CHANGES revision 101613
155714Skris 255714Skris OpenSSL CHANGES 355714Skris _______________ 455714Skris 5101613Snectar Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 6101613Snectar 7101613Snectar *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 8101613Snectar and get fix the header length calculation. 9101613Snectar [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 10101613Snectar Alon Kantor <alonk@checkpoint.com> (and others), 11101613Snectar Steve Henson] 12101613Snectar 13101613Snectar *) Use proper error handling instead of 'assertions' in buffer 14101613Snectar overflow checks added in 0.9.6e. This prevents DoS (the 15101613Snectar assertions could call abort()). 16101613Snectar [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] 17101613Snectar 18100936Snectar Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 19100936Snectar 20100936Snectar *) Fix cipher selection routines: ciphers without encryption had no flags 21100936Snectar for the cipher strength set and where therefore not handled correctly 22100936Snectar by the selection routines (PR #130). 23100936Snectar [Lutz Jaenicke] 24100936Snectar 25100936Snectar *) Fix EVP_dsa_sha macro. 26100936Snectar [Nils Larsch] 27100936Snectar 28100936Snectar *) New option 29100936Snectar SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 30100936Snectar for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 31100936Snectar that was added in OpenSSL 0.9.6d. 32100936Snectar 33100936Snectar As the countermeasure turned out to be incompatible with some 34100936Snectar broken SSL implementations, the new option is part of SSL_OP_ALL. 35100936Snectar SSL_OP_ALL is usually employed when compatibility with weird SSL 36100936Snectar implementations is desired (e.g. '-bugs' option to 's_client' and 37100936Snectar 's_server'), so the new option is automatically set in many 38100936Snectar applications. 39100936Snectar [Bodo Moeller] 40100936Snectar 41100936Snectar *) Changes in security patch: 42100936Snectar 43100936Snectar Changes marked "(CHATS)" were sponsored by the Defense Advanced 44100936Snectar Research Projects Agency (DARPA) and Air Force Research Laboratory, 45100936Snectar Air Force Materiel Command, USAF, under agreement number 46100936Snectar F30602-01-2-0537. 47100936Snectar 48100936Snectar *) Add various sanity checks to asn1_get_length() to reject 49100936Snectar the ASN1 length bytes if they exceed sizeof(long), will appear 50100936Snectar negative or the content length exceeds the length of the 51100936Snectar supplied buffer. 52100936Snectar [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 53100936Snectar 54100936Snectar *) Assertions for various potential buffer overflows, not known to 55100936Snectar happen in practice. 56100936Snectar [Ben Laurie (CHATS)] 57100936Snectar 58100936Snectar *) Various temporary buffers to hold ASCII versions of integers were 59100936Snectar too small for 64 bit platforms. (CAN-2002-0655) 60100936Snectar [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 61100936Snectar 62100936Snectar *) Remote buffer overflow in SSL3 protocol - an attacker could 63100936Snectar supply an oversized session ID to a client. (CAN-2002-0656) 64100936Snectar [Ben Laurie (CHATS)] 65100936Snectar 66100936Snectar *) Remote buffer overflow in SSL2 protocol - an attacker could 67100936Snectar supply an oversized client master key. (CAN-2002-0656) 68100936Snectar [Ben Laurie (CHATS)] 69100936Snectar 70100928Snectar Changes between 0.9.6c and 0.9.6d [9 May 2002] 71100928Snectar 72100928Snectar *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 73100928Snectar encoded as NULL) with id-dsa-with-sha1. 74100928Snectar [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 75100928Snectar 76100928Snectar *) Check various X509_...() return values in apps/req.c. 77100928Snectar [Nils Larsch <nla@trustcenter.de>] 78100928Snectar 79100928Snectar *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 80100928Snectar an end-of-file condition would erronously be flagged, when the CRLF 81100928Snectar was just at the end of a processed block. The bug was discovered when 82100928Snectar processing data through a buffering memory BIO handing the data to a 83100928Snectar BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 84100928Snectar <ptsekov@syntrex.com> and Nedelcho Stanev. 85100928Snectar [Lutz Jaenicke] 86100928Snectar 87100928Snectar *) Implement a countermeasure against a vulnerability recently found 88100928Snectar in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 89100928Snectar before application data chunks to avoid the use of known IVs 90100928Snectar with data potentially chosen by the attacker. 91100928Snectar [Bodo Moeller] 92100928Snectar 93100928Snectar *) Fix length checks in ssl3_get_client_hello(). 94100928Snectar [Bodo Moeller] 95100928Snectar 96100928Snectar *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 97100928Snectar to prevent ssl3_read_internal() from incorrectly assuming that 98100928Snectar ssl3_read_bytes() found application data while handshake 99100928Snectar processing was enabled when in fact s->s3->in_read_app_data was 100100928Snectar merely automatically cleared during the initial handshake. 101100928Snectar [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 102100928Snectar 103100928Snectar *) Fix object definitions for Private and Enterprise: they were not 104100928Snectar recognized in their shortname (=lowercase) representation. Extend 105100928Snectar obj_dat.pl to issue an error when using undefined keywords instead 106100928Snectar of silently ignoring the problem (Svenning Sorensen 107100928Snectar <sss@sss.dnsalias.net>). 108100928Snectar [Lutz Jaenicke] 109100928Snectar 110100928Snectar *) Fix DH_generate_parameters() so that it works for 'non-standard' 111100928Snectar generators, i.e. generators other than 2 and 5. (Previously, the 112100928Snectar code did not properly initialise the 'add' and 'rem' values to 113100928Snectar BN_generate_prime().) 114100928Snectar 115100928Snectar In the new general case, we do not insist that 'generator' is 116100928Snectar actually a primitive root: This requirement is rather pointless; 117100928Snectar a generator of the order-q subgroup is just as good, if not 118100928Snectar better. 119100928Snectar [Bodo Moeller] 120100928Snectar 121100928Snectar *) Map new X509 verification errors to alerts. Discovered and submitted by 122100928Snectar Tom Wu <tom@arcot.com>. 123100928Snectar [Lutz Jaenicke] 124100928Snectar 125100928Snectar *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 126100928Snectar returning non-zero before the data has been completely received 127100928Snectar when using non-blocking I/O. 128100928Snectar [Bodo Moeller; problem pointed out by John Hughes] 129100928Snectar 130100928Snectar *) Some of the ciphers missed the strength entry (SSL_LOW etc). 131100928Snectar [Ben Laurie, Lutz Jaenicke] 132100928Snectar 133100928Snectar *) Fix bug in SSL_clear(): bad sessions were not removed (found by 134100928Snectar Yoram Zahavi <YoramZ@gilian.com>). 135100928Snectar [Lutz Jaenicke] 136100928Snectar 137100928Snectar *) Add information about CygWin 1.3 and on, and preserve proper 138100928Snectar configuration for the versions before that. 139100928Snectar [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 140100928Snectar 141100928Snectar *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 142100928Snectar check whether we deal with a copy of a session and do not delete from 143100928Snectar the cache in this case. Problem reported by "Izhar Shoshani Levi" 144100928Snectar <izhar@checkpoint.com>. 145100928Snectar [Lutz Jaenicke] 146100928Snectar 147100928Snectar *) Do not store session data into the internal session cache, if it 148100928Snectar is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 149100928Snectar flag is set). Proposed by Aslam <aslam@funk.com>. 150100928Snectar [Lutz Jaenicke] 151100928Snectar 152100928Snectar *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 153100928Snectar value is 0. 154100928Snectar [Richard Levitte] 155100928Snectar 156100928Snectar *) [In 0.9.6c-engine release:] 157100928Snectar Fix a crashbug and a logic bug in hwcrhk_load_pubkey() 158100928Snectar [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 159100928Snectar 160100928Snectar *) Add the configuration target linux-s390x. 161100928Snectar [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 162100928Snectar 163100928Snectar *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 164100928Snectar ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 165100928Snectar variable as an indication that a ClientHello message has been 166100928Snectar received. As the flag value will be lost between multiple 167100928Snectar invocations of ssl3_accept when using non-blocking I/O, the 168100928Snectar function may not be aware that a handshake has actually taken 169100928Snectar place, thus preventing a new session from being added to the 170100928Snectar session cache. 171100928Snectar 172100928Snectar To avoid this problem, we now set s->new_session to 2 instead of 173100928Snectar using a local variable. 174100928Snectar [Lutz Jaenicke, Bodo Moeller] 175100928Snectar 176100928Snectar *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 177100928Snectar if the SSL_R_LENGTH_MISMATCH error is detected. 178100928Snectar [Geoff Thorpe, Bodo Moeller] 179100928Snectar 180100928Snectar *) New 'shared_ldflag' column in Configure platform table. 181100928Snectar [Richard Levitte] 182100928Snectar 183100928Snectar *) Fix EVP_CIPHER_mode macro. 184100928Snectar ["Dan S. Camper" <dan@bti.net>] 185100928Snectar 186100928Snectar *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 187100928Snectar type, we must throw them away by setting rr->length to 0. 188100928Snectar [D P Chang <dpc@qualys.com>] 189100928Snectar 19089837Skris Changes between 0.9.6b and 0.9.6c [21 dec 2001] 19189837Skris 19289837Skris *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 19389837Skris <Dominikus.Scherkl@biodata.com>. (The previous implementation 19489837Skris worked incorrectly for those cases where range = 10..._2 and 19589837Skris 3*range is two bits longer than range.) 19689837Skris [Bodo Moeller] 19789837Skris 19889837Skris *) Only add signing time to PKCS7 structures if it is not already 19989837Skris present. 20089837Skris [Steve Henson] 20189837Skris 20289837Skris *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 20389837Skris OBJ_ld_ce should be OBJ_id_ce. 20489837Skris Also some ip-pda OIDs in crypto/objects/objects.txt were 20589837Skris incorrect (cf. RFC 3039). 20689837Skris [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 20789837Skris 20889837Skris *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 20989837Skris returns early because it has nothing to do. 21089837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 21189837Skris 21289837Skris *) [In 0.9.6c-engine release:] 21389837Skris Fix mutex callback return values in crypto/engine/hw_ncipher.c. 21489837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 21589837Skris 21689837Skris *) [In 0.9.6c-engine release:] 21789837Skris Add support for Cryptographic Appliance's keyserver technology. 21889837Skris (Use engine 'keyclient') 21989837Skris [Cryptographic Appliances and Geoff Thorpe] 22089837Skris 22189837Skris *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 22289837Skris is called via tools/c89.sh because arguments have to be 22389837Skris rearranged (all '-L' options must appear before the first object 22489837Skris modules). 22589837Skris [Richard Shapiro <rshapiro@abinitio.com>] 22689837Skris 22789837Skris *) [In 0.9.6c-engine release:] 22889837Skris Add support for Broadcom crypto accelerator cards, backported 22989837Skris from 0.9.7. 23089837Skris [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 23189837Skris 23289837Skris *) [In 0.9.6c-engine release:] 23389837Skris Add support for SureWare crypto accelerator cards from 23489837Skris Baltimore Technologies. (Use engine 'sureware') 23589837Skris [Baltimore Technologies and Mark Cox] 23689837Skris 23789837Skris *) [In 0.9.6c-engine release:] 23889837Skris Add support for crypto accelerator cards from Accelerated 23989837Skris Encryption Processing, www.aep.ie. (Use engine 'aep') 24089837Skris [AEP Inc. and Mark Cox] 24189837Skris 24289837Skris *) Add a configuration entry for gcc on UnixWare. 24389837Skris [Gary Benson <gbenson@redhat.com>] 24489837Skris 24589837Skris *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 24689837Skris messages are stored in a single piece (fixed-length part and 24789837Skris variable-length part combined) and fix various bugs found on the way. 24889837Skris [Bodo Moeller] 24989837Skris 25089837Skris *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 25189837Skris instead. BIO_gethostbyname() does not know what timeouts are 25289837Skris appropriate, so entries would stay in cache even when they have 25389837Skris become invalid. 25489837Skris [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 25589837Skris 25689837Skris *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 25789837Skris faced with a pathologically small ClientHello fragment that does 25889837Skris not contain client_version: Instead of aborting with an error, 25989837Skris simply choose the highest available protocol version (i.e., 26089837Skris TLS 1.0 unless it is disabled). In practice, ClientHello 26189837Skris messages are never sent like this, but this change gives us 26289837Skris strictly correct behaviour at least for TLS. 26389837Skris [Bodo Moeller] 26489837Skris 26589837Skris *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 26689837Skris never resets s->method to s->ctx->method when called from within 26789837Skris one of the SSL handshake functions. 26889837Skris [Bodo Moeller; problem pointed out by Niko Baric] 26989837Skris 27089837Skris *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 27189837Skris (sent using the client's version number) if client_version is 27289837Skris smaller than the protocol version in use. Also change 27389837Skris ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 27489837Skris the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 27589837Skris the client will at least see that alert. 27689837Skris [Bodo Moeller] 27789837Skris 27889837Skris *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 27989837Skris correctly. 28089837Skris [Bodo Moeller] 28189837Skris 28289837Skris *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 28389837Skris client receives HelloRequest while in a handshake. 28489837Skris [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 28589837Skris 28689837Skris *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 28789837Skris should end in 'break', not 'goto end' which circuments various 28889837Skris cleanups done in state SSL_ST_OK. But session related stuff 28989837Skris must be disabled for SSL_ST_OK in the case that we just sent a 29089837Skris HelloRequest. 29189837Skris 29289837Skris Also avoid some overhead by not calling ssl_init_wbio_buffer() 29389837Skris before just sending a HelloRequest. 29489837Skris [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 29589837Skris 29689837Skris *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 29789837Skris reveal whether illegal block cipher padding was found or a MAC 29889837Skris verification error occured. (Neither SSLerr() codes nor alerts 29989837Skris are directly visible to potential attackers, but the information 30089837Skris may leak via logfiles.) 30189837Skris 30289837Skris Similar changes are not required for the SSL 2.0 implementation 30389837Skris because the number of padding bytes is sent in clear for SSL 2.0, 30489837Skris and the extra bytes are just ignored. However ssl/s2_pkt.c 30589837Skris failed to verify that the purported number of padding bytes is in 30689837Skris the legal range. 30789837Skris [Bodo Moeller] 30889837Skris 30989837Skris *) Add OpenUNIX-8 support including shared libraries 31089837Skris (Boyd Lynn Gerber <gerberb@zenez.com>). 31189837Skris [Lutz Jaenicke] 31289837Skris 31389837Skris *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 31489837Skris 'wristwatch attack' using huge encoding parameters (cf. 31589837Skris James H. Manger's CRYPTO 2001 paper). Note that the 31689837Skris RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 31789837Skris encoding parameters and hence was not vulnerable. 31889837Skris [Bodo Moeller] 31989837Skris 32089837Skris *) BN_sqr() bug fix. 32189837Skris [Ulf M�ller, reported by Jim Ellis <jim.ellis@cavium.com>] 32289837Skris 32389837Skris *) Rabin-Miller test analyses assume uniformly distributed witnesses, 32489837Skris so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 32589837Skris followed by modular reduction. 32689837Skris [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 32789837Skris 32889837Skris *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 32989837Skris equivalent based on BN_pseudo_rand() instead of BN_rand(). 33089837Skris [Bodo Moeller] 33189837Skris 33289837Skris *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 33389837Skris This function was broken, as the check for a new client hello message 33489837Skris to handle SGC did not allow these large messages. 33589837Skris (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 33689837Skris [Lutz Jaenicke] 33789837Skris 33889837Skris *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 33989837Skris [Lutz Jaenicke] 34089837Skris 34189837Skris *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 34289837Skris for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 34389837Skris [Lutz Jaenicke] 34489837Skris 34589837Skris *) Rework the configuration and shared library support for Tru64 Unix. 34689837Skris The configuration part makes use of modern compiler features and 34789837Skris still retains old compiler behavior for those that run older versions 34889837Skris of the OS. The shared library support part includes a variant that 34989837Skris uses the RPATH feature, and is available through the special 35089837Skris configuration target "alpha-cc-rpath", which will never be selected 35189837Skris automatically. 35289837Skris [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 35389837Skris 35489837Skris *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 35589837Skris with the same message size as in ssl3_get_certificate_request(). 35689837Skris Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 35789837Skris messages might inadvertently be reject as too long. 35889837Skris [Petr Lampa <lampa@fee.vutbr.cz>] 35989837Skris 36089837Skris *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 36189837Skris [Andy Polyakov] 36289837Skris 36389837Skris *) Modified SSL library such that the verify_callback that has been set 36489837Skris specificly for an SSL object with SSL_set_verify() is actually being 36589837Skris used. Before the change, a verify_callback set with this function was 36689837Skris ignored and the verify_callback() set in the SSL_CTX at the time of 36789837Skris the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 36889837Skris to allow the necessary settings. 36989837Skris [Lutz Jaenicke] 37089837Skris 37189837Skris *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 37289837Skris explicitly to NULL, as at least on Solaris 8 this seems not always to be 37389837Skris done automatically (in contradiction to the requirements of the C 37489837Skris standard). This made problems when used from OpenSSH. 37589837Skris [Lutz Jaenicke] 37689837Skris 37789837Skris *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 37889837Skris dh->length and always used 37989837Skris 38089837Skris BN_rand_range(priv_key, dh->p). 38189837Skris 38289837Skris BN_rand_range() is not necessary for Diffie-Hellman, and this 38389837Skris specific range makes Diffie-Hellman unnecessarily inefficient if 38489837Skris dh->length (recommended exponent length) is much smaller than the 38589837Skris length of dh->p. We could use BN_rand_range() if the order of 38689837Skris the subgroup was stored in the DH structure, but we only have 38789837Skris dh->length. 38889837Skris 38989837Skris So switch back to 39089837Skris 39189837Skris BN_rand(priv_key, l, ...) 39289837Skris 39389837Skris where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 39489837Skris otherwise. 39589837Skris [Bodo Moeller] 39689837Skris 39789837Skris *) In 39889837Skris 39989837Skris RSA_eay_public_encrypt 40089837Skris RSA_eay_private_decrypt 40189837Skris RSA_eay_private_encrypt (signing) 40289837Skris RSA_eay_public_decrypt (signature verification) 40389837Skris 40489837Skris (default implementations for RSA_public_encrypt, 40589837Skris RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 40689837Skris always reject numbers >= n. 40789837Skris [Bodo Moeller] 40889837Skris 40989837Skris *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 41089837Skris to synchronize access to 'locking_thread'. This is necessary on 41189837Skris systems where access to 'locking_thread' (an 'unsigned long' 41289837Skris variable) is not atomic. 41389837Skris [Bodo Moeller] 41489837Skris 41589837Skris *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 41689837Skris *before* setting the 'crypto_lock_rand' flag. The previous code had 41789837Skris a race condition if 0 is a valid thread ID. 41889837Skris [Travis Vitek <vitek@roguewave.com>] 41989837Skris 42089837Skris *) Add support for shared libraries under Irix. 42189837Skris [Albert Chin-A-Young <china@thewrittenword.com>] 42289837Skris 42389837Skris *) Add configuration option to build on Linux on both big-endian and 42489837Skris little-endian MIPS. 42589837Skris [Ralf Baechle <ralf@uni-koblenz.de>] 42689837Skris 42789837Skris *) Add the possibility to create shared libraries on HP-UX. 42889837Skris [Richard Levitte] 42989837Skris 43079998Skris Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 43179998Skris 43279998Skris *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 43379998Skris to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 43479998Skris Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 43579998Skris PRNG state recovery was possible based on the output of 43679998Skris one PRNG request appropriately sized to gain knowledge on 43779998Skris 'md' followed by enough consecutive 1-byte PRNG requests 43879998Skris to traverse all of 'state'. 43979998Skris 44079998Skris 1. When updating 'md_local' (the current thread's copy of 'md') 44179998Skris during PRNG output generation, hash all of the previous 44279998Skris 'md_local' value, not just the half used for PRNG output. 44379998Skris 44479998Skris 2. Make the number of bytes from 'state' included into the hash 44579998Skris independent from the number of PRNG bytes requested. 44679998Skris 44779998Skris The first measure alone would be sufficient to avoid 44879998Skris Markku-Juhani's attack. (Actually it had never occurred 44979998Skris to me that the half of 'md_local' used for chaining was the 45079998Skris half from which PRNG output bytes were taken -- I had always 45179998Skris assumed that the secret half would be used.) The second 45279998Skris measure makes sure that additional data from 'state' is never 45379998Skris mixed into 'md_local' in small portions; this heuristically 45479998Skris further strengthens the PRNG. 45579998Skris [Bodo Moeller] 45679998Skris 45779998Skris *) Fix crypto/bn/asm/mips3.s. 45879998Skris [Andy Polyakov] 45979998Skris 46079998Skris *) When only the key is given to "enc", the IV is undefined. Print out 46179998Skris an error message in this case. 46279998Skris [Lutz Jaenicke] 46379998Skris 46479998Skris *) Handle special case when X509_NAME is empty in X509 printing routines. 46579998Skris [Steve Henson] 46679998Skris 46779998Skris *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 46879998Skris positive and less than q. 46979998Skris [Bodo Moeller] 47079998Skris 47179998Skris *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 47279998Skris used: it isn't thread safe and the add_lock_callback should handle 47379998Skris that itself. 47479998Skris [Paul Rose <Paul.Rose@bridge.com>] 47579998Skris 47679998Skris *) Verify that incoming data obeys the block size in 47779998Skris ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 47879998Skris [Bodo Moeller] 47979998Skris 48079998Skris *) Fix OAEP check. 48179998Skris [Ulf M�ller, Bodo M�ller] 48279998Skris 48379998Skris *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 48489837Skris RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 48579998Skris when fixing the server behaviour for backwards-compatible 'client 48679998Skris hello' messages. (Note that the attack is impractical against 48779998Skris SSL 3.0 and TLS 1.0 anyway because length and version checking 48879998Skris means that the probability of guessing a valid ciphertext is 48979998Skris around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 49079998Skris paper.) 49179998Skris 49279998Skris Before 0.9.5, the countermeasure (hide the error by generating a 49379998Skris random 'decryption result') did not work properly because 49479998Skris ERR_clear_error() was missing, meaning that SSL_get_error() would 49579998Skris detect the supposedly ignored error. 49679998Skris 49779998Skris Both problems are now fixed. 49879998Skris [Bodo Moeller] 49979998Skris 50079998Skris *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 50179998Skris (previously it was 1024). 50279998Skris [Bodo Moeller] 50379998Skris 50479998Skris *) Fix for compatibility mode trust settings: ignore trust settings 50579998Skris unless some valid trust or reject settings are present. 50679998Skris [Steve Henson] 50779998Skris 50879998Skris *) Fix for blowfish EVP: its a variable length cipher. 50979998Skris [Steve Henson] 51079998Skris 51179998Skris *) Fix various bugs related to DSA S/MIME verification. Handle missing 51279998Skris parameters in DSA public key structures and return an error in the 51379998Skris DSA routines if parameters are absent. 51479998Skris [Steve Henson] 51579998Skris 51679998Skris *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 51779998Skris in the current directory if neither $RANDFILE nor $HOME was set. 51879998Skris RAND_file_name() in 0.9.6a returned NULL in this case. This has 51979998Skris caused some confusion to Windows users who haven't defined $HOME. 52079998Skris Thus RAND_file_name() is changed again: e_os.h can define a 52179998Skris DEFAULT_HOME, which will be used if $HOME is not set. 52279998Skris For Windows, we use "C:"; on other platforms, we still require 52379998Skris environment variables. 52479998Skris 52579998Skris *) Move 'if (!initialized) RAND_poll()' into regions protected by 52679998Skris CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 52779998Skris having multiple threads call RAND_poll() concurrently. 52879998Skris [Bodo Moeller] 52979998Skris 53079998Skris *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 53179998Skris combination of a flag and a thread ID variable. 53279998Skris Otherwise while one thread is in ssleay_rand_bytes (which sets the 53379998Skris flag), *other* threads can enter ssleay_add_bytes without obeying 53489837Skris the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 53579998Skris that they do not hold after the first thread unsets add_do_not_lock). 53679998Skris [Bodo Moeller] 53779998Skris 53879998Skris *) Change bctest again: '-x' expressions are not available in all 53979998Skris versions of 'test'. 54079998Skris [Bodo Moeller] 54179998Skris 54276866Skris Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 54372613Skris 54476866Skris *) Fix a couple of memory leaks in PKCS7_dataDecode() 54576866Skris [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 54676866Skris 54776866Skris *) Change Configure and Makefiles to provide EXE_EXT, which will contain 54876866Skris the default extension for executables, if any. Also, make the perl 54976866Skris scripts that use symlink() to test if it really exists and use "cp" 55076866Skris if it doesn't. All this made OpenSSL compilable and installable in 55176866Skris CygWin. 55276866Skris [Richard Levitte] 55376866Skris 55476866Skris *) Fix for asn1_GetSequence() for indefinite length constructed data. 55576866Skris If SEQUENCE is length is indefinite just set c->slen to the total 55676866Skris amount of data available. 55776866Skris [Steve Henson, reported by shige@FreeBSD.org] 55876866Skris [This change does not apply to 0.9.7.] 55976866Skris 56076866Skris *) Change bctest to avoid here-documents inside command substitution 56176866Skris (workaround for FreeBSD /bin/sh bug). 56276866Skris For compatibility with Ultrix, avoid shell functions (introduced 56376866Skris in the bctest version that searches along $PATH). 56476866Skris [Bodo Moeller] 56576866Skris 56676866Skris *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 56776866Skris with des_encrypt() defined on some operating systems, like Solaris 56876866Skris and UnixWare. 56976866Skris [Richard Levitte] 57076866Skris 57176866Skris *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 57276866Skris On the Importance of Eliminating Errors in Cryptographic 57376866Skris Computations, J. Cryptology 14 (2001) 2, 101-119, 57476866Skris http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 57576866Skris [Ulf Moeller] 57676866Skris 57776866Skris *) MIPS assembler BIGNUM division bug fix. 57876866Skris [Andy Polyakov] 57976866Skris 58076866Skris *) Disabled incorrect Alpha assembler code. 58176866Skris [Richard Levitte] 58276866Skris 58376866Skris *) Fix PKCS#7 decode routines so they correctly update the length 58476866Skris after reading an EOC for the EXPLICIT tag. 58576866Skris [Steve Henson] 58676866Skris [This change does not apply to 0.9.7.] 58776866Skris 58876866Skris *) Fix bug in PKCS#12 key generation routines. This was triggered 58976866Skris if a 3DES key was generated with a 0 initial byte. Include 59076866Skris PKCS12_BROKEN_KEYGEN compilation option to retain the old 59176866Skris (but broken) behaviour. 59276866Skris [Steve Henson] 59376866Skris 59476866Skris *) Enhance bctest to search for a working bc along $PATH and print 59576866Skris it when found. 59676866Skris [Tim Rice <tim@multitalents.net> via Richard Levitte] 59776866Skris 59876866Skris *) Fix memory leaks in err.c: free err_data string if necessary; 59976866Skris don't write to the wrong index in ERR_set_error_data. 60076866Skris [Bodo Moeller] 60176866Skris 60276866Skris *) Implement ssl23_peek (analogous to ssl23_read), which previously 60376866Skris did not exist. 60476866Skris [Bodo Moeller] 60576866Skris 60676866Skris *) Replace rdtsc with _emit statements for VC++ version 5. 60776866Skris [Jeremy Cooper <jeremy@baymoo.org>] 60876866Skris 60976866Skris *) Make it possible to reuse SSLv2 sessions. 61076866Skris [Richard Levitte] 61176866Skris 61276866Skris *) In copy_email() check for >= 0 as a return value for 61376866Skris X509_NAME_get_index_by_NID() since 0 is a valid index. 61476866Skris [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 61576866Skris 61676866Skris *) Avoid coredump with unsupported or invalid public keys by checking if 61776866Skris X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 61876866Skris PKCS7_verify() fails with non detached data. 61976866Skris [Steve Henson] 62076866Skris 62176866Skris *) Don't use getenv in library functions when run as setuid/setgid. 62276866Skris New function OPENSSL_issetugid(). 62376866Skris [Ulf Moeller] 62476866Skris 62576866Skris *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 62676866Skris due to incorrect handling of multi-threading: 62776866Skris 62876866Skris 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 62976866Skris 63076866Skris 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 63176866Skris 63276866Skris 3. Count how many times MemCheck_off() has been called so that 63376866Skris nested use can be treated correctly. This also avoids 63476866Skris inband-signalling in the previous code (which relied on the 63576866Skris assumption that thread ID 0 is impossible). 63676866Skris [Bodo Moeller] 63776866Skris 63876866Skris *) Add "-rand" option also to s_client and s_server. 63976866Skris [Lutz Jaenicke] 64076866Skris 64176866Skris *) Fix CPU detection on Irix 6.x. 64276866Skris [Kurt Hockenbury <khockenb@stevens-tech.edu> and 64376866Skris "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 64476866Skris 64576866Skris *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 64676866Skris was empty. 64776866Skris [Steve Henson] 64876866Skris [This change does not apply to 0.9.7.] 64976866Skris 65076866Skris *) Use the cached encoding of an X509_NAME structure rather than 65176866Skris copying it. This is apparently the reason for the libsafe "errors" 65276866Skris but the code is actually correct. 65376866Skris [Steve Henson] 65476866Skris 65572613Skris *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 65672613Skris Bleichenbacher's DSA attack. 65776866Skris Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 65876866Skris to be set and top=0 forces the highest bit to be set; top=-1 is new 65976866Skris and leaves the highest bit random. 66076866Skris [Ulf Moeller, Bodo Moeller] 66172613Skris 66272613Skris *) In the NCONF_...-based implementations for CONF_... queries 66372613Skris (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 66472613Skris a temporary CONF structure with the data component set to NULL 66572613Skris (which gives segmentation faults in lh_retrieve). 66672613Skris Instead, use NULL for the CONF pointer in CONF_get_string and 66772613Skris CONF_get_number (which may use environment variables) and directly 66872613Skris return NULL from CONF_get_section. 66972613Skris [Bodo Moeller] 67072613Skris 67172613Skris *) Fix potential buffer overrun for EBCDIC. 67272613Skris [Ulf Moeller] 67372613Skris 67472613Skris *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 67572613Skris keyUsage if basicConstraints absent for a CA. 67672613Skris [Steve Henson] 67772613Skris 67872613Skris *) Make SMIME_write_PKCS7() write mail header values with a format that 67972613Skris is more generally accepted (no spaces before the semicolon), since 68072613Skris some programs can't parse those values properly otherwise. Also make 68172613Skris sure BIO's that break lines after each write do not create invalid 68272613Skris headers. 68372613Skris [Richard Levitte] 68472613Skris 68572613Skris *) Make the CRL encoding routines work with empty SEQUENCE OF. The 68672613Skris macros previously used would not encode an empty SEQUENCE OF 68772613Skris and break the signature. 68872613Skris [Steve Henson] 68976866Skris [This change does not apply to 0.9.7.] 69072613Skris 69172613Skris *) Zero the premaster secret after deriving the master secret in 69272613Skris DH ciphersuites. 69372613Skris [Steve Henson] 69472613Skris 69572613Skris *) Add some EVP_add_digest_alias registrations (as found in 69672613Skris OpenSSL_add_all_digests()) to SSL_library_init() 69772613Skris aka OpenSSL_add_ssl_algorithms(). This provides improved 69872613Skris compatibility with peers using X.509 certificates 69972613Skris with unconventional AlgorithmIdentifier OIDs. 70072613Skris [Bodo Moeller] 70172613Skris 70272613Skris *) Fix for Irix with NO_ASM. 70372613Skris ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 70472613Skris 70572613Skris *) ./config script fixes. 70672613Skris [Ulf Moeller, Richard Levitte] 70772613Skris 70872613Skris *) Fix 'openssl passwd -1'. 70972613Skris [Bodo Moeller] 71072613Skris 71172613Skris *) Change PKCS12_key_gen_asc() so it can cope with non null 71272613Skris terminated strings whose length is passed in the passlen 71372613Skris parameter, for example from PEM callbacks. This was done 71472613Skris by adding an extra length parameter to asc2uni(). 71572613Skris [Steve Henson, reported by <oddissey@samsung.co.kr>] 71672613Skris 71772613Skris *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 71872613Skris call failed, free the DSA structure. 71972613Skris [Bodo Moeller] 72072613Skris 72172613Skris *) Fix to uni2asc() to cope with zero length Unicode strings. 72272613Skris These are present in some PKCS#12 files. 72372613Skris [Steve Henson] 72472613Skris 72572613Skris *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 72672613Skris Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 72772613Skris when writing a 32767 byte record. 72872613Skris [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 72972613Skris 73072613Skris *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 73172613Skris obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 73272613Skris 73372613Skris (RSA objects have a reference count access to which is protected 73472613Skris by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 73572613Skris so they are meant to be shared between threads.) 73672613Skris [Bodo Moeller, Geoff Thorpe; original patch submitted by 73772613Skris "Reddie, Steven" <Steven.Reddie@ca.com>] 73872613Skris 73972613Skris *) Fix a deadlock in CRYPTO_mem_leaks(). 74072613Skris [Bodo Moeller] 74172613Skris 74276866Skris *) Use better test patterns in bntest. 74376866Skris [Ulf M�ller] 74476866Skris 74572613Skris *) rand_win.c fix for Borland C. 74672613Skris [Ulf M�ller] 74772613Skris 74872613Skris *) BN_rshift bugfix for n == 0. 74972613Skris [Bodo Moeller] 75072613Skris 75176866Skris *) Add a 'bctest' script that checks for some known 'bc' bugs 75276866Skris so that 'make test' does not abort just because 'bc' is broken. 75376866Skris [Bodo Moeller] 75476866Skris 75572613Skris *) Store verify_result within SSL_SESSION also for client side to 75672613Skris avoid potential security hole. (Re-used sessions on the client side 75772613Skris always resulted in verify_result==X509_V_OK, not using the original 75872613Skris result of the server certificate verification.) 75972613Skris [Lutz Jaenicke] 76072613Skris 76172613Skris *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 76272613Skris SSL3_RT_APPLICATION_DATA, return 0. 76372613Skris Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 76472613Skris [Bodo Moeller] 76572613Skris 76672613Skris *) Fix SSL_peek: 76772613Skris Both ssl2_peek and ssl3_peek, which were totally broken in earlier 76872613Skris releases, have been re-implemented by renaming the previous 76972613Skris implementations of ssl2_read and ssl3_read to ssl2_read_internal 77072613Skris and ssl3_read_internal, respectively, and adding 'peek' parameters 77172613Skris to them. The new ssl[23]_{read,peek} functions are calls to 77272613Skris ssl[23]_read_internal with the 'peek' flag set appropriately. 77372613Skris A 'peek' parameter has also been added to ssl3_read_bytes, which 77472613Skris does the actual work for ssl3_read_internal. 77572613Skris [Bodo Moeller] 77672613Skris 77776866Skris *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 77876866Skris the method-specific "init()" handler. Also clean up ex_data after 77976866Skris calling the method-specific "finish()" handler. Previously, this was 78076866Skris happening the other way round. 78176866Skris [Geoff Thorpe] 78276866Skris 78372613Skris *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 78472613Skris The previous value, 12, was not always sufficient for BN_mod_exp(). 78572613Skris [Bodo Moeller] 78672613Skris 78776866Skris *) Make sure that shared libraries get the internal name engine with 78876866Skris the full version number and not just 0. This should mark the 78976866Skris shared libraries as not backward compatible. Of course, this should 79076866Skris be changed again when we can guarantee backward binary compatibility. 79176866Skris [Richard Levitte] 79276866Skris 79372613Skris *) Fix typo in get_cert_by_subject() in by_dir.c 79472613Skris [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 79572613Skris 79676866Skris *) Rework the system to generate shared libraries: 79776866Skris 79876866Skris - Make note of the expected extension for the shared libraries and 79976866Skris if there is a need for symbolic links from for example libcrypto.so.0 80076866Skris to libcrypto.so.0.9.7. There is extended info in Configure for 80176866Skris that. 80276866Skris 80376866Skris - Make as few rebuilds of the shared libraries as possible. 80476866Skris 80576866Skris - Still avoid linking the OpenSSL programs with the shared libraries. 80676866Skris 80776866Skris - When installing, install the shared libraries separately from the 80876866Skris static ones. 80976866Skris [Richard Levitte] 81076866Skris 81172613Skris *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 81272613Skris 81372613Skris Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 81472613Skris and not in SSL_clear because the latter is also used by the 81572613Skris accept/connect functions; previously, the settings made by 81672613Skris SSL_set_read_ahead would be lost during the handshake. 81772613Skris [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 81872613Skris 81972613Skris *) Correct util/mkdef.pl to be selective about disabled algorithms. 82072613Skris Previously, it would create entries for disableed algorithms no 82172613Skris matter what. 82272613Skris [Richard Levitte] 82372613Skris 82472613Skris *) Added several new manual pages for SSL_* function. 82572613Skris [Lutz Jaenicke] 82672613Skris 82768651Skris Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 82868651Skris 82968651Skris *) In ssl23_get_client_hello, generate an error message when faced 83068651Skris with an initial SSL 3.0/TLS record that is too small to contain the 83168651Skris first two bytes of the ClientHello message, i.e. client_version. 83268651Skris (Note that this is a pathologic case that probably has never happened 83368651Skris in real life.) The previous approach was to use the version number 83468651Skris from the record header as a substitute; but our protocol choice 83568651Skris should not depend on that one because it is not authenticated 83668651Skris by the Finished messages. 83768651Skris [Bodo Moeller] 83868651Skris 83968651Skris *) More robust randomness gathering functions for Windows. 84068651Skris [Jeffrey Altman <jaltman@columbia.edu>] 84168651Skris 84268651Skris *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 84368651Skris not set then we don't setup the error code for issuer check errors 84468651Skris to avoid possibly overwriting other errors which the callback does 84568651Skris handle. If an application does set the flag then we assume it knows 84668651Skris what it is doing and can handle the new informational codes 84768651Skris appropriately. 84868651Skris [Steve Henson] 84968651Skris 85068651Skris *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 85168651Skris a general "ANY" type, as such it should be able to decode anything 85268651Skris including tagged types. However it didn't check the class so it would 85368651Skris wrongly interpret tagged types in the same way as their universal 85468651Skris counterpart and unknown types were just rejected. Changed so that the 85568651Skris tagged and unknown types are handled in the same way as a SEQUENCE: 85668651Skris that is the encoding is stored intact. There is also a new type 85768651Skris "V_ASN1_OTHER" which is used when the class is not universal, in this 85868651Skris case we have no idea what the actual type is so we just lump them all 85968651Skris together. 86068651Skris [Steve Henson] 86168651Skris 86268651Skris *) On VMS, stdout may very well lead to a file that is written to 86368651Skris in a record-oriented fashion. That means that every write() will 86468651Skris write a separate record, which will be read separately by the 86568651Skris programs trying to read from it. This can be very confusing. 86668651Skris 86768651Skris The solution is to put a BIO filter in the way that will buffer 86868651Skris text until a linefeed is reached, and then write everything a 86968651Skris line at a time, so every record written will be an actual line, 87068651Skris not chunks of lines and not (usually doesn't happen, but I've 87168651Skris seen it once) several lines in one record. BIO_f_linebuffer() is 87268651Skris the answer. 87368651Skris 87468651Skris Currently, it's a VMS-only method, because that's where it has 87568651Skris been tested well enough. 87668651Skris [Richard Levitte] 87768651Skris 87868651Skris *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 87968651Skris it can return incorrect results. 88068651Skris (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 88168651Skris but it was in 0.9.6-beta[12].) 88268651Skris [Bodo Moeller] 88368651Skris 88468651Skris *) Disable the check for content being present when verifying detached 88568651Skris signatures in pk7_smime.c. Some versions of Netscape (wrongly) 88668651Skris include zero length content when signing messages. 88768651Skris [Steve Henson] 88868651Skris 88968651Skris *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 89068651Skris BIO_ctrl (for BIO pairs). 89168651Skris [Bodo M�ller] 89268651Skris 89368651Skris *) Add DSO method for VMS. 89468651Skris [Richard Levitte] 89568651Skris 89668651Skris *) Bug fix: Montgomery multiplication could produce results with the 89768651Skris wrong sign. 89868651Skris [Ulf M�ller] 89968651Skris 90068651Skris *) Add RPM specification openssl.spec and modify it to build three 90168651Skris packages. The default package contains applications, application 90268651Skris documentation and run-time libraries. The devel package contains 90368651Skris include files, static libraries and function documentation. The 90468651Skris doc package contains the contents of the doc directory. The original 90568651Skris openssl.spec was provided by Damien Miller <djm@mindrot.org>. 90668651Skris [Richard Levitte] 90768651Skris 90868651Skris *) Add a large number of documentation files for many SSL routines. 90968651Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 91068651Skris 91168651Skris *) Add a configuration entry for Sony News 4. 91268651Skris [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 91368651Skris 91468651Skris *) Don't set the two most significant bits to one when generating a 91568651Skris random number < q in the DSA library. 91668651Skris [Ulf M�ller] 91768651Skris 91868651Skris *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 91968651Skris behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 92068651Skris the underlying transport is blocking) if a handshake took place. 92168651Skris (The default behaviour is needed by applications such as s_client 92268651Skris and s_server that use select() to determine when to use SSL_read; 92368651Skris but for applications that know in advance when to expect data, it 92468651Skris just makes things more complicated.) 92568651Skris [Bodo Moeller] 92668651Skris 92768651Skris *) Add RAND_egd_bytes(), which gives control over the number of bytes read 92868651Skris from EGD. 92968651Skris [Ben Laurie] 93068651Skris 93168651Skris *) Add a few more EBCDIC conditionals that make `req' and `x509' 93268651Skris work better on such systems. 93368651Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 93468651Skris 93568651Skris *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 93668651Skris Update PKCS12_parse() so it copies the friendlyName and the 93768651Skris keyid to the certificates aux info. 93868651Skris [Steve Henson] 93968651Skris 94068651Skris *) Fix bug in PKCS7_verify() which caused an infinite loop 94168651Skris if there was more than one signature. 94268651Skris [Sven Uszpelkat <su@celocom.de>] 94368651Skris 94468651Skris *) Major change in util/mkdef.pl to include extra information 94568651Skris about each symbol, as well as presentig variables as well 94668651Skris as functions. This change means that there's n more need 94768651Skris to rebuild the .num files when some algorithms are excluded. 94868651Skris [Richard Levitte] 94968651Skris 95068651Skris *) Allow the verify time to be set by an application, 95168651Skris rather than always using the current time. 95268651Skris [Steve Henson] 95368651Skris 95468651Skris *) Phase 2 verify code reorganisation. The certificate 95568651Skris verify code now looks up an issuer certificate by a 95668651Skris number of criteria: subject name, authority key id 95768651Skris and key usage. It also verifies self signed certificates 95868651Skris by the same criteria. The main comparison function is 95968651Skris X509_check_issued() which performs these checks. 96068651Skris 96168651Skris Lot of changes were necessary in order to support this 96268651Skris without completely rewriting the lookup code. 96368651Skris 96468651Skris Authority and subject key identifier are now cached. 96568651Skris 96668651Skris The LHASH 'certs' is X509_STORE has now been replaced 96768651Skris by a STACK_OF(X509_OBJECT). This is mainly because an 96868651Skris LHASH can't store or retrieve multiple objects with 96968651Skris the same hash value. 97068651Skris 97168651Skris As a result various functions (which were all internal 97268651Skris use only) have changed to handle the new X509_STORE 97368651Skris structure. This will break anything that messed round 97468651Skris with X509_STORE internally. 97568651Skris 97668651Skris The functions X509_STORE_add_cert() now checks for an 97768651Skris exact match, rather than just subject name. 97868651Skris 97968651Skris The X509_STORE API doesn't directly support the retrieval 98068651Skris of multiple certificates matching a given criteria, however 98168651Skris this can be worked round by performing a lookup first 98268651Skris (which will fill the cache with candidate certificates) 98368651Skris and then examining the cache for matches. This is probably 98468651Skris the best we can do without throwing out X509_LOOKUP 98568651Skris entirely (maybe later...). 98668651Skris 98768651Skris The X509_VERIFY_CTX structure has been enhanced considerably. 98868651Skris 98968651Skris All certificate lookup operations now go via a get_issuer() 99068651Skris callback. Although this currently uses an X509_STORE it 99168651Skris can be replaced by custom lookups. This is a simple way 99268651Skris to bypass the X509_STORE hackery necessary to make this 99368651Skris work and makes it possible to use more efficient techniques 99468651Skris in future. A very simple version which uses a simple 99568651Skris STACK for its trusted certificate store is also provided 99668651Skris using X509_STORE_CTX_trusted_stack(). 99768651Skris 99868651Skris The verify_cb() and verify() callbacks now have equivalents 99968651Skris in the X509_STORE_CTX structure. 100068651Skris 100168651Skris X509_STORE_CTX also has a 'flags' field which can be used 100268651Skris to customise the verify behaviour. 100368651Skris [Steve Henson] 100468651Skris 100568651Skris *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 100668651Skris excludes S/MIME capabilities. 100768651Skris [Steve Henson] 100868651Skris 100968651Skris *) When a certificate request is read in keep a copy of the 101068651Skris original encoding of the signed data and use it when outputing 101168651Skris again. Signatures then use the original encoding rather than 101268651Skris a decoded, encoded version which may cause problems if the 101368651Skris request is improperly encoded. 101468651Skris [Steve Henson] 101568651Skris 101668651Skris *) For consistency with other BIO_puts implementations, call 101768651Skris buffer_write(b, ...) directly in buffer_puts instead of calling 101868651Skris BIO_write(b, ...). 101968651Skris 102068651Skris In BIO_puts, increment b->num_write as in BIO_write. 102168651Skris [Peter.Sylvester@EdelWeb.fr] 102268651Skris 102368651Skris *) Fix BN_mul_word for the case where the word is 0. (We have to use 102468651Skris BN_zero, we may not return a BIGNUM with an array consisting of 102568651Skris words set to zero.) 102668651Skris [Bodo Moeller] 102768651Skris 102868651Skris *) Avoid calling abort() from within the library when problems are 102968651Skris detected, except if preprocessor symbols have been defined 103068651Skris (such as REF_CHECK, BN_DEBUG etc.). 103168651Skris [Bodo Moeller] 103268651Skris 103368651Skris *) New openssl application 'rsautl'. This utility can be 103468651Skris used for low level RSA operations. DER public key 103568651Skris BIO/fp routines also added. 103668651Skris [Steve Henson] 103768651Skris 103868651Skris *) New Configure entry and patches for compiling on QNX 4. 103968651Skris [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 104068651Skris 104168651Skris *) A demo state-machine implementation was sponsored by 104268651Skris Nuron (http://www.nuron.com/) and is now available in 104368651Skris demos/state_machine. 104468651Skris [Ben Laurie] 104568651Skris 104668651Skris *) New options added to the 'dgst' utility for signature 104768651Skris generation and verification. 104868651Skris [Steve Henson] 104968651Skris 105068651Skris *) Unrecognized PKCS#7 content types are now handled via a 105168651Skris catch all ASN1_TYPE structure. This allows unsupported 105268651Skris types to be stored as a "blob" and an application can 105368651Skris encode and decode it manually. 105468651Skris [Steve Henson] 105568651Skris 105668651Skris *) Fix various signed/unsigned issues to make a_strex.c 105768651Skris compile under VC++. 105868651Skris [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 105968651Skris 106068651Skris *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 106168651Skris length if passed a buffer. ASN1_INTEGER_to_BN failed 106268651Skris if passed a NULL BN and its argument was negative. 106368651Skris [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 106468651Skris 106568651Skris *) Modification to PKCS#7 encoding routines to output definite 106668651Skris length encoding. Since currently the whole structures are in 106768651Skris memory there's not real point in using indefinite length 106868651Skris constructed encoding. However if OpenSSL is compiled with 106968651Skris the flag PKCS7_INDEFINITE_ENCODING the old form is used. 107068651Skris [Steve Henson] 107168651Skris 107268651Skris *) Added BIO_vprintf() and BIO_vsnprintf(). 107368651Skris [Richard Levitte] 107468651Skris 107568651Skris *) Added more prefixes to parse for in the the strings written 107668651Skris through a logging bio, to cover all the levels that are available 107768651Skris through syslog. The prefixes are now: 107868651Skris 107968651Skris PANIC, EMERG, EMR => LOG_EMERG 108068651Skris ALERT, ALR => LOG_ALERT 108168651Skris CRIT, CRI => LOG_CRIT 108268651Skris ERROR, ERR => LOG_ERR 108368651Skris WARNING, WARN, WAR => LOG_WARNING 108468651Skris NOTICE, NOTE, NOT => LOG_NOTICE 108568651Skris INFO, INF => LOG_INFO 108668651Skris DEBUG, DBG => LOG_DEBUG 108768651Skris 108868651Skris and as before, if none of those prefixes are present at the 108968651Skris beginning of the string, LOG_ERR is chosen. 109068651Skris 109168651Skris On Win32, the LOG_* levels are mapped according to this: 109268651Skris 109368651Skris LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 109468651Skris LOG_WARNING => EVENTLOG_WARNING_TYPE 109568651Skris LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 109668651Skris 109768651Skris [Richard Levitte] 109868651Skris 109968651Skris *) Made it possible to reconfigure with just the configuration 110068651Skris argument "reconf" or "reconfigure". The command line arguments 110168651Skris are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 110268651Skris and are retrieved from there when reconfiguring. 110368651Skris [Richard Levitte] 110468651Skris 110568651Skris *) MD4 implemented. 110668651Skris [Assar Westerlund <assar@sics.se>, Richard Levitte] 110768651Skris 110868651Skris *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 110968651Skris [Richard Levitte] 111068651Skris 111168651Skris *) The obj_dat.pl script was messing up the sorting of object 111268651Skris names. The reason was that it compared the quoted version 111368651Skris of strings as a result "OCSP" > "OCSP Signing" because 111468651Skris " > SPACE. Changed script to store unquoted versions of 111568651Skris names and add quotes on output. It was also omitting some 111668651Skris names from the lookup table if they were given a default 111768651Skris value (that is if SN is missing it is given the same 111868651Skris value as LN and vice versa), these are now added on the 111968651Skris grounds that if an object has a name we should be able to 112068651Skris look it up. Finally added warning output when duplicate 112168651Skris short or long names are found. 112268651Skris [Steve Henson] 112368651Skris 112468651Skris *) Changes needed for Tandem NSK. 112568651Skris [Scott Uroff <scott@xypro.com>] 112668651Skris 112768651Skris *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 112868651Skris RSA_padding_check_SSLv23(), special padding was never detected 112968651Skris and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 113068651Skris version rollback attacks was not effective. 113168651Skris 113268651Skris In s23_clnt.c, don't use special rollback-attack detection padding 113368651Skris (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 113468651Skris client; similarly, in s23_srvr.c, don't do the rollback check if 113568651Skris SSL 2.0 is the only protocol enabled in the server. 113668651Skris [Bodo Moeller] 113768651Skris 113868651Skris *) Make it possible to get hexdumps of unprintable data with 'openssl 113968651Skris asn1parse'. By implication, the functions ASN1_parse_dump() and 114068651Skris BIO_dump_indent() are added. 114168651Skris [Richard Levitte] 114268651Skris 114368651Skris *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 114468651Skris these print out strings and name structures based on various 114568651Skris flags including RFC2253 support and proper handling of 114668651Skris multibyte characters. Added options to the 'x509' utility 114768651Skris to allow the various flags to be set. 114868651Skris [Steve Henson] 114968651Skris 115068651Skris *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 115168651Skris Also change the functions X509_cmp_current_time() and 115268651Skris X509_gmtime_adj() work with an ASN1_TIME structure, 115368651Skris this will enable certificates using GeneralizedTime in validity 115468651Skris dates to be checked. 115568651Skris [Steve Henson] 115668651Skris 115768651Skris *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 115868651Skris negative public key encodings) on by default, 115968651Skris NO_NEG_PUBKEY_BUG can be set to disable it. 116068651Skris [Steve Henson] 116168651Skris 116268651Skris *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 116368651Skris content octets. An i2c_ASN1_OBJECT is unnecessary because 116468651Skris the encoding can be trivially obtained from the structure. 116568651Skris [Steve Henson] 116668651Skris 116768651Skris *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 116868651Skris not read locks (CRYPTO_r_[un]lock). 116968651Skris [Bodo Moeller] 117068651Skris 117168651Skris *) A first attempt at creating official support for shared 117268651Skris libraries through configuration. I've kept it so the 117368651Skris default is static libraries only, and the OpenSSL programs 117468651Skris are always statically linked for now, but there are 117568651Skris preparations for dynamic linking in place. 117689837Skris This has been tested on Linux and Tru64. 117768651Skris [Richard Levitte] 117868651Skris 117968651Skris *) Randomness polling function for Win9x, as described in: 118068651Skris Peter Gutmann, Software Generation of Practically Strong 118168651Skris Random Numbers. 118268651Skris [Ulf M�ller] 118368651Skris 118468651Skris *) Fix so PRNG is seeded in req if using an already existing 118568651Skris DSA key. 118668651Skris [Steve Henson] 118768651Skris 118868651Skris *) New options to smime application. -inform and -outform 118968651Skris allow alternative formats for the S/MIME message including 119068651Skris PEM and DER. The -content option allows the content to be 119168651Skris specified separately. This should allow things like Netscape 119268651Skris form signing output easier to verify. 119368651Skris [Steve Henson] 119468651Skris 119568651Skris *) Fix the ASN1 encoding of tags using the 'long form'. 119668651Skris [Steve Henson] 119768651Skris 119868651Skris *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 119968651Skris STRING types. These convert content octets to and from the 120068651Skris underlying type. The actual tag and length octets are 120168651Skris already assumed to have been read in and checked. These 120268651Skris are needed because all other string types have virtually 120368651Skris identical handling apart from the tag. By having versions 120468651Skris of the ASN1 functions that just operate on content octets 120568651Skris IMPLICIT tagging can be handled properly. It also allows 120668651Skris the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 120768651Skris and ASN1_INTEGER are identical apart from the tag. 120868651Skris [Steve Henson] 120968651Skris 121068651Skris *) Change the handling of OID objects as follows: 121168651Skris 121268651Skris - New object identifiers are inserted in objects.txt, following 121368651Skris the syntax given in objects.README. 121468651Skris - objects.pl is used to process obj_mac.num and create a new 121568651Skris obj_mac.h. 121668651Skris - obj_dat.pl is used to create a new obj_dat.h, using the data in 121768651Skris obj_mac.h. 121868651Skris 121968651Skris This is currently kind of a hack, and the perl code in objects.pl 122068651Skris isn't very elegant, but it works as I intended. The simplest way 122168651Skris to check that it worked correctly is to look in obj_dat.h and 122268651Skris check the array nid_objs and make sure the objects haven't moved 122368651Skris around (this is important!). Additions are OK, as well as 122468651Skris consistent name changes. 122568651Skris [Richard Levitte] 122668651Skris 122768651Skris *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 122868651Skris [Bodo Moeller] 122968651Skris 123068651Skris *) Addition of the command line parameter '-rand file' to 'openssl req'. 123168651Skris The given file adds to whatever has already been seeded into the 123268651Skris random pool through the RANDFILE configuration file option or 123368651Skris environment variable, or the default random state file. 123468651Skris [Richard Levitte] 123568651Skris 123668651Skris *) mkstack.pl now sorts each macro group into lexical order. 123768651Skris Previously the output order depended on the order the files 123868651Skris appeared in the directory, resulting in needless rewriting 123968651Skris of safestack.h . 124068651Skris [Steve Henson] 124168651Skris 124268651Skris *) Patches to make OpenSSL compile under Win32 again. Mostly 124368651Skris work arounds for the VC++ problem that it treats func() as 124468651Skris func(void). Also stripped out the parts of mkdef.pl that 124568651Skris added extra typesafe functions: these no longer exist. 124668651Skris [Steve Henson] 124768651Skris 124868651Skris *) Reorganisation of the stack code. The macros are now all 124968651Skris collected in safestack.h . Each macro is defined in terms of 125068651Skris a "stack macro" of the form SKM_<name>(type, a, b). The 125168651Skris DEBUG_SAFESTACK is now handled in terms of function casts, 125268651Skris this has the advantage of retaining type safety without the 125368651Skris use of additional functions. If DEBUG_SAFESTACK is not defined 125468651Skris then the non typesafe macros are used instead. Also modified the 125568651Skris mkstack.pl script to handle the new form. Needs testing to see 125668651Skris if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 125768651Skris the default if no major problems. Similar behaviour for ASN1_SET_OF 125868651Skris and PKCS12_STACK_OF. 125968651Skris [Steve Henson] 126068651Skris 126168651Skris *) When some versions of IIS use the 'NET' form of private key the 126268651Skris key derivation algorithm is different. Normally MD5(password) is 126368651Skris used as a 128 bit RC4 key. In the modified case 126468651Skris MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 126568651Skris new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 126668651Skris as the old Netscape_RSA functions except they have an additional 126768651Skris 'sgckey' parameter which uses the modified algorithm. Also added 126868651Skris an -sgckey command line option to the rsa utility. Thanks to 126968651Skris Adrian Peck <bertie@ncipher.com> for posting details of the modified 127068651Skris algorithm to openssl-dev. 127168651Skris [Steve Henson] 127268651Skris 127368651Skris *) The evp_local.h macros were using 'c.##kname' which resulted in 127468651Skris invalid expansion on some systems (SCO 5.0.5 for example). 127568651Skris Corrected to 'c.kname'. 127668651Skris [Phillip Porch <root@theporch.com>] 127768651Skris 127868651Skris *) New X509_get1_email() and X509_REQ_get1_email() functions that return 127968651Skris a STACK of email addresses from a certificate or request, these look 128068651Skris in the subject name and the subject alternative name extensions and 128168651Skris omit any duplicate addresses. 128268651Skris [Steve Henson] 128368651Skris 128468651Skris *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 128568651Skris This makes DSA verification about 2 % faster. 128668651Skris [Bodo Moeller] 128768651Skris 128868651Skris *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 128968651Skris (meaning that now 2^5 values will be precomputed, which is only 4 KB 129068651Skris plus overhead for 1024 bit moduli). 129168651Skris This makes exponentiations about 0.5 % faster for 1024 bit 129268651Skris exponents (as measured by "openssl speed rsa2048"). 129368651Skris [Bodo Moeller] 129468651Skris 129568651Skris *) Rename memory handling macros to avoid conflicts with other 129668651Skris software: 129768651Skris Malloc => OPENSSL_malloc 129868651Skris Malloc_locked => OPENSSL_malloc_locked 129968651Skris Realloc => OPENSSL_realloc 130068651Skris Free => OPENSSL_free 130168651Skris [Richard Levitte] 130268651Skris 130368651Skris *) New function BN_mod_exp_mont_word for small bases (roughly 15% 130468651Skris faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 130568651Skris [Bodo Moeller] 130668651Skris 130768651Skris *) CygWin32 support. 130868651Skris [John Jarvie <jjarvie@newsguy.com>] 130968651Skris 131068651Skris *) The type-safe stack code has been rejigged. It is now only compiled 131168651Skris in when OpenSSL is configured with the DEBUG_SAFESTACK option and 131268651Skris by default all type-specific stack functions are "#define"d back to 131368651Skris standard stack functions. This results in more streamlined output 131468651Skris but retains the type-safety checking possibilities of the original 131568651Skris approach. 131668651Skris [Geoff Thorpe] 131768651Skris 131868651Skris *) The STACK code has been cleaned up, and certain type declarations 131968651Skris that didn't make a lot of sense have been brought in line. This has 132068651Skris also involved a cleanup of sorts in safestack.h to more correctly 132168651Skris map type-safe stack functions onto their plain stack counterparts. 132268651Skris This work has also resulted in a variety of "const"ifications of 132368651Skris lots of the code, especially "_cmp" operations which should normally 132468651Skris be prototyped with "const" parameters anyway. 132568651Skris [Geoff Thorpe] 132668651Skris 132768651Skris *) When generating bytes for the first time in md_rand.c, 'stir the pool' 132868651Skris by seeding with STATE_SIZE dummy bytes (with zero entropy count). 132968651Skris (The PRNG state consists of two parts, the large pool 'state' and 'md', 133068651Skris where all of 'md' is used each time the PRNG is used, but 'state' 133168651Skris is used only indexed by a cyclic counter. As entropy may not be 133268651Skris well distributed from the beginning, 'md' is important as a 133368651Skris chaining variable. However, the output function chains only half 133468651Skris of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 133568651Skris all of 'md', and seeding with STATE_SIZE dummy bytes will result 133668651Skris in all of 'state' being rewritten, with the new values depending 133768651Skris on virtually all of 'md'. This overcomes the 80 bit limitation.) 133868651Skris [Bodo Moeller] 133968651Skris 134068651Skris *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 134168651Skris the handshake is continued after ssl_verify_cert_chain(); 134268651Skris otherwise, if SSL_VERIFY_NONE is set, remaining error codes 134368651Skris can lead to 'unexplainable' connection aborts later. 134468651Skris [Bodo Moeller; problem tracked down by Lutz Jaenicke] 134568651Skris 134668651Skris *) Major EVP API cipher revision. 134768651Skris Add hooks for extra EVP features. This allows various cipher 134868651Skris parameters to be set in the EVP interface. Support added for variable 134968651Skris key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 135068651Skris setting of RC2 and RC5 parameters. 135168651Skris 135268651Skris Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 135368651Skris ciphers. 135468651Skris 135568651Skris Remove lots of duplicated code from the EVP library. For example *every* 135668651Skris cipher init() function handles the 'iv' in the same way according to the 135768651Skris cipher mode. They also all do nothing if the 'key' parameter is NULL and 135868651Skris for CFB and OFB modes they zero ctx->num. 135968651Skris 136068651Skris New functionality allows removal of S/MIME code RC2 hack. 136168651Skris 136268651Skris Most of the routines have the same form and so can be declared in terms 136368651Skris of macros. 136468651Skris 136568651Skris By shifting this to the top level EVP_CipherInit() it can be removed from 136668651Skris all individual ciphers. If the cipher wants to handle IVs or keys 136768651Skris differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 136868651Skris flags. 136968651Skris 137068651Skris Change lots of functions like EVP_EncryptUpdate() to now return a 137168651Skris value: although software versions of the algorithms cannot fail 137268651Skris any installed hardware versions can. 137368651Skris [Steve Henson] 137468651Skris 137568651Skris *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 137668651Skris this option is set, tolerate broken clients that send the negotiated 137768651Skris protocol version number instead of the requested protocol version 137868651Skris number. 137968651Skris [Bodo Moeller] 138068651Skris 138168651Skris *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 138268651Skris i.e. non-zero for export ciphersuites, zero otherwise. 138368651Skris Previous versions had this flag inverted, inconsistent with 138468651Skris rsa_tmp_cb (..._TMP_RSA_CB). 138568651Skris [Bodo Moeller; problem reported by Amit Chopra] 138668651Skris 138768651Skris *) Add missing DSA library text string. Work around for some IIS 138868651Skris key files with invalid SEQUENCE encoding. 138968651Skris [Steve Henson] 139068651Skris 139168651Skris *) Add a document (doc/standards.txt) that list all kinds of standards 139268651Skris and so on that are implemented in OpenSSL. 139368651Skris [Richard Levitte] 139468651Skris 139568651Skris *) Enhance c_rehash script. Old version would mishandle certificates 139668651Skris with the same subject name hash and wouldn't handle CRLs at all. 139768651Skris Added -fingerprint option to crl utility, to support new c_rehash 139868651Skris features. 139968651Skris [Steve Henson] 140068651Skris 140168651Skris *) Eliminate non-ANSI declarations in crypto.h and stack.h. 140268651Skris [Ulf M�ller] 140368651Skris 140468651Skris *) Fix for SSL server purpose checking. Server checking was 140568651Skris rejecting certificates which had extended key usage present 140668651Skris but no ssl client purpose. 140768651Skris [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 140868651Skris 140968651Skris *) Make PKCS#12 code work with no password. The PKCS#12 spec 141068651Skris is a little unclear about how a blank password is handled. 141168651Skris Since the password in encoded as a BMPString with terminating 141268651Skris double NULL a zero length password would end up as just the 141368651Skris double NULL. However no password at all is different and is 141468651Skris handled differently in the PKCS#12 key generation code. NS 141568651Skris treats a blank password as zero length. MSIE treats it as no 141668651Skris password on export: but it will try both on import. We now do 141768651Skris the same: PKCS12_parse() tries zero length and no password if 141868651Skris the password is set to "" or NULL (NULL is now a valid password: 141968651Skris it wasn't before) as does the pkcs12 application. 142068651Skris [Steve Henson] 142168651Skris 142268651Skris *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 142368651Skris perror when PEM_read_bio_X509_REQ fails, the error message must 142468651Skris be obtained from the error queue. 142568651Skris [Bodo Moeller] 142668651Skris 142768651Skris *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 142868651Skris it in ERR_remove_state if appropriate, and change ERR_get_state 142968651Skris accordingly to avoid race conditions (this is necessary because 143068651Skris thread_hash is no longer constant once set). 143168651Skris [Bodo Moeller] 143268651Skris 143368651Skris *) Bugfix for linux-elf makefile.one. 143468651Skris [Ulf M�ller] 143568651Skris 143668651Skris *) RSA_get_default_method() will now cause a default 143768651Skris RSA_METHOD to be chosen if one doesn't exist already. 143868651Skris Previously this was only set during a call to RSA_new() 143968651Skris or RSA_new_method(NULL) meaning it was possible for 144068651Skris RSA_get_default_method() to return NULL. 144168651Skris [Geoff Thorpe] 144268651Skris 144368651Skris *) Added native name translation to the existing DSO code 144468651Skris that will convert (if the flag to do so is set) filenames 144568651Skris that are sufficiently small and have no path information 144668651Skris into a canonical native form. Eg. "blah" converted to 144768651Skris "libblah.so" or "blah.dll" etc. 144868651Skris [Geoff Thorpe] 144968651Skris 145068651Skris *) New function ERR_error_string_n(e, buf, len) which is like 145168651Skris ERR_error_string(e, buf), but writes at most 'len' bytes 145268651Skris including the 0 terminator. For ERR_error_string_n, 'buf' 145368651Skris may not be NULL. 145468651Skris [Damien Miller <djm@mindrot.org>, Bodo Moeller] 145568651Skris 145668651Skris *) CONF library reworked to become more general. A new CONF 145768651Skris configuration file reader "class" is implemented as well as a 145868651Skris new functions (NCONF_*, for "New CONF") to handle it. The now 145968651Skris old CONF_* functions are still there, but are reimplemented to 146068651Skris work in terms of the new functions. Also, a set of functions 146168651Skris to handle the internal storage of the configuration data is 146268651Skris provided to make it easier to write new configuration file 146368651Skris reader "classes" (I can definitely see something reading a 146468651Skris configuration file in XML format, for example), called _CONF_*, 146568651Skris or "the configuration storage API"... 146668651Skris 146768651Skris The new configuration file reading functions are: 146868651Skris 146968651Skris NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 147068651Skris NCONF_get_section, NCONF_get_string, NCONF_get_numbre 147168651Skris 147268651Skris NCONF_default, NCONF_WIN32 147368651Skris 147468651Skris NCONF_dump_fp, NCONF_dump_bio 147568651Skris 147668651Skris NCONF_default and NCONF_WIN32 are method (or "class") choosers, 147768651Skris NCONF_new creates a new CONF object. This works in the same way 147868651Skris as other interfaces in OpenSSL, like the BIO interface. 147968651Skris NCONF_dump_* dump the internal storage of the configuration file, 148068651Skris which is useful for debugging. All other functions take the same 148168651Skris arguments as the old CONF_* functions wth the exception of the 148268651Skris first that must be a `CONF *' instead of a `LHASH *'. 148368651Skris 148468651Skris To make it easer to use the new classes with the old CONF_* functions, 148568651Skris the function CONF_set_default_method is provided. 148668651Skris [Richard Levitte] 148768651Skris 148868651Skris *) Add '-tls1' option to 'openssl ciphers', which was already 148968651Skris mentioned in the documentation but had not been implemented. 149068651Skris (This option is not yet really useful because even the additional 149168651Skris experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 149268651Skris [Bodo Moeller] 149368651Skris 149468651Skris *) Initial DSO code added into libcrypto for letting OpenSSL (and 149568651Skris OpenSSL-based applications) load shared libraries and bind to 149668651Skris them in a portable way. 149768651Skris [Geoff Thorpe, with contributions from Richard Levitte] 149868651Skris 149959191Skris Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 150059191Skris 150159191Skris *) Make sure _lrotl and _lrotr are only used with MSVC. 150259191Skris 150359191Skris *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 150459191Skris (the default implementation of RAND_status). 150559191Skris 150659191Skris *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 150759191Skris to '-clrext' (= clear extensions), as intended and documented. 150859191Skris [Bodo Moeller; inconsistency pointed out by Michael Attili 150959191Skris <attili@amaxo.com>] 151059191Skris 151159191Skris *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 151259191Skris was larger than the MD block size. 151359191Skris [Steve Henson, pointed out by Yost William <YostW@tce.com>] 151459191Skris 151559191Skris *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 151659191Skris fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 151759191Skris using the passed key: if the passed key was a private key the result 151859191Skris of X509_print(), for example, would be to print out all the private key 151959191Skris components. 152059191Skris [Steve Henson] 152159191Skris 152259191Skris *) des_quad_cksum() byte order bug fix. 152359191Skris [Ulf M�ller, using the problem description in krb4-0.9.7, where 152459191Skris the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 152559191Skris 152659191Skris *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 152759191Skris discouraged. 152859191Skris [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 152959191Skris 153059191Skris *) For easily testing in shell scripts whether some command 153159191Skris 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 153259191Skris returns with exit code 0 iff no command of the given name is available. 153359191Skris 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 153459191Skris the output goes to stdout and nothing is printed to stderr. 153559191Skris Additional arguments are always ignored. 153659191Skris 153759191Skris Since for each cipher there is a command of the same name, 153859191Skris the 'no-cipher' compilation switches can be tested this way. 153959191Skris 154059191Skris ('openssl no-XXX' is not able to detect pseudo-commands such 154159191Skris as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 154259191Skris [Bodo Moeller] 154359191Skris 154459191Skris *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 154559191Skris [Bodo Moeller] 154659191Skris 154759191Skris *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 154859191Skris is set; it will be thrown away anyway because each handshake creates 154959191Skris its own key. 155059191Skris ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 155159191Skris to parameters -- in previous versions (since OpenSSL 0.9.3) the 155259191Skris 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 155359191Skris you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 155459191Skris [Bodo Moeller] 155559191Skris 155659191Skris *) New s_client option -ign_eof: EOF at stdin is ignored, and 155759191Skris 'Q' and 'R' lose their special meanings (quit/renegotiate). 155859191Skris This is part of what -quiet does; unlike -quiet, -ign_eof 155959191Skris does not suppress any output. 156059191Skris [Richard Levitte] 156159191Skris 156259191Skris *) Add compatibility options to the purpose and trust code. The 156359191Skris purpose X509_PURPOSE_ANY is "any purpose" which automatically 156459191Skris accepts a certificate or CA, this was the previous behaviour, 156559191Skris with all the associated security issues. 156659191Skris 156759191Skris X509_TRUST_COMPAT is the old trust behaviour: only and 156859191Skris automatically trust self signed roots in certificate store. A 156959191Skris new trust setting X509_TRUST_DEFAULT is used to specify that 157059191Skris a purpose has no associated trust setting and it should instead 157159191Skris use the value in the default purpose. 157259191Skris [Steve Henson] 157359191Skris 157459191Skris *) Fix the PKCS#8 DSA private key code so it decodes keys again 157559191Skris and fix a memory leak. 157659191Skris [Steve Henson] 157759191Skris 157859191Skris *) In util/mkerr.pl (which implements 'make errors'), preserve 157959191Skris reason strings from the previous version of the .c file, as 158059191Skris the default to have only downcase letters (and digits) in 158159191Skris automatically generated reasons codes is not always appropriate. 158259191Skris [Bodo Moeller] 158359191Skris 158459191Skris *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 158559191Skris using strerror. Previously, ERR_reason_error_string() returned 158659191Skris library names as reason strings for SYSerr; but SYSerr is a special 158759191Skris case where small numbers are errno values, not library numbers. 158859191Skris [Bodo Moeller] 158959191Skris 159059191Skris *) Add '-dsaparam' option to 'openssl dhparam' application. This 159159191Skris converts DSA parameters into DH parameters. (When creating parameters, 159259191Skris DSA_generate_parameters is used.) 159359191Skris [Bodo Moeller] 159459191Skris 159559191Skris *) Include 'length' (recommended exponent length) in C code generated 159659191Skris by 'openssl dhparam -C'. 159759191Skris [Bodo Moeller] 159859191Skris 159959191Skris *) The second argument to set_label in perlasm was already being used 160059191Skris so couldn't be used as a "file scope" flag. Moved to third argument 160159191Skris which was free. 160259191Skris [Steve Henson] 160359191Skris 160459191Skris *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 160559191Skris instead of RAND_bytes for encryption IVs and salts. 160659191Skris [Bodo Moeller] 160759191Skris 160859191Skris *) Include RAND_status() into RAND_METHOD instead of implementing 160959191Skris it only for md_rand.c Otherwise replacing the PRNG by calling 161059191Skris RAND_set_rand_method would be impossible. 161159191Skris [Bodo Moeller] 161259191Skris 161359191Skris *) Don't let DSA_generate_key() enter an infinite loop if the random 161459191Skris number generation fails. 161559191Skris [Bodo Moeller] 161659191Skris 161759191Skris *) New 'rand' application for creating pseudo-random output. 161859191Skris [Bodo Moeller] 161959191Skris 162059191Skris *) Added configuration support for Linux/IA64 162159191Skris [Rolf Haberrecker <rolf@suse.de>] 162259191Skris 162359191Skris *) Assembler module support for Mingw32. 162459191Skris [Ulf M�ller] 162559191Skris 162659191Skris *) Shared library support for HPUX (in shlib/). 162759191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 162859191Skris 162959191Skris *) Shared library support for Solaris gcc. 163059191Skris [Lutz Behnke <behnke@trustcenter.de>] 163159191Skris 163259191Skris Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 163359191Skris 163459191Skris *) PKCS7_encrypt() was adding text MIME headers twice because they 163559191Skris were added manually and by SMIME_crlf_copy(). 163659191Skris [Steve Henson] 163759191Skris 163859191Skris *) In bntest.c don't call BN_rand with zero bits argument. 163959191Skris [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 164059191Skris 164159191Skris *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 164259191Skris case was implemented. This caused BN_div_recp() to fail occasionally. 164359191Skris [Ulf M�ller] 164459191Skris 164559191Skris *) Add an optional second argument to the set_label() in the perl 164659191Skris assembly language builder. If this argument exists and is set 164759191Skris to 1 it signals that the assembler should use a symbol whose 164859191Skris scope is the entire file, not just the current function. This 164959191Skris is needed with MASM which uses the format label:: for this scope. 165059191Skris [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 165159191Skris 165259191Skris *) Change the ASN1 types so they are typedefs by default. Before 165359191Skris almost all types were #define'd to ASN1_STRING which was causing 165459191Skris STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 165559191Skris for example. 165659191Skris [Steve Henson] 165759191Skris 165859191Skris *) Change names of new functions to the new get1/get0 naming 165959191Skris convention: After 'get1', the caller owns a reference count 166059191Skris and has to call ..._free; 'get0' returns a pointer to some 166159191Skris data structure without incrementing reference counters. 166259191Skris (Some of the existing 'get' functions increment a reference 166359191Skris counter, some don't.) 166459191Skris Similarly, 'set1' and 'add1' functions increase reference 166559191Skris counters or duplicate objects. 166659191Skris [Steve Henson] 166759191Skris 166859191Skris *) Allow for the possibility of temp RSA key generation failure: 166959191Skris the code used to assume it always worked and crashed on failure. 167059191Skris [Steve Henson] 167159191Skris 167259191Skris *) Fix potential buffer overrun problem in BIO_printf(). 167359191Skris [Ulf M�ller, using public domain code by Patrick Powell; problem 167459191Skris pointed out by David Sacerdote <das33@cornell.edu>] 167559191Skris 167659191Skris *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 167759191Skris RAND_egd() and RAND_status(). In the command line application, 167859191Skris the EGD socket can be specified like a seed file using RANDFILE 167959191Skris or -rand. 168059191Skris [Ulf M�ller] 168159191Skris 168259191Skris *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 168359191Skris Some CAs (e.g. Verisign) distribute certificates in this form. 168459191Skris [Steve Henson] 168559191Skris 168659191Skris *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 168759191Skris list to exclude them. This means that no special compilation option 168859191Skris is needed to use anonymous DH: it just needs to be included in the 168959191Skris cipher list. 169059191Skris [Steve Henson] 169159191Skris 169259191Skris *) Change the EVP_MD_CTX_type macro so its meaning consistent with 169359191Skris EVP_MD_type. The old functionality is available in a new macro called 169459191Skris EVP_MD_md(). Change code that uses it and update docs. 169559191Skris [Steve Henson] 169659191Skris 169759191Skris *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 169859191Skris where the 'void *' argument is replaced by a function pointer argument. 169959191Skris Previously 'void *' was abused to point to functions, which works on 170059191Skris many platforms, but is not correct. As these functions are usually 170159191Skris called by macros defined in OpenSSL header files, most source code 170259191Skris should work without changes. 170359191Skris [Richard Levitte] 170459191Skris 170559191Skris *) <openssl/opensslconf.h> (which is created by Configure) now contains 170659191Skris sections with information on -D... compiler switches used for 170759191Skris compiling the library so that applications can see them. To enable 170859191Skris one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 170959191Skris must be defined. E.g., 171059191Skris #define OPENSSL_ALGORITHM_DEFINES 171159191Skris #include <openssl/opensslconf.h> 171259191Skris defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 171359191Skris [Richard Levitte, Ulf and Bodo M�ller] 171459191Skris 171559191Skris *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 171659191Skris record layer. 171759191Skris [Bodo Moeller] 171859191Skris 171959191Skris *) Change the 'other' type in certificate aux info to a STACK_OF 172059191Skris X509_ALGOR. Although not an AlgorithmIdentifier as such it has 172159191Skris the required ASN1 format: arbitrary types determined by an OID. 172259191Skris [Steve Henson] 172359191Skris 172459191Skris *) Add some PEM_write_X509_REQ_NEW() functions and a command line 172559191Skris argument to 'req'. This is not because the function is newer or 172659191Skris better than others it just uses the work 'NEW' in the certificate 172759191Skris request header lines. Some software needs this. 172859191Skris [Steve Henson] 172959191Skris 173059191Skris *) Reorganise password command line arguments: now passwords can be 173159191Skris obtained from various sources. Delete the PEM_cb function and make 173259191Skris it the default behaviour: i.e. if the callback is NULL and the 173359191Skris usrdata argument is not NULL interpret it as a null terminated pass 173459191Skris phrase. If usrdata and the callback are NULL then the pass phrase 173559191Skris is prompted for as usual. 173659191Skris [Steve Henson] 173759191Skris 173859191Skris *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 173959191Skris the support is automatically enabled. The resulting binaries will 174059191Skris autodetect the card and use it if present. 174159191Skris [Ben Laurie and Compaq Inc.] 174259191Skris 174359191Skris *) Work around for Netscape hang bug. This sends certificate request 174459191Skris and server done in one record. Since this is perfectly legal in the 174559191Skris SSL/TLS protocol it isn't a "bug" option and is on by default. See 174659191Skris the bugs/SSLv3 entry for more info. 174759191Skris [Steve Henson] 174859191Skris 174959191Skris *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 175059191Skris [Andy Polyakov] 175159191Skris 175259191Skris *) Add -rand argument to smime and pkcs12 applications and read/write 175359191Skris of seed file. 175459191Skris [Steve Henson] 175559191Skris 175659191Skris *) New 'passwd' tool for crypt(3) and apr1 password hashes. 175759191Skris [Bodo Moeller] 175859191Skris 175959191Skris *) Add command line password options to the remaining applications. 176059191Skris [Steve Henson] 176159191Skris 176259191Skris *) Bug fix for BN_div_recp() for numerators with an even number of 176359191Skris bits. 176459191Skris [Ulf M�ller] 176559191Skris 176659191Skris *) More tests in bntest.c, and changed test_bn output. 176759191Skris [Ulf M�ller] 176859191Skris 176959191Skris *) ./config recognizes MacOS X now. 177059191Skris [Andy Polyakov] 177159191Skris 177259191Skris *) Bug fix for BN_div() when the first words of num and divsor are 177359191Skris equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 177459191Skris [Ulf M�ller] 177559191Skris 177659191Skris *) Add support for various broken PKCS#8 formats, and command line 177759191Skris options to produce them. 177859191Skris [Steve Henson] 177959191Skris 178059191Skris *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 178159191Skris get temporary BIGNUMs from a BN_CTX. 178259191Skris [Ulf M�ller] 178359191Skris 178459191Skris *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 178559191Skris for p == 0. 178659191Skris [Ulf M�ller] 178759191Skris 178859191Skris *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 178959191Skris include a #define from the old name to the new. The original intent 179059191Skris was that statically linked binaries could for example just call 179159191Skris SSLeay_add_all_ciphers() to just add ciphers to the table and not 179259191Skris link with digests. This never worked becayse SSLeay_add_all_digests() 179359191Skris and SSLeay_add_all_ciphers() were in the same source file so calling 179459191Skris one would link with the other. They are now in separate source files. 179559191Skris [Steve Henson] 179659191Skris 179759191Skris *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 179859191Skris [Steve Henson] 179959191Skris 180059191Skris *) Use a less unusual form of the Miller-Rabin primality test (it used 180159191Skris a binary algorithm for exponentiation integrated into the Miller-Rabin 180259191Skris loop, our standard modexp algorithms are faster). 180359191Skris [Bodo Moeller] 180459191Skris 180559191Skris *) Support for the EBCDIC character set completed. 180659191Skris [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 180759191Skris 180859191Skris *) Source code cleanups: use const where appropriate, eliminate casts, 180959191Skris use void * instead of char * in lhash. 181059191Skris [Ulf M�ller] 181159191Skris 181259191Skris *) Bugfix: ssl3_send_server_key_exchange was not restartable 181359191Skris (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 181459191Skris this the server could overwrite ephemeral keys that the client 181559191Skris has already seen). 181659191Skris [Bodo Moeller] 181759191Skris 181859191Skris *) Turn DSA_is_prime into a macro that calls BN_is_prime, 181959191Skris using 50 iterations of the Rabin-Miller test. 182059191Skris 182159191Skris DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 182259191Skris iterations of the Rabin-Miller test as required by the appendix 182359191Skris to FIPS PUB 186[-1]) instead of DSA_is_prime. 182459191Skris As BN_is_prime_fasttest includes trial division, DSA parameter 182559191Skris generation becomes much faster. 182659191Skris 182759191Skris This implies a change for the callback functions in DSA_is_prime 182859191Skris and DSA_generate_parameters: The callback function is called once 182959191Skris for each positive witness in the Rabin-Miller test, not just 183059191Skris occasionally in the inner loop; and the parameters to the 183159191Skris callback function now provide an iteration count for the outer 183259191Skris loop rather than for the current invocation of the inner loop. 183359191Skris DSA_generate_parameters additionally can call the callback 183459191Skris function with an 'iteration count' of -1, meaning that a 183559191Skris candidate has passed the trial division test (when q is generated 183659191Skris from an application-provided seed, trial division is skipped). 183759191Skris [Bodo Moeller] 183859191Skris 183959191Skris *) New function BN_is_prime_fasttest that optionally does trial 184059191Skris division before starting the Rabin-Miller test and has 184159191Skris an additional BN_CTX * argument (whereas BN_is_prime always 184259191Skris has to allocate at least one BN_CTX). 184359191Skris 'callback(1, -1, cb_arg)' is called when a number has passed the 184459191Skris trial division stage. 184559191Skris [Bodo Moeller] 184659191Skris 184759191Skris *) Fix for bug in CRL encoding. The validity dates weren't being handled 184859191Skris as ASN1_TIME. 184959191Skris [Steve Henson] 185059191Skris 185159191Skris *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 185259191Skris [Steve Henson] 185359191Skris 185459191Skris *) New function BN_pseudo_rand(). 185559191Skris [Ulf M�ller] 185659191Skris 185759191Skris *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 185859191Skris bignum version of BN_from_montgomery() with the working code from 185959191Skris SSLeay 0.9.0 (the word based version is faster anyway), and clean up 186059191Skris the comments. 186159191Skris [Ulf M�ller] 186259191Skris 186359191Skris *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 186459191Skris made it impossible to use the same SSL_SESSION data structure in 186559191Skris SSL2 clients in multiple threads. 186659191Skris [Bodo Moeller] 186759191Skris 186859191Skris *) The return value of RAND_load_file() no longer counts bytes obtained 186959191Skris by stat(). RAND_load_file(..., -1) is new and uses the complete file 187059191Skris to seed the PRNG (previously an explicit byte count was required). 187159191Skris [Ulf M�ller, Bodo M�ller] 187259191Skris 187359191Skris *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 187459191Skris used (char *) instead of (void *) and had casts all over the place. 187559191Skris [Steve Henson] 187659191Skris 187759191Skris *) Make BN_generate_prime() return NULL on error if ret!=NULL. 187859191Skris [Ulf M�ller] 187959191Skris 188059191Skris *) Retain source code compatibility for BN_prime_checks macro: 188159191Skris BN_is_prime(..., BN_prime_checks, ...) now uses 188259191Skris BN_prime_checks_for_size to determine the appropriate number of 188359191Skris Rabin-Miller iterations. 188459191Skris [Ulf M�ller] 188559191Skris 188659191Skris *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 188759191Skris DH_CHECK_P_NOT_SAFE_PRIME. 188859191Skris (Check if this is true? OpenPGP calls them "strong".) 188959191Skris [Ulf M�ller] 189059191Skris 189159191Skris *) Merge the functionality of "dh" and "gendh" programs into a new program 189259191Skris "dhparam". The old programs are retained for now but will handle DH keys 189359191Skris (instead of parameters) in future. 189459191Skris [Steve Henson] 189559191Skris 189659191Skris *) Make the ciphers, s_server and s_client programs check the return values 189759191Skris when a new cipher list is set. 189859191Skris [Steve Henson] 189959191Skris 190059191Skris *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 190159191Skris ciphers. Before when the 56bit ciphers were enabled the sorting was 190259191Skris wrong. 190359191Skris 190459191Skris The syntax for the cipher sorting has been extended to support sorting by 190559191Skris cipher-strength (using the strength_bits hard coded in the tables). 190659191Skris The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 190759191Skris 190859191Skris Fix a bug in the cipher-command parser: when supplying a cipher command 190959191Skris string with an "undefined" symbol (neither command nor alphanumeric 191059191Skris [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 191159191Skris an error is flagged. 191259191Skris 191359191Skris Due to the strength-sorting extension, the code of the 191459191Skris ssl_create_cipher_list() function was completely rearranged. I hope that 191559191Skris the readability was also increased :-) 191659191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 191759191Skris 191859191Skris *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 191959191Skris for the first serial number and places 2 in the serial number file. This 192059191Skris avoids problems when the root CA is created with serial number zero and 192159191Skris the first user certificate has the same issuer name and serial number 192259191Skris as the root CA. 192359191Skris [Steve Henson] 192459191Skris 192559191Skris *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 192659191Skris the new code. Add documentation for this stuff. 192759191Skris [Steve Henson] 192859191Skris 192959191Skris *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 193059191Skris X509_*() to X509at_*() on the grounds that they don't handle X509 193159191Skris structures and behave in an analagous way to the X509v3 functions: 193259191Skris they shouldn't be called directly but wrapper functions should be used 193359191Skris instead. 193459191Skris 193559191Skris So we also now have some wrapper functions that call the X509at functions 193659191Skris when passed certificate requests. (TO DO: similar things can be done with 193759191Skris PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 193859191Skris things. Some of these need some d2i or i2d and print functionality 193959191Skris because they handle more complex structures.) 194059191Skris [Steve Henson] 194159191Skris 194259191Skris *) Add missing #ifndefs that caused missing symbols when building libssl 194359191Skris as a shared library without RSA. Use #ifndef NO_SSL2 instead of 194459191Skris NO_RSA in ssl/s2*.c. 194559191Skris [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 194659191Skris 194759191Skris *) Precautions against using the PRNG uninitialized: RAND_bytes() now 194859191Skris has a return value which indicates the quality of the random data 194959191Skris (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 195059191Skris error queue. New function RAND_pseudo_bytes() generates output that is 195159191Skris guaranteed to be unique but not unpredictable. RAND_add is like 195259191Skris RAND_seed, but takes an extra argument for an entropy estimate 195359191Skris (RAND_seed always assumes full entropy). 195459191Skris [Ulf M�ller] 195559191Skris 195659191Skris *) Do more iterations of Rabin-Miller probable prime test (specifically, 195759191Skris 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 195859191Skris instead of only 2 for all lengths; see BN_prime_checks_for_size definition 195959191Skris in crypto/bn/bn_prime.c for the complete table). This guarantees a 196059191Skris false-positive rate of at most 2^-80 for random input. 196159191Skris [Bodo Moeller] 196259191Skris 196359191Skris *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 196459191Skris [Bodo Moeller] 196559191Skris 196659191Skris *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 196759191Skris in the 0.9.5 release), this returns the chain 196859191Skris from an X509_CTX structure with a dup of the stack and all 196959191Skris the X509 reference counts upped: so the stack will exist 197059191Skris after X509_CTX_cleanup() has been called. Modify pkcs12.c 197159191Skris to use this. 197259191Skris 197359191Skris Also make SSL_SESSION_print() print out the verify return 197459191Skris code. 197559191Skris [Steve Henson] 197659191Skris 197759191Skris *) Add manpage for the pkcs12 command. Also change the default 197859191Skris behaviour so MAC iteration counts are used unless the new 197959191Skris -nomaciter option is used. This improves file security and 198059191Skris only older versions of MSIE (4.0 for example) need it. 198159191Skris [Steve Henson] 198259191Skris 198359191Skris *) Honor the no-xxx Configure options when creating .DEF files. 198459191Skris [Ulf M�ller] 198559191Skris 198659191Skris *) Add PKCS#10 attributes to field table: challengePassword, 198759191Skris unstructuredName and unstructuredAddress. These are taken from 198859191Skris draft PKCS#9 v2.0 but are compatible with v1.2 provided no 198959191Skris international characters are used. 199059191Skris 199159191Skris More changes to X509_ATTRIBUTE code: allow the setting of types 199259191Skris based on strings. Remove the 'loc' parameter when adding 199359191Skris attributes because these will be a SET OF encoding which is sorted 199459191Skris in ASN1 order. 199559191Skris [Steve Henson] 199659191Skris 199759191Skris *) Initial changes to the 'req' utility to allow request generation 199859191Skris automation. This will allow an application to just generate a template 199959191Skris file containing all the field values and have req construct the 200059191Skris request. 200159191Skris 200259191Skris Initial support for X509_ATTRIBUTE handling. Stacks of these are 200359191Skris used all over the place including certificate requests and PKCS#7 200459191Skris structures. They are currently handled manually where necessary with 200559191Skris some primitive wrappers for PKCS#7. The new functions behave in a 200659191Skris manner analogous to the X509 extension functions: they allow 200759191Skris attributes to be looked up by NID and added. 200859191Skris 200959191Skris Later something similar to the X509V3 code would be desirable to 201059191Skris automatically handle the encoding, decoding and printing of the 201159191Skris more complex types. The string types like challengePassword can 201259191Skris be handled by the string table functions. 201359191Skris 201459191Skris Also modified the multi byte string table handling. Now there is 201559191Skris a 'global mask' which masks out certain types. The table itself 201659191Skris can use the flag STABLE_NO_MASK to ignore the mask setting: this 201759191Skris is useful when for example there is only one permissible type 201859191Skris (as in countryName) and using the mask might result in no valid 201959191Skris types at all. 202059191Skris [Steve Henson] 202159191Skris 202259191Skris *) Clean up 'Finished' handling, and add functions SSL_get_finished and 202359191Skris SSL_get_peer_finished to allow applications to obtain the latest 202459191Skris Finished messages sent to the peer or expected from the peer, 202559191Skris respectively. (SSL_get_peer_finished is usually the Finished message 202659191Skris actually received from the peer, otherwise the protocol will be aborted.) 202759191Skris 202859191Skris As the Finished message are message digests of the complete handshake 202959191Skris (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 203059191Skris be used for external authentication procedures when the authentication 203159191Skris provided by SSL/TLS is not desired or is not enough. 203259191Skris [Bodo Moeller] 203359191Skris 203459191Skris *) Enhanced support for Alpha Linux is added. Now ./config checks if 203559191Skris the host supports BWX extension and if Compaq C is present on the 203659191Skris $PATH. Just exploiting of the BWX extension results in 20-30% 203759191Skris performance kick for some algorithms, e.g. DES and RC4 to mention 203859191Skris a couple. Compaq C in turn generates ~20% faster code for MD5 and 203959191Skris SHA1. 204059191Skris [Andy Polyakov] 204159191Skris 204259191Skris *) Add support for MS "fast SGC". This is arguably a violation of the 204359191Skris SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 204459191Skris weak crypto and after checking the certificate is SGC a second one 204559191Skris with strong crypto. MS SGC stops the first handshake after receiving 204659191Skris the server certificate message and sends a second client hello. Since 204759191Skris a server will typically do all the time consuming operations before 204859191Skris expecting any further messages from the client (server key exchange 204959191Skris is the most expensive) there is little difference between the two. 205059191Skris 205159191Skris To get OpenSSL to support MS SGC we have to permit a second client 205259191Skris hello message after we have sent server done. In addition we have to 205359191Skris reset the MAC if we do get this second client hello. 205459191Skris [Steve Henson] 205559191Skris 205659191Skris *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 205759191Skris if a DER encoded private key is RSA or DSA traditional format. Changed 205859191Skris d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 205959191Skris format DER encoded private key. Newer code should use PKCS#8 format which 206059191Skris has the key type encoded in the ASN1 structure. Added DER private key 206159191Skris support to pkcs8 application. 206259191Skris [Steve Henson] 206359191Skris 206459191Skris *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 206559191Skris ciphersuites has been selected (as required by the SSL 3/TLS 1 206659191Skris specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 206759191Skris is set, we interpret this as a request to violate the specification 206859191Skris (the worst that can happen is a handshake failure, and 'correct' 206959191Skris behaviour would result in a handshake failure anyway). 207059191Skris [Bodo Moeller] 207159191Skris 207259191Skris *) In SSL_CTX_add_session, take into account that there might be multiple 207359191Skris SSL_SESSION structures with the same session ID (e.g. when two threads 207459191Skris concurrently obtain them from an external cache). 207559191Skris The internal cache can handle only one SSL_SESSION with a given ID, 207659191Skris so if there's a conflict, we now throw out the old one to achieve 207759191Skris consistency. 207859191Skris [Bodo Moeller] 207959191Skris 208059191Skris *) Add OIDs for idea and blowfish in CBC mode. This will allow both 208159191Skris to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 208259191Skris some routines that use cipher OIDs: some ciphers do not have OIDs 208359191Skris defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 208459191Skris example. 208559191Skris [Steve Henson] 208659191Skris 208759191Skris *) Simplify the trust setting structure and code. Now we just have 208859191Skris two sequences of OIDs for trusted and rejected settings. These will 208959191Skris typically have values the same as the extended key usage extension 209059191Skris and any application specific purposes. 209159191Skris 209259191Skris The trust checking code now has a default behaviour: it will just 209359191Skris check for an object with the same NID as the passed id. Functions can 209459191Skris be provided to override either the default behaviour or the behaviour 209559191Skris for a given id. SSL client, server and email already have functions 209659191Skris in place for compatibility: they check the NID and also return "trusted" 209759191Skris if the certificate is self signed. 209859191Skris [Steve Henson] 209959191Skris 210059191Skris *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 210159191Skris traditional format into an EVP_PKEY structure. 210259191Skris [Steve Henson] 210359191Skris 210459191Skris *) Add a password callback function PEM_cb() which either prompts for 210559191Skris a password if usr_data is NULL or otherwise assumes it is a null 210659191Skris terminated password. Allow passwords to be passed on command line 210759191Skris environment or config files in a few more utilities. 210859191Skris [Steve Henson] 210959191Skris 211059191Skris *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 211159191Skris keys. Add some short names for PKCS#8 PBE algorithms and allow them 211259191Skris to be specified on the command line for the pkcs8 and pkcs12 utilities. 211359191Skris Update documentation. 211459191Skris [Steve Henson] 211559191Skris 211659191Skris *) Support for ASN1 "NULL" type. This could be handled before by using 211759191Skris ASN1_TYPE but there wasn't any function that would try to read a NULL 211859191Skris and produce an error if it couldn't. For compatibility we also have 211959191Skris ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 212059191Skris don't allocate anything because they don't need to. 212159191Skris [Steve Henson] 212259191Skris 212359191Skris *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 212459191Skris for details. 212559191Skris [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 212659191Skris 212759191Skris *) Rebuild of the memory allocation routines used by OpenSSL code and 212859191Skris possibly others as well. The purpose is to make an interface that 212959191Skris provide hooks so anyone can build a separate set of allocation and 213059191Skris deallocation routines to be used by OpenSSL, for example memory 213159191Skris pool implementations, or something else, which was previously hard 213259191Skris since Malloc(), Realloc() and Free() were defined as macros having 213359191Skris the values malloc, realloc and free, respectively (except for Win32 213459191Skris compilations). The same is provided for memory debugging code. 213559191Skris OpenSSL already comes with functionality to find memory leaks, but 213659191Skris this gives people a chance to debug other memory problems. 213759191Skris 213859191Skris With these changes, a new set of functions and macros have appeared: 213959191Skris 214068651Skris CRYPTO_set_mem_debug_functions() [F] 214168651Skris CRYPTO_get_mem_debug_functions() [F] 214268651Skris CRYPTO_dbg_set_options() [F] 214368651Skris CRYPTO_dbg_get_options() [F] 214468651Skris CRYPTO_malloc_debug_init() [M] 214559191Skris 214659191Skris The memory debug functions are NULL by default, unless the library 214759191Skris is compiled with CRYPTO_MDEBUG or friends is defined. If someone 214859191Skris wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 214959191Skris gives the standard debugging functions that come with OpenSSL) or 215059191Skris CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 215159191Skris provided by the library user) must be used. When the standard 215259191Skris debugging functions are used, CRYPTO_dbg_set_options can be used to 215359191Skris request additional information: 215459191Skris CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 215559191Skris the CRYPTO_MDEBUG_xxx macro when compiling the library. 215659191Skris 215759191Skris Also, things like CRYPTO_set_mem_functions will always give the 215859191Skris expected result (the new set of functions is used for allocation 215959191Skris and deallocation) at all times, regardless of platform and compiler 216059191Skris options. 216159191Skris 216259191Skris To finish it up, some functions that were never use in any other 216359191Skris way than through macros have a new API and new semantic: 216459191Skris 216559191Skris CRYPTO_dbg_malloc() 216659191Skris CRYPTO_dbg_realloc() 216759191Skris CRYPTO_dbg_free() 216859191Skris 216959191Skris All macros of value have retained their old syntax. 217059191Skris [Richard Levitte and Bodo Moeller] 217159191Skris 217259191Skris *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 217359191Skris ordering of SMIMECapabilities wasn't in "strength order" and there 217459191Skris was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 217559191Skris algorithm. 217659191Skris [Steve Henson] 217759191Skris 217859191Skris *) Some ASN1 types with illegal zero length encoding (INTEGER, 217959191Skris ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 218059191Skris [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 218159191Skris 218259191Skris *) Merge in my S/MIME library for OpenSSL. This provides a simple 218359191Skris S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 218459191Skris functionality to handle multipart/signed properly) and a utility 218559191Skris called 'smime' to call all this stuff. This is based on code I 218659191Skris originally wrote for Celo who have kindly allowed it to be 218759191Skris included in OpenSSL. 218859191Skris [Steve Henson] 218959191Skris 219059191Skris *) Add variants des_set_key_checked and des_set_key_unchecked of 219159191Skris des_set_key (aka des_key_sched). Global variable des_check_key 219259191Skris decides which of these is called by des_set_key; this way 219359191Skris des_check_key behaves as it always did, but applications and 219459191Skris the library itself, which was buggy for des_check_key == 1, 219559191Skris have a cleaner way to pick the version they need. 219659191Skris [Bodo Moeller] 219759191Skris 219859191Skris *) New function PKCS12_newpass() which changes the password of a 219959191Skris PKCS12 structure. 220059191Skris [Steve Henson] 220159191Skris 220259191Skris *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 220359191Skris dynamic mix. In both cases the ids can be used as an index into the 220459191Skris table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 220559191Skris functions so they accept a list of the field values and the 220659191Skris application doesn't need to directly manipulate the X509_TRUST 220759191Skris structure. 220859191Skris [Steve Henson] 220959191Skris 221059191Skris *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 221159191Skris need initialising. 221259191Skris [Steve Henson] 221359191Skris 221459191Skris *) Modify the way the V3 extension code looks up extensions. This now 221559191Skris works in a similar way to the object code: we have some "standard" 221659191Skris extensions in a static table which is searched with OBJ_bsearch() 221759191Skris and the application can add dynamic ones if needed. The file 221859191Skris crypto/x509v3/ext_dat.h now has the info: this file needs to be 221959191Skris updated whenever a new extension is added to the core code and kept 222059191Skris in ext_nid order. There is a simple program 'tabtest.c' which checks 222159191Skris this. New extensions are not added too often so this file can readily 222259191Skris be maintained manually. 222359191Skris 222459191Skris There are two big advantages in doing things this way. The extensions 222559191Skris can be looked up immediately and no longer need to be "added" using 222659191Skris X509V3_add_standard_extensions(): this function now does nothing. 222759191Skris [Side note: I get *lots* of email saying the extension code doesn't 222859191Skris work because people forget to call this function] 222959191Skris Also no dynamic allocation is done unless new extensions are added: 223059191Skris so if we don't add custom extensions there is no need to call 223159191Skris X509V3_EXT_cleanup(). 223259191Skris [Steve Henson] 223359191Skris 223459191Skris *) Modify enc utility's salting as follows: make salting the default. Add a 223559191Skris magic header, so unsalted files fail gracefully instead of just decrypting 223659191Skris to garbage. This is because not salting is a big security hole, so people 223759191Skris should be discouraged from doing it. 223859191Skris [Ben Laurie] 223959191Skris 224059191Skris *) Fixes and enhancements to the 'x509' utility. It allowed a message 224159191Skris digest to be passed on the command line but it only used this 224259191Skris parameter when signing a certificate. Modified so all relevant 224359191Skris operations are affected by the digest parameter including the 224459191Skris -fingerprint and -x509toreq options. Also -x509toreq choked if a 224559191Skris DSA key was used because it didn't fix the digest. 224659191Skris [Steve Henson] 224759191Skris 224859191Skris *) Initial certificate chain verify code. Currently tests the untrusted 224959191Skris certificates for consistency with the verify purpose (which is set 225059191Skris when the X509_STORE_CTX structure is set up) and checks the pathlength. 225159191Skris 225259191Skris There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 225359191Skris this is because it will reject chains with invalid extensions whereas 225459191Skris every previous version of OpenSSL and SSLeay made no checks at all. 225559191Skris 225659191Skris Trust code: checks the root CA for the relevant trust settings. Trust 225759191Skris settings have an initial value consistent with the verify purpose: e.g. 225859191Skris if the verify purpose is for SSL client use it expects the CA to be 225959191Skris trusted for SSL client use. However the default value can be changed to 226059191Skris permit custom trust settings: one example of this would be to only trust 226159191Skris certificates from a specific "secure" set of CAs. 226259191Skris 226359191Skris Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 226459191Skris which should be used for version portability: especially since the 226559191Skris verify structure is likely to change more often now. 226659191Skris 226759191Skris SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 226859191Skris to set them. If not set then assume SSL clients will verify SSL servers 226959191Skris and vice versa. 227059191Skris 227159191Skris Two new options to the verify program: -untrusted allows a set of 227259191Skris untrusted certificates to be passed in and -purpose which sets the 227359191Skris intended purpose of the certificate. If a purpose is set then the 227459191Skris new chain verify code is used to check extension consistency. 227559191Skris [Steve Henson] 227659191Skris 227759191Skris *) Support for the authority information access extension. 227859191Skris [Steve Henson] 227959191Skris 228059191Skris *) Modify RSA and DSA PEM read routines to transparently handle 228159191Skris PKCS#8 format private keys. New *_PUBKEY_* functions that handle 228259191Skris public keys in a format compatible with certificate 228359191Skris SubjectPublicKeyInfo structures. Unfortunately there were already 228459191Skris functions called *_PublicKey_* which used various odd formats so 228559191Skris these are retained for compatibility: however the DSA variants were 228659191Skris never in a public release so they have been deleted. Changed dsa/rsa 228759191Skris utilities to handle the new format: note no releases ever handled public 228859191Skris keys so we should be OK. 228959191Skris 229059191Skris The primary motivation for this change is to avoid the same fiasco 229159191Skris that dogs private keys: there are several incompatible private key 229259191Skris formats some of which are standard and some OpenSSL specific and 229359191Skris require various evil hacks to allow partial transparent handling and 229459191Skris even then it doesn't work with DER formats. Given the option anything 229559191Skris other than PKCS#8 should be dumped: but the other formats have to 229659191Skris stay in the name of compatibility. 229759191Skris 229859191Skris With public keys and the benefit of hindsight one standard format 229959191Skris is used which works with EVP_PKEY, RSA or DSA structures: though 230059191Skris it clearly returns an error if you try to read the wrong kind of key. 230159191Skris 230259191Skris Added a -pubkey option to the 'x509' utility to output the public key. 230359191Skris Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 230459191Skris (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 230559191Skris EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 230659191Skris that do the same as the EVP_PKEY_assign_*() except they up the 230759191Skris reference count of the added key (they don't "swallow" the 230859191Skris supplied key). 230959191Skris [Steve Henson] 231059191Skris 231159191Skris *) Fixes to crypto/x509/by_file.c the code to read in certificates and 231259191Skris CRLs would fail if the file contained no certificates or no CRLs: 231359191Skris added a new function to read in both types and return the number 231459191Skris read: this means that if none are read it will be an error. The 231559191Skris DER versions of the certificate and CRL reader would always fail 231659191Skris because it isn't possible to mix certificates and CRLs in DER format 231759191Skris without choking one or the other routine. Changed this to just read 231859191Skris a certificate: this is the best we can do. Also modified the code 231959191Skris in apps/verify.c to take notice of return codes: it was previously 232059191Skris attempting to read in certificates from NULL pointers and ignoring 232159191Skris any errors: this is one reason why the cert and CRL reader seemed 232259191Skris to work. It doesn't check return codes from the default certificate 232359191Skris routines: these may well fail if the certificates aren't installed. 232459191Skris [Steve Henson] 232559191Skris 232659191Skris *) Code to support otherName option in GeneralName. 232759191Skris [Steve Henson] 232859191Skris 232959191Skris *) First update to verify code. Change the verify utility 233059191Skris so it warns if it is passed a self signed certificate: 233159191Skris for consistency with the normal behaviour. X509_verify 233259191Skris has been modified to it will now verify a self signed 233359191Skris certificate if *exactly* the same certificate appears 233459191Skris in the store: it was previously impossible to trust a 233559191Skris single self signed certificate. This means that: 233659191Skris openssl verify ss.pem 233759191Skris now gives a warning about a self signed certificate but 233859191Skris openssl verify -CAfile ss.pem ss.pem 233959191Skris is OK. 234059191Skris [Steve Henson] 234159191Skris 234259191Skris *) For servers, store verify_result in SSL_SESSION data structure 234359191Skris (and add it to external session representation). 234459191Skris This is needed when client certificate verifications fails, 234559191Skris but an application-provided verification callback (set by 234659191Skris SSL_CTX_set_cert_verify_callback) allows accepting the session 234759191Skris anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 234859191Skris but returns 1): When the session is reused, we have to set 234959191Skris ssl->verify_result to the appropriate error code to avoid 235059191Skris security holes. 235159191Skris [Bodo Moeller, problem pointed out by Lutz Jaenicke] 235259191Skris 235359191Skris *) Fix a bug in the new PKCS#7 code: it didn't consider the 235459191Skris case in PKCS7_dataInit() where the signed PKCS7 structure 235559191Skris didn't contain any existing data because it was being created. 235659191Skris [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 235759191Skris 235859191Skris *) Add a salt to the key derivation routines in enc.c. This 235959191Skris forms the first 8 bytes of the encrypted file. Also add a 236059191Skris -S option to allow a salt to be input on the command line. 236159191Skris [Steve Henson] 236259191Skris 236359191Skris *) New function X509_cmp(). Oddly enough there wasn't a function 236459191Skris to compare two certificates. We do this by working out the SHA1 236559191Skris hash and comparing that. X509_cmp() will be needed by the trust 236659191Skris code. 236759191Skris [Steve Henson] 236859191Skris 236959191Skris *) SSL_get1_session() is like SSL_get_session(), but increments 237059191Skris the reference count in the SSL_SESSION returned. 237159191Skris [Geoff Thorpe <geoff@eu.c2.net>] 237259191Skris 237359191Skris *) Fix for 'req': it was adding a null to request attributes. 237459191Skris Also change the X509_LOOKUP and X509_INFO code to handle 237559191Skris certificate auxiliary information. 237659191Skris [Steve Henson] 237759191Skris 237859191Skris *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 237959191Skris the 'enc' command. 238059191Skris [Steve Henson] 238159191Skris 238259191Skris *) Add the possibility to add extra information to the memory leak 238359191Skris detecting output, to form tracebacks, showing from where each 238459191Skris allocation was originated: CRYPTO_push_info("constant string") adds 238559191Skris the string plus current file name and line number to a per-thread 238659191Skris stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 238759191Skris is like calling CYRPTO_pop_info() until the stack is empty. 238859191Skris Also updated memory leak detection code to be multi-thread-safe. 238959191Skris [Richard Levitte] 239059191Skris 239159191Skris *) Add options -text and -noout to pkcs7 utility and delete the 239259191Skris encryption options which never did anything. Update docs. 239359191Skris [Steve Henson] 239459191Skris 239559191Skris *) Add options to some of the utilities to allow the pass phrase 239659191Skris to be included on either the command line (not recommended on 239759191Skris OSes like Unix) or read from the environment. Update the 239859191Skris manpages and fix a few bugs. 239959191Skris [Steve Henson] 240059191Skris 240159191Skris *) Add a few manpages for some of the openssl commands. 240259191Skris [Steve Henson] 240359191Skris 240459191Skris *) Fix the -revoke option in ca. It was freeing up memory twice, 240559191Skris leaking and not finding already revoked certificates. 240659191Skris [Steve Henson] 240759191Skris 240859191Skris *) Extensive changes to support certificate auxiliary information. 240959191Skris This involves the use of X509_CERT_AUX structure and X509_AUX 241059191Skris functions. An X509_AUX function such as PEM_read_X509_AUX() 241159191Skris can still read in a certificate file in the usual way but it 241259191Skris will also read in any additional "auxiliary information". By 241359191Skris doing things this way a fair degree of compatibility can be 241459191Skris retained: existing certificates can have this information added 241559191Skris using the new 'x509' options. 241659191Skris 241759191Skris Current auxiliary information includes an "alias" and some trust 241859191Skris settings. The trust settings will ultimately be used in enhanced 241959191Skris certificate chain verification routines: currently a certificate 242059191Skris can only be trusted if it is self signed and then it is trusted 242159191Skris for all purposes. 242259191Skris [Steve Henson] 242359191Skris 242459191Skris *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 242559191Skris The problem was that one of the replacement routines had not been working 242659191Skris since SSLeay releases. For now the offending routine has been replaced 242759191Skris with non-optimised assembler. Even so, this now gives around 95% 242859191Skris performance improvement for 1024 bit RSA signs. 242959191Skris [Mark Cox] 243059191Skris 243159191Skris *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 243259191Skris handling. Most clients have the effective key size in bits equal to 243359191Skris the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 243459191Skris A few however don't do this and instead use the size of the decrypted key 243559191Skris to determine the RC2 key length and the AlgorithmIdentifier to determine 243659191Skris the effective key length. In this case the effective key length can still 243759191Skris be 40 bits but the key length can be 168 bits for example. This is fixed 243859191Skris by manually forcing an RC2 key into the EVP_PKEY structure because the 243959191Skris EVP code can't currently handle unusual RC2 key sizes: it always assumes 244059191Skris the key length and effective key length are equal. 244159191Skris [Steve Henson] 244259191Skris 244359191Skris *) Add a bunch of functions that should simplify the creation of 244459191Skris X509_NAME structures. Now you should be able to do: 244559191Skris X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 244659191Skris and have it automatically work out the correct field type and fill in 244759191Skris the structures. The more adventurous can try: 244859191Skris X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 244959191Skris and it will (hopefully) work out the correct multibyte encoding. 245059191Skris [Steve Henson] 245159191Skris 245259191Skris *) Change the 'req' utility to use the new field handling and multibyte 245359191Skris copy routines. Before the DN field creation was handled in an ad hoc 245459191Skris way in req, ca, and x509 which was rather broken and didn't support 245559191Skris BMPStrings or UTF8Strings. Since some software doesn't implement 245659191Skris BMPStrings or UTF8Strings yet, they can be enabled using the config file 245759191Skris using the dirstring_type option. See the new comment in the default 245859191Skris openssl.cnf for more info. 245959191Skris [Steve Henson] 246059191Skris 246159191Skris *) Make crypto/rand/md_rand.c more robust: 246259191Skris - Assure unique random numbers after fork(). 246359191Skris - Make sure that concurrent threads access the global counter and 246459191Skris md serializably so that we never lose entropy in them 246559191Skris or use exactly the same state in multiple threads. 246659191Skris Access to the large state is not always serializable because 246759191Skris the additional locking could be a performance killer, and 246859191Skris md should be large enough anyway. 246959191Skris [Bodo Moeller] 247059191Skris 247159191Skris *) New file apps/app_rand.c with commonly needed functionality 247259191Skris for handling the random seed file. 247359191Skris 247459191Skris Use the random seed file in some applications that previously did not: 247559191Skris ca, 247659191Skris dsaparam -genkey (which also ignored its '-rand' option), 247759191Skris s_client, 247859191Skris s_server, 247959191Skris x509 (when signing). 248059191Skris Except on systems with /dev/urandom, it is crucial to have a random 248159191Skris seed file at least for key creation, DSA signing, and for DH exchanges; 248259191Skris for RSA signatures we could do without one. 248359191Skris 248459191Skris gendh and gendsa (unlike genrsa) used to read only the first byte 248559191Skris of each file listed in the '-rand' option. The function as previously 248659191Skris found in genrsa is now in app_rand.c and is used by all programs 248759191Skris that support '-rand'. 248859191Skris [Bodo Moeller] 248959191Skris 249059191Skris *) In RAND_write_file, use mode 0600 for creating files; 249159191Skris don't just chmod when it may be too late. 249259191Skris [Bodo Moeller] 249359191Skris 249459191Skris *) Report an error from X509_STORE_load_locations 249559191Skris when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 249659191Skris [Bill Perry] 249759191Skris 249859191Skris *) New function ASN1_mbstring_copy() this copies a string in either 249959191Skris ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 250059191Skris into an ASN1_STRING type. A mask of permissible types is passed 250159191Skris and it chooses the "minimal" type to use or an error if not type 250259191Skris is suitable. 250359191Skris [Steve Henson] 250459191Skris 250559191Skris *) Add function equivalents to the various macros in asn1.h. The old 250659191Skris macros are retained with an M_ prefix. Code inside the library can 250759191Skris use the M_ macros. External code (including the openssl utility) 250859191Skris should *NOT* in order to be "shared library friendly". 250959191Skris [Steve Henson] 251059191Skris 251159191Skris *) Add various functions that can check a certificate's extensions 251259191Skris to see if it usable for various purposes such as SSL client, 251359191Skris server or S/MIME and CAs of these types. This is currently 251459191Skris VERY EXPERIMENTAL but will ultimately be used for certificate chain 251559191Skris verification. Also added a -purpose flag to x509 utility to 251659191Skris print out all the purposes. 251759191Skris [Steve Henson] 251859191Skris 251959191Skris *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 252059191Skris functions. 252159191Skris [Steve Henson] 252259191Skris 252359191Skris *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 252459191Skris for, obtain and decode and extension and obtain its critical flag. 252559191Skris This allows all the necessary extension code to be handled in a 252659191Skris single function call. 252759191Skris [Steve Henson] 252859191Skris 252959191Skris *) RC4 tune-up featuring 30-40% performance improvement on most RISC 253059191Skris platforms. See crypto/rc4/rc4_enc.c for further details. 253159191Skris [Andy Polyakov] 253259191Skris 253359191Skris *) New -noout option to asn1parse. This causes no output to be produced 253459191Skris its main use is when combined with -strparse and -out to extract data 253559191Skris from a file (which may not be in ASN.1 format). 253659191Skris [Steve Henson] 253759191Skris 253859191Skris *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 253959191Skris when producing the local key id. 254059191Skris [Richard Levitte <levitte@stacken.kth.se>] 254159191Skris 254259191Skris *) New option -dhparam in s_server. This allows a DH parameter file to be 254359191Skris stated explicitly. If it is not stated then it tries the first server 254459191Skris certificate file. The previous behaviour hard coded the filename 254559191Skris "server.pem". 254659191Skris [Steve Henson] 254759191Skris 254859191Skris *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 254959191Skris a public key to be input or output. For example: 255059191Skris openssl rsa -in key.pem -pubout -out pubkey.pem 255159191Skris Also added necessary DSA public key functions to handle this. 255259191Skris [Steve Henson] 255359191Skris 255459191Skris *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 255559191Skris in the message. This was handled by allowing 255659191Skris X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 255759191Skris [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 255859191Skris 255959191Skris *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 256059191Skris to the end of the strings whereas this didn't. This would cause problems 256159191Skris if strings read with d2i_ASN1_bytes() were later modified. 256259191Skris [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 256359191Skris 256459191Skris *) Fix for base64 decode bug. When a base64 bio reads only one line of 256559191Skris data and it contains EOF it will end up returning an error. This is 256659191Skris caused by input 46 bytes long. The cause is due to the way base64 256759191Skris BIOs find the start of base64 encoded data. They do this by trying a 256859191Skris trial decode on each line until they find one that works. When they 256959191Skris do a flag is set and it starts again knowing it can pass all the 257059191Skris data directly through the decoder. Unfortunately it doesn't reset 257159191Skris the context it uses. This means that if EOF is reached an attempt 257259191Skris is made to pass two EOFs through the context and this causes the 257359191Skris resulting error. This can also cause other problems as well. As is 257459191Skris usual with these problems it takes *ages* to find and the fix is 257559191Skris trivial: move one line. 257659191Skris [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 257759191Skris 257859191Skris *) Ugly workaround to get s_client and s_server working under Windows. The 257959191Skris old code wouldn't work because it needed to select() on sockets and the 258059191Skris tty (for keypresses and to see if data could be written). Win32 only 258159191Skris supports select() on sockets so we select() with a 1s timeout on the 258259191Skris sockets and then see if any characters are waiting to be read, if none 258359191Skris are present then we retry, we also assume we can always write data to 258459191Skris the tty. This isn't nice because the code then blocks until we've 258559191Skris received a complete line of data and it is effectively polling the 258659191Skris keyboard at 1s intervals: however it's quite a bit better than not 258759191Skris working at all :-) A dedicated Windows application might handle this 258859191Skris with an event loop for example. 258959191Skris [Steve Henson] 259059191Skris 259159191Skris *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 259259191Skris and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 259359191Skris will be called when RSA_sign() and RSA_verify() are used. This is useful 259459191Skris if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 259559191Skris For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 259659191Skris should *not* be used: RSA_sign() and RSA_verify() must be used instead. 259759191Skris This necessitated the support of an extra signature type NID_md5_sha1 259859191Skris for SSL signatures and modifications to the SSL library to use it instead 259959191Skris of calling RSA_public_decrypt() and RSA_private_encrypt(). 260059191Skris [Steve Henson] 260159191Skris 260259191Skris *) Add new -verify -CAfile and -CApath options to the crl program, these 260359191Skris will lookup a CRL issuers certificate and verify the signature in a 260459191Skris similar way to the verify program. Tidy up the crl program so it 260559191Skris no longer accesses structures directly. Make the ASN1 CRL parsing a bit 260659191Skris less strict. It will now permit CRL extensions even if it is not 260759191Skris a V2 CRL: this will allow it to tolerate some broken CRLs. 260859191Skris [Steve Henson] 260959191Skris 261059191Skris *) Initialize all non-automatic variables each time one of the openssl 261159191Skris sub-programs is started (this is necessary as they may be started 261259191Skris multiple times from the "OpenSSL>" prompt). 261359191Skris [Lennart Bang, Bodo Moeller] 261459191Skris 261559191Skris *) Preliminary compilation option RSA_NULL which disables RSA crypto without 261659191Skris removing all other RSA functionality (this is what NO_RSA does). This 261759191Skris is so (for example) those in the US can disable those operations covered 261859191Skris by the RSA patent while allowing storage and parsing of RSA keys and RSA 261959191Skris key generation. 262059191Skris [Steve Henson] 262159191Skris 262259191Skris *) Non-copying interface to BIO pairs. 262359191Skris (still largely untested) 262459191Skris [Bodo Moeller] 262559191Skris 262659191Skris *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 262759191Skris ASCII string. This was handled independently in various places before. 262859191Skris [Steve Henson] 262959191Skris 263059191Skris *) New functions UTF8_getc() and UTF8_putc() that parse and generate 263159191Skris UTF8 strings a character at a time. 263259191Skris [Steve Henson] 263359191Skris 263459191Skris *) Use client_version from client hello to select the protocol 263559191Skris (s23_srvr.c) and for RSA client key exchange verification 263659191Skris (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 263759191Skris [Bodo Moeller] 263859191Skris 263959191Skris *) Add various utility functions to handle SPKACs, these were previously 264059191Skris handled by poking round in the structure internals. Added new function 264159191Skris NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 264259191Skris print, verify and generate SPKACs. Based on an original idea from 264359191Skris Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 264459191Skris [Steve Henson] 264559191Skris 264659191Skris *) RIPEMD160 is operational on all platforms and is back in 'make test'. 264759191Skris [Andy Polyakov] 264859191Skris 264959191Skris *) Allow the config file extension section to be overwritten on the 265059191Skris command line. Based on an original idea from Massimiliano Pala 265159191Skris <madwolf@comune.modena.it>. The new option is called -extensions 265259191Skris and can be applied to ca, req and x509. Also -reqexts to override 265359191Skris the request extensions in req and -crlexts to override the crl extensions 265459191Skris in ca. 265559191Skris [Steve Henson] 265659191Skris 265759191Skris *) Add new feature to the SPKAC handling in ca. Now you can include 265859191Skris the same field multiple times by preceding it by "XXXX." for example: 265959191Skris 1.OU="Unit name 1" 266059191Skris 2.OU="Unit name 2" 266159191Skris this is the same syntax as used in the req config file. 266259191Skris [Steve Henson] 266359191Skris 266459191Skris *) Allow certificate extensions to be added to certificate requests. These 266559191Skris are specified in a 'req_extensions' option of the req section of the 266659191Skris config file. They can be printed out with the -text option to req but 266759191Skris are otherwise ignored at present. 266859191Skris [Steve Henson] 266959191Skris 267059191Skris *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 267159191Skris data read consists of only the final block it would not decrypted because 267259191Skris EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 267359191Skris A misplaced 'break' also meant the decrypted final block might not be 267459191Skris copied until the next read. 267559191Skris [Steve Henson] 267659191Skris 267759191Skris *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 267859191Skris a few extra parameters to the DH structure: these will be useful if 267959191Skris for example we want the value of 'q' or implement X9.42 DH. 268059191Skris [Steve Henson] 268159191Skris 268259191Skris *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 268359191Skris provides hooks that allow the default DSA functions or functions on a 268459191Skris "per key" basis to be replaced. This allows hardware acceleration and 268559191Skris hardware key storage to be handled without major modification to the 268659191Skris library. Also added low level modexp hooks and CRYPTO_EX structure and 268759191Skris associated functions. 268859191Skris [Steve Henson] 268959191Skris 269059191Skris *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 269159191Skris as "read only": it can't be written to and the buffer it points to will 269259191Skris not be freed. Reading from a read only BIO is much more efficient than 269359191Skris a normal memory BIO. This was added because there are several times when 269459191Skris an area of memory needs to be read from a BIO. The previous method was 269559191Skris to create a memory BIO and write the data to it, this results in two 269659191Skris copies of the data and an O(n^2) reading algorithm. There is a new 269759191Skris function BIO_new_mem_buf() which creates a read only memory BIO from 269859191Skris an area of memory. Also modified the PKCS#7 routines to use read only 269959191Skris memory BIOs. 270059191Skris [Steve Henson] 270159191Skris 270259191Skris *) Bugfix: ssl23_get_client_hello did not work properly when called in 270359191Skris state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 270459191Skris a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 270559191Skris but a retry condition occured while trying to read the rest. 270659191Skris [Bodo Moeller] 270759191Skris 270859191Skris *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 270959191Skris NID_pkcs7_encrypted by default: this was wrong since this should almost 271059191Skris always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 271159191Skris the encrypted data type: this is a more sensible place to put it and it 271259191Skris allows the PKCS#12 code to be tidied up that duplicated this 271359191Skris functionality. 271459191Skris [Steve Henson] 271559191Skris 271659191Skris *) Changed obj_dat.pl script so it takes its input and output files on 271759191Skris the command line. This should avoid shell escape redirection problems 271859191Skris under Win32. 271959191Skris [Steve Henson] 272059191Skris 272159191Skris *) Initial support for certificate extension requests, these are included 272259191Skris in things like Xenroll certificate requests. Included functions to allow 272359191Skris extensions to be obtained and added. 272459191Skris [Steve Henson] 272559191Skris 272659191Skris *) -crlf option to s_client and s_server for sending newlines as 272759191Skris CRLF (as required by many protocols). 272859191Skris [Bodo Moeller] 272959191Skris 273055714Skris Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 273155714Skris 273255714Skris *) Install libRSAglue.a when OpenSSL is built with RSAref. 273355714Skris [Ralf S. Engelschall] 273455714Skris 273555714Skris *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 273655714Skris [Andrija Antonijevic <TheAntony2@bigfoot.com>] 273755714Skris 273855714Skris *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 273955714Skris program. 274055714Skris [Steve Henson] 274155714Skris 274255714Skris *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 274355714Skris DH parameters/keys (q is lost during that conversion, but the resulting 274455714Skris DH parameters contain its length). 274555714Skris 274655714Skris For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 274755714Skris much faster than DH_generate_parameters (which creates parameters 274855714Skris where p = 2*q + 1), and also the smaller q makes DH computations 274955714Skris much more efficient (160-bit exponentiation instead of 1024-bit 275055714Skris exponentiation); so this provides a convenient way to support DHE 275155714Skris ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 275255714Skris utter importance to use 275355714Skris SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 275455714Skris or 275555714Skris SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 275655714Skris when such DH parameters are used, because otherwise small subgroup 275755714Skris attacks may become possible! 275855714Skris [Bodo Moeller] 275955714Skris 276055714Skris *) Avoid memory leak in i2d_DHparams. 276155714Skris [Bodo Moeller] 276255714Skris 276355714Skris *) Allow the -k option to be used more than once in the enc program: 276455714Skris this allows the same encrypted message to be read by multiple recipients. 276555714Skris [Steve Henson] 276655714Skris 276755714Skris *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 276855714Skris an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 276955714Skris it will always use the numerical form of the OID, even if it has a short 277055714Skris or long name. 277155714Skris [Steve Henson] 277255714Skris 277355714Skris *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 277455714Skris method only got called if p,q,dmp1,dmq1,iqmp components were present, 277555714Skris otherwise bn_mod_exp was called. In the case of hardware keys for example 277655714Skris no private key components need be present and it might store extra data 277759191Skris in the RSA structure, which cannot be accessed from bn_mod_exp. 277859191Skris By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 277959191Skris private key operations. 278055714Skris [Steve Henson] 278155714Skris 278255714Skris *) Added support for SPARC Linux. 278355714Skris [Andy Polyakov] 278455714Skris 278555714Skris *) pem_password_cb function type incompatibly changed from 278655714Skris typedef int pem_password_cb(char *buf, int size, int rwflag); 278755714Skris to 278855714Skris ....(char *buf, int size, int rwflag, void *userdata); 278955714Skris so that applications can pass data to their callbacks: 279055714Skris The PEM[_ASN1]_{read,write}... functions and macros now take an 279155714Skris additional void * argument, which is just handed through whenever 279255714Skris the password callback is called. 279359191Skris [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 279455714Skris 279555714Skris New function SSL_CTX_set_default_passwd_cb_userdata. 279655714Skris 279755714Skris Compatibility note: As many C implementations push function arguments 279855714Skris onto the stack in reverse order, the new library version is likely to 279955714Skris interoperate with programs that have been compiled with the old 280055714Skris pem_password_cb definition (PEM_whatever takes some data that 280155714Skris happens to be on the stack as its last argument, and the callback 280255714Skris just ignores this garbage); but there is no guarantee whatsoever that 280355714Skris this will work. 280455714Skris 280555714Skris *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 280655714Skris (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 280755714Skris problems not only on Windows, but also on some Unix platforms. 280855714Skris To avoid problematic command lines, these definitions are now in an 280955714Skris auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 281055714Skris for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 281155714Skris [Bodo Moeller] 281255714Skris 281355714Skris *) MIPS III/IV assembler module is reimplemented. 281455714Skris [Andy Polyakov] 281555714Skris 281655714Skris *) More DES library cleanups: remove references to srand/rand and 281755714Skris delete an unused file. 281855714Skris [Ulf M�ller] 281955714Skris 282055714Skris *) Add support for the the free Netwide assembler (NASM) under Win32, 282155714Skris since not many people have MASM (ml) and it can be hard to obtain. 282255714Skris This is currently experimental but it seems to work OK and pass all 282355714Skris the tests. Check out INSTALL.W32 for info. 282455714Skris [Steve Henson] 282555714Skris 282655714Skris *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 282755714Skris without temporary keys kept an extra copy of the server key, 282855714Skris and connections with temporary keys did not free everything in case 282955714Skris of an error. 283055714Skris [Bodo Moeller] 283155714Skris 283255714Skris *) New function RSA_check_key and new openssl rsa option -check 283355714Skris for verifying the consistency of RSA keys. 283455714Skris [Ulf Moeller, Bodo Moeller] 283555714Skris 283655714Skris *) Various changes to make Win32 compile work: 283755714Skris 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 283855714Skris 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 283955714Skris comparison" warnings. 284055714Skris 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 284155714Skris [Steve Henson] 284255714Skris 284355714Skris *) Add a debugging option to PKCS#5 v2 key generation function: when 284455714Skris you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 284555714Skris derived keys are printed to stderr. 284655714Skris [Steve Henson] 284755714Skris 284855714Skris *) Copy the flags in ASN1_STRING_dup(). 284955714Skris [Roman E. Pavlov <pre@mo.msk.ru>] 285055714Skris 285155714Skris *) The x509 application mishandled signing requests containing DSA 285255714Skris keys when the signing key was also DSA and the parameters didn't match. 285355714Skris 285455714Skris It was supposed to omit the parameters when they matched the signing key: 285555714Skris the verifying software was then supposed to automatically use the CA's 285655714Skris parameters if they were absent from the end user certificate. 285755714Skris 285855714Skris Omitting parameters is no longer recommended. The test was also 285955714Skris the wrong way round! This was probably due to unusual behaviour in 286055714Skris EVP_cmp_parameters() which returns 1 if the parameters match. 286155714Skris This meant that parameters were omitted when they *didn't* match and 286255714Skris the certificate was useless. Certificates signed with 'ca' didn't have 286355714Skris this bug. 286455714Skris [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 286555714Skris 286655714Skris *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 286755714Skris The interface is as follows: 286855714Skris Applications can use 286955714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 287055714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 287155714Skris "off" is now the default. 287255714Skris The library internally uses 287355714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 287455714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 287555714Skris to disable memory-checking temporarily. 287655714Skris 287755714Skris Some inconsistent states that previously were possible (and were 287855714Skris even the default) are now avoided. 287955714Skris 288055714Skris -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 288155714Skris with each memory chunk allocated; this is occasionally more helpful 288255714Skris than just having a counter. 288355714Skris 288455714Skris -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 288555714Skris 288655714Skris -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 288755714Skris extensions. 288855714Skris [Bodo Moeller] 288955714Skris 289055714Skris *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 289155714Skris which largely parallels "options", but is for changing API behaviour, 289255714Skris whereas "options" are about protocol behaviour. 289355714Skris Initial "mode" flags are: 289455714Skris 289555714Skris SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 289655714Skris a single record has been written. 289755714Skris SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 289855714Skris retries use the same buffer location. 289955714Skris (But all of the contents must be 290055714Skris copied!) 290155714Skris [Bodo Moeller] 290255714Skris 290379998Skris *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 290455714Skris worked. 290555714Skris 290655714Skris *) Fix problems with no-hmac etc. 290755714Skris [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 290855714Skris 290955714Skris *) New functions RSA_get_default_method(), RSA_set_method() and 291055714Skris RSA_get_method(). These allows replacement of RSA_METHODs without having 291155714Skris to mess around with the internals of an RSA structure. 291255714Skris [Steve Henson] 291355714Skris 291455714Skris *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 291555714Skris Also really enable memory leak checks in openssl.c and in some 291655714Skris test programs. 291755714Skris [Chad C. Mulligan, Bodo Moeller] 291855714Skris 291955714Skris *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 292055714Skris up the length of negative integers. This has now been simplified to just 292155714Skris store the length when it is first determined and use it later, rather 292255714Skris than trying to keep track of where data is copied and updating it to 292355714Skris point to the end. 292455714Skris [Steve Henson, reported by Brien Wheeler 292555714Skris <bwheeler@authentica-security.com>] 292655714Skris 292755714Skris *) Add a new function PKCS7_signatureVerify. This allows the verification 292855714Skris of a PKCS#7 signature but with the signing certificate passed to the 292955714Skris function itself. This contrasts with PKCS7_dataVerify which assumes the 293055714Skris certificate is present in the PKCS#7 structure. This isn't always the 293155714Skris case: certificates can be omitted from a PKCS#7 structure and be 293255714Skris distributed by "out of band" means (such as a certificate database). 293355714Skris [Steve Henson] 293455714Skris 293555714Skris *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 293655714Skris function prototypes in pem.h, also change util/mkdef.pl to add the 293755714Skris necessary function names. 293855714Skris [Steve Henson] 293955714Skris 294055714Skris *) mk1mf.pl (used by Windows builds) did not properly read the 294155714Skris options set by Configure in the top level Makefile, and Configure 294255714Skris was not even able to write more than one option correctly. 294355714Skris Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 294455714Skris [Bodo Moeller] 294555714Skris 294655714Skris *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 294755714Skris file to be loaded from a BIO or FILE pointer. The BIO version will 294855714Skris for example allow memory BIOs to contain config info. 294955714Skris [Steve Henson] 295055714Skris 295155714Skris *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 295255714Skris Whoever hopes to achieve shared-library compatibility across versions 295355714Skris must use this, not the compile-time macro. 295455714Skris (Exercise 0.9.4: Which is the minimum library version required by 295555714Skris such programs?) 295655714Skris Note: All this applies only to multi-threaded programs, others don't 295755714Skris need locks. 295855714Skris [Bodo Moeller] 295955714Skris 296055714Skris *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 296155714Skris through a BIO pair triggered the default case, i.e. 296255714Skris SSLerr(...,SSL_R_UNKNOWN_STATE). 296355714Skris [Bodo Moeller] 296455714Skris 296555714Skris *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 296655714Skris can use the SSL library even if none of the specific BIOs is 296755714Skris appropriate. 296855714Skris [Bodo Moeller] 296955714Skris 297055714Skris *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 297155714Skris for the encoded length. 297255714Skris [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 297355714Skris 297455714Skris *) Add initial documentation of the X509V3 functions. 297555714Skris [Steve Henson] 297655714Skris 297755714Skris *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 297855714Skris PEM_write_bio_PKCS8PrivateKey() that are equivalent to 297955714Skris PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 298055714Skris secure PKCS#8 private key format with a high iteration count. 298155714Skris [Steve Henson] 298255714Skris 298355714Skris *) Fix determination of Perl interpreter: A perl or perl5 298455714Skris _directory_ in $PATH was also accepted as the interpreter. 298555714Skris [Ralf S. Engelschall] 298655714Skris 298755714Skris *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 298855714Skris wrong with it but it was very old and did things like calling 298955714Skris PEM_ASN1_read() directly and used MD5 for the hash not to mention some 299055714Skris unusual formatting. 299155714Skris [Steve Henson] 299255714Skris 299355714Skris *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 299455714Skris to use the new extension code. 299555714Skris [Steve Henson] 299655714Skris 299755714Skris *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 299855714Skris with macros. This should make it easier to change their form, add extra 299955714Skris arguments etc. Fix a few PEM prototypes which didn't have cipher as a 300055714Skris constant. 300155714Skris [Steve Henson] 300255714Skris 300355714Skris *) Add to configuration table a new entry that can specify an alternative 300455714Skris name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 300555714Skris according to Mark Crispin <MRC@Panda.COM>. 300655714Skris [Bodo Moeller] 300755714Skris 300855714Skris#if 0 300955714Skris *) DES CBC did not update the IV. Weird. 301055714Skris [Ben Laurie] 301155714Skris#else 301255714Skris des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 301355714Skris Changing the behaviour of the former might break existing programs -- 301455714Skris where IV updating is needed, des_ncbc_encrypt can be used. 301555714Skris#endif 301655714Skris 301755714Skris *) When bntest is run from "make test" it drives bc to check its 301855714Skris calculations, as well as internally checking them. If an internal check 301955714Skris fails, it needs to cause bc to give a non-zero result or make test carries 302055714Skris on without noticing the failure. Fixed. 302155714Skris [Ben Laurie] 302255714Skris 302355714Skris *) DES library cleanups. 302455714Skris [Ulf M�ller] 302555714Skris 302655714Skris *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 302755714Skris used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 302855714Skris ciphers. NOTE: although the key derivation function has been verified 302955714Skris against some published test vectors it has not been extensively tested 303055714Skris yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 303155714Skris of v2.0. 303255714Skris [Steve Henson] 303355714Skris 303455714Skris *) Instead of "mkdir -p", which is not fully portable, use new 303555714Skris Perl script "util/mkdir-p.pl". 303655714Skris [Bodo Moeller] 303755714Skris 303855714Skris *) Rewrite the way password based encryption (PBE) is handled. It used to 303955714Skris assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 304055714Skris structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 304155714Skris but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 304255714Skris the 'parameter' field of the AlgorithmIdentifier is passed to the 304355714Skris underlying key generation function so it must do its own ASN1 parsing. 304455714Skris This has also changed the EVP_PBE_CipherInit() function which now has a 304555714Skris 'parameter' argument instead of literal salt and iteration count values 304655714Skris and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 304755714Skris [Steve Henson] 304855714Skris 304955714Skris *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 305055714Skris and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 305155714Skris Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 305255714Skris KEY" because this clashed with PKCS#8 unencrypted string. Since this 305355714Skris value was just used as a "magic string" and not used directly its 305455714Skris value doesn't matter. 305555714Skris [Steve Henson] 305655714Skris 305755714Skris *) Introduce some semblance of const correctness to BN. Shame C doesn't 305855714Skris support mutable. 305955714Skris [Ben Laurie] 306055714Skris 306155714Skris *) "linux-sparc64" configuration (ultrapenguin). 306255714Skris [Ray Miller <ray.miller@oucs.ox.ac.uk>] 306355714Skris "linux-sparc" configuration. 306455714Skris [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 306555714Skris 306655714Skris *) config now generates no-xxx options for missing ciphers. 306755714Skris [Ulf M�ller] 306855714Skris 306955714Skris *) Support the EBCDIC character set (work in progress). 307055714Skris File ebcdic.c not yet included because it has a different license. 307155714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 307255714Skris 307355714Skris *) Support BS2000/OSD-POSIX. 307455714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 307555714Skris 307655714Skris *) Make callbacks for key generation use void * instead of char *. 307755714Skris [Ben Laurie] 307855714Skris 307955714Skris *) Make S/MIME samples compile (not yet tested). 308055714Skris [Ben Laurie] 308155714Skris 308255714Skris *) Additional typesafe stacks. 308355714Skris [Ben Laurie] 308455714Skris 308555714Skris *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 308655714Skris [Bodo Moeller] 308755714Skris 308855714Skris 308955714Skris Changes between 0.9.3 and 0.9.3a [29 May 1999] 309055714Skris 309155714Skris *) New configuration variant "sco5-gcc". 309255714Skris 309355714Skris *) Updated some demos. 309455714Skris [Sean O Riordain, Wade Scholine] 309555714Skris 309655714Skris *) Add missing BIO_free at exit of pkcs12 application. 309755714Skris [Wu Zhigang] 309855714Skris 309955714Skris *) Fix memory leak in conf.c. 310055714Skris [Steve Henson] 310155714Skris 310255714Skris *) Updates for Win32 to assembler version of MD5. 310355714Skris [Steve Henson] 310455714Skris 310555714Skris *) Set #! path to perl in apps/der_chop to where we found it 310655714Skris instead of using a fixed path. 310755714Skris [Bodo Moeller] 310855714Skris 310955714Skris *) SHA library changes for irix64-mips4-cc. 311055714Skris [Andy Polyakov] 311155714Skris 311255714Skris *) Improvements for VMS support. 311355714Skris [Richard Levitte] 311455714Skris 311555714Skris 311655714Skris Changes between 0.9.2b and 0.9.3 [24 May 1999] 311755714Skris 311855714Skris *) Bignum library bug fix. IRIX 6 passes "make test" now! 311955714Skris This also avoids the problems with SC4.2 and unpatched SC5. 312055714Skris [Andy Polyakov <appro@fy.chalmers.se>] 312155714Skris 312255714Skris *) New functions sk_num, sk_value and sk_set to replace the previous macros. 312355714Skris These are required because of the typesafe stack would otherwise break 312455714Skris existing code. If old code used a structure member which used to be STACK 312555714Skris and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 312655714Skris sk_num or sk_value it would produce an error because the num, data members 312755714Skris are not present in STACK_OF. Now it just produces a warning. sk_set 312855714Skris replaces the old method of assigning a value to sk_value 312955714Skris (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 313055714Skris that does this will no longer work (and should use sk_set instead) but 313155714Skris this could be regarded as a "questionable" behaviour anyway. 313255714Skris [Steve Henson] 313355714Skris 313455714Skris *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 313555714Skris correctly handle encrypted S/MIME data. 313655714Skris [Steve Henson] 313755714Skris 313855714Skris *) Change type of various DES function arguments from des_cblock 313955714Skris (which means, in function argument declarations, pointer to char) 314055714Skris to des_cblock * (meaning pointer to array with 8 char elements), 314155714Skris which allows the compiler to do more typechecking; it was like 314255714Skris that back in SSLeay, but with lots of ugly casts. 314355714Skris 314455714Skris Introduce new type const_des_cblock. 314555714Skris [Bodo Moeller] 314655714Skris 314755714Skris *) Reorganise the PKCS#7 library and get rid of some of the more obvious 314855714Skris problems: find RecipientInfo structure that matches recipient certificate 314955714Skris and initialise the ASN1 structures properly based on passed cipher. 315055714Skris [Steve Henson] 315155714Skris 315255714Skris *) Belatedly make the BN tests actually check the results. 315355714Skris [Ben Laurie] 315455714Skris 315555714Skris *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 315655714Skris to and from BNs: it was completely broken. New compilation option 315755714Skris NEG_PUBKEY_BUG to allow for some broken certificates that encode public 315855714Skris key elements as negative integers. 315955714Skris [Steve Henson] 316055714Skris 316155714Skris *) Reorganize and speed up MD5. 316255714Skris [Andy Polyakov <appro@fy.chalmers.se>] 316355714Skris 316455714Skris *) VMS support. 316555714Skris [Richard Levitte <richard@levitte.org>] 316655714Skris 316755714Skris *) New option -out to asn1parse to allow the parsed structure to be 316855714Skris output to a file. This is most useful when combined with the -strparse 316955714Skris option to examine the output of things like OCTET STRINGS. 317055714Skris [Steve Henson] 317155714Skris 317255714Skris *) Make SSL library a little more fool-proof by not requiring any longer 317355714Skris that SSL_set_{accept,connect}_state be called before 317455714Skris SSL_{accept,connect} may be used (SSL_set_..._state is omitted 317555714Skris in many applications because usually everything *appeared* to work as 317655714Skris intended anyway -- now it really works as intended). 317755714Skris [Bodo Moeller] 317855714Skris 317955714Skris *) Move openssl.cnf out of lib/. 318055714Skris [Ulf M�ller] 318155714Skris 318255714Skris *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 318355714Skris -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 318455714Skris -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 318555714Skris [Ralf S. Engelschall] 318655714Skris 318755714Skris *) Various fixes to the EVP and PKCS#7 code. It may now be able to 318855714Skris handle PKCS#7 enveloped data properly. 318955714Skris [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 319055714Skris 319155714Skris *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 319255714Skris copying pointers. The cert_st handling is changed by this in 319355714Skris various ways (and thus what used to be known as ctx->default_cert 319455714Skris is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 319555714Skris any longer when s->cert does not give us what we need). 319655714Skris ssl_cert_instantiate becomes obsolete by this change. 319755714Skris As soon as we've got the new code right (possibly it already is?), 319855714Skris we have solved a couple of bugs of the earlier code where s->cert 319955714Skris was used as if it could not have been shared with other SSL structures. 320055714Skris 320155714Skris Note that using the SSL API in certain dirty ways now will result 320255714Skris in different behaviour than observed with earlier library versions: 320355714Skris Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 320455714Skris does not influence s as it used to. 320555714Skris 320655714Skris In order to clean up things more thoroughly, inside SSL_SESSION 320755714Skris we don't use CERT any longer, but a new structure SESS_CERT 320855714Skris that holds per-session data (if available); currently, this is 320955714Skris the peer's certificate chain and, for clients, the server's certificate 321055714Skris and temporary key. CERT holds only those values that can have 321155714Skris meaningful defaults in an SSL_CTX. 321255714Skris [Bodo Moeller] 321355714Skris 321455714Skris *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 321555714Skris from the internal representation. Various PKCS#7 fixes: remove some 321655714Skris evil casts and set the enc_dig_alg field properly based on the signing 321755714Skris key type. 321855714Skris [Steve Henson] 321955714Skris 322055714Skris *) Allow PKCS#12 password to be set from the command line or the 322155714Skris environment. Let 'ca' get its config file name from the environment 322255714Skris variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 322355714Skris and 'x509'). 322455714Skris [Steve Henson] 322555714Skris 322655714Skris *) Allow certificate policies extension to use an IA5STRING for the 322755714Skris organization field. This is contrary to the PKIX definition but 322855714Skris VeriSign uses it and IE5 only recognises this form. Document 'x509' 322955714Skris extension option. 323055714Skris [Steve Henson] 323155714Skris 323255714Skris *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 323355714Skris without disallowing inline assembler and the like for non-pedantic builds. 323455714Skris [Ben Laurie] 323555714Skris 323655714Skris *) Support Borland C++ builder. 323755714Skris [Janez Jere <jj@void.si>, modified by Ulf M�ller] 323855714Skris 323955714Skris *) Support Mingw32. 324055714Skris [Ulf M�ller] 324155714Skris 324255714Skris *) SHA-1 cleanups and performance enhancements. 324355714Skris [Andy Polyakov <appro@fy.chalmers.se>] 324455714Skris 324555714Skris *) Sparc v8plus assembler for the bignum library. 324655714Skris [Andy Polyakov <appro@fy.chalmers.se>] 324755714Skris 324855714Skris *) Accept any -xxx and +xxx compiler options in Configure. 324955714Skris [Ulf M�ller] 325055714Skris 325155714Skris *) Update HPUX configuration. 325255714Skris [Anonymous] 325355714Skris 325455714Skris *) Add missing sk_<type>_unshift() function to safestack.h 325555714Skris [Ralf S. Engelschall] 325655714Skris 325755714Skris *) New function SSL_CTX_use_certificate_chain_file that sets the 325855714Skris "extra_cert"s in addition to the certificate. (This makes sense 325955714Skris only for "PEM" format files, as chains as a whole are not 326055714Skris DER-encoded.) 326155714Skris [Bodo Moeller] 326255714Skris 326355714Skris *) Support verify_depth from the SSL API. 326455714Skris x509_vfy.c had what can be considered an off-by-one-error: 326555714Skris Its depth (which was not part of the external interface) 326655714Skris was actually counting the number of certificates in a chain; 326755714Skris now it really counts the depth. 326855714Skris [Bodo Moeller] 326955714Skris 327055714Skris *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 327155714Skris instead of X509err, which often resulted in confusing error 327255714Skris messages since the error codes are not globally unique 327355714Skris (e.g. an alleged error in ssl3_accept when a certificate 327455714Skris didn't match the private key). 327555714Skris 327655714Skris *) New function SSL_CTX_set_session_id_context that allows to set a default 327755714Skris value (so that you don't need SSL_set_session_id_context for each 327855714Skris connection using the SSL_CTX). 327955714Skris [Bodo Moeller] 328055714Skris 328155714Skris *) OAEP decoding bug fix. 328255714Skris [Ulf M�ller] 328355714Skris 328455714Skris *) Support INSTALL_PREFIX for package builders, as proposed by 328555714Skris David Harris. 328655714Skris [Bodo Moeller] 328755714Skris 328855714Skris *) New Configure options "threads" and "no-threads". For systems 328955714Skris where the proper compiler options are known (currently Solaris 329055714Skris and Linux), "threads" is the default. 329155714Skris [Bodo Moeller] 329255714Skris 329355714Skris *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 329455714Skris [Bodo Moeller] 329555714Skris 329655714Skris *) Install various scripts to $(OPENSSLDIR)/misc, not to 329755714Skris $(INSTALLTOP)/bin -- they shouldn't clutter directories 329855714Skris such as /usr/local/bin. 329955714Skris [Bodo Moeller] 330055714Skris 330155714Skris *) "make linux-shared" to build shared libraries. 330255714Skris [Niels Poppe <niels@netbox.org>] 330355714Skris 330455714Skris *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 330555714Skris [Ulf M�ller] 330655714Skris 330755714Skris *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 330855714Skris extension adding in x509 utility. 330955714Skris [Steve Henson] 331055714Skris 331155714Skris *) Remove NOPROTO sections and error code comments. 331255714Skris [Ulf M�ller] 331355714Skris 331455714Skris *) Partial rewrite of the DEF file generator to now parse the ANSI 331555714Skris prototypes. 331655714Skris [Steve Henson] 331755714Skris 331855714Skris *) New Configure options --prefix=DIR and --openssldir=DIR. 331955714Skris [Ulf M�ller] 332055714Skris 332155714Skris *) Complete rewrite of the error code script(s). It is all now handled 332255714Skris by one script at the top level which handles error code gathering, 332355714Skris header rewriting and C source file generation. It should be much better 332455714Skris than the old method: it now uses a modified version of Ulf's parser to 332555714Skris read the ANSI prototypes in all header files (thus the old K&R definitions 332655714Skris aren't needed for error creation any more) and do a better job of 332755714Skris translating function codes into names. The old 'ASN1 error code imbedded 332855714Skris in a comment' is no longer necessary and it doesn't use .err files which 332955714Skris have now been deleted. Also the error code call doesn't have to appear all 333055714Skris on one line (which resulted in some large lines...). 333155714Skris [Steve Henson] 333255714Skris 333355714Skris *) Change #include filenames from <foo.h> to <openssl/foo.h>. 333455714Skris [Bodo Moeller] 333555714Skris 333655714Skris *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 333755714Skris 0 (which usually indicates a closed connection), but continue reading. 333855714Skris [Bodo Moeller] 333955714Skris 334055714Skris *) Fix some race conditions. 334155714Skris [Bodo Moeller] 334255714Skris 334355714Skris *) Add support for CRL distribution points extension. Add Certificate 334455714Skris Policies and CRL distribution points documentation. 334555714Skris [Steve Henson] 334655714Skris 334755714Skris *) Move the autogenerated header file parts to crypto/opensslconf.h. 334855714Skris [Ulf M�ller] 334955714Skris 335055714Skris *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 335155714Skris 8 of keying material. Merlin has also confirmed interop with this fix 335255714Skris between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 335355714Skris [Merlin Hughes <merlin@baltimore.ie>] 335455714Skris 335555714Skris *) Fix lots of warnings. 335655714Skris [Richard Levitte <levitte@stacken.kth.se>] 335755714Skris 335855714Skris *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 335955714Skris the directory spec didn't end with a LIST_SEPARATOR_CHAR. 336055714Skris [Richard Levitte <levitte@stacken.kth.se>] 336155714Skris 336255714Skris *) Fix problems with sizeof(long) == 8. 336355714Skris [Andy Polyakov <appro@fy.chalmers.se>] 336455714Skris 336555714Skris *) Change functions to ANSI C. 336655714Skris [Ulf M�ller] 336755714Skris 336855714Skris *) Fix typos in error codes. 336955714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 337055714Skris 337155714Skris *) Remove defunct assembler files from Configure. 337255714Skris [Ulf M�ller] 337355714Skris 337455714Skris *) SPARC v8 assembler BIGNUM implementation. 337555714Skris [Andy Polyakov <appro@fy.chalmers.se>] 337655714Skris 337755714Skris *) Support for Certificate Policies extension: both print and set. 337855714Skris Various additions to support the r2i method this uses. 337955714Skris [Steve Henson] 338055714Skris 338155714Skris *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 338255714Skris return a const string when you are expecting an allocated buffer. 338355714Skris [Ben Laurie] 338455714Skris 338555714Skris *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 338655714Skris types DirectoryString and DisplayText. 338755714Skris [Steve Henson] 338855714Skris 338955714Skris *) Add code to allow r2i extensions to access the configuration database, 339055714Skris add an LHASH database driver and add several ctx helper functions. 339155714Skris [Steve Henson] 339255714Skris 339355714Skris *) Fix an evil bug in bn_expand2() which caused various BN functions to 339455714Skris fail when they extended the size of a BIGNUM. 339555714Skris [Steve Henson] 339655714Skris 339755714Skris *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 339855714Skris support typesafe stack. 339955714Skris [Steve Henson] 340055714Skris 340155714Skris *) Fix typo in SSL_[gs]et_options(). 340255714Skris [Nils Frostberg <nils@medcom.se>] 340355714Skris 340455714Skris *) Delete various functions and files that belonged to the (now obsolete) 340555714Skris old X509V3 handling code. 340655714Skris [Steve Henson] 340755714Skris 340855714Skris *) New Configure option "rsaref". 340955714Skris [Ulf M�ller] 341055714Skris 341155714Skris *) Don't auto-generate pem.h. 341255714Skris [Bodo Moeller] 341355714Skris 341455714Skris *) Introduce type-safe ASN.1 SETs. 341555714Skris [Ben Laurie] 341655714Skris 341755714Skris *) Convert various additional casted stacks to type-safe STACK_OF() variants. 341855714Skris [Ben Laurie, Ralf S. Engelschall, Steve Henson] 341955714Skris 342055714Skris *) Introduce type-safe STACKs. This will almost certainly break lots of code 342155714Skris that links with OpenSSL (well at least cause lots of warnings), but fear 342255714Skris not: the conversion is trivial, and it eliminates loads of evil casts. A 342355714Skris few STACKed things have been converted already. Feel free to convert more. 342455714Skris In the fullness of time, I'll do away with the STACK type altogether. 342555714Skris [Ben Laurie] 342655714Skris 342755714Skris *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 342855714Skris specified in <certfile> by updating the entry in the index.txt file. 342955714Skris This way one no longer has to edit the index.txt file manually for 343055714Skris revoking a certificate. The -revoke option does the gory details now. 343155714Skris [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 343255714Skris 343355714Skris *) Fix `openssl crl -noout -text' combination where `-noout' killed the 343455714Skris `-text' option at all and this way the `-noout -text' combination was 343555714Skris inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 343655714Skris [Ralf S. Engelschall] 343755714Skris 343855714Skris *) Make sure a corresponding plain text error message exists for the 343955714Skris X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 344055714Skris verify callback function determined that a certificate was revoked. 344155714Skris [Ralf S. Engelschall] 344255714Skris 344355714Skris *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 344455714Skris ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 344555714Skris all available cipers including rc5, which was forgotten until now. 344655714Skris In order to let the testing shell script know which algorithms 344755714Skris are available, a new (up to now undocumented) command 344855714Skris "openssl list-cipher-commands" is used. 344955714Skris [Bodo Moeller] 345055714Skris 345155714Skris *) Bugfix: s_client occasionally would sleep in select() when 345255714Skris it should have checked SSL_pending() first. 345355714Skris [Bodo Moeller] 345455714Skris 345555714Skris *) New functions DSA_do_sign and DSA_do_verify to provide access to 345655714Skris the raw DSA values prior to ASN.1 encoding. 345755714Skris [Ulf M�ller] 345855714Skris 345955714Skris *) Tweaks to Configure 346055714Skris [Niels Poppe <niels@netbox.org>] 346155714Skris 346255714Skris *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 346355714Skris yet... 346455714Skris [Steve Henson] 346555714Skris 346655714Skris *) New variables $(RANLIB) and $(PERL) in the Makefiles. 346755714Skris [Ulf M�ller] 346855714Skris 346955714Skris *) New config option to avoid instructions that are illegal on the 80386. 347055714Skris The default code is faster, but requires at least a 486. 347155714Skris [Ulf M�ller] 347255714Skris 347355714Skris *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 347455714Skris SSL2_SERVER_VERSION (not used at all) macros, which are now the 347555714Skris same as SSL2_VERSION anyway. 347655714Skris [Bodo Moeller] 347755714Skris 347855714Skris *) New "-showcerts" option for s_client. 347955714Skris [Bodo Moeller] 348055714Skris 348155714Skris *) Still more PKCS#12 integration. Add pkcs12 application to openssl 348255714Skris application. Various cleanups and fixes. 348355714Skris [Steve Henson] 348455714Skris 348555714Skris *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 348655714Skris modify error routines to work internally. Add error codes and PBE init 348755714Skris to library startup routines. 348855714Skris [Steve Henson] 348955714Skris 349055714Skris *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 349155714Skris packing functions to asn1 and evp. Changed function names and error 349255714Skris codes along the way. 349355714Skris [Steve Henson] 349455714Skris 349555714Skris *) PKCS12 integration: and so it begins... First of several patches to 349655714Skris slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 349755714Skris objects to objects.h 349855714Skris [Steve Henson] 349955714Skris 350055714Skris *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 350155714Skris and display support for Thawte strong extranet extension. 350255714Skris [Steve Henson] 350355714Skris 350455714Skris *) Add LinuxPPC support. 350555714Skris [Jeff Dubrule <igor@pobox.org>] 350655714Skris 350755714Skris *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 350855714Skris bn_div_words in alpha.s. 350955714Skris [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 351055714Skris 351155714Skris *) Make sure the RSA OAEP test is skipped under -DRSAref because 351255714Skris OAEP isn't supported when OpenSSL is built with RSAref. 351355714Skris [Ulf Moeller <ulf@fitug.de>] 351455714Skris 351555714Skris *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 351655714Skris so they no longer are missing under -DNOPROTO. 351755714Skris [Soren S. Jorvang <soren@t.dk>] 351855714Skris 351955714Skris 352055714Skris Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 352155714Skris 352255714Skris *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 352355714Skris doesn't work when the session is reused. Coming soon! 352455714Skris [Ben Laurie] 352555714Skris 352655714Skris *) Fix a security hole, that allows sessions to be reused in the wrong 352755714Skris context thus bypassing client cert protection! All software that uses 352855714Skris client certs and session caches in multiple contexts NEEDS PATCHING to 352955714Skris allow session reuse! A fuller solution is in the works. 353055714Skris [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 353155714Skris 353255714Skris *) Some more source tree cleanups (removed obsolete files 353355714Skris crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 353455714Skris permission on "config" script to be executable) and a fix for the INSTALL 353555714Skris document. 353655714Skris [Ulf Moeller <ulf@fitug.de>] 353755714Skris 353855714Skris *) Remove some legacy and erroneous uses of malloc, free instead of 353955714Skris Malloc, Free. 354055714Skris [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 354155714Skris 354255714Skris *) Make rsa_oaep_test return non-zero on error. 354355714Skris [Ulf Moeller <ulf@fitug.de>] 354455714Skris 354555714Skris *) Add support for native Solaris shared libraries. Configure 354655714Skris solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 354755714Skris if someone would make that last step automatic. 354855714Skris [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 354955714Skris 355055714Skris *) ctx_size was not built with the right compiler during "make links". Fixed. 355155714Skris [Ben Laurie] 355255714Skris 355355714Skris *) Change the meaning of 'ALL' in the cipher list. It now means "everything 355455714Skris except NULL ciphers". This means the default cipher list will no longer 355555714Skris enable NULL ciphers. They need to be specifically enabled e.g. with 355655714Skris the string "DEFAULT:eNULL". 355755714Skris [Steve Henson] 355855714Skris 355955714Skris *) Fix to RSA private encryption routines: if p < q then it would 356055714Skris occasionally produce an invalid result. This will only happen with 356155714Skris externally generated keys because OpenSSL (and SSLeay) ensure p > q. 356255714Skris [Steve Henson] 356355714Skris 356455714Skris *) Be less restrictive and allow also `perl util/perlpath.pl 356555714Skris /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 356655714Skris because this way one can also use an interpreter named `perl5' (which is 356755714Skris usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 356855714Skris installed as `perl'). 356955714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 357055714Skris 357155714Skris *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 357255714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 357355714Skris 357455714Skris *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 357555714Skris advapi32.lib to Win32 build and change the pem test comparision 357655714Skris to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 357755714Skris suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 357855714Skris and crypto/des/ede_cbcm_enc.c. 357955714Skris [Steve Henson] 358055714Skris 358155714Skris *) DES quad checksum was broken on big-endian architectures. Fixed. 358255714Skris [Ben Laurie] 358355714Skris 358455714Skris *) Comment out two functions in bio.h that aren't implemented. Fix up the 358555714Skris Win32 test batch file so it (might) work again. The Win32 test batch file 358655714Skris is horrible: I feel ill.... 358755714Skris [Steve Henson] 358855714Skris 358955714Skris *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 359055714Skris in e_os.h. Audit of header files to check ANSI and non ANSI 359155714Skris sections: 10 functions were absent from non ANSI section and not exported 359255714Skris from Windows DLLs. Fixed up libeay.num for new functions. 359355714Skris [Steve Henson] 359455714Skris 359555714Skris *) Make `openssl version' output lines consistent. 359655714Skris [Ralf S. Engelschall] 359755714Skris 359855714Skris *) Fix Win32 symbol export lists for BIO functions: Added 359955714Skris BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 360055714Skris to ms/libeay{16,32}.def. 360155714Skris [Ralf S. Engelschall] 360255714Skris 360355714Skris *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 360455714Skris fine under Unix and passes some trivial tests I've now added. But the 360555714Skris whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 360655714Skris added to make sure no one expects that this stuff really works in the 360755714Skris OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 360855714Skris up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 360955714Skris openssl_bio.xs. 361055714Skris [Ralf S. Engelschall] 361155714Skris 361255714Skris *) Fix the generation of two part addresses in perl. 361355714Skris [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 361455714Skris 361555714Skris *) Add config entry for Linux on MIPS. 361655714Skris [John Tobey <jtobey@channel1.com>] 361755714Skris 361855714Skris *) Make links whenever Configure is run, unless we are on Windoze. 361955714Skris [Ben Laurie] 362055714Skris 362155714Skris *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 362255714Skris Currently only issuerAltName and AuthorityKeyIdentifier make any sense 362355714Skris in CRLs. 362455714Skris [Steve Henson] 362555714Skris 362655714Skris *) Add a useful kludge to allow package maintainers to specify compiler and 362755714Skris other platforms details on the command line without having to patch the 362855714Skris Configure script everytime: One now can use ``perl Configure 362955714Skris <id>:<details>'', i.e. platform ids are allowed to have details appended 363055714Skris to them (seperated by colons). This is treated as there would be a static 363155714Skris pre-configured entry in Configure's %table under key <id> with value 363255714Skris <details> and ``perl Configure <id>'' is called. So, when you want to 363355714Skris perform a quick test-compile under FreeBSD 3.1 with pgcc and without 363455714Skris assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 363555714Skris now, which overrides the FreeBSD-elf entry on-the-fly. 363655714Skris [Ralf S. Engelschall] 363755714Skris 363855714Skris *) Disable new TLS1 ciphersuites by default: they aren't official yet. 363955714Skris [Ben Laurie] 364055714Skris 364155714Skris *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 364255714Skris on the `perl Configure ...' command line. This way one can compile 364355714Skris OpenSSL libraries with Position Independent Code (PIC) which is needed 364455714Skris for linking it into DSOs. 364555714Skris [Ralf S. Engelschall] 364655714Skris 364755714Skris *) Remarkably, export ciphers were totally broken and no-one had noticed! 364855714Skris Fixed. 364955714Skris [Ben Laurie] 365055714Skris 365155714Skris *) Cleaned up the LICENSE document: The official contact for any license 365255714Skris questions now is the OpenSSL core team under openssl-core@openssl.org. 365355714Skris And add a paragraph about the dual-license situation to make sure people 365455714Skris recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 365555714Skris to the OpenSSL toolkit. 365655714Skris [Ralf S. Engelschall] 365755714Skris 365855714Skris *) General source tree makefile cleanups: Made `making xxx in yyy...' 365955714Skris display consistent in the source tree and replaced `/bin/rm' by `rm'. 366055714Skris Additonally cleaned up the `make links' target: Remove unnecessary 366155714Skris semicolons, subsequent redundant removes, inline point.sh into mklink.sh 366255714Skris to speed processing and no longer clutter the display with confusing 366355714Skris stuff. Instead only the actually done links are displayed. 366455714Skris [Ralf S. Engelschall] 366555714Skris 366655714Skris *) Permit null encryption ciphersuites, used for authentication only. It used 366755714Skris to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 366855714Skris It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 366955714Skris encryption. 367055714Skris [Ben Laurie] 367155714Skris 367255714Skris *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 367355714Skris signed attributes when verifying signatures (this would break them), 367455714Skris the detached data encoding was wrong and public keys obtained using 367555714Skris X509_get_pubkey() weren't freed. 367655714Skris [Steve Henson] 367755714Skris 367855714Skris *) Add text documentation for the BUFFER functions. Also added a work around 367955714Skris to a Win95 console bug. This was triggered by the password read stuff: the 368055714Skris last character typed gets carried over to the next fread(). If you were 368155714Skris generating a new cert request using 'req' for example then the last 368255714Skris character of the passphrase would be CR which would then enter the first 368355714Skris field as blank. 368455714Skris [Steve Henson] 368555714Skris 368655714Skris *) Added the new `Includes OpenSSL Cryptography Software' button as 368755714Skris doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 368855714Skris button and can be used by applications based on OpenSSL to show the 368955714Skris relationship to the OpenSSL project. 369055714Skris [Ralf S. Engelschall] 369155714Skris 369255714Skris *) Remove confusing variables in function signatures in files 369355714Skris ssl/ssl_lib.c and ssl/ssl.h. 369455714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 369555714Skris 369655714Skris *) Don't install bss_file.c under PREFIX/include/ 369755714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 369855714Skris 369955714Skris *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 370055714Skris functions that return function pointers and has support for NT specific 370155714Skris stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 370255714Skris #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 370355714Skris unsigned to signed types: this was killing the Win32 compile. 370455714Skris [Steve Henson] 370555714Skris 370655714Skris *) Add new certificate file to stack functions, 370755714Skris SSL_add_dir_cert_subjects_to_stack() and 370855714Skris SSL_add_file_cert_subjects_to_stack(). These largely supplant 370955714Skris SSL_load_client_CA_file(), and can be used to add multiple certs easily 371055714Skris to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 371155714Skris This means that Apache-SSL and similar packages don't have to mess around 371255714Skris to add as many CAs as they want to the preferred list. 371355714Skris [Ben Laurie] 371455714Skris 371555714Skris *) Experiment with doxygen documentation. Currently only partially applied to 371655714Skris ssl/ssl_lib.c. 371755714Skris See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 371855714Skris openssl.doxy as the configuration file. 371955714Skris [Ben Laurie] 372055714Skris 372155714Skris *) Get rid of remaining C++-style comments which strict C compilers hate. 372255714Skris [Ralf S. Engelschall, pointed out by Carlos Amengual] 372355714Skris 372455714Skris *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 372555714Skris compiled in by default: it has problems with large keys. 372655714Skris [Steve Henson] 372755714Skris 372855714Skris *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 372955714Skris DH private keys and/or callback functions which directly correspond to 373055714Skris their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 373155714Skris is needed for applications which have to configure certificates on a 373255714Skris per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 373355714Skris (e.g. s_server). 373455714Skris For the RSA certificate situation is makes no difference, but 373555714Skris for the DSA certificate situation this fixes the "no shared cipher" 373655714Skris problem where the OpenSSL cipher selection procedure failed because the 373755714Skris temporary keys were not overtaken from the context and the API provided 373855714Skris no way to reconfigure them. 373955714Skris The new functions now let applications reconfigure the stuff and they 374055714Skris are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 374155714Skris SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 374255714Skris non-public-API function ssl_cert_instantiate() is used as a helper 374355714Skris function and also to reduce code redundancy inside ssl_rsa.c. 374455714Skris [Ralf S. Engelschall] 374555714Skris 374655714Skris *) Move s_server -dcert and -dkey options out of the undocumented feature 374755714Skris area because they are useful for the DSA situation and should be 374855714Skris recognized by the users. 374955714Skris [Ralf S. Engelschall] 375055714Skris 375155714Skris *) Fix the cipher decision scheme for export ciphers: the export bits are 375255714Skris *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 375355714Skris SSL_EXP_MASK. So, the original variable has to be used instead of the 375455714Skris already masked variable. 375555714Skris [Richard Levitte <levitte@stacken.kth.se>] 375655714Skris 375755714Skris *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 375855714Skris [Richard Levitte <levitte@stacken.kth.se>] 375955714Skris 376055714Skris *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 376155714Skris from `int' to `unsigned int' because it's a length and initialized by 376255714Skris EVP_DigestFinal() which expects an `unsigned int *'. 376355714Skris [Richard Levitte <levitte@stacken.kth.se>] 376455714Skris 376555714Skris *) Don't hard-code path to Perl interpreter on shebang line of Configure 376655714Skris script. Instead use the usual Shell->Perl transition trick. 376755714Skris [Ralf S. Engelschall] 376855714Skris 376955714Skris *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 377055714Skris (in addition to RSA certificates) to match the behaviour of `openssl dsa 377155714Skris -noout -modulus' as it's already the case for `openssl rsa -noout 377255714Skris -modulus'. For RSA the -modulus is the real "modulus" while for DSA 377355714Skris currently the public key is printed (a decision which was already done by 377455714Skris `openssl dsa -modulus' in the past) which serves a similar purpose. 377555714Skris Additionally the NO_RSA no longer completely removes the whole -modulus 377655714Skris option; it now only avoids using the RSA stuff. Same applies to NO_DSA 377755714Skris now, too. 377855714Skris [Ralf S. Engelschall] 377955714Skris 378055714Skris *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 378155714Skris BIO. See the source (crypto/evp/bio_ok.c) for more info. 378255714Skris [Arne Ansper <arne@ats.cyber.ee>] 378355714Skris 378455714Skris *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 378555714Skris to be added. Now both 'req' and 'ca' can use new objects defined in the 378655714Skris config file. 378755714Skris [Steve Henson] 378855714Skris 378955714Skris *) Add cool BIO that does syslog (or event log on NT). 379055714Skris [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 379155714Skris 379255714Skris *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 379355714Skris TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 379455714Skris TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 379555714Skris Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 379655714Skris [Ben Laurie] 379755714Skris 379855714Skris *) Add preliminary config info for new extension code. 379955714Skris [Steve Henson] 380055714Skris 380155714Skris *) Make RSA_NO_PADDING really use no padding. 380255714Skris [Ulf Moeller <ulf@fitug.de>] 380355714Skris 380455714Skris *) Generate errors when private/public key check is done. 380555714Skris [Ben Laurie] 380655714Skris 380755714Skris *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 380855714Skris for some CRL extensions and new objects added. 380955714Skris [Steve Henson] 381055714Skris 381155714Skris *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 381255714Skris key usage extension and fuller support for authority key id. 381355714Skris [Steve Henson] 381455714Skris 381555714Skris *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 381655714Skris padding method for RSA, which is recommended for new applications in PKCS 381755714Skris #1 v2.0 (RFC 2437, October 1998). 381855714Skris OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 381955714Skris foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 382055714Skris against Bleichbacher's attack on RSA. 382155714Skris [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 382255714Skris Ben Laurie] 382355714Skris 382455714Skris *) Updates to the new SSL compression code 382555714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 382655714Skris 382755714Skris *) Fix so that the version number in the master secret, when passed 382855714Skris via RSA, checks that if TLS was proposed, but we roll back to SSLv3 382955714Skris (because the server will not accept higher), that the version number 383055714Skris is 0x03,0x01, not 0x03,0x00 383155714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 383255714Skris 383355714Skris *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 383455714Skris leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 383555714Skris in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 383655714Skris [Steve Henson] 383755714Skris 383855714Skris *) Support for RAW extensions where an arbitrary extension can be 383955714Skris created by including its DER encoding. See apps/openssl.cnf for 384055714Skris an example. 384155714Skris [Steve Henson] 384255714Skris 384355714Skris *) Make sure latest Perl versions don't interpret some generated C array 384455714Skris code as Perl array code in the crypto/err/err_genc.pl script. 384555714Skris [Lars Weber <3weber@informatik.uni-hamburg.de>] 384655714Skris 384755714Skris *) Modify ms/do_ms.bat to not generate assembly language makefiles since 384855714Skris not many people have the assembler. Various Win32 compilation fixes and 384955714Skris update to the INSTALL.W32 file with (hopefully) more accurate Win32 385055714Skris build instructions. 385155714Skris [Steve Henson] 385255714Skris 385355714Skris *) Modify configure script 'Configure' to automatically create crypto/date.h 385455714Skris file under Win32 and also build pem.h from pem.org. New script 385555714Skris util/mkfiles.pl to create the MINFO file on environments that can't do a 385655714Skris 'make files': perl util/mkfiles.pl >MINFO should work. 385755714Skris [Steve Henson] 385855714Skris 385955714Skris *) Major rework of DES function declarations, in the pursuit of correctness 386055714Skris and purity. As a result, many evil casts evaporated, and some weirdness, 386155714Skris too. You may find this causes warnings in your code. Zapping your evil 386255714Skris casts will probably fix them. Mostly. 386355714Skris [Ben Laurie] 386455714Skris 386555714Skris *) Fix for a typo in asn1.h. Bug fix to object creation script 386655714Skris obj_dat.pl. It considered a zero in an object definition to mean 386755714Skris "end of object": none of the objects in objects.h have any zeros 386855714Skris so it wasn't spotted. 386955714Skris [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 387055714Skris 387155714Skris *) Add support for Triple DES Cipher Block Chaining with Output Feedback 387255714Skris Masking (CBCM). In the absence of test vectors, the best I have been able 387355714Skris to do is check that the decrypt undoes the encrypt, so far. Send me test 387455714Skris vectors if you have them. 387555714Skris [Ben Laurie] 387655714Skris 387755714Skris *) Correct calculation of key length for export ciphers (too much space was 387855714Skris allocated for null ciphers). This has not been tested! 387955714Skris [Ben Laurie] 388055714Skris 388155714Skris *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 388255714Skris message is now correct (it understands "crypto" and "ssl" on its 388355714Skris command line). There is also now an "update" option. This will update 388455714Skris the util/ssleay.num and util/libeay.num files with any new functions. 388555714Skris If you do a: 388655714Skris perl util/mkdef.pl crypto ssl update 388755714Skris it will update them. 388855714Skris [Steve Henson] 388955714Skris 389055714Skris *) Overhauled the Perl interface (perl/*): 389155714Skris - ported BN stuff to OpenSSL's different BN library 389255714Skris - made the perl/ source tree CVS-aware 389355714Skris - renamed the package from SSLeay to OpenSSL (the files still contain 389455714Skris their history because I've copied them in the repository) 389555714Skris - removed obsolete files (the test scripts will be replaced 389655714Skris by better Test::Harness variants in the future) 389755714Skris [Ralf S. Engelschall] 389855714Skris 389955714Skris *) First cut for a very conservative source tree cleanup: 390055714Skris 1. merge various obsolete readme texts into doc/ssleay.txt 390155714Skris where we collect the old documents and readme texts. 390255714Skris 2. remove the first part of files where I'm already sure that we no 390355714Skris longer need them because of three reasons: either they are just temporary 390455714Skris files which were left by Eric or they are preserved original files where 390555714Skris I've verified that the diff is also available in the CVS via "cvs diff 390655714Skris -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 390755714Skris the crypto/md/ stuff). 390855714Skris [Ralf S. Engelschall] 390955714Skris 391055714Skris *) More extension code. Incomplete support for subject and issuer alt 391155714Skris name, issuer and authority key id. Change the i2v function parameters 391255714Skris and add an extra 'crl' parameter in the X509V3_CTX structure: guess 391355714Skris what that's for :-) Fix to ASN1 macro which messed up 391455714Skris IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 391555714Skris [Steve Henson] 391655714Skris 391755714Skris *) Preliminary support for ENUMERATED type. This is largely copied from the 391855714Skris INTEGER code. 391955714Skris [Steve Henson] 392055714Skris 392155714Skris *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 392255714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 392355714Skris 392455714Skris *) Make sure `make rehash' target really finds the `openssl' program. 392555714Skris [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 392655714Skris 392755714Skris *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 392855714Skris like to hear about it if this slows down other processors. 392955714Skris [Ben Laurie] 393055714Skris 393155714Skris *) Add CygWin32 platform information to Configure script. 393255714Skris [Alan Batie <batie@aahz.jf.intel.com>] 393355714Skris 393455714Skris *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 393555714Skris [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 393655714Skris 393755714Skris *) New program nseq to manipulate netscape certificate sequences 393855714Skris [Steve Henson] 393955714Skris 394055714Skris *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 394155714Skris few typos. 394255714Skris [Steve Henson] 394355714Skris 394455714Skris *) Fixes to BN code. Previously the default was to define BN_RECURSION 394555714Skris but the BN code had some problems that would cause failures when 394655714Skris doing certificate verification and some other functions. 394755714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 394855714Skris 394955714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 395055714Skris [Steve Henson] 395155714Skris 395255714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 395355714Skris [Steve Henson] 395455714Skris 395555714Skris *) Add several PKIX and private extended key usage OIDs. 395655714Skris [Steve Henson] 395755714Skris 395855714Skris *) Modify the 'ca' program to handle the new extension code. Modify 395955714Skris openssl.cnf for new extension format, add comments. 396055714Skris [Steve Henson] 396155714Skris 396255714Skris *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 396355714Skris and add a sample to openssl.cnf so req -x509 now adds appropriate 396455714Skris CA extensions. 396555714Skris [Steve Henson] 396655714Skris 396755714Skris *) Continued X509 V3 changes. Add to other makefiles, integrate with the 396855714Skris error code, add initial support to X509_print() and x509 application. 396955714Skris [Steve Henson] 397055714Skris 397155714Skris *) Takes a deep breath and start addding X509 V3 extension support code. Add 397255714Skris files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 397355714Skris stuff is currently isolated and isn't even compiled yet. 397455714Skris [Steve Henson] 397555714Skris 397655714Skris *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 397755714Skris ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 397855714Skris Removed the versions check from X509 routines when loading extensions: 397955714Skris this allows certain broken certificates that don't set the version 398055714Skris properly to be processed. 398155714Skris [Steve Henson] 398255714Skris 398355714Skris *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 398455714Skris Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 398555714Skris can still be regenerated with "make depend". 398655714Skris [Ben Laurie] 398755714Skris 398855714Skris *) Spelling mistake in C version of CAST-128. 398955714Skris [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 399055714Skris 399155714Skris *) Changes to the error generation code. The perl script err-code.pl 399255714Skris now reads in the old error codes and retains the old numbers, only 399355714Skris adding new ones if necessary. It also only changes the .err files if new 399455714Skris codes are added. The makefiles have been modified to only insert errors 399555714Skris when needed (to avoid needlessly modifying header files). This is done 399655714Skris by only inserting errors if the .err file is newer than the auto generated 399755714Skris C file. To rebuild all the error codes from scratch (the old behaviour) 399855714Skris either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 399955714Skris or delete all the .err files. 400055714Skris [Steve Henson] 400155714Skris 400255714Skris *) CAST-128 was incorrectly implemented for short keys. The C version has 400355714Skris been fixed, but is untested. The assembler versions are also fixed, but 400455714Skris new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 400555714Skris to regenerate it if needed. 400655714Skris [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 400755714Skris Hagino <itojun@kame.net>] 400855714Skris 400955714Skris *) File was opened incorrectly in randfile.c. 401055714Skris [Ulf M�ller <ulf@fitug.de>] 401155714Skris 401255714Skris *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 401355714Skris functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 401455714Skris GeneralizedTime. ASN1_TIME is the proper type used in certificates et 401555714Skris al: it's just almost always a UTCTime. Note this patch adds new error 401655714Skris codes so do a "make errors" if there are problems. 401755714Skris [Steve Henson] 401855714Skris 401955714Skris *) Correct Linux 1 recognition in config. 402055714Skris [Ulf M�ller <ulf@fitug.de>] 402155714Skris 402255714Skris *) Remove pointless MD5 hash when using DSA keys in ca. 402355714Skris [Anonymous <nobody@replay.com>] 402455714Skris 402555714Skris *) Generate an error if given an empty string as a cert directory. Also 402655714Skris generate an error if handed NULL (previously returned 0 to indicate an 402755714Skris error, but didn't set one). 402855714Skris [Ben Laurie, reported by Anonymous <nobody@replay.com>] 402955714Skris 403055714Skris *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 403155714Skris [Ben Laurie] 403255714Skris 403355714Skris *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 403455714Skris parameters. This was causing a warning which killed off the Win32 compile. 403555714Skris [Steve Henson] 403655714Skris 403755714Skris *) Remove C++ style comments from crypto/bn/bn_local.h. 403855714Skris [Neil Costigan <neil.costigan@celocom.com>] 403955714Skris 404055714Skris *) The function OBJ_txt2nid was broken. It was supposed to return a nid 404155714Skris based on a text string, looking up short and long names and finally 404255714Skris "dot" format. The "dot" format stuff didn't work. Added new function 404355714Skris OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 404455714Skris OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 404555714Skris OID is not part of the table. 404655714Skris [Steve Henson] 404755714Skris 404855714Skris *) Add prototypes to X509 lookup/verify methods, fixing a bug in 404955714Skris X509_LOOKUP_by_alias(). 405055714Skris [Ben Laurie] 405155714Skris 405255714Skris *) Sort openssl functions by name. 405355714Skris [Ben Laurie] 405455714Skris 405555714Skris *) Get the gendsa program working (hopefully) and add it to app list. Remove 405655714Skris encryption from sample DSA keys (in case anyone is interested the password 405755714Skris was "1234"). 405855714Skris [Steve Henson] 405955714Skris 406055714Skris *) Make _all_ *_free functions accept a NULL pointer. 406155714Skris [Frans Heymans <fheymans@isaserver.be>] 406255714Skris 406355714Skris *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 406455714Skris NULL pointers. 406555714Skris [Anonymous <nobody@replay.com>] 406655714Skris 406755714Skris *) s_server should send the CAfile as acceptable CAs, not its own cert. 406855714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 406955714Skris 407055714Skris *) Don't blow it for numeric -newkey arguments to apps/req. 407155714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 407255714Skris 407355714Skris *) Temp key "for export" tests were wrong in s3_srvr.c. 407455714Skris [Anonymous <nobody@replay.com>] 407555714Skris 407655714Skris *) Add prototype for temp key callback functions 407755714Skris SSL_CTX_set_tmp_{rsa,dh}_callback(). 407855714Skris [Ben Laurie] 407955714Skris 408055714Skris *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 408155714Skris DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 408255714Skris [Steve Henson] 408355714Skris 408455714Skris *) X509_name_add_entry() freed the wrong thing after an error. 408555714Skris [Arne Ansper <arne@ats.cyber.ee>] 408655714Skris 408755714Skris *) rsa_eay.c would attempt to free a NULL context. 408855714Skris [Arne Ansper <arne@ats.cyber.ee>] 408955714Skris 409055714Skris *) BIO_s_socket() had a broken should_retry() on Windoze. 409155714Skris [Arne Ansper <arne@ats.cyber.ee>] 409255714Skris 409355714Skris *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 409455714Skris [Arne Ansper <arne@ats.cyber.ee>] 409555714Skris 409655714Skris *) Make sure the already existing X509_STORE->depth variable is initialized 409755714Skris in X509_STORE_new(), but document the fact that this variable is still 409855714Skris unused in the certificate verification process. 409955714Skris [Ralf S. Engelschall] 410055714Skris 410155714Skris *) Fix the various library and apps files to free up pkeys obtained from 410255714Skris X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 410355714Skris [Steve Henson] 410455714Skris 410555714Skris *) Fix reference counting in X509_PUBKEY_get(). This makes 410655714Skris demos/maurice/example2.c work, amongst others, probably. 410755714Skris [Steve Henson and Ben Laurie] 410855714Skris 410955714Skris *) First cut of a cleanup for apps/. First the `ssleay' program is now named 411055714Skris `openssl' and second, the shortcut symlinks for the `openssl <command>' 411155714Skris are no longer created. This way we have a single and consistent command 411255714Skris line interface `openssl <command>', similar to `cvs <command>'. 411355714Skris [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 411455714Skris 411555714Skris *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 411655714Skris BIT STRING wrapper always have zero unused bits. 411755714Skris [Steve Henson] 411855714Skris 411955714Skris *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 412055714Skris [Steve Henson] 412155714Skris 412255714Skris *) Make the top-level INSTALL documentation easier to understand. 412355714Skris [Paul Sutton] 412455714Skris 412555714Skris *) Makefiles updated to exit if an error occurs in a sub-directory 412655714Skris make (including if user presses ^C) [Paul Sutton] 412755714Skris 412855714Skris *) Make Montgomery context stuff explicit in RSA data structure. 412955714Skris [Ben Laurie] 413055714Skris 413155714Skris *) Fix build order of pem and err to allow for generated pem.h. 413255714Skris [Ben Laurie] 413355714Skris 413455714Skris *) Fix renumbering bug in X509_NAME_delete_entry(). 413555714Skris [Ben Laurie] 413655714Skris 413755714Skris *) Enhanced the err-ins.pl script so it makes the error library number 413855714Skris global and can add a library name. This is needed for external ASN1 and 413955714Skris other error libraries. 414055714Skris [Steve Henson] 414155714Skris 414255714Skris *) Fixed sk_insert which never worked properly. 414355714Skris [Steve Henson] 414455714Skris 414555714Skris *) Fix ASN1 macros so they can handle indefinite length construted 414655714Skris EXPLICIT tags. Some non standard certificates use these: they can now 414755714Skris be read in. 414855714Skris [Steve Henson] 414955714Skris 415055714Skris *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 415155714Skris into a single doc/ssleay.txt bundle. This way the information is still 415255714Skris preserved but no longer messes up this directory. Now it's new room for 415355714Skris the new set of documenation files. 415455714Skris [Ralf S. Engelschall] 415555714Skris 415655714Skris *) SETs were incorrectly DER encoded. This was a major pain, because they 415755714Skris shared code with SEQUENCEs, which aren't coded the same. This means that 415855714Skris almost everything to do with SETs or SEQUENCEs has either changed name or 415955714Skris number of arguments. 416055714Skris [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 416155714Skris 416255714Skris *) Fix test data to work with the above. 416355714Skris [Ben Laurie] 416455714Skris 416555714Skris *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 416655714Skris was already fixed by Eric for 0.9.1 it seems. 416755714Skris [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 416855714Skris 416955714Skris *) Autodetect FreeBSD3. 417055714Skris [Ben Laurie] 417155714Skris 417255714Skris *) Fix various bugs in Configure. This affects the following platforms: 417355714Skris nextstep 417455714Skris ncr-scde 417555714Skris unixware-2.0 417655714Skris unixware-2.0-pentium 417755714Skris sco5-cc. 417855714Skris [Ben Laurie] 417955714Skris 418055714Skris *) Eliminate generated files from CVS. Reorder tests to regenerate files 418155714Skris before they are needed. 418255714Skris [Ben Laurie] 418355714Skris 418455714Skris *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 418555714Skris [Ben Laurie] 418655714Skris 418755714Skris 418855714Skris Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 418955714Skris 419055714Skris *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 419155714Skris changed SSLeay to OpenSSL in version strings. 419255714Skris [Ralf S. Engelschall] 419355714Skris 419455714Skris *) Some fixups to the top-level documents. 419555714Skris [Paul Sutton] 419655714Skris 419755714Skris *) Fixed the nasty bug where rsaref.h was not found under compile-time 419855714Skris because the symlink to include/ was missing. 419955714Skris [Ralf S. Engelschall] 420055714Skris 420155714Skris *) Incorporated the popular no-RSA/DSA-only patches 420255714Skris which allow to compile a RSA-free SSLeay. 420355714Skris [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 420455714Skris 420555714Skris *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 420655714Skris when "ssleay" is still not found. 420755714Skris [Ralf S. Engelschall] 420855714Skris 420955714Skris *) Added more platforms to Configure: Cray T3E, HPUX 11, 421055714Skris [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 421155714Skris 421255714Skris *) Updated the README file. 421355714Skris [Ralf S. Engelschall] 421455714Skris 421555714Skris *) Added various .cvsignore files in the CVS repository subdirs 421655714Skris to make a "cvs update" really silent. 421755714Skris [Ralf S. Engelschall] 421855714Skris 421955714Skris *) Recompiled the error-definition header files and added 422055714Skris missing symbols to the Win32 linker tables. 422155714Skris [Ralf S. Engelschall] 422255714Skris 422355714Skris *) Cleaned up the top-level documents; 422455714Skris o new files: CHANGES and LICENSE 422555714Skris o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 422655714Skris o merged COPYRIGHT into LICENSE 422755714Skris o removed obsolete TODO file 422855714Skris o renamed MICROSOFT to INSTALL.W32 422955714Skris [Ralf S. Engelschall] 423055714Skris 423155714Skris *) Removed dummy files from the 0.9.1b source tree: 423255714Skris crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 423355714Skris crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 423455714Skris crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 423555714Skris crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 423655714Skris util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 423755714Skris [Ralf S. Engelschall] 423855714Skris 423955714Skris *) Added various platform portability fixes. 424055714Skris [Mark J. Cox] 424155714Skris 424255714Skris *) The Genesis of the OpenSSL rpject: 424355714Skris We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 424455714Skris Young and Tim J. Hudson created while they were working for C2Net until 424555714Skris summer 1998. 424655714Skris [The OpenSSL Project] 424755714Skris 424855714Skris 424955714Skris Changes between 0.9.0b and 0.9.1b [not released] 425055714Skris 425155714Skris *) Updated a few CA certificates under certs/ 425255714Skris [Eric A. Young] 425355714Skris 425455714Skris *) Changed some BIGNUM api stuff. 425555714Skris [Eric A. Young] 425655714Skris 425755714Skris *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 425855714Skris DGUX x86, Linux Alpha, etc. 425955714Skris [Eric A. Young] 426055714Skris 426155714Skris *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 426255714Skris RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 426355714Skris available). 426455714Skris [Eric A. Young] 426555714Skris 426655714Skris *) Add -strparse option to asn1pars program which parses nested 426755714Skris binary structures 426855714Skris [Dr Stephen Henson <shenson@bigfoot.com>] 426955714Skris 427055714Skris *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 427155714Skris [Eric A. Young] 427255714Skris 427355714Skris *) DSA fix for "ca" program. 427455714Skris [Eric A. Young] 427555714Skris 427655714Skris *) Added "-genkey" option to "dsaparam" program. 427755714Skris [Eric A. Young] 427855714Skris 427955714Skris *) Added RIPE MD160 (rmd160) message digest. 428055714Skris [Eric A. Young] 428155714Skris 428255714Skris *) Added -a (all) option to "ssleay version" command. 428355714Skris [Eric A. Young] 428455714Skris 428555714Skris *) Added PLATFORM define which is the id given to Configure. 428655714Skris [Eric A. Young] 428755714Skris 428855714Skris *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 428955714Skris [Eric A. Young] 429055714Skris 429155714Skris *) Extended the ASN.1 parser routines. 429255714Skris [Eric A. Young] 429355714Skris 429455714Skris *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 429555714Skris [Eric A. Young] 429655714Skris 429755714Skris *) Added a BN_CTX to the BN library. 429855714Skris [Eric A. Young] 429955714Skris 430055714Skris *) Fixed the weak key values in DES library 430155714Skris [Eric A. Young] 430255714Skris 430355714Skris *) Changed API in EVP library for cipher aliases. 430455714Skris [Eric A. Young] 430555714Skris 430655714Skris *) Added support for RC2/64bit cipher. 430755714Skris [Eric A. Young] 430855714Skris 430955714Skris *) Converted the lhash library to the crypto/mem.c functions. 431055714Skris [Eric A. Young] 431155714Skris 431255714Skris *) Added more recognized ASN.1 object ids. 431355714Skris [Eric A. Young] 431455714Skris 431555714Skris *) Added more RSA padding checks for SSL/TLS. 431655714Skris [Eric A. Young] 431755714Skris 431855714Skris *) Added BIO proxy/filter functionality. 431955714Skris [Eric A. Young] 432055714Skris 432155714Skris *) Added extra_certs to SSL_CTX which can be used 432255714Skris send extra CA certificates to the client in the CA cert chain sending 432355714Skris process. It can be configured with SSL_CTX_add_extra_chain_cert(). 432455714Skris [Eric A. Young] 432555714Skris 432655714Skris *) Now Fortezza is denied in the authentication phase because 432755714Skris this is key exchange mechanism is not supported by SSLeay at all. 432855714Skris [Eric A. Young] 432955714Skris 433055714Skris *) Additional PKCS1 checks. 433155714Skris [Eric A. Young] 433255714Skris 433355714Skris *) Support the string "TLSv1" for all TLS v1 ciphers. 433455714Skris [Eric A. Young] 433555714Skris 433655714Skris *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 433755714Skris ex_data index of the SSL context in the X509_STORE_CTX ex_data. 433855714Skris [Eric A. Young] 433955714Skris 434055714Skris *) Fixed a few memory leaks. 434155714Skris [Eric A. Young] 434255714Skris 434355714Skris *) Fixed various code and comment typos. 434455714Skris [Eric A. Young] 434555714Skris 434655714Skris *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 434755714Skris bytes sent in the client random. 434855714Skris [Edward Bishop <ebishop@spyglass.com>] 434955714Skris 4350