On Fri, Nov 30, 2012 at 9:17 PM, Peter Gavin <[email protected]> wrote:
> On Fri, Nov 30, 2012 at 4:03 AM, Julius Baxter <[email protected]>
> wrote:
>>
>> Cool. I shall whip up a new arch spec.
>
>
> I started working on modifying binutils, too.  For now I've chosen 0x15 and
> 0x16 for the AECR and AESR regs, respectively, but it would be no problem to
> change it.

In the manual I've got AECR and 12 and AESR at 13.

>
> One nit: CYMACADDE should come before OVMACADDE, for consistency with the
> others.

I agree. Fixed.

>
> Another nit:  l.addc probably shouldn't trigger an exception when CY is set,
> even if the CYADDE flag is on.  That instruction is meant to be chained with
> other l.addc instructions, so it doesn't seem useful to raise the exception
> in that case.

I reckon this can happen and should cause an exception, because say
you're adding the two most significant operand words, which overflow,
but you did it with l.addc, you would still expect it to set CY and
trigger the exception if you told it to.

Cheers

Julius
_______________________________________________
OpenRISC mailing list
[email protected]
http://lists.openrisc.net/listinfo/openrisc

Reply via email to