From: Marek Olšák <marek.olsak@amd.com>
Date: Tue, 8 Oct 2019 02:46:00 +0000 (-0400)
Subject: nir: add nir_shader_compiler_options::lower_to_scalar
X-Git-Url: https://git.libre-soc.org/?p=mesa.git;a=commitdiff_plain;h=cebc38ff602c662851b601bc7a48a7fad2700aff

nir: add nir_shader_compiler_options::lower_to_scalar

This will replace PIPE_SHADER_CAP_SCALAR_ISA.

Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
---

diff --git a/src/broadcom/compiler/nir_to_vir.c b/src/broadcom/compiler/nir_to_vir.c
index b640dcc341b..a982778f8b1 100644
--- a/src/broadcom/compiler/nir_to_vir.c
+++ b/src/broadcom/compiler/nir_to_vir.c
@@ -2668,6 +2668,7 @@ const nir_shader_compiler_options v3d_nir_options = {
         .lower_mul_high = true,
         .lower_wpos_pntc = true,
         .lower_rotate = true,
+        .lower_to_scalar = true,
 };
 
 /**
diff --git a/src/compiler/nir/nir.h b/src/compiler/nir/nir.h
index 90b91390a64..960184b7f3c 100644
--- a/src/compiler/nir/nir.h
+++ b/src/compiler/nir/nir.h
@@ -2736,6 +2736,7 @@ typedef struct nir_shader_compiler_options {
     * for IO purposes and would prefer loads/stores be vectorized.
     */
    bool vectorize_io;
+   bool lower_to_scalar;
 
    /**
     * Should nir_lower_io() create load_interpolated_input intrinsics?
diff --git a/src/freedreno/ir3/ir3_nir.c b/src/freedreno/ir3/ir3_nir.c
index 9342c9d299b..b729692bdff 100644
--- a/src/freedreno/ir3/ir3_nir.c
+++ b/src/freedreno/ir3/ir3_nir.c
@@ -56,6 +56,7 @@ static const nir_shader_compiler_options options = {
 		.lower_bitfield_extract_to_shifts = true,
 		.use_interpolated_input_intrinsics = true,
 		.lower_rotate = true,
+		.lower_to_scalar = true,
 };
 
 /* we don't want to lower vertex_id to _zero_based on newer gpus: */
@@ -82,6 +83,7 @@ static const nir_shader_compiler_options options_a6xx = {
 		.use_interpolated_input_intrinsics = true,
 		.lower_rotate = true,
 		.vectorize_io = true,
+		.lower_to_scalar = true,
 };
 
 const nir_shader_compiler_options *
diff --git a/src/gallium/drivers/nouveau/nv50/nv50_screen.c b/src/gallium/drivers/nouveau/nv50/nv50_screen.c
index f0846867621..fe1f628a5ab 100644
--- a/src/gallium/drivers/nouveau/nv50/nv50_screen.c
+++ b/src/gallium/drivers/nouveau/nv50/nv50_screen.c
@@ -919,6 +919,7 @@ static const nir_shader_compiler_options nir_options = {
    .lower_all_io_to_temps = false,
    .lower_cs_local_index_from_id = true,
    .lower_rotate = true,
+   .lower_to_scalar = true,
    .use_interpolated_input_intrinsics = true,
    .max_unroll_iterations = 32,
 };
diff --git a/src/gallium/drivers/nouveau/nvc0/nvc0_screen.c b/src/gallium/drivers/nouveau/nvc0/nvc0_screen.c
index 06f1fb2b708..8c04cda17aa 100644
--- a/src/gallium/drivers/nouveau/nvc0/nvc0_screen.c
+++ b/src/gallium/drivers/nouveau/nvc0/nvc0_screen.c
@@ -976,6 +976,7 @@ static const nir_shader_compiler_options nir_options = {
    .max_unroll_iterations = 32,
    .lower_int64_options = nir_lower_divmod64, // TODO
    .lower_doubles_options = nir_lower_dmod, // TODO
+   .lower_to_scalar = true,
 };
 
 static const void *
diff --git a/src/gallium/drivers/radeonsi/si_get.c b/src/gallium/drivers/radeonsi/si_get.c
index 111279674e6..e37319750f4 100644
--- a/src/gallium/drivers/radeonsi/si_get.c
+++ b/src/gallium/drivers/radeonsi/si_get.c
@@ -524,6 +524,7 @@ static const struct nir_shader_compiler_options nir_options = {
 	.lower_extract_byte = true,
 	.lower_extract_word = true,
 	.lower_rotate = true,
+	.lower_to_scalar = true,
 	.optimize_sample_mask_in = true,
 	.max_unroll_iterations = 32,
 	.use_interpolated_input_intrinsics = true,
diff --git a/src/gallium/drivers/vc4/vc4_program.c b/src/gallium/drivers/vc4/vc4_program.c
index 7d88444c673..79edc6b1c2f 100644
--- a/src/gallium/drivers/vc4/vc4_program.c
+++ b/src/gallium/drivers/vc4/vc4_program.c
@@ -2191,6 +2191,7 @@ static const nir_shader_compiler_options nir_options = {
         .lower_ldexp = true,
         .lower_negate = true,
         .lower_rotate = true,
+        .lower_to_scalar = true,
         .max_unroll_iterations = 32,
 };
 
diff --git a/src/intel/compiler/brw_compiler.c b/src/intel/compiler/brw_compiler.c
index c04264cba32..20660d6fb28 100644
--- a/src/intel/compiler/brw_compiler.c
+++ b/src/intel/compiler/brw_compiler.c
@@ -50,6 +50,7 @@
    .lower_base_vertex = true
 
 #define COMMON_SCALAR_OPTIONS                                                 \
+   .lower_to_scalar = true,                                                   \
    .lower_pack_half_2x16 = true,                                              \
    .lower_pack_snorm_2x16 = true,                                             \
    .lower_pack_snorm_4x8 = true,                                              \