This is the mail archive of the
binutils@sourceware.org
mailing list for the binutils project.
Re: PPC code generation question
- From: Alan Modra <amodra at bigpond dot net dot au>
- To: Jan Beulich <jbeulich at novell dot com>
- Cc: binutils at sourceware dot org
- Date: Mon, 25 Feb 2008 21:03:58 +1030
- Subject: Re: PPC code generation question
- References: <47C281A8.76E4.0078.0@novell.com>
On Mon, Feb 25, 2008 at 07:51:52AM +0000, Jan Beulich wrote:
> Is it intentional, unavoidable, or a bug that this code
>
> .text
> _start:
> bge+ .
> bge- .-4
>
> generates different a binary different from this
>
> .text
> _start:
> bge+ .
> bge- .-4
> .machine power6
>
> While I suppose that the difference is not very significant, it still seems
> odd.
I'd say it is a bug. Fixup processing accesses ppc_cpu (*), so gets
the value at the end of the file. We probably ought to save the
current value of ppc_cpu in the fixup and use that.
*) md_apply_fix calls ppc_insert_operand which uses ppc_cpu.
--
Alan Modra
Australia Development Lab, IBM