i965: enable component packing for vs and fs

Rather than trying to work out the total number of components
used at a location we simply treat all outputs as vec4s. This
removes the need for complex code looping over varyings to match
packed locations and the need for storing the total number of
components used at each location.

Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

4 files changed, 16 insertions, 25 deletions

diff --git a/src/mesa/drivers/dri/i965/brw_fs.h b/src/mesa/drivers/dri/i965/brw_fs.h
index 574475f071..fc1e1c4a2c 100644
--- a/src/mesa/drivers/dri/i965/brw_fs.h
+++ b/src/mesa/drivers/dri/i965/brw_fs.h
@@ -317,7 +317,6 @@ public:
    fs_reg frag_stencil;
    fs_reg sample_mask;
    fs_reg outputs[VARYING_SLOT_MAX];
-   unsigned output_components[VARYING_SLOT_MAX];
    fs_reg dual_src_output;
    bool do_dual_src;
    int first_non_payload_grf;
diff --git a/src/mesa/drivers/dri/i965/brw_fs_nir.cpp b/src/mesa/drivers/dri/i965/brw_fs_nir.cpp
index 4d4c94a55b..65bca6d6db 100644
--- a/src/mesa/drivers/dri/i965/brw_fs_nir.cpp
+++ b/src/mesa/drivers/dri/i965/brw_fs_nir.cpp
@@ -67,13 +67,12 @@ fs_visitor::nir_setup_single_output_varying(fs_reg *reg,
       }
    } else {
       assert(type->is_scalar() || type->is_vector());
-      unsigned num_elements = type->vector_elements;
+      unsigned num_iter = 1;
       if (type->is_double())
-         num_elements *= 2;
-      for (unsigned count = 0; count < num_elements; count += 4) {
+         num_iter = 2;
+      for (unsigned count = 0; count < num_iter; count++) {
          this->outputs[*location] = *reg;
-         this->output_components[*location] = MIN2(4, num_elements - count);
-         *reg = offset(*reg, bld, this->output_components[*location]);
+         *reg = offset(*reg, bld, 4);
          (*location)++;
       }
    }
@@ -114,7 +113,6 @@ fs_visitor::nir_setup_outputs()
             /* Writing gl_FragColor outputs to all color regions. */
             for (unsigned int i = 0; i < MAX2(key->nr_color_regions, 1); i++) {
                this->outputs[i] = reg;
-               this->output_components[i] = 4;
             }
          } else if (var->data.location == FRAG_RESULT_DEPTH) {
             this->frag_depth = reg;
@@ -123,8 +121,6 @@ fs_visitor::nir_setup_outputs()
          } else if (var->data.location == FRAG_RESULT_SAMPLE_MASK) {
             this->sample_mask = reg;
          } else {
-            int vector_elements = var->type->without_array()->vector_elements;
-
             /* gl_FragData or a user-defined FS output */
             assert(var->data.location >= FRAG_RESULT_DATA0 &&
                    var->data.location < FRAG_RESULT_DATA0+BRW_MAX_DRAW_BUFFERS);
@@ -132,8 +128,7 @@ fs_visitor::nir_setup_outputs()
             /* General color output. */
             for (unsigned int i = 0; i < MAX2(1, var->type->length); i++) {
                int output = var->data.location - FRAG_RESULT_DATA0 + i;
-               this->outputs[output] = offset(reg, bld, vector_elements * i);
-               this->output_components[output] = vector_elements;
+               this->outputs[output] = offset(reg, bld, 4 * i);
             }
          }
          break;
@@ -2360,6 +2355,7 @@ fs_visitor::nir_emit_vs_intrinsic(const fs_builder &bld,
 
    case nir_intrinsic_load_input: {
       fs_reg src = fs_reg(ATTR, instr->const_index[0], dest.type);
+      unsigned first_component = nir_intrinsic_component(instr);
       unsigned num_components = instr->num_components;
       enum brw_reg_type type = dest.type;
 
@@ -2368,7 +2364,7 @@ fs_visitor::nir_emit_vs_intrinsic(const fs_builder &bld,
       src = offset(src, bld, const_offset->u32[0]);
 
       for (unsigned j = 0; j < num_components; j++) {
-         bld.MOV(offset(dest, bld, j), offset(src, bld, j));
+         bld.MOV(offset(dest, bld, j), offset(src, bld, j + first_component));
       }
 
       if (type == BRW_REGISTER_TYPE_DF) {
@@ -4103,6 +4099,7 @@ fs_visitor::nir_emit_intrinsic(const fs_builder &bld, nir_intrinsic_instr *instr
       new_dest = offset(new_dest, bld, const_offset->u32[0]);
 
       unsigned num_components = instr->num_components;
+      unsigned first_component = nir_intrinsic_component(instr);
       unsigned bit_size = instr->src[0].is_ssa ?
          instr->src[0].ssa->bit_size : instr->src[0].reg.reg->bit_size;
       if (bit_size == 64) {
@@ -4116,7 +4113,8 @@ fs_visitor::nir_emit_intrinsic(const fs_builder &bld, nir_intrinsic_instr *instr
       }
 
       for (unsigned j = 0; j < num_components; j++) {
-         bld.MOV(offset(new_dest, bld, j), offset(src, bld, j));
+         bld.MOV(offset(new_dest, bld, j + first_component),
+                 offset(src, bld, j));
       }
       break;
    }
diff --git a/src/mesa/drivers/dri/i965/brw_fs_visitor.cpp b/src/mesa/drivers/dri/i965/brw_fs_visitor.cpp
index 156a6306b7..6d843749f4 100644
--- a/src/mesa/drivers/dri/i965/brw_fs_visitor.cpp
+++ b/src/mesa/drivers/dri/i965/brw_fs_visitor.cpp
@@ -459,8 +459,7 @@ fs_visitor::emit_fb_writes()
             src0_alpha = offset(outputs[0], bld, 3);
 
          inst = emit_single_fb_write(abld, this->outputs[target], reg_undef,
-                                     src0_alpha,
-                                     this->output_components[target]);
+                                     src0_alpha, 4);
          inst->target = target;
       }
    }
@@ -545,9 +544,7 @@ void fs_visitor::compute_clip_distance(gl_clip_plane *clip_planes)
    const fs_builder abld = bld.annotate("user clip distances");
 
    this->outputs[VARYING_SLOT_CLIP_DIST0] = vgrf(glsl_type::vec4_type);
-   this->output_components[VARYING_SLOT_CLIP_DIST0] = 4;
    this->outputs[VARYING_SLOT_CLIP_DIST1] = vgrf(glsl_type::vec4_type);
-   this->output_components[VARYING_SLOT_CLIP_DIST1] = 4;
 
    for (int i = 0; i < key->nr_userclip_plane_consts; i++) {
       fs_reg u = userplane[i];
@@ -724,10 +721,8 @@ fs_visitor::emit_urb_writes(const fs_reg &gs_vertex_count)
                sources[length++] = reg;
             }
          } else {
-            for (unsigned i = 0; i < output_components[varying]; i++)
+            for (unsigned i = 0; i < 4; i++)
                sources[length++] = offset(this->outputs[varying], bld, i);
-            for (unsigned i = output_components[varying]; i < 4; i++)
-               sources[length++] = brw_imm_d(0);
          }
          break;
       }
@@ -901,7 +896,6 @@ fs_visitor::init()
    this->nir_ssa_values = NULL;
 
    memset(&this->payload, 0, sizeof(this->payload));
-   memset(this->output_components, 0, sizeof(this->output_components));
    this->source_depth_to_render_target = false;
    this->runtime_check_aads_emit = false;
    this->first_non_payload_grf = 0;
diff --git a/src/mesa/drivers/dri/i965/brw_nir.c b/src/mesa/drivers/dri/i965/brw_nir.c
index fe7653137e..388ae9e038 100644
--- a/src/mesa/drivers/dri/i965/brw_nir.c
+++ b/src/mesa/drivers/dri/i965/brw_nir.c
@@ -302,8 +302,8 @@ brw_nir_lower_vue_outputs(nir_shader *nir,
    if (is_scalar) {
       nir_assign_var_locations(&nir->outputs, &nir->num_outputs,
                                VARYING_SLOT_VAR0,
-                               type_size_scalar);
-      nir_lower_io(nir, nir_var_shader_out, type_size_scalar);
+                               type_size_vec4_times_4);
+      nir_lower_io(nir, nir_var_shader_out, type_size_vec4_times_4);
    } else {
       nir_foreach_variable(var, &nir->outputs)
          var->data.driver_location = var->data.location;
@@ -340,8 +340,8 @@ void
 brw_nir_lower_fs_outputs(nir_shader *nir)
 {
    nir_assign_var_locations(&nir->outputs, &nir->num_outputs,
-                            FRAG_RESULT_DATA0, type_size_scalar);
-   nir_lower_io(nir, nir_var_shader_out, type_size_scalar);
+                            FRAG_RESULT_DATA0, type_size_vec4_times_4);
+   nir_lower_io(nir, nir_var_shader_out, type_size_vec4_times_4);
 }
 
 void