[Mesa-dev] [PATCH 25/57] i965/fs: Stop using fs_reg::in_range() in favor of regions_overlap().

[Iago Toral]

On Wed, 2016-09-07 at 18:48 -0700, Francisco Jerez wrote:
> Its only use left in the FS back-end should be using
> regions_overlap()
> instead to avoid getting a false negative result in cases where
> source
> and destination overlap but the former starts before the latter in
> the
> VGRF file.
> ---
>  src/mesa/drivers/dri/i965/brw_fs_nir.cpp | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/src/mesa/drivers/dri/i965/brw_fs_nir.cpp
> b/src/mesa/drivers/dri/i965/brw_fs_nir.cpp
> index 42ed131..b06606b 100644
> --- a/src/mesa/drivers/dri/i965/brw_fs_nir.cpp
> +++ b/src/mesa/drivers/dri/i965/brw_fs_nir.cpp
> @@ -4820,7 +4820,8 @@ shuffle_64bit_data_for_32bit_write(const
> fs_builder &bld,
>     assert(type_sz(src.type) == 8);
>     assert(type_sz(dst.type) == 4);
>  
> -   assert(!src.in_range(dst, 2 * components * bld.dispatch_width() /
> 8));
> +   assert(!regions_overlap(dst, 2 *
> dst.component_size(bld.dispatch_width()),
> +                           src,
> src.component_size(bld.dispatch_width())));

I think you need to multiply the sizes of the regions for dst and src
by the number of components being shuffled, like in the original
assert, otherwise we would only be testing if there is overlap for the
first component.

>     for (unsigned i = 0; i < components; i++) {
>        const fs_reg component_i = offset(src, bld, i);