On Tue, Nov 3, 2015 at 12:31 AM, Kenneth Graunke <kenn...@whitecape.org> wrote:
> We can't preserve dominance or live variable information.
>
> This also begs the question: what about globals? Metadata only exists
> at the nir_function_impl level, so it would seem there is no metadata
> about global variables for us to invalidate.
>
> Signed-off-by: Kenneth Graunke <kenn...@whitecape.org>
> ---
> src/glsl/nir/nir_remove_dead_variables.c | 9 +++++++--
> 1 file changed, 7 insertions(+), 2 deletions(-)
>
> diff --git a/src/glsl/nir/nir_remove_dead_variables.c
> b/src/glsl/nir/nir_remove_dead_variables.c
> index d6783e7..77b6c13 100644
> --- a/src/glsl/nir/nir_remove_dead_variables.c
> +++ b/src/glsl/nir/nir_remove_dead_variables.c
> @@ -126,8 +126,13 @@ nir_remove_dead_variables(nir_shader *shader)
> progress = remove_dead_vars(&shader->globals, live) || progress;
>
> nir_foreach_overload(shader, overload) {
> - if (overload->impl)
> - progress = remove_dead_vars(&overload->impl->locals, live) ||
> progress;
> + if (overload->impl) {
> + if (remove_dead_vars(&overload->impl->locals, live)) {
> + nir_metadata_preserve(overload->impl, nir_metadata_block_index |
> + nir_metadata_dominance);
This doesn't invalidate live_variables
> + progress = true;
> + }
> + }
> }
>
> _mesa_set_destroy(live, NULL);
> --
> 2.6.2
>
> _______________________________________________
> mesa-dev mailing list
> mesa-dev@lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/mesa-dev
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev
