[Mesa-dev] [PATCH 3/3] radeonsi/compute: Enable PIPE_SHADER_IR_NATIVE for compute shaders v2

Michel Dänzer michel at daenzer.net
Sun Oct 19 21:06:34 PDT 2014


On 17.10.2014 08:41, Tom Stellard wrote:
> v2:
>    - Drop dependency on LLVM >= 3.5.1

Have you tested this patch with LLVM < 3.6?


> @@ -61,10 +66,7 @@ static void *si_create_compute_state(
>   	struct si_context *sctx = (struct si_context *)ctx;
>   	struct si_compute *program = CALLOC_STRUCT(si_compute);

[...]

> +	memset(&program->binary, 0, sizeof(program->binary));

Looks like this memset is redundant.


> @@ -388,22 +415,12 @@ static void si_delete_compute_state(struct pipe_context *ctx, void* state){
>   		return;
>   	}
>
> -	if (program->kernels) {
> -		for (int i = 0; i < program->num_kernels; i++){
> -			if (program->kernels[i].bo){
> -				si_shader_destroy(ctx, &program->kernels[i]);
> -			}
> -		}
> -		FREE(program->kernels);
> -	}
> -
> -	if (program->llvm_ctx){
> -		LLVMContextDispose(program->llvm_ctx);
> -	}

Is this really not necessary anymore? If not, it might be better to 
remove it in a separate patch.


Other than that, this patch looks good to me, and patch 1 is

Reviewed-by: Michel Dänzer <michel.daenzer at amd.com>

I haven't looked at patch 2 in detail.


-- 
Earthling Michel Dänzer            |                  http://www.amd.com
Libre software enthusiast          |                Mesa and X developer


More information about the mesa-dev mailing list