[aarch64] Use better dwarf2 encoding factors

[Andrew Pinski]

On Wed, Aug 6, 2014 at 3:39 PM, Richard Henderson <rth@redhat.com> wrote:
> These constants were copied from the 32-bit ARM port.  But for AArch64, we only
> have 32-bit instructions, and (even for ilp32) 64-bit frame related stores.

Yes that is correct about ILP32.

>
> Using the proper factors results in smaller unwind info.
>
> Ok?

>From my point of view this is fine.

Thanks,
Andrew

>
>
> r~