Re: bootstrap failure on amd64

[Andrew Pinski <pinskia at physics dot uc dot edu>]

On Jul 27, 2004, at 12:15 AM, Steven Bosscher wrote:

> Hi,
>
> The bootstrap failure on AMD64 is caused by your patch:
> http://gcc.gnu.org/ml/gcc-cvs/2004-07/msg01260.html
>
>         * calls.c (combine_pending_stack_adjustment_and_call): Make
>         preferred_unit_stack_boundary argument unsigned.  Make
>         unadjusted_alignment unsigned.
>         (expand_call): Make preferred_stack_boundary and
>         preferred_unit_stack_boundary variables unsigned.
>         * function.c (assign_stack_local_1): Make alignment unsigned.
>         * function.h (struct function): Make stack_alignment_needed,
>         preferred_stack_boundary unsigned.
>         * config/i386/i386.c (ix86_preferred_stack_boundary): Make 
> unsigned.
>         (ix86_compute_frame_layout): Make stack_alignment_needed,
>         preferred_alignment variables unsigned.
>         * config/i386/i386.h (ix86_preferred_stack_boundary): Make 
> unsigned.
>
> Reverting it with the attached patch makes bootstrapping
> possible again for me.

[00:44] < norwoodites> it looks like the change in calls.c, 
combine_pending_stack_adjustment_and_call
[00:44] < norwoodites> mainly   HOST_WIDE_INT unadjusted_alignment;
[00:44] < norwoodites> this is just from looking into the code and 
nothing else
[00:45] < norwoodites>       if (unadjusted_alignment > 0)
[00:45] < norwoodites>         adjustment -= 
preferred_unit_stack_boundary - unadjusted_alignment;
[00:45] < norwoodites>       else
[00:45] < norwoodites>         adjustment += unadjusted_alignment;


Thanks,
Andrew Pinski