Patch to fix gcc.c-torture/compile/20010102-1.c on IA64 HP-UX

[Jakub Jelinek]

On Tue, Sep 16, 2008 at 02:02:37PM -0500, Peter Bergner wrote:
> On Tue, 2008-09-16 at 11:47 -0700, Steve Ellcey wrote:
> > On Tue, 2008-09-16 at 13:20 -0500, Peter Bergner wrote:
> > I guess I am not sure why setting REG_POINTER on a hard register is bad.
> > I looked at PR 36533 but it did not enlighten me.
> 
> Jakub, what was the exact problem having the REG_POINTER attribute set on
> a hard register cause on x86 again?

The problem is that unlike pseudo registers, a hard register can contain
different kinds of variables over the lifetime of a function.
So, it can be a pointer for part of a function and non-pointer for another
part of the function.  Or it can contain two different pointers with
different alignments, yet REGNO_POINTER_ALIGN is a shared value for
particular REGNO inside of the whole function.

	Jakub