|
|||||||||||
|
openssh krb5 password-based auth
From: Buck Huppmann <buckh(at)pobox.com>
Date: Thu Dec 26 2002 - 16:23:44 EST
Index: auth-krb5.c RCS file: /cvs/src/usr.bin/ssh/auth-krb5.c,v retrieving revision 1.9 diff -u -w -r1.9 auth-krb5.c --- auth-krb5.c 9 Sep 2002 06:48:06 -0000 1.9 +++ auth-krb5.c 26 Dec 2002 21:03:16 -0000 @@ -223,7 +223,7 @@ if (problem) goto out; - problem = krb5_cc_gen_new(authctxt->krb5_ctx, &krb5_mcc_ops, which is sorta analogous to what goes on in auth-krb4.c:
/* Try to get TGT using our password. */
r = krb_get_pw_in_tkt((char *) pw->pw_name, "", realm,
"krbtgt", realm, DEFAULT_TKT_LIFE, (char *)password);
if (r != INTK_OK) {
debug("Kerberos v4 password authentication for %s "
"failed: %s", pw->pw_name, krb_err_txt[r]);
goto failure;
}
/* Successful authentication. */
chown(tkt_string(), pw->pw_uid, pw->pw_gid);
but i'm sure if this were safe (aside from making it optional) it would be that way already, so can somebody explain the folly (apart from the spurious ID switching) of doing this? it looks like the krb5_cc_gen_new(..., &krb5_mcc_ops, ...) call goes back to the original check-in of this file, so do i need to ask the heimdal guys? thanks for the help Received on Sun Dec 29 20:21:37 2002 This archive was generated by hypermail 2.1.8 : Wed Aug 23 2006 - 13:48:28 EDT |
||||||||||
|
|||||||||||