matrix_cubesphere: use separate light shader

src/light.fs.asm

@@ -0,0 +1,6 @@
+-- temp[0] -- color
+
+OUT TEX_SEM_WAIT
+src0.rgb = temp[0] :
+  out[0].a    = MAX src0.1 src0.1 ,
+  out[0].rgb  = MAX src0.rgb src0.rgb ;

src/light.vs.asm

@@ -0,0 +1,13 @@
+-- input[0] -- position
+-- input[1] -- color
+
+-- position clip space
+temp[1].x   = VE_DOT   const[0].xyzw    input[0].xyzw ;
+temp[1].y   = VE_DOT   const[1].xyzw    input[0].xyzw ;
+temp[1].z   = VE_DOT   const[2].xyzw    input[0].xyzw ;
+temp[1].w   = VE_DOT   const[3].xyzw    input[0].xyzw ;
+
+-- position (clip space)
+out[0].xyzw = VE_ADD    temp[1].xyzw    const[0].0000 ;
+-- color
+out[1].xyzw = VE_ADD   const[4].xyzw    const[4].0000 ;

src/matrix_cubesphere.cpp

@@ -28,17 +28,20 @@
 #include "../model/model2.h"
 #include "../model/cubesphere.h"
 
-#define CUBESPHERE_SHADER 0
+#define CLEAR_SHADER 0
-#define CLEAR_SHADER 1
+#define CUBESPHERE_SHADER 1
+#define LIGHT_SHADER 2
 
 const char * vertex_shader_paths[] = {
-  "matrix_cubesphere.vs.bin",
   "clear.vs.bin",
+  "matrix_cubesphere.vs.bin",
+  "light.vs.bin",
 };
 const int vertex_shader_paths_length = (sizeof (vertex_shader_paths)) / (sizeof (vertex_shader_paths[0]));
 const char * fragment_shader_paths[] = {
-  "matrix_cubesphere.fs.bin",
   "clear.fs.bin",
+  "matrix_cubesphere.fs.bin",
+  "light.fs.bin"
 };
 const int fragment_shader_paths_length = (sizeof (fragment_shader_paths)) / (sizeof (fragment_shader_paths[0]));
 
@@ -60,7 +63,7 @@ void _3d_clear(struct shaders& shaders)
   T0V(VAP_OUT_VTX_FMT_0
       , VAP_OUT_VTX_FMT_0__VTX_POS_PRESENT(1));
   T0V(VAP_OUT_VTX_FMT_1
-      , 0x0);
+      , 0);
 
   //
 
@@ -107,10 +110,6 @@ void _3d_clear(struct shaders& shaders)
       , VAP_CNTL_STATUS__PVS_BYPASS(0)
       );
 
-  T0V(VAP_VTX_SIZE
-      , VAP_VTX_SIZE__DWORDS_PER_VTX(2)
-      );
-
   //////////////////////////////////////////////////////////////////////////////
   // GA POINT SIZE
   //////////////////////////////////////////////////////////////////////////////
@@ -124,17 +123,24 @@ void _3d_clear(struct shaders& shaders)
   // 3D_DRAW
   //////////////////////////////////////////////////////////////////////////////
 
+  const int dwords_per_vtx = 2;
+
+  T0V(VAP_VTX_SIZE
+      , VAP_VTX_SIZE__DWORDS_PER_VTX(dwords_per_vtx)
+      );
+
   const float center[] = {
     800.0f, 600.0f,
   };
-  T3(_3D_DRAW_IMMD_2, (1 + 2) - 1);
+  const int vertex_count = 1;
+  T3(_3D_DRAW_IMMD_2, (1 + vertex_count * dwords_per_vtx) - 1);
   TU( VAP_VF_CNTL__PRIM_TYPE(1) // point list
     | VAP_VF_CNTL__PRIM_WALK(3)
     | VAP_VF_CNTL__INDEX_SIZE(0)
     | VAP_VF_CNTL__VTX_REUSE_DIS(0)
     | VAP_VF_CNTL__DUAL_INDEX_MODE(0)
     | VAP_VF_CNTL__USE_ALT_NUM_VERTS(0)
-    | VAP_VF_CNTL__NUM_VERTICES(1)
+    | VAP_VF_CNTL__NUM_VERTICES(vertex_count)
     );
   for (int i = 0; i < 2; i++) {
     TF(center[i]);
@@ -200,7 +206,13 @@ void _3d_cube_inner(mat4x4 trans,
   const int triangle_count = obj->triangle_count;
   const int vertex_count = triangle_count * 3;
 
-  T3(_3D_DRAW_IMMD_2, (1 + vertex_count * 8) - 1);
+  const int dwords_per_vtx = 8;
+
+  T0V(VAP_VTX_SIZE
+      , VAP_VTX_SIZE__DWORDS_PER_VTX(dwords_per_vtx)
+      );
+
+  T3(_3D_DRAW_IMMD_2, (1 + vertex_count * dwords_per_vtx) - 1);
   TU( VAP_VF_CNTL__PRIM_TYPE(4)
     | VAP_VF_CNTL__PRIM_WALK(3)
     | VAP_VF_CNTL__INDEX_SIZE(0)
@@ -228,8 +240,152 @@ void _3d_cube_inner(mat4x4 trans,
   }
 }
 
+void _3d_light_inner(mat4x4 trans)
+{
+  T0V(VAP_PVS_STATE_FLUSH_REG, 0x00000000);
+
+  //////////////////////////////////////////////////////////////////////////////
+  // VAP_PVS
+  //////////////////////////////////////////////////////////////////////////////
+
+  const vec4 color = {1, 1, 0, 1};
+
+  const float consts[] = {
+    // 0
+    trans[0][0], trans[0][1], trans[0][2], trans[0][3],
+    trans[1][0], trans[1][1], trans[1][2], trans[1][3],
+    trans[2][0], trans[2][1], trans[2][2], trans[2][3],
+    trans[3][0], trans[3][1], trans[3][2], trans[3][3],
+
+    // 4
+    color[0], color[1], color[2], color[2],
+  };
+  ib_vap_pvs_const_cntl(consts, (sizeof (consts)));
+
+  //////////////////////////////////////////////////////////////////////////////
+  // 3D_DRAW
+  //////////////////////////////////////////////////////////////////////////////
+
+  const model * model = &cubesphere_model;
+  const object * obj = model->object[0];
+  const int triangle_count = obj->triangle_count;
+  const int vertex_count = triangle_count * 3;
+
+  int dwords_per_vtx = 3;
+
+  T0V(VAP_VTX_SIZE
+      , VAP_VTX_SIZE__DWORDS_PER_VTX(dwords_per_vtx)
+      );
+
+  T3(_3D_DRAW_IMMD_2, (1 + vertex_count * dwords_per_vtx) - 1);
+  TU( VAP_VF_CNTL__PRIM_TYPE(4)
+    | VAP_VF_CNTL__PRIM_WALK(3)
+    | VAP_VF_CNTL__INDEX_SIZE(0)
+    | VAP_VF_CNTL__VTX_REUSE_DIS(0)
+    | VAP_VF_CNTL__DUAL_INDEX_MODE(0)
+    | VAP_VF_CNTL__USE_ALT_NUM_VERTS(0)
+    | VAP_VF_CNTL__NUM_VERTICES(vertex_count)
+    );
+
+  for (int i = 0; i < triangle_count; i++) {
+    for (int j = 0; j < 3; j++) {
+      vec3 p = model->position[obj->triangle[i][j].position];
+
+      TF(p.x);
+      TF(p.y);
+      TF(p.z);
+    }
+  }
+}
+
+vec3 _3d_light(struct shaders& shaders,
+               const mat4x4& view_to_clip,
+               float theta)
+{
+  ib_rs_instructions(1);
+
+  //////////////////////////////////////////////////////////////////////////////
+  // VAP OUT
+  //////////////////////////////////////////////////////////////////////////////
+
+  T0V(VAP_OUT_VTX_FMT_0
+      , VAP_OUT_VTX_FMT_0__VTX_POS_PRESENT(1));
+  T0V(VAP_OUT_VTX_FMT_1
+      , VAP_OUT_VTX_FMT_1__TEX_0_COMP_CNT(4));
+
+  //
+
+  ib_zbuffer(ZBUFFER_RELOC_INDEX, 1); // less
+
+  ib_texture__0();
+
+  ib_vap_stream_cntl__3();
+
+  // shaders
+  T0V(US_PIXSIZE
+      , US_PIXSIZE__PIX_SIZE(1)
+      );
+  ib_ga_us(&shaders.fragment[LIGHT_SHADER]);
+  ib_vap_pvs(&shaders.vertex[LIGHT_SHADER]);
+
+  //////////////////////////////////////////////////////////////////////////////
+  // VAP
+  //////////////////////////////////////////////////////////////////////////////
+
+  T0V(VAP_CLIP_CNTL
+      , VAP_CLIP_CNTL__PS_UCP_MODE(3)
+      );
+
+  T0V(VAP_VTE_CNTL
+      , VAP_VTE_CNTL__VPORT_X_SCALE_ENA(1)
+      | VAP_VTE_CNTL__VPORT_X_OFFSET_ENA(1)
+      | VAP_VTE_CNTL__VPORT_Y_SCALE_ENA(1)
+      | VAP_VTE_CNTL__VPORT_Y_OFFSET_ENA(1)
+      | VAP_VTE_CNTL__VPORT_Z_SCALE_ENA(0)
+      | VAP_VTE_CNTL__VPORT_Z_OFFSET_ENA(0)
+      | VAP_VTE_CNTL__VTX_XY_FMT(0) // enable W division
+      | VAP_VTE_CNTL__VTX_Z_FMT(1)  // disable W division
+      | VAP_VTE_CNTL__VTX_W0_FMT(1)
+      | VAP_VTE_CNTL__SERIAL_PROC_ENA(0)
+      );
+
+  T0V(VAP_CNTL_STATUS
+      , VAP_CNTL_STATUS__PVS_BYPASS(0)
+      );
+
+  //////////////////////////////////////////////////////////////////////////////
+  // VAP INDEX
+  //////////////////////////////////////////////////////////////////////////////
+
+  T0V(VAP_INDEX_OFFSET, 0);
+
+  T0V(VAP_VF_MAX_VTX_INDX
+      , VAP_VF_MAX_VTX_INDX__MAX_INDX(0)
+      );
+  T0V(VAP_VF_MIN_VTX_INDX
+      , VAP_VF_MIN_VTX_INDX__MIN_INDX(0)
+      );
+
+  // light
+  mat4x4 t1 = translate(vec3(1, 0, 0));
+  mat4x4 s = scale(0.1f);
+  mat4x4 rz = rotate_y(theta * 2.f);
+
+  mat4x4 world_trans = rz * t1 * s;
+
+  mat4x4 trans = view_to_clip * world_trans;
+
+  _3d_light_inner(trans);
+
+  vec3 light_pos = world_trans * light_pos;
+
+  return light_pos;
+}
+
 void _3d_cube(struct shaders& shaders,
-              float theta)
+              const mat4x4& view_to_clip,
+              float theta,
+              const vec3& light_pos)
 {
   ib_rs_instructions(4);
 
@@ -264,10 +420,6 @@ void _3d_cube(struct shaders& shaders,
   // VAP
   //////////////////////////////////////////////////////////////////////////////
 
-  T0V(VAP_VTX_SIZE
-      , VAP_VTX_SIZE__DWORDS_PER_VTX(8)
-      );
-
   T0V(VAP_CLIP_CNTL
       , VAP_CLIP_CNTL__PS_UCP_MODE(3)
       );
@@ -306,51 +458,16 @@ void _3d_cube(struct shaders& shaders,
   // matrix
   //////////////////////////////////////////////////////////////////////////////
 
-  float theta1 = theta;
+  // cube
-  float theta2 = theta;
+  mat4x4 rx = rotate_x(1 * theta * 0.5f);
+  mat4x4 ry = rotate_y(0 * theta * 0.8f + 1.4f);
+  mat4x4 s = scale(0.9f);
 
-  mat4x4 aspect = scale(vec3(3.0f/4.0f, 1, 1));
+  mat4x4 world_trans = rx * ry * s;
 
-  mat4x4 p = perspective(0.01f, 5.0f,
+  mat4x4 trans = view_to_clip * world_trans;
-                         0.001f, 0.999f,
-                         0.5f, 2.0f);
 
-  vec4 light_pos = vec4(0, 0, 0, 1.0f);
+  _3d_cube_inner(trans, world_trans, light_pos);
-
-  // light
-  if (1) {
-    mat4x4 t = translate(vec3(0, 0, 3));
-    mat4x4 t1 = translate(vec3(1, 0, 0));
-    mat4x4 s = scale(0.1f);
-    mat4x4 rz = rotate_y(theta * 2.f);
-
-    mat4x4 world_trans = rz * t1 * s;
-
-    //mat3x3 normal_trans = transpose(inverse(submatrix(world_trans, 3, 3)));
-
-    mat4x4 trans = aspect * p * t * world_trans;
-
-    light_pos = world_trans * light_pos;
-
-    _3d_cube_inner(trans, world_trans, light_pos);
-  }
-
-
-  // object
-  if (1) {
-    mat4x4 t = translate(vec3(0, 0, 3));
-    mat4x4 rx = rotate_x(1 * theta1 * 0.5f);
-    mat4x4 ry = rotate_y(0 * theta2 * 0.8f + 1.4f);
-    mat4x4 s = scale(0.9f);
-
-    mat4x4 world_trans = rx * ry * s;
-
-    //mat3x3 normal_trans = transpose(inverse(submatrix(world_trans, 3, 3)));
-
-    mat4x4 trans = aspect * p * t * world_trans;
-
-    _3d_cube_inner(trans, world_trans, light_pos);
-  }
 }
 
 int indirect_buffer(shaders& shaders,
@@ -387,8 +504,16 @@ int indirect_buffer(shaders& shaders,
   // DRAW
   //////////////////////////////////////////////////////////////////////////////
 
+  mat4x4 aspect = scale(vec3(3.0f/4.0f, 1, 1));
+  mat4x4 p = perspective(0.01f, 5.0f,
+                         0.001f, 0.999f,
+                         0.5f, 2.0f);
+  mat4x4 t = translate(vec3(0, 0, 3));
+  mat4x4 view_to_clip = aspect * p * t;
+
   _3d_clear(shaders);
-  _3d_cube(shaders, theta);
+  vec3 light_pos = _3d_light(shaders, view_to_clip, theta);
+  _3d_cube(shaders, view_to_clip, theta, light_pos);
 
   //////////////////////////////////////////////////////////////////////////////
   // padding

src/r500/indirect_buffer.c

@@ -650,6 +650,27 @@ void ib_vap_stream_cntl__2()
       );
 }
 
+void ib_vap_stream_cntl__3()
+{
+  //////////////////////////////////////////////////////////////////////////////
+  // VAP_PROG_STREAM_CNTL
+  //////////////////////////////////////////////////////////////////////////////
+
+  T0V(VAP_PROG_STREAM_CNTL_0
+      , VAP_PROG_STREAM_CNTL__DATA_TYPE_0__FLOAT_3
+      | VAP_PROG_STREAM_CNTL__SKIP_DWORDS_0(0)
+      | VAP_PROG_STREAM_CNTL__DST_VEC_LOC_0(0)
+      | VAP_PROG_STREAM_CNTL__LAST_VEC_0(1)
+      );
+  T0V(VAP_PROG_STREAM_CNTL_EXT_0
+      , VAP_PROG_STREAM_CNTL_EXT__SWIZZLE_SELECT_X_0__SELECT_X
+      | VAP_PROG_STREAM_CNTL_EXT__SWIZZLE_SELECT_Y_0__SELECT_Y
+      | VAP_PROG_STREAM_CNTL_EXT__SWIZZLE_SELECT_Z_0__SELECT_Z
+      | VAP_PROG_STREAM_CNTL_EXT__SWIZZLE_SELECT_W_0__SELECT_FP_ONE
+      | VAP_PROG_STREAM_CNTL_EXT__WRITE_ENA_0(0b1111)
+      );
+}
+
 void ib_vap_stream_cntl__323()
 {
   //////////////////////////////////////////////////////////////////////////////

src/r500/indirect_buffer.h

@@ -64,6 +64,7 @@ void ib_vap_pvs(struct shader_offset * offset);
 void ib_ga_us(struct shader_offset * offset);
 void ib_vap_pvs_const_cntl(const float * consts, int size);
 void ib_vap_stream_cntl__2();
+void ib_vap_stream_cntl__3();
 void ib_vap_stream_cntl__323();
 
 #ifdef __cplusplus