Re: Bug#873508: sparse test failures on ppc32le (and other not so common archs)

[Luc Van Oostenryck <luc.vanoostenryck@gmail.com>]

On Wed, Sep 27, 2017 at 10:00 AM, Uwe Kleine-König
<uwe@kleine-koenig.org> wrote:
> Hello,
>
> On Tue, Sep 26, 2017 at 08:11:01PM +0200, Uwe Kleine-König wrote:
>> And ppc64 and x32 need the respective cpp defines added I think. If
>> noone beats me to it, I will look into the latter at least during the
>> next few days.
>
> Looking at ppc64, the following fixes the build:
>
> diff --git a/cgcc b/cgcc
> index a8d7b4f217fe..a1c02899c623 100755
> --- a/cgcc
> +++ b/cgcc
> @@ -286,7 +286,7 @@ sub add_specs {
>      } elsif ($spec eq 'ppc64') {
>         return (' -D__powerpc__=1 -D__PPC__=1 -D_STRING_ARCH_unaligned=1' .
>                 ' -D__powerpc64__=1 -D__PPC64__=1' .
> -               ' -D_CALL_ELF=2' .
> +               ' -D_CALL_ELF=1' .
>                 ' -m64' .
>                 &float_types (1, 1, 21, [24,8], [53,11], [113,15]));
>      } elsif ($spec eq 's390x') {
>
> I wonder if that could be right. Luc, you added =2 in
> e0306fe0b725af6e2e7ff59d7f0d99c96315791a, maybe you can comment?

Neither =2 or =1 is correct, it depends on which version of the ELF
ABI you're using.
This in turn (as I understood) is most of the time tied to fact that
you're using ppc64le
(which normally needs ELFv2) or not.

I can't look at this now but will do this evening.

-- Luc