[Mesa-dev] [PATCH 9/9] glsl: optimize (0 cmp x + y) into (-x cmp y).

[Ian Romanick]

This patch is

Reviewed-by: Ian Romanick <ian.d.romanick at intel.com>

Please also add the following to the commit message:

Cc: "10.4 10.5" <mesa-stable at lists.freedesktop.org>

If you guys have the time, I'd also really appreciate a piglit test or
two for this case.

On 02/24/2015 10:02 AM, Eduardo Lima Mitev wrote:
> From: Samuel Iglesias Gonsalvez <siglesias at igalia.com>
> 
> The optimization done by commit 34ec1a24d did not take it into account.
> 
> Fixes:
> 
> dEQP-GLES3.functional.shaders.random.all_features.fragment.20
> 
> Signed-off-by: Samuel Iglesias Gonsalvez <siglesias at igalia.com>
> ---
>  src/glsl/opt_algebraic.cpp | 15 ++++++++++++---
>  1 file changed, 12 insertions(+), 3 deletions(-)
> 
> diff --git a/src/glsl/opt_algebraic.cpp b/src/glsl/opt_algebraic.cpp
> index c6f4a9c..8183d57 100644
> --- a/src/glsl/opt_algebraic.cpp
> +++ b/src/glsl/opt_algebraic.cpp
> @@ -578,9 +578,18 @@ ir_algebraic_visitor::handle_expression(ir_expression *ir)
>           if (!is_vec_zero(zero))
>              continue;
>  
> -         return new(mem_ctx) ir_expression(ir->operation,
> -                                           add->operands[0],
> -                                           neg(add->operands[1]));
> +         /* Depending of the zero position we want to optimize
> +          * (0 cmp x+y) into (-x cmp y) or (x+y cmp 0) into (x cmp -y)
> +          */
> +         if (add_pos == 1) {
> +            return new(mem_ctx) ir_expression(ir->operation,
> +                                              neg(add->operands[0]),
> +                                              add->operands[1]);
> +         } else {
> +            return new(mem_ctx) ir_expression(ir->operation,
> +                                              add->operands[0],
> +                                              neg(add->operands[1]));
> +         }
>        }
>        break;
>  
>