On Mon, Jul 16, 2018 at 8:32 AM Karol Herbst <kher...@redhat.com> wrote:
> On Mon, Jul 16, 2018 at 5:24 PM, Jason Ekstrand <ja...@jlekstrand.net>
> wrote:
> > Do we need to do something similar for specialization constants?
> >
>
> I don't think so. The problem here wasn't constants, but simply that
> we can end up with non 32 bit shift values inside the SPIR-V, but we
> only support 32 bit shifts in nir. It is legal SPIR-V, so we might
> just want to fix that.
>
Right, but you can have 64 or 16-bit specialization constants as well.
> > On Mon, Jul 16, 2018 at 7:28 AM Karol Herbst <kher...@redhat.com> wrote:
> >>
> >> Signed-off-by: Karol Herbst <kher...@redhat.com>
> >> ---
> >> src/compiler/spirv/vtn_alu.c | 10 ++++++++++
> >> 1 file changed, 10 insertions(+)
> >>
> >> diff --git a/src/compiler/spirv/vtn_alu.c b/src/compiler/spirv/vtn_alu.c
> >> index 5db6c7f0a87..d6f149d12e9 100644
> >> --- a/src/compiler/spirv/vtn_alu.c
> >> +++ b/src/compiler/spirv/vtn_alu.c
> >> @@ -743,6 +743,16 @@ vtn_handle_alu(struct vtn_builder *b, SpvOp opcode,
> >> src[1] = tmp;
> >> }
> >>
> >> + switch (op) {
> >> + case nir_op_ishl:
> >> + case nir_op_ishr:
> >> + case nir_op_ushr:
> >> + src[1] = nir_u2u32(&b->nb, src[1]);
> >> + break;
> >> + default:
> >> + break;
> >> + }
> >> +
> >> val->ssa->def = nir_build_alu(&b->nb, op, src[0], src[1], src[2],
> >> src[3]);
> >> break;
> >> } /* default */
> >> --
> >> 2.17.1
> >>
> >> _______________________________________________
> >> 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
