only issuerAltName and AuthorityKeyIdentifier make any sense 361055714Skris in CRLs. 361155714Skris [Steve Henson] 361255714Skris 361355714Skris *) Add a useful kludge to allow package maintainers to specify compiler and 361455714Skris other platforms details on the command line without having to patch the 361555714Skris Configure script everytime: One now can use ``perl Configure 361655714Skris <id>:<details>'', i.e. platform ids are allowed to have details appended 361755714Skris to them (seperated by colons). This is treated as there would be a static 361855714Skris pre-configured entry in Configure's %table under key <id> with value 361955714Skris <details> and ``perl Configure <id>'' is called. So, when you want to 362055714Skris perform a quick test-compile under FreeBSD 3.1 with pgcc and without 362155714Skris assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 362255714Skris now, which overrides the FreeBSD-elf entry on-the-fly. 362355714Skris [Ralf S. Engelschall] 362455714Skris 362555714Skris *) Disable new TLS1 ciphersuites by default: they aren't official yet. 362655714Skris [Ben Laurie] 362755714Skris 362855714Skris *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 362955714Skris on the `perl Configure ...' command line. This way one can compile 363055714Skris OpenSSL libraries with Position Independent Code (PIC) which is needed 363155714Skris for linking it into DSOs. 363255714Skris [Ralf S. Engelschall] 363355714Skris 363455714Skris *) Remarkably, export ciphers were totally broken and no-one had noticed! 363555714Skris Fixed. 363655714Skris [Ben Laurie] 363755714Skris 363855714Skris *) Cleaned up the LICENSE document: The official contact for any license 363955714Skris questions now is the OpenSSL core team under openssl-core@openssl.org. 364055714Skris And add a paragraph about the dual-license situation to make sure people 364155714Skris recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 364255714Skris to the OpenSSL toolkit. 364355714Skris [Ralf S. Engelschall] 364455714Skris 364555714Skris *) General source tree makefile cleanups: Made `making xxx in yyy...' 364655714Skris display consistent in the source tree and replaced `/bin/rm' by `rm'. 364755714Skris Additonally cleaned up the `make links' target: Remove unnecessary 364855714Skris semicolons, subsequent redundant removes, inline point.sh into mklink.sh 364955714Skris to speed processing and no longer clutter the display with confusing 365055714Skris stuff. Instead only the actually done links are displayed. 365155714Skris [Ralf S. Engelschall] 365255714Skris 365355714Skris *) Permit null encryption ciphersuites, used for authentication only. It used 365455714Skris to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 365555714Skris It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 365655714Skris encryption. 365755714Skris [Ben Laurie] 365855714Skris 365955714Skris *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 366055714Skris signed attributes when verifying signatures (this would break them), 366155714Skris the detached data encoding was wrong and public keys obtained using 366255714Skris X509_get_pubkey() weren't freed. 366355714Skris [Steve Henson] 366455714Skris 366555714Skris *) Add text documentation for the BUFFER functions. Also added a work around 366655714Skris to a Win95 console bug. This was triggered by the password read stuff: the 366755714Skris last character typed gets carried over to the next fread(). If you were 366855714Skris generating a new cert request using 'req' for example then the last 366955714Skris character of the passphrase would be CR which would then enter the first 367055714Skris field as blank. 367155714Skris [Steve Henson] 367255714Skris 367355714Skris *) Added the new `Includes OpenSSL Cryptography Software' button as 367455714Skris doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 367555714Skris button and can be used by applications based on OpenSSL to show the 367655714Skris relationship to the OpenSSL project. 367755714Skris [Ralf S. Engelschall] 367855714Skris 367955714Skris *) Remove confusing variables in function signatures in files 368055714Skris ssl/ssl_lib.c and ssl/ssl.h. 368155714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 368255714Skris 368355714Skris *) Don't install bss_file.c under PREFIX/include/ 368455714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 368555714Skris 368655714Skris *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 368755714Skris functions that return function pointers and has support for NT specific 368855714Skris stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 368955714Skris #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 369055714Skris unsigned to signed types: this was killing the Win32 compile. 369155714Skris [Steve Henson] 369255714Skris 369355714Skris *) Add new certificate file to stack functions, 369455714Skris SSL_add_dir_cert_subjects_to_stack() and 369555714Skris SSL_add_file_cert_subjects_to_stack(). These largely supplant 369655714Skris SSL_load_client_CA_file(), and can be used to add multiple certs easily 369755714Skris to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 369855714Skris This means that Apache-SSL and similar packages don't have to mess around 369955714Skris to add as many CAs as they want to the preferred list. 370055714Skris [Ben Laurie] 370155714Skris 370255714Skris *) Experiment with doxygen documentation. Currently only partially applied to 370355714Skris ssl/ssl_lib.c. 370455714Skris See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 370555714Skris openssl.doxy as the configuration file. 370655714Skris [Ben Laurie] 370755714Skris 370855714Skris *) Get rid of remaining C++-style comments which strict C compilers hate. 370955714Skris [Ralf S. Engelschall, pointed out by Carlos Amengual] 371055714Skris 371155714Skris *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 371255714Skris compiled in by default: it has problems with large keys. 371355714Skris [Steve Henson] 371455714Skris 371555714Skris *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 371655714Skris DH private keys and/or callback functions which directly correspond to 371755714Skris their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 371855714Skris is needed for applications which have to configure certificates on a 371955714Skris per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 372055714Skris (e.g. s_server). 372155714Skris For the RSA certificate situation is makes no difference, but 372255714Skris for the DSA certificate situation this fixes the "no shared cipher" 372355714Skris problem where the OpenSSL cipher selection procedure failed because the 372455714Skris temporary keys were not overtaken from the context and the API provided 372555714Skris no way to reconfigure them. 372655714Skris The new functions now let applications reconfigure the stuff and they 372755714Skris are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 372855714Skris SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 372955714Skris non-public-API function ssl_cert_instantiate() is used as a helper 373055714Skris function and also to reduce code redundancy inside ssl_rsa.c. 373155714Skris [Ralf S. Engelschall] 373255714Skris 373355714Skris *) Move s_server -dcert and -dkey options out of the undocumented feature 373455714Skris area because they are useful for the DSA situation and should be 373555714Skris recognized by the users. 373655714Skris [Ralf S. Engelschall] 373755714Skris 373855714Skris *) Fix the cipher decision scheme for export ciphers: the export bits are 373955714Skris *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 374055714Skris SSL_EXP_MASK. So, the original variable has to be used instead of the 374155714Skris already masked variable. 374255714Skris [Richard Levitte <levitte@stacken.kth.se>] 374355714Skris 374455714Skris *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 374555714Skris [Richard Levitte <levitte@stacken.kth.se>] 374655714Skris 374755714Skris *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 374855714Skris from `int' to `unsigned int' because it's a length and initialized by 374955714Skris EVP_DigestFinal() which expects an `unsigned int *'. 375055714Skris [Richard Levitte <levitte@stacken.kth.se>] 375155714Skris 375255714Skris *) Don't hard-code path to Perl interpreter on shebang line of Configure 375355714Skris script. Instead use the usual Shell->Perl transition trick. 375455714Skris [Ralf S. Engelschall] 375555714Skris 375655714Skris *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 375755714Skris (in addition to RSA certificates) to match the behaviour of `openssl dsa 375855714Skris -noout -modulus' as it's already the case for `openssl rsa -noout 375955714Skris -modulus'. For RSA the -modulus is the real "modulus" while for DSA 376055714Skris currently the public key is printed (a decision which was already done by 376155714Skris `openssl dsa -modulus' in the past) which serves a similar purpose. 376255714Skris Additionally the NO_RSA no longer completely removes the whole -modulus 376355714Skris option; it now only avoids using the RSA stuff. Same applies to NO_DSA 376455714Skris now, too. 376555714Skris [Ralf S. Engelschall] 376655714Skris 376755714Skris *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 376855714Skris BIO. See the source (crypto/evp/bio_ok.c) for more info. 376955714Skris [Arne Ansper <arne@ats.cyber.ee>] 377055714Skris 377155714Skris *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 377255714Skris to be added. Now both 'req' and 'ca' can use new objects defined in the 377355714Skris config file. 377455714Skris [Steve Henson] 377555714Skris 377655714Skris *) Add cool BIO that does syslog (or event log on NT). 377755714Skris [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 377855714Skris 377955714Skris *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 378055714Skris TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 378155714Skris TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 378255714Skris Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 378355714Skris [Ben Laurie] 378455714Skris 378555714Skris *) Add preliminary config info for new extension code. 378655714Skris [Steve Henson] 378755714Skris 378855714Skris *) Make RSA_NO_PADDING really use no padding. 378955714Skris [Ulf Moeller <ulf@fitug.de>] 379055714Skris 379155714Skris *) Generate errors when private/public key check is done. 379255714Skris [Ben Laurie] 379355714Skris 379455714Skris *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 379555714Skris for some CRL extensions and new objects added. 379655714Skris [Steve Henson] 379755714Skris 379855714Skris *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 379955714Skris key usage extension and fuller support for authority key id. 380055714Skris [Steve Henson] 380155714Skris 380255714Skris *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 380355714Skris padding method for RSA, which is recommended for new applications in PKCS 380455714Skris #1 v2.0 (RFC 2437, October 1998). 380555714Skris OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 380655714Skris foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 380755714Skris against Bleichbacher's attack on RSA. 380855714Skris [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 380955714Skris Ben Laurie] 381055714Skris 381155714Skris *) Updates to the new SSL compression code 381255714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 381355714Skris 381455714Skris *) Fix so that the version number in the master secret, when passed 381555714Skris via RSA, checks that if TLS was proposed, but we roll back to SSLv3 381655714Skris (because the server will not accept higher), that the version number 381755714Skris is 0x03,0x01, not 0x03,0x00 381855714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 381955714Skris 382055714Skris *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 382155714Skris leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 382255714Skris in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 382355714Skris [Steve Henson] 382455714Skris 382555714Skris *) Support for RAW extensions where an arbitrary extension can be 382655714Skris created by including its DER encoding. See apps/openssl.cnf for 382755714Skris an example. 382855714Skris [Steve Henson] 382955714Skris 383055714Skris *) Make sure latest Perl versions don't interpret some generated C array 383155714Skris code as Perl array code in the crypto/err/err_genc.pl script. 383255714Skris [Lars Weber <3weber@informatik.uni-hamburg.de>] 383355714Skris 383455714Skris *) Modify ms/do_ms.bat to not generate assembly language makefiles since 383555714Skris not many people have the assembler. Various Win32 compilation fixes and 383655714Skris update to the INSTALL.W32 file with (hopefully) more accurate Win32 383755714Skris build instructions. 383855714Skris [Steve Henson] 383955714Skris 384055714Skris *) Modify configure script 'Configure' to automatically create crypto/date.h 384155714Skris file under Win32 and also build pem.h from pem.org. New script 384255714Skris util/mkfiles.pl to create the MINFO file on environments that can't do a 384355714Skris 'make files': perl util/mkfiles.pl >MINFO should work. 384455714Skris [Steve Henson] 384555714Skris 384655714Skris *) Major rework of DES function declarations, in the pursuit of correctness 384755714Skris and purity. As a result, many evil casts evaporated, and some weirdness, 384855714Skris too. You may find this causes warnings in your code. Zapping your evil 384955714Skris casts will probably fix them. Mostly. 385055714Skris [Ben Laurie] 385155714Skris 385255714Skris *) Fix for a typo in asn1.h. Bug fix to object creation script 385355714Skris obj_dat.pl. It considered a zero in an object definition to mean 385455714Skris "end of object": none of the objects in objects.h have any zeros 385555714Skris so it wasn't spotted. 385655714Skris [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 385755714Skris 385855714Skris *) Add support for Triple DES Cipher Block Chaining with Output Feedback 385955714Skris Masking (CBCM). In the absence of test vectors, the best I have been able 386055714Skris to do is check that the decrypt undoes the encrypt, so far. Send me test 386155714Skris vectors if you have them. 386255714Skris [Ben Laurie] 386355714Skris 386455714Skris *) Correct calculation of key length for export ciphers (too much space was 386555714Skris allocated for null ciphers). This has not been tested! 386655714Skris [Ben Laurie] 386755714Skris 386855714Skris *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 386955714Skris message is now correct (it understands "crypto" and "ssl" on its 387055714Skris command line). There is also now an "update" option. This will update 387155714Skris the util/ssleay.num and util/libeay.num files with any new functions. 387255714Skris If you do a: 387355714Skris perl util/mkdef.pl crypto ssl update 387455714Skris it will update them. 387555714Skris [Steve Henson] 387655714Skris 387755714Skris *) Overhauled the Perl interface (perl/*): 387855714Skris - ported BN stuff to OpenSSL's different BN library 387955714Skris - made the perl/ source tree CVS-aware 388055714Skris - renamed the package from SSLeay to OpenSSL (the files still contain 388155714Skris their history because I've copied them in the repository) 388255714Skris - removed obsolete files (the test scripts will be replaced 388355714Skris by better Test::Harness variants in the future) 388455714Skris [Ralf S. Engelschall] 388555714Skris 388655714Skris *) First cut for a very conservative source tree cleanup: 388755714Skris 1. merge various obsolete readme texts into doc/ssleay.txt 388855714Skris where we collect the old documents and readme texts. 388955714Skris 2. remove the first part of files where I'm already sure that we no 389055714Skris longer need them because of three reasons: either they are just temporary 389155714Skris files which were left by Eric or they are preserved original files where 389255714Skris I've verified that the diff is also available in the CVS via "cvs diff 389355714Skris -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 389455714Skris the crypto/md/ stuff). 389555714Skris [Ralf S. Engelschall] 389655714Skris 389755714Skris *) More extension code. Incomplete support for subject and issuer alt 389855714Skris name, issuer and authority key id. Change the i2v function parameters 389955714Skris and add an extra 'crl' parameter in the X509V3_CTX structure: guess 390055714Skris what that's for :-) Fix to ASN1 macro which messed up 390155714Skris IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 390255714Skris [Steve Henson] 390355714Skris 390455714Skris *) Preliminary support for ENUMERATED type. This is largely copied from the 390555714Skris INTEGER code. 390655714Skris [Steve Henson] 390755714Skris 390855714Skris *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 390955714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 391055714Skris 391155714Skris *) Make sure `make rehash' target really finds the `openssl' program. 391255714Skris [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 391355714Skris 391455714Skris *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 391555714Skris like to hear about it if this slows down other processors. 391655714Skris [Ben Laurie] 391755714Skris 391855714Skris *) Add CygWin32 platform information to Configure script. 391955714Skris [Alan Batie <batie@aahz.jf.intel.com>] 392055714Skris 392155714Skris *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 392255714Skris [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 392355714Skris 392455714Skris *) New program nseq to manipulate netscape certificate sequences 392555714Skris [Steve Henson] 392655714Skris 392755714Skris *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 392855714Skris few typos. 392955714Skris [Steve Henson] 393055714Skris 393155714Skris *) Fixes to BN code. Previously the default was to define BN_RECURSION 393255714Skris but the BN code had some problems that would cause failures when 393355714Skris doing certificate verification and some other functions. 393455714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 393555714Skris 393655714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 393755714Skris [Steve Henson] 393855714Skris 393955714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 394055714Skris [Steve Henson] 394155714Skris 394255714Skris *) Add several PKIX and private extended key usage OIDs. 394355714Skris [Steve Henson] 394455714Skris 394555714Skris *) Modify the 'ca' program to handle the new extension code. Modify 394655714Skris openssl.cnf for new extension format, add comments. 394755714Skris [Steve Henson] 394855714Skris 394955714Skris *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 395055714Skris and add a sample to openssl.cnf so req -x509 now adds appropriate 395155714Skris CA extensions. 395255714Skris [Steve Henson] 395355714Skris 395455714Skris *) Continued X509 V3 changes. Add to other makefiles, integrate with the 395555714Skris error code, add initial support to X509_print() and x509 application. 395655714Skris [Steve Henson] 395755714Skris 395855714Skris *) Takes a deep breath and start addding X509 V3 extension support code. Add 395955714Skris files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 396055714Skris stuff is currently isolated and isn't even compiled yet. 396155714Skris [Steve Henson] 396255714Skris 396355714Skris *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 396455714Skris ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 396555714Skris Removed the versions check from X509 routines when loading extensions: 396655714Skris this allows certain broken certificates that don't set the version 396755714Skris properly to be processed. 396855714Skris [Steve Henson] 396955714Skris 397055714Skris *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 397155714Skris Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 397255714Skris can still be regenerated with "make depend". 397355714Skris [Ben Laurie] 397455714Skris 397555714Skris *) Spelling mistake in C version of CAST-128. 397655714Skris [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 397755714Skris 397855714Skris *) Changes to the error generation code. The perl script err-code.pl 397955714Skris now reads in the old error codes and retains the old numbers, only 398055714Skris adding new ones if necessary. It also only changes the .err files if new 398155714Skris codes are added. The makefiles have been modified to only insert errors 398255714Skris when needed (to avoid needlessly modifying header files). This is done 398355714Skris by only inserting errors if the .err file is newer than the auto generated 398455714Skris C file. To rebuild all the error codes from scratch (the old behaviour) 398555714Skris either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 398655714Skris or delete all the .err files. 398755714Skris [Steve Henson] 398855714Skris 398955714Skris *) CAST-128 was incorrectly implemented for short keys. The C version has 399055714Skris been fixed, but is untested. The assembler versions are also fixed, but 399155714Skris new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 399255714Skris to regenerate it if needed. 399355714Skris [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 399455714Skris Hagino <itojun@kame.net>] 399555714Skris 399655714Skris *) File was opened incorrectly in randfile.c. 399755714Skris [Ulf M�ller <ulf@fitug.de>] 399855714Skris 399955714Skris *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 400055714Skris functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 400155714Skris GeneralizedTime. ASN1_TIME is the proper type used in certificates et 400255714Skris al: it's just almost always a UTCTime. Note this patch adds new error 400355714Skris codes so do a "make errors" if there are problems. 400455714Skris [Steve Henson] 400555714Skris 400655714Skris *) Correct Linux 1 recognition in config. 400755714Skris [Ulf M�ller <ulf@fitug.de>] 400855714Skris 400955714Skris *) Remove pointless MD5 hash when using DSA keys in ca. 401055714Skris [Anonymous <nobody@replay.com>] 401155714Skris 401255714Skris *) Generate an error if given an empty string as a cert directory. Also 401355714Skris generate an error if handed NULL (previously returned 0 to indicate an 401455714Skris error, but didn't set one). 401555714Skris [Ben Laurie, reported by Anonymous <nobody@replay.com>] 401655714Skris 401755714Skris *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 401855714Skris [Ben Laurie] 401955714Skris 402055714Skris *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 402155714Skris parameters. This was causing a warning which killed off the Win32 compile. 402255714Skris [Steve Henson] 402355714Skris 402455714Skris *) Remove C++ style comments from crypto/bn/bn_local.h. 402555714Skris [Neil Costigan <neil.costigan@celocom.com>] 402655714Skris 402755714Skris *) The function OBJ_txt2nid was broken. It was supposed to return a nid 402855714Skris based on a text string, looking up short and long names and finally 402955714Skris "dot" format. The "dot" format stuff didn't work. Added new function 403055714Skris OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 403155714Skris OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 403255714Skris OID is not part of the table. 403355714Skris [Steve Henson] 403455714Skris 403555714Skris *) Add prototypes to X509 lookup/verify methods, fixing a bug in 403655714Skris X509_LOOKUP_by_alias(). 403755714Skris [Ben Laurie] 403855714Skris 403955714Skris *) Sort openssl functions by name. 404055714Skris [Ben Laurie] 404155714Skris 404255714Skris *) Get the gendsa program working (hopefully) and add it to app list. Remove 404355714Skris encryption from sample DSA keys (in case anyone is interested the password 404455714Skris was "1234"). 404555714Skris [Steve Henson] 404655714Skris 404755714Skris *) Make _all_ *_free functions accept a NULL pointer. 404855714Skris [Frans Heymans <fheymans@isaserver.be>] 404955714Skris 405055714Skris *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 405155714Skris NULL pointers. 405255714Skris [Anonymous <nobody@replay.com>] 405355714Skris 405455714Skris *) s_server should send the CAfile as acceptable CAs, not its own cert. 405555714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 405655714Skris 405755714Skris *) Don't blow it for numeric -newkey arguments to apps/req. 405855714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 405955714Skris 406055714Skris *) Temp key "for export" tests were wrong in s3_srvr.c. 406155714Skris [Anonymous <nobody@replay.com>] 406255714Skris 406355714Skris *) Add prototype for temp key callback functions 406455714Skris SSL_CTX_set_tmp_{rsa,dh}_callback(). 406555714Skris [Ben Laurie] 406655714Skris 406755714Skris *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 406855714Skris DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 406955714Skris [Steve Henson] 407055714Skris 407155714Skris *) X509_name_add_entry() freed the wrong thing after an error. 407255714Skris [Arne Ansper <arne@ats.cyber.ee>] 407355714Skris 407455714Skris *) rsa_eay.c would attempt to free a NULL context. 407555714Skris [Arne Ansper <arne@ats.cyber.ee>] 407655714Skris 407755714Skris *) BIO_s_socket() had a broken should_retry() on Windoze. 407855714Skris [Arne Ansper <arne@ats.cyber.ee>] 407955714Skris 408055714Skris *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 408155714Skris [Arne Ansper <arne@ats.cyber.ee>] 408255714Skris 408355714Skris *) Make sure the already existing X509_STORE->depth variable is initialized 408455714Skris in X509_STORE_new(), but document the fact that this variable is still 408555714Skris unused in the certificate verification process. 408655714Skris [Ralf S. Engelschall] 408755714Skris 408855714Skris *) Fix the various library and apps files to free up pkeys obtained from 408955714Skris X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 409055714Skris [Steve Henson] 409155714Skris 409255714Skris *) Fix reference counting in X509_PUBKEY_get(). This makes 409355714Skris demos/maurice/example2.c work, amongst others, probably. 409455714Skris [Steve Henson and Ben Laurie] 409555714Skris 409655714Skris *) First cut of a cleanup for apps/. First the `ssleay' program is now named 409755714Skris `openssl' and second, the shortcut symlinks for the `openssl <command>' 409855714Skris are no longer created. This way we have a single and consistent command 409955714Skris line interface `openssl <command>', similar to `cvs <command>'. 410055714Skris [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 410155714Skris 410255714Skris *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 410355714Skris BIT STRING wrapper always have zero unused bits. 410455714Skris [Steve Henson] 410555714Skris 410655714Skris *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 410755714Skris [Steve Henson] 410855714Skris 410955714Skris *) Make the top-level INSTALL documentation easier to understand. 411055714Skris [Paul Sutton] 411155714Skris 411255714Skris *) Makefiles updated to exit if an error occurs in a sub-directory 411355714Skris make (including if user presses ^C) [Paul Sutton] 411455714Skris 411555714Skris *) Make Montgomery context stuff explicit in RSA data structure. 411655714Skris [Ben Laurie] 411755714Skris 411855714Skris *) Fix build order of pem and err to allow for generated pem.h. 411955714Skris [Ben Laurie] 412055714Skris 412155714Skris *) Fix renumbering bug in X509_NAME_delete_entry(). 412255714Skris [Ben Laurie] 412355714Skris 412455714Skris *) Enhanced the err-ins.pl script so it makes the error library number 412555714Skris global and can add a library name. This is needed for external ASN1 and 412655714Skris other error libraries. 412755714Skris [Steve Henson] 412855714Skris 412955714Skris *) Fixed sk_insert which never worked properly. 413055714Skris [Steve Henson] 413155714Skris 413255714Skris *) Fix ASN1 macros so they can handle indefinite length construted 413355714Skris EXPLICIT tags. Some non standard certificates use these: they can now 413455714Skris be read in. 413555714Skris [Steve Henson] 413655714Skris 413755714Skris *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 413855714Skris into a single doc/ssleay.txt bundle. This way the information is still 413955714Skris preserved but no longer messes up this directory. Now it's new room for 414055714Skris the new set of documenation files. 414155714Skris [Ralf S. Engelschall] 414255714Skris 414355714Skris *) SETs were incorrectly DER encoded. This was a major pain, because they 414455714Skris shared code with SEQUENCEs, which aren't coded the same. This means that 414555714Skris almost everything to do with SETs or SEQUENCEs has either changed name or 414655714Skris number of arguments. 414755714Skris [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 414855714Skris 414955714Skris *) Fix test data to work with the above. 415055714Skris [Ben Laurie] 415155714Skris 415255714Skris *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 415355714Skris was already fixed by Eric for 0.9.1 it seems. 415455714Skris [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 415555714Skris 415655714Skris *) Autodetect FreeBSD3. 415755714Skris [Ben Laurie] 415855714Skris 415955714Skris *) Fix various bugs in Configure. This affects the following platforms: 416055714Skris nextstep 416155714Skris ncr-scde 416255714Skris unixware-2.0 416355714Skris unixware-2.0-pentium 416455714Skris sco5-cc. 416555714Skris [Ben Laurie] 416655714Skris 416755714Skris *) Eliminate generated files from CVS. Reorder tests to regenerate files 416855714Skris before they are needed. 416955714Skris [Ben Laurie] 417055714Skris 417155714Skris *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 417255714Skris [Ben Laurie] 417355714Skris 417455714Skris 417555714Skris Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 417655714Skris 417755714Skris *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 417855714Skris changed SSLeay to OpenSSL in version strings. 417955714Skris [Ralf S. Engelschall] 418055714Skris 418155714Skris *) Some fixups to the top-level documents. 418255714Skris [Paul Sutton] 418355714Skris 418455714Skris *) Fixed the nasty bug where rsaref.h was not found under compile-time 418555714Skris because the symlink to include/ was missing. 418655714Skris [Ralf S. Engelschall] 418755714Skris 418855714Skris *) Incorporated the popular no-RSA/DSA-only patches 418955714Skris which allow to compile a RSA-free SSLeay. 419055714Skris [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 419155714Skris 419255714Skris *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 419355714Skris when "ssleay" is still not found. 419455714Skris [Ralf S. Engelschall] 419555714Skris 419655714Skris *) Added more platforms to Configure: Cray T3E, HPUX 11, 419755714Skris [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 419855714Skris 419955714Skris *) Updated the README file. 420055714Skris [Ralf S. Engelschall] 420155714Skris 420255714Skris *) Added various .cvsignore files in the CVS repository subdirs 420355714Skris to make a "cvs update" really silent. 420455714Skris [Ralf S. Engelschall] 420555714Skris 420655714Skris *) Recompiled the error-definition header files and added 420755714Skris missing symbols to the Win32 linker tables. 420855714Skris [Ralf S. Engelschall] 420955714Skris 421055714Skris *) Cleaned up the top-level documents; 421155714Skris o new files: CHANGES and LICENSE 421255714Skris o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 421355714Skris o merged COPYRIGHT into LICENSE 421455714Skris o removed obsolete TODO file 421555714Skris o renamed MICROSOFT to INSTALL.W32 421655714Skris [Ralf S. Engelschall] 421755714Skris 421855714Skris *) Removed dummy files from the 0.9.1b source tree: 421955714Skris crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 422055714Skris crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 422155714Skris crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 422255714Skris crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 422355714Skris util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 422455714Skris [Ralf S. Engelschall] 422555714Skris 422655714Skris *) Added various platform portability fixes. 422755714Skris [Mark J. Cox] 422855714Skris 422955714Skris *) The Genesis of the OpenSSL rpject: 423055714Skris We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 423155714Skris Young and Tim J. Hudson created while they were working for C2Net until 423255714Skris summer 1998. 423355714Skris [The OpenSSL Project] 423455714Skris 423555714Skris 423655714Skris Changes between 0.9.0b and 0.9.1b [not released] 423755714Skris 423855714Skris *) Updated a few CA certificates under certs/ 423955714Skris [Eric A. Young] 424055714Skris 424155714Skris *) Changed some BIGNUM api stuff. 424255714Skris [Eric A. Young] 424355714Skris 424455714Skris *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 424555714Skris DGUX x86, Linux Alpha, etc. 424655714Skris [Eric A. Young] 424755714Skris 424855714Skris *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 424955714Skris RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 425055714Skris available). 425155714Skris [Eric A. Young] 425255714Skris 425355714Skris *) Add -strparse option to asn1pars program which parses nested 425455714Skris binary structures 425555714Skris [Dr Stephen Henson <shenson@bigfoot.com>] 425655714Skris 425755714Skris *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 425855714Skris [Eric A. Young] 425955714Skris 426055714Skris *) DSA fix for "ca" program. 426155714Skris [Eric A. Young] 426255714Skris 426355714Skris *) Added "-genkey" option to "dsaparam" program. 426455714Skris [Eric A. Young] 426555714Skris 426655714Skris *) Added RIPE MD160 (rmd160) message digest. 426755714Skris [Eric A. Young] 426855714Skris 426955714Skris *) Added -a (all) option to "ssleay version" command. 427055714Skris [Eric A. Young] 427155714Skris 427255714Skris *) Added PLATFORM define which is the id given to Configure. 427355714Skris [Eric A. Young] 427455714Skris 427555714Skris *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 427655714Skris [Eric A. Young] 427755714Skris 427855714Skris *) Extended the ASN.1 parser routines. 427955714Skris [Eric A. Young] 428055714Skris 428155714Skris *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 428255714Skris [Eric A. Young] 428355714Skris 428455714Skris *) Added a BN_CTX to the BN library. 428555714Skris [Eric A. Young] 428655714Skris 428755714Skris *) Fixed the weak key values in DES library 428855714Skris [Eric A. Young] 428955714Skris 429055714Skris *) Changed API in EVP library for cipher aliases. 429155714Skris [Eric A. Young] 429255714Skris 429355714Skris *) Added support for RC2/64bit cipher. 429455714Skris [Eric A. Young] 429555714Skris 429655714Skris *) Converted the lhash library to the crypto/mem.c functions. 429755714Skris [Eric A. Young] 429855714Skris 429955714Skris *) Added more recognized ASN.1 object ids. 430055714Skris [Eric A. Young] 430155714Skris 430255714Skris *) Added more RSA padding checks for SSL/TLS. 430355714Skris [Eric A. Young] 430455714Skris 430555714Skris *) Added BIO proxy/filter functionality. 430655714Skris [Eric A. Young] 430755714Skris 430855714Skris *) Added extra_certs to SSL_CTX which can be used 430955714Skris send extra CA certificates to the client in the CA cert chain sending 431055714Skris process. It can be configured with SSL_CTX_add_extra_chain_cert(). 431155714Skris [Eric A. Young] 431255714Skris 431355714Skris *) Now Fortezza is denied in the authentication phase because 431455714Skris this is key exchange mechanism is not supported by SSLeay at all. 431555714Skris [Eric A. Young] 431655714Skris 431755714Skris *) Additional PKCS1 checks. 431855714Skris [Eric A. Young] 431955714Skris 432055714Skris *) Support the string "TLSv1" for all TLS v1 ciphers. 432155714Skris [Eric A. Young] 432255714Skris 432355714Skris *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 432455714Skris ex_data index of the SSL context in the X509_STORE_CTX ex_data. 432555714Skris [Eric A. Young] 432655714Skris 432755714Skris *) Fixed a few memory leaks. 432855714Skris [Eric A. Young] 432955714Skris 433055714Skris *) Fixed various code and comment typos. 433155714Skris [Eric A. Young] 433255714Skris 433355714Skris *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 433455714Skris bytes sent in the client random. 433555714Skris [Edward Bishop <ebishop@spyglass.com>] 433655714Skris 4337