[Mesa-dev] [PATCH v2 25/42] nir: Translate glsl shared var store intrinsic to nir intrinsic

[Iago Toral]

On Tue, 2015-11-17 at 21:54 -0800, Jordan Justen wrote:
> Signed-off-by: Jordan Justen <jordan.l.justen at intel.com>
> ---
>  src/glsl/nir/glsl_to_nir.cpp  | 33 +++++++++++++++++++++++++++++++++
>  src/glsl/nir/nir_intrinsics.h |  3 ++-
>  2 files changed, 35 insertions(+), 1 deletion(-)
> 
> diff --git a/src/glsl/nir/glsl_to_nir.cpp b/src/glsl/nir/glsl_to_nir.cpp
> index a59d09c..a832570 100644
> --- a/src/glsl/nir/glsl_to_nir.cpp
> +++ b/src/glsl/nir/glsl_to_nir.cpp
> @@ -729,6 +729,8 @@ nir_visitor::visit(ir_call *ir)
>           op = nir_intrinsic_memory_barrier_shared;
>        } else if (strcmp(ir->callee_name(), "__intrinsic_load_shared") == 0) {
>           op = nir_intrinsic_load_shared;
> +      } else if (strcmp(ir->callee_name(), "__intrinsic_store_shared") == 0) {
> +         op = nir_intrinsic_store_shared;
>        } else {
>           unreachable("not reached");
>        }
> @@ -1003,6 +1005,37 @@ nir_visitor::visit(ir_call *ir)
>           nir_builder_instr_insert(&b, &instr->instr);
>           break;
>        }
> +      case nir_intrinsic_store_shared: {
> +         exec_node *param = ir->actual_parameters.get_head();
> +         ir_rvalue *offset = ((ir_instruction *)param)->as_rvalue();
> +
> +         param = param->get_next();
> +         ir_rvalue *val = ((ir_instruction *)param)->as_rvalue();
> +
> +         param = param->get_next();
> +         ir_constant *write_mask = ((ir_instruction *)param)->as_constant();
> +         assert(write_mask);
> +
> +         /* Check if we need the indirect version */
> +         ir_constant *const_offset = offset->as_constant();
> +         if (!const_offset) {
> +            op = nir_intrinsic_store_shared_indirect;
> +            ralloc_free(instr);
> +            instr = nir_intrinsic_instr_create(shader, op);
> +            instr->src[1] = nir_src_for_ssa(evaluate_rvalue(offset));
> +            instr->const_index[0] = 0;
> +         } else {
> +            instr->const_index[0] = const_offset->value.u[0];
> +         }
> +
> +         instr->const_index[1] = write_mask->value.u[0];
> +
> +         instr->src[0] = nir_src_for_ssa(evaluate_rvalue(val));
> +         instr->num_components = val->type->vector_elements;
> +
> +         nir_builder_instr_insert(&b, &instr->instr);
> +         break;
> +      }
>        default:
>           unreachable("not reached");
>        }
> diff --git a/src/glsl/nir/nir_intrinsics.h b/src/glsl/nir/nir_intrinsics.h
> index de15128..6912258 100644
> --- a/src/glsl/nir/nir_intrinsics.h
> +++ b/src/glsl/nir/nir_intrinsics.h
> @@ -277,5 +277,6 @@ LOAD(shared, 0, 1, NIR_INTRINSIC_CAN_ELIMINATE)
>  STORE(output, 0, 0, 0, 0)
>  STORE(per_vertex_output, 1, 1, 0, 0)
>  STORE(ssbo, 1, 1, 1, 0)
> +STORE(shared, 0, 0, 0, 0)

This should be:

STORE(shared, 0, 0, 1, 0)

to account for the for the writemask, which always goes in
const_index[1]. With that change:

Reviewed-by: Iago Toral Quiroga <itoral at igalia.com>

> -LAST_INTRINSIC(store_ssbo_indirect)
> +LAST_INTRINSIC(store_shared_indirect)