[Mesa-dev] [PATCH v2 1/3] nir/split_var_copies: handle IMAGE and SAMPLER for bindless vars

Jason Ekstrand jason at jlekstrand.net
Tue Apr 10 15:12:14 UTC 2018


I still don't see anything to make nir_validate not fail out on you if it
sees a read or a write to/from an IMAGE or SAMPLER.

On Tue, Apr 10, 2018 at 8:05 AM, Karol Herbst <kherbst at redhat.com> wrote:

> v2: fix assertion for bindless to non bindless assignments
>
> Signed-off-by: Karol Herbst <kherbst at redhat.com>
> ---
>  src/compiler/nir/nir_split_var_copies.c | 8 +++++++-
>  1 file changed, 7 insertions(+), 1 deletion(-)
>
> diff --git a/src/compiler/nir/nir_split_var_copies.c
> b/src/compiler/nir/nir_split_var_copies.c
> index bc3ceedbdb8..e592754d770 100644
> --- a/src/compiler/nir/nir_split_var_copies.c
> +++ b/src/compiler/nir/nir_split_var_copies.c
> @@ -222,8 +222,9 @@ split_var_copies_block(nir_block *block, struct
> split_var_copies_state *state)
>        nir_deref_var *src_head = intrinsic->variables[1];
>        nir_deref *dest_tail = nir_deref_tail(&dest_head->deref);
>        nir_deref *src_tail = nir_deref_tail(&src_head->deref);
> +      enum glsl_base_type base_type = glsl_get_base_type(src_tail->type);
>
> -      switch (glsl_get_base_type(src_tail->type)) {
> +      switch (base_type) {
>        case GLSL_TYPE_ARRAY:
>        case GLSL_TYPE_STRUCT:
>           split_var_copy_instr(intrinsic, dest_head, src_head,
> @@ -241,6 +242,11 @@ split_var_copies_block(nir_block *block, struct
> split_var_copies_state *state)
>              ralloc_steal(state->dead_ctx, instr);
>           }
>           break;
> +      /* for bindless those are uint64 */
> +      case GLSL_TYPE_IMAGE:
> +      case GLSL_TYPE_SAMPLER:
> +         assert(src_head->var->data.bindless ||
> +                glsl_get_base_type(src_head->var->type) == base_type);
>        case GLSL_TYPE_INT:
>        case GLSL_TYPE_UINT:
>        case GLSL_TYPE_INT16:
> --
> 2.14.3
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/mesa-dev/attachments/20180410/b5fbefea/attachment.html>


More information about the mesa-dev mailing list