PointCloud: Initial rendering support for Workbenchtmp-pointcloud-render

Also includes outline overlays. Removes the temp overlay drawing

We make the geometry follow camera like billboards this uses less
geometry. Currently we use half octahedron for now. Goal would be
to use icospheres.

This patch also optimize the case when pointcloud has uniform radius.
However we should premultiply the radius prop by the default radius
beforehand to avoid a multiplication on CPU.

Differential Revision: https://developer.blender.org/D8301

8 files changed, 83 insertions, 51 deletions

diff --git a/source/blender/draw/intern/DRW_render.h b/source/blender/draw/intern/DRW_render.h
index 555043ab408..1ab5ec18f65 100644
--- a/source/blender/draw/intern/DRW_render.h
+++ b/source/blender/draw/intern/DRW_render.h
@@ -623,7 +623,7 @@ void DRW_render_object_iter(void *vedata,
                                              struct RenderEngine *engine,
                                              struct Depsgraph *depsgraph));
 void DRW_render_instance_buffer_finish(void);
-void DRW_render_viewport_size_set(int size[2]);
+void DRW_render_viewport_size_set(const int size[2]);
 
 void DRW_custom_pipeline(DrawEngineType *draw_engine_type,
                          struct Depsgraph *depsgraph,
diff --git a/source/blender/draw/intern/draw_cache_extract.h b/source/blender/draw/intern/draw_cache_extract.h
index 302f9a0d3a8..f05e8e2f9d6 100644
--- a/source/blender/draw/intern/draw_cache_extract.h
+++ b/source/blender/draw/intern/draw_cache_extract.h
@@ -60,6 +60,10 @@ typedef struct DRW_MeshCDMask {
    *  modifiers could remove it. (see T68857) */
   uint32_t edit_uv : 1;
 } DRW_MeshCDMask;
+/* Keep `DRW_MeshCDMask` struct within an `uint64_t`.
+ * bit-wise and atomic operations are used to compare and update the struct.
+ * See `mesh_cd_layers_type_*` functions. */
+BLI_STATIC_ASSERT(sizeof(DRW_MeshCDMask) <= sizeof(uint64_t), "DRW_MeshCDMask exceeds 64 bits")
 
 typedef enum eMRIterType {
   MR_ITER_LOOPTRI = 1 << 0,
diff --git a/source/blender/draw/intern/draw_cache_extract_mesh.c b/source/blender/draw/intern/draw_cache_extract_mesh.c
index 98da668f78f..a27ee90b148 100644
--- a/source/blender/draw/intern/draw_cache_extract_mesh.c
+++ b/source/blender/draw/intern/draw_cache_extract_mesh.c
@@ -2534,26 +2534,28 @@ static void *extract_vcol_init(const MeshRenderData *mr, void *buf)
   }
 
   /* Sculpt Vertex Colors */
-  for (int i = 0; i < 8; i++) {
-    if (svcol_layers & (1 << i)) {
-      char attr_name[32], attr_safe_name[GPU_MAX_SAFE_ATTR_NAME];
-      const char *layer_name = CustomData_get_layer_name(cd_vdata, CD_PROP_COLOR, i);
-      GPU_vertformat_safe_attr_name(layer_name, attr_safe_name, GPU_MAX_SAFE_ATTR_NAME);
-
-      BLI_snprintf(attr_name, sizeof(attr_name), "c%s", attr_safe_name);
-      GPU_vertformat_attr_add(&format, attr_name, GPU_COMP_U16, 4, GPU_FETCH_INT_TO_FLOAT_UNIT);
-
-      if (i == CustomData_get_render_layer(cd_vdata, CD_PROP_COLOR)) {
-        GPU_vertformat_alias_add(&format, "c");
-      }
-      if (i == CustomData_get_active_layer(cd_vdata, CD_PROP_COLOR)) {
-        GPU_vertformat_alias_add(&format, "ac");
-      }
-      /* Gather number of auto layers. */
-      /* We only do `vcols` that are not overridden by `uvs`. */
-      if (CustomData_get_named_layer_index(cd_ldata, CD_MLOOPUV, layer_name) == -1) {
-        BLI_snprintf(attr_name, sizeof(attr_name), "a%s", attr_safe_name);
-        GPU_vertformat_alias_add(&format, attr_name);
+  if (U.experimental.use_sculpt_vertex_colors) {
+    for (int i = 0; i < 8; i++) {
+      if (svcol_layers & (1 << i)) {
+        char attr_name[32], attr_safe_name[GPU_MAX_SAFE_ATTR_NAME];
+        const char *layer_name = CustomData_get_layer_name(cd_vdata, CD_PROP_COLOR, i);
+        GPU_vertformat_safe_attr_name(layer_name, attr_safe_name, GPU_MAX_SAFE_ATTR_NAME);
+
+        BLI_snprintf(attr_name, sizeof(attr_name), "c%s", attr_safe_name);
+        GPU_vertformat_attr_add(&format, attr_name, GPU_COMP_U16, 4, GPU_FETCH_INT_TO_FLOAT_UNIT);
+
+        if (i == CustomData_get_render_layer(cd_vdata, CD_PROP_COLOR)) {
+          GPU_vertformat_alias_add(&format, "c");
+        }
+        if (i == CustomData_get_active_layer(cd_vdata, CD_PROP_COLOR)) {
+          GPU_vertformat_alias_add(&format, "ac");
+        }
+        /* Gather number of auto layers. */
+        /* We only do `vcols` that are not overridden by `uvs`. */
+        if (CustomData_get_named_layer_index(cd_ldata, CD_MLOOPUV, layer_name) == -1) {
+          BLI_snprintf(attr_name, sizeof(attr_name), "a%s", attr_safe_name);
+          GPU_vertformat_alias_add(&format, attr_name);
+        }
       }
     }
   }
@@ -2599,7 +2601,7 @@ static void *extract_vcol_init(const MeshRenderData *mr, void *buf)
       }
     }
 
-    if (svcol_layers & (1 << i)) {
+    if (svcol_layers & (1 << i) && U.experimental.use_sculpt_vertex_colors) {
       if (mr->extract_type == MR_EXTRACT_BMESH) {
         int cd_ofs = CustomData_get_n_offset(cd_vdata, CD_PROP_COLOR, i);
         BMIter f_iter;
@@ -4384,10 +4386,6 @@ static void *extract_fdots_pos_init(const MeshRenderData *mr, void *buf)
   GPUVertBuf *vbo = buf;
   GPU_vertbuf_init_with_format(vbo, &format);
   GPU_vertbuf_data_alloc(vbo, mr->poly_len);
-  if (!mr->use_subsurf_fdots) {
-    /* Clear so we can accumulate on it. */
-    memset(vbo->data, 0x0, mr->poly_len * vbo->format.stride);
-  }
   return vbo->data;
 }
 
@@ -4396,12 +4394,20 @@ static void extract_fdots_pos_iter_poly_bm(const MeshRenderData *mr,
                                            void *data)
 {
   float(*center)[3] = data;
-  EXTRACT_POLY_AND_LOOP_FOREACH_BM_BEGIN(l, l_index, params, mr)
+
+  EXTRACT_POLY_FOREACH_BM_BEGIN(f, f_index, params, mr)
   {
-    float w = 1.0f / (float)l->f->len;
-    madd_v3_v3fl(center[BM_elem_index_get(l->f)], bm_vert_co_get(mr, l->v), w);
+    float *co = center[f_index];
+    zero_v3(co);
+
+    BMLoop *l_iter, *l_first;
+    l_iter = l_first = BM_FACE_FIRST_LOOP(f);
+    do {
+      add_v3_v3(co, bm_vert_co_get(mr, l_iter->v));
+    } while ((l_iter = l_iter->next) != l_first);
+    mul_v3_fl(co, 1.0f / (float)f->len);
   }
-  EXTRACT_POLY_AND_LOOP_FOREACH_BM_END(l);
+  EXTRACT_POLY_FOREACH_BM_END;
 }
 
 static void extract_fdots_pos_iter_poly_mesh(const MeshRenderData *mr,
@@ -4409,20 +4415,34 @@ static void extract_fdots_pos_iter_poly_mesh(const MeshRenderData *mr,
                                              void *data)
 {
   float(*center)[3] = (float(*)[3])data;
-  EXTRACT_POLY_AND_LOOP_FOREACH_MESH_BEGIN(mp, mp_index, ml, ml_index, params, mr)
-  {
-    const MVert *mv = &mr->mvert[ml->v];
-    if (mr->use_subsurf_fdots) {
+  const MVert *mvert = mr->mvert;
+  const MLoop *mloop = mr->mloop;
+
+  if (mr->use_subsurf_fdots) {
+    EXTRACT_POLY_AND_LOOP_FOREACH_MESH_BEGIN(mp, mp_index, ml, ml_index, params, mr)
+    {
+      const MVert *mv = &mr->mvert[ml->v];
       if (mv->flag & ME_VERT_FACEDOT) {
         copy_v3_v3(center[mp_index], mv->co);
       }
     }
-    else {
-      float w = 1.0f / (float)mp->totloop;
-      madd_v3_v3fl(center[mp_index], mv->co, w);
+    EXTRACT_POLY_AND_LOOP_FOREACH_MESH_END;
+  }
+  else {
+    EXTRACT_POLY_FOREACH_MESH_BEGIN(mp, mp_index, params, mr)
+    {
+      float *co = center[mp_index];
+      zero_v3(co);
+
+      const MLoop *ml = &mloop[mp->loopstart];
+      for (int i = 0; i < mp->totloop; i++, ml++) {
+        const MVert *mv = &mvert[ml->v];
+        add_v3_v3(center[mp_index], mv->co);
+      }
+      mul_v3_fl(co, 1.0f / (float)mp->totloop);
     }
+    EXTRACT_POLY_FOREACH_MESH_END;
   }
-  EXTRACT_POLY_AND_LOOP_FOREACH_MESH_END;
 }
 
 static const MeshExtract extract_fdots_pos = {
diff --git a/source/blender/draw/intern/draw_cache_impl_mesh.c b/source/blender/draw/intern/draw_cache_impl_mesh.c
index ea1717f0684..40de0794b9e 100644
--- a/source/blender/draw/intern/draw_cache_impl_mesh.c
+++ b/source/blender/draw/intern/draw_cache_impl_mesh.c
@@ -74,22 +74,25 @@ static void mesh_batch_cache_clear(Mesh *me);
 /* Return true is all layers in _b_ are inside _a_. */
 BLI_INLINE bool mesh_cd_layers_type_overlap(DRW_MeshCDMask a, DRW_MeshCDMask b)
 {
-  return (*((uint32_t *)&a) & *((uint32_t *)&b)) == *((uint32_t *)&b);
+  return (*((uint64_t *)&a) & *((uint64_t *)&b)) == *((uint64_t *)&b);
 }
 
 BLI_INLINE bool mesh_cd_layers_type_equal(DRW_MeshCDMask a, DRW_MeshCDMask b)
 {
-  return *((uint32_t *)&a) == *((uint32_t *)&b);
+  return *((uint64_t *)&a) == *((uint64_t *)&b);
 }
 
 BLI_INLINE void mesh_cd_layers_type_merge(DRW_MeshCDMask *a, DRW_MeshCDMask b)
 {
-  atomic_fetch_and_or_uint32((uint32_t *)a, *(uint32_t *)&b);
+  uint32_t *a_p = (uint32_t *)a;
+  uint32_t *b_p = (uint32_t *)&b;
+  atomic_fetch_and_or_uint32(a_p, *b_p);
+  atomic_fetch_and_or_uint32(a_p + 1, *(b_p + 1));
 }
 
 BLI_INLINE void mesh_cd_layers_type_clear(DRW_MeshCDMask *a)
 {
-  *((uint32_t *)a) = 0;
+  *((uint64_t *)a) = 0;
 }
 
 static void mesh_cd_calc_edit_uv_layer(const Mesh *UNUSED(me), DRW_MeshCDMask *cd_used)
@@ -205,8 +208,10 @@ static DRW_MeshCDMask mesh_cd_calc_used_gpu_layers(const Mesh *me,
             }
 
             if (layer == -1) {
-              layer = CustomData_get_named_layer(cd_vdata, CD_PROP_COLOR, name);
-              type = CD_PROP_COLOR;
+              if (U.experimental.use_sculpt_vertex_colors) {
+                layer = CustomData_get_named_layer(cd_vdata, CD_PROP_COLOR, name);
+                type = CD_PROP_COLOR;
+              }
             }
 #if 0 /* Tangents are always from UV's - this will never happen. */
             if (layer == -1) {
diff --git a/source/blender/draw/intern/draw_manager.c b/source/blender/draw/intern/draw_manager.c
index 3e42c4cdb23..5359f649d6b 100644
--- a/source/blender/draw/intern/draw_manager.c
+++ b/source/blender/draw/intern/draw_manager.c
@@ -369,7 +369,7 @@ void DRW_engine_viewport_data_size_get(
 }
 
 /* WARNING: only use for custom pipeline. 99% of the time, you don't want to use this. */
-void DRW_render_viewport_size_set(int size[2])
+void DRW_render_viewport_size_set(const int size[2])
 {
   DST.size[0] = size[0];
   DST.size[1] = size[1];
diff --git a/source/blender/draw/intern/draw_manager.h b/source/blender/draw/intern/draw_manager.h
index 31a2dd7f0fe..ab87760942c 100644
--- a/source/blender/draw/intern/draw_manager.h
+++ b/source/blender/draw/intern/draw_manager.h
@@ -583,7 +583,7 @@ void drw_state_set(DRWState state);
 void drw_debug_draw(void);
 void drw_debug_init(void);
 
-eDRWCommandType command_type_get(uint64_t *command_type_bits, int index);
+eDRWCommandType command_type_get(const uint64_t *command_type_bits, int index);
 
 void drw_batch_cache_validate(Object *ob);
 void drw_batch_cache_generate_requested(struct Object *ob);
diff --git a/source/blender/draw/intern/draw_manager_data.c b/source/blender/draw/intern/draw_manager_data.c
index 14d77ab1b21..43fec503301 100644
--- a/source/blender/draw/intern/draw_manager_data.c
+++ b/source/blender/draw/intern/draw_manager_data.c
@@ -623,7 +623,7 @@ static void command_type_set(uint64_t *command_type_bits, int index, eDRWCommand
   command_type_bits[index / 16] |= ((uint64_t)type) << ((index % 16) * 4);
 }
 
-eDRWCommandType command_type_get(uint64_t *command_type_bits, int index)
+eDRWCommandType command_type_get(const uint64_t *command_type_bits, int index)
 {
   return ((command_type_bits[index / 16] >> ((index % 16) * 4)) & 0xF);
 }
diff --git a/source/blender/draw/intern/draw_manager_shader.c b/source/blender/draw/intern/draw_manager_shader.c
index 6304b707cb9..592b030c45e 100644
--- a/source/blender/draw/intern/draw_manager_shader.c
+++ b/source/blender/draw/intern/draw_manager_shader.c
@@ -91,10 +91,13 @@ static void drw_deferred_shader_queue_free(ListBase *queue)
   }
 }
 
-static void drw_deferred_shader_compilation_exec(void *custom_data,
-                                                 short *stop,
-                                                 short *do_update,
-                                                 float *progress)
+static void drw_deferred_shader_compilation_exec(
+    void *custom_data,
+    /* Cannot be const, this function implements wm_jobs_start_callback.
+     * NOLINTNEXTLINE: readability-non-const-parameter. */
+    short *stop,
+    short *do_update,
+    float *progress)
 {
   DRWShaderCompiler *comp = (DRWShaderCompiler *)custom_data;
   void *gl_context = comp->gl_context;