diff options
Diffstat (limited to 'src/kdc/ChangeLog')
-rw-r--r-- | src/kdc/ChangeLog | 135 |
1 files changed, 135 insertions, 0 deletions
diff --git a/src/kdc/ChangeLog b/src/kdc/ChangeLog index eb42736..8711472 100644 --- a/src/kdc/ChangeLog +++ b/src/kdc/ChangeLog @@ -1,3 +1,138 @@ +2004-07-23 Tom Yu <tlyu@mit.edu> + + * kerberos_v4.c (kerberos_v4): Duplicate backdating fix for + APPL_REQUEST as well. Fix comments. + +2004-03-22 Ken Raeburn <raeburn@mit.edu> + + * network.c (delete_fd): Free pointed-to data after removing it + from the connection set. + (kill_tcp_connection): Move delete_fd call to the end. + (accept_tcp_connection): Decrement connection counter again if we + drop the incoming connection for lack of buffer space. + +2004-02-19 Ken Hornstein <kenh@mit.edu> + + * fakeka.c (main): Bug from David Thompson <thomas@cs.wisc.edu>. + Bug originally introduced during conversion from bcopy() to + memcpy(). + +2004-02-12 Sam Hartman <hartmans@mit.edu> + + * kdc_util.c (validate_tgs_request): Fix logic error + +2004-02-09 Tom Yu <tlyu@mit.edu> + + * main.c (init_realm): Apply patch from Will Fiveash to use + correct TCP listening ports. + +2004-02-06 Sam Hartman <hartmans@avalanche-breakdown.mit.edu> + + * kdc_util.c: Ignore unknown TGS options + +2004-01-05 Sam Hartman <hartmans@mit.edu> + + * kerberos_v4.c (kerberos_v4): Only backdate the rquest in the + issued ticket. Client libraries tend to verify that the + backdating falls within clockskew. a + +2003-06-03 Tom Yu <tlyu@mit.edu> + + * extern.h (master_princ): Remove realm_mkvno, realm_tgskey, + realm_tgskvno, realm_kstypes, realm_nkstypes. They're not needed + anymore. + + * main.c (finish_realm): Remove references to realm_kstypes, + realm_tgskey. + (init_realm): Don't bother with realm_kstypes. Don't bother + looking up the master kvno. Don't bother caching the TGS key. + None of these were being used. + +2003-05-30 Ken Raeburn <raeburn@mit.edu> + + * main.c (init_realm): Use KRB5_KDB_MAX_RLIFE, not + KRB5_KDB_MAX_LIFE, as default for realm's max renewable lifetime. + +2003-05-23 Sam Hartman <hartmans@mit.edu> + + * kdc_preauth.c (_make_etype_info_entry): Add flag to know if we + are producing etype_info2 so we know whether filling in s2kparams + is allowed. In the etype_info2 case support afs3 salts. + (etype_info_helper): Pass in flag + (return_etype_info2): And here + +2003-05-23 Ezra Peisach <epeisach@mit.edu> + + * kdc_preauth.c (return_etype_info2): After encoding the + etype_info2 and copying the pointers to the pa_data, free the + krb5_data pointer. + +2003-05-22 Sam Hartman <hartmans@mit.edu> + + * kdc_util.c (validate_as_request): Only reject options we + understand and believe are inappropriate for AS requests. Per + spec, unknown options are ignored. + +2003-05-14 Sam Hartman <hartmans@mit.edu> + + * kdc_preauth.c (check_padata): Allow bad_integrity to be returned to a client + +2003-05-08 Sam Hartman <hartmans@mit.edu> + + * kdc_preauth.c (return_pw_salt): Don't return pw-salt if the + client's enctype list mandates it supports enctype-info2 + +2003-05-09 Tom Yu <tlyu@mit.edu> + + * kdc_util.c (kdc_process_tgs_req): Rename getremotesubkey -> + getrecvsubkey. + +2003-05-07 Sam Hartman <hartmans@mit.edu> + + * kdc_preauth.c (get_etype_info): Patch from Sun to reorganize + code and make sure that even for md5 the database order is + preserved. + (enctype_requires_etype_info_2): new function; determines wether a + particular enctype in a client request means that the client is + required to support etype_info2 by Kerberos clarifications. + (etype_info_helper): Renamed from get_etype_info to abstract out + code in common between etype_info and etype_info2 + (get_enctype_info): Return etype info only if request contains no + enctypes that require etype_info2 + (return_etype_info2): New function. + +2003-04-02 Sam Hartman <hartmans@mit.edu> + + * kdc_preauth.c (get_etype_info): Avoid infinite loop if request + does not contain des-cbc-crc and database does + +2003-04-01 Nalin Dahyabhai <nalin@redhat.com> + + * do_tgs_req.c (process_tgs_req): Check that principal name + component 1 is present before examining it. + * kdc_util.c (krb5_is_tgs_principal, validate_tgs_request): Check + principal name length before examining components. + +2003-03-28 Tom Yu <tlyu@mit.edu> + + * kdc_preauth.c (verify_enc_timestamp): Save decryption error, in + case we get NO_MATCHING_KEY later. This allows us to log a more + sane error if an incorrect password is used for encrypting the + enc-timestamp preauth. + +2003-03-16 Sam Hartman <hartmans@mit.edu> + + * main.c (initialize_realms): Add support to call + enable_v4_crossrealm if the user wants insecure operation + + * kerberos_v4.c: Add enable_v4_crossrealm. By default krb4 + cross-realm is not allowed as it is insecure. Also, remove + support for generating krb4 tickets encrypted in 3DES as they are + insecure. + + * kdc_util.h: Define enable_v4_crossrealm, new function to enable + secure krb4 cross-realm authentication + 2003-03-05 Tom Yu <tlyu@mit.edu> * main.c (init_realm): Update call to krb5_ktdb_resolve(). |