[60] in Kerberos-V5-bugs
re: alpha2 request asn.1 problem
daemon@ATHENA.MIT.EDU (John T Kohl)
Wed Nov 7 11:21:46 1990
Date: Wed, 7 Nov 90 10:29:38 -0500
From: John T Kohl <jtkohl@ATHENA.MIT.EDU>
To: marks@Eng.Sun.COM, krb5-testers@ATHENA.MIT.EDU
In-Reply-To: [0058]
>When using kinit to obtain a TGT, krb5kdc core dumps with a SIGSEGV
>in some asn.1 free_translation code. Specifically:
I've seen another manifestation of this problem in testing. Some code
from KRB5_KDC__REQ__BODY2krb5_kdc_req:
for (i = 0; (i < tptr->nelem) && tptr->Ticket[i]; i++) {
The "&& <not null>" was put in because of a similar problem.
I believe it's an ASN.1/ISODE problem, since nelem *always* seems to get
set to something bogus, even when you encode something with no
additional tickets. I just ran some tests, and running
encode_KRB5_TGS__REQ and then decode_KRB5_TGS__REQ yields a different
structure, particularly the element counts on the additional__tickets is
bogus.
I don't understand enough about ISODE to figure this one out, although I
think the problem lies in the generated code for
decode_KRB5_KDC__REQ__BODY. It seems to count the # of elements in the
entire body, and use that as the count of additional tickets. Anybody
else care to examine it?
