I thought I'd just fixed this...
Just sent a patch which fixes this and a few other image bugs. I've
had it sitting in a FutureTech branch for about a week and forgot to
send it. :(
On July 14, 2018 16:26:17 Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl> wrote:
> nir_validate complains in case less than 4 components are
> provided.
>
> CC: 18.1 <mesa-sta...@lists.freedesktop.org>
> ---
> src/compiler/spirv/spirv_to_nir.c | 16 +++++++++++++---
> 1 file changed, 13 insertions(+), 3 deletions(-)
>
> diff --git a/src/compiler/spirv/spirv_to_nir.c
> b/src/compiler/spirv/spirv_to_nir.c
> index 48154303ff2..b082af57b10 100644
> --- a/src/compiler/spirv/spirv_to_nir.c
> +++ b/src/compiler/spirv/spirv_to_nir.c
> @@ -2450,10 +2450,20 @@ vtn_handle_image(struct vtn_builder *b, SpvOp opcode,
> case SpvOpAtomicStore:
> intrin->src[3] = nir_src_for_ssa(vtn_ssa_value(b, w[4])->def);
> break;
> - case SpvOpImageWrite:
> - intrin->src[3] = nir_src_for_ssa(vtn_ssa_value(b, w[3])->def);
> - break;
> + case SpvOpImageWrite: {
> + nir_ssa_def *value = vtn_ssa_value(b, w[3])->def;
>
> + if (value->num_components < 4) {
> + unsigned swiz[4];
> + for (unsigned i = 0; i < 4; i++)
> + swiz[i] = i < value->num_components ? i : 0;
> +
> + value = nir_swizzle(&b->nb, value, swiz, 4, false);
> + }
> +
> + intrin->src[3] = nir_src_for_ssa(value);
> + break;
> + }
> case SpvOpAtomicCompareExchange:
> case SpvOpAtomicIIncrement:
> case SpvOpAtomicIDecrement:
> --
> 2.18.0
>
> _______________________________________________
> mesa-dev mailing list
> mesa-dev@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/mesa-dev
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev
