On Tue, Sep 23, 2003 at 11:42:13AM -0700, Guy Harris wrote:
>
> On Sep 23, 2003, at 4:58 AM, Graeme Reid wrote:
> > The only known problem with it so far is that an invoke Id in a
> > return result or return error of >32k
> > will be displayed as a negative number.
>
> Well, the constrained invokeId is an unsigned integer (FT_UINT32), but
> the other invokeID is signed (FT_INT32), so values with the uppermost
> bit set will be treated as negative numbers.
>
> The first of those is handled by "dissect_per_constrained_integer()",
> which treats the integer as unsigned. The latter is handled by
> "dissect_per_integer()", which treats it as signed; I don't have any of
> the ASN.1 specs handy, so I don't know whether there needs to be, for
> example, a "dissect_per_unsigned_integer()" routine.
Well, X.680 doesn't seem to have any unsigned integer type, it just has
a signed integer type.
SNMP defines the unsigned Counter and Gauge types as
Counter ::=
[APPLICATION 1]
IMPLICIT INTEGER (0..4294967295)
Gauge ::=
[APPLICATION 2]
IMPLICIT INTEGER (0..4294967295)
Perhaps I'm misreading it, but I infer from X.690 that you might need 5
data bytes to represent that - it sounds as if the value 4294967295
requires 33 bits, given the "and then reducing this value by the
numerical value assigned to bit 8 of the first octet if that bit is set
to one" in clause 8.3.3 of X.690 (BER, DER), so the uppermost bit of a
positive integer has to be 0, not 1.
X.691 (PER), however, has both "2's-complement-binary-integer encoding"
and "non-negative-binary-integer encoding". Constrained integers are,
it appears, encoded as an index into the range of allowed values, which
are always non-negative. Unconstrained integers are (as per 10.8)
encoded as a 2's-complement-binary-integer, i.e. they're signed.
The page at
http://www.itu.int/ITU-T/asn1/database/itu-t/h/h450.1/1998/Remote-Operations-Apdus.html
says
ReturnResult{OPERATION:Operations} ::= SEQUENCE {
invokeId
InvokeId
(CONSTRAINED BY {-- must be that for an outstanding operation -- } !
RejectProblem:returnResult-unrecognizedInvocation)
(CONSTRAINED BY {-- which returns a result -- } !
RejectProblem:returnResult-resultResponseUnexpected),
result
SEQUENCE {opcode
OPERATION.&operationCode({Operations})
(CONSTRAINED BY {-- identified by invokeId -- } !
RejectProblem:returnResult-unrecognizedInvocation),
result
OPERATION.&ResultType
({Operations}{@.opcode} !
RejectProblem:returnResult-mistypedResult)} OPTIONAL
}
(CONSTRAINED BY { -- must conform to the above definition --} !
RejectProblem:general-mistypedPDU)
and "InvokeID" links to
http://www.itu.int/ITU-T/asn1/database/itu-t/h/h450.1/1998/Remote-Operations-Apdus.html#Remote-Operations-Apdus.InvokeId
which says
InvokeId ::= INTEGER
so that's a signed value, as it's unconstrained.
Is the uppermost bit of the value 1? If so, then it's negative, and
should be displayed as such.
