nir,amd: optimize front_face ? a : -a

Foz-DB Navi31:
Totals from 3345 (4.21% of 79395) affected shaders:
MaxWaves: 96182 -> 96174 (-0.01%)
Instrs: 3135439 -> 3129508 (-0.19%); split: -0.24%, +0.05%
CodeSize: 16776088 -> 16718048 (-0.35%); split: -0.38%, +0.03%
VGPRs: 190884 -> 190848 (-0.02%); split: -0.03%, +0.01%
Latency: 32624132 -> 32621734 (-0.01%); split: -0.16%, +0.16%
InvThroughput: 5759987 -> 5749957 (-0.17%); split: -0.23%, +0.05%
VClause: 51044 -> 51086 (+0.08%); split: -0.12%, +0.20%
SClause: 103415 -> 103223 (-0.19%); split: -0.64%, +0.45%
Copies: 170398 -> 170555 (+0.09%); split: -0.64%, +0.74%
PreSGPRs: 135567 -> 133887 (-1.24%)
PreVGPRs: 140569 -> 141317 (+0.53%)
VALU: 1959144 -> 1953839 (-0.27%); split: -0.30%, +0.03%
SALU: 217956 -> 217676 (-0.13%); split: -0.20%, +0.07%

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32791>

src/compiler/nir/nir.h

@@ -4016,6 +4016,11 @@ typedef struct nir_shader_compiler_options {
     */
    bool optimize_sample_mask_in;
 
+   /**
+    * Optimize load_front_face ? a : -a to load_front_face_fsign * a
+    */
+   bool optimize_load_front_face_fsign;
+
    /**
     * Optimize boolean reductions of quad broadcasts. This should only be enabled if
     * nir_intrinsic_reduce supports INCLUDE_HELPERS.