On Sep 14, 2010, at 18:14, robin wrote:
> From: "Paul Gilmartin" <[email protected]>
> Sent: Sunday, 12 September 2010 4:57 PM
>
>> Another test:
>>
>> Active Usings: None
>> Loc Object Code Addr1 Addr2 Stmt Source Statement
>> 000000 00000 00004 1 NEGDISP CSECT
>> 2 *
>> R:1 FFFFF0 3 USING X'7FFFFFF0',R1
>> 000000 4132 1020 00010 4 LA R3,-X'7FFFFFF0'(R2)
>
> Did I miss something here?
> Should the displacement in the object code be x'010', not x'020' ?
x'020' is at least more correct than x'010'. In decimal:
-(2**31-16) - (2**31-16) = -2**32 + 32 = x'...FF00000020'
This is an overflow, not reported by HLASM, which takes the lowest
32 bits; observes that it's within displacement range, and uses
the lowest 12 bits as the displacement: x'020'.
The result is correct modulo 2**24 and modulo 2**31, but incorrect
in 64-bit addressing. Disregarding modulo arithmetic, the result
is mathematically incorrect. I'd be happier if HLASM detected
the overflow and reported Addressability Error.
-- gil
