[Mesa-dev] [PATCH] glsl: Fix loop bounds detection.

[Kenneth Graunke]

On 01/07/2013 06:23 PM, Paul Berry wrote:
> When analyzing a loop where the loop condition is expressed in the
> non-standard order (e.g. "4 > i" instead of "i < 4"), we were
> reversing the condition incorrectly, leading to a loop bound that was
> off by 1.
>
> Fixes piglit tests {vs,fs}-loop-bounds-unrolled.shader_test.
> ---
>   src/glsl/loop_controls.cpp | 8 ++++----
>   1 file changed, 4 insertions(+), 4 deletions(-)
>
> diff --git a/src/glsl/loop_controls.cpp b/src/glsl/loop_controls.cpp
> index 9acbadc..79c8204 100644
> --- a/src/glsl/loop_controls.cpp
> +++ b/src/glsl/loop_controls.cpp
> @@ -222,10 +222,10 @@ loop_control_visitor::visit_leave(ir_loop *ir)
>   	    limit = cond->operands[0]->as_constant();
>
>   	    switch (cmp) {
> -	    case ir_binop_less:    cmp = ir_binop_gequal;  break;
> -	    case ir_binop_greater: cmp = ir_binop_lequal;  break;
> -	    case ir_binop_lequal:  cmp = ir_binop_greater; break;
> -	    case ir_binop_gequal:  cmp = ir_binop_less;    break;
> +	    case ir_binop_less:    cmp = ir_binop_greater; break;
> +	    case ir_binop_greater: cmp = ir_binop_less;    break;
> +	    case ir_binop_lequal:  cmp = ir_binop_gequal;  break;
> +	    case ir_binop_gequal:  cmp = ir_binop_lequal;  break;
>   	    default: assert(!"Should not get here.");
>   	    }
>   	 }

Reviewed-by: Kenneth Graunke <kenneth at whitecape.org>