On 03/12/13 21:35, Cole Robinson wrote:
> Add arm32 to the whitelist for handling lack of int128 type variants.
> Otherwise we get errors like:
>
> /target/include/asm/types.h:79:1: error: unknown type name â__int128_tâ
> typedef __int128_t dcell;
>
> This may introduce issues building on aarch64 but I have no way to test.
> ---
>
> Recently, Fedora's build system had many arm hosts added. Since we always
> cross build openbios, the RPM build is scheduled independent of the
> build machine architecture. So whenever an openbios build was scheduled
> to an arm host, things fell over.
>
> config/scripts/switch-arch | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/config/scripts/switch-arch b/config/scripts/switch-arch
> index 44f2df4..86cab3c 100755
> --- a/config/scripts/switch-arch
> +++ b/config/scripts/switch-arch
> @@ -77,7 +77,7 @@ crosscflags()
> if test "$target" = "sparc64" -o "$target" = "ia64" \
> -o "$target" = "amd64" -o "$target" = "x86_64" \
> -o "$target" = "alpha"; then
> - if test "$host" = "x86"; then
> + if test "$host" = "x86" -o "$host" = "arm"; then
> cflags="$cflags -DNEED_FAKE_INT128_T"
> elif test "$host" = "ppc" -a `uname -s` = "Darwin"; then
> cflags="$cflags -DNEED_FAKE_INT128_T"
Hi Cole,
Thanks for the patch - I don't tend to do much with the build system but
it looks fairly reasonable to me. One minor stylistic nit is that
switch-arch seems to be grouped by arch, so would it be possible to
rework so that the overall result looks like this:
if test "$host" = "x86"; then
if test "$host" = "arm"; then
cflags="$cflags -DNEED_FAKE_INT128_T"
elif test "$host" = "ppc" -a `uname -s` = "Darwin"; then
cflags="$cflags -DNEED_FAKE_INT128_T"
If you can resubmit like this (and also with a SOB) then I'm happy to
apply it.
ATB,
Mark.