GPU: Make nodetree GLSL Codegen render engine agnostic

This commit removes all EEVEE specific code from the `gpu_shader_material*.glsl`
files. It defines a clear interface to evaluate the closure nodes leaving
more flexibility to the render engine.

Some of the long standing workaround are fixed:
- bump mapping support is no longer duplicating a lot of node and is instead
  compiled into a function call.
- bump rewiring to Normal socket is no longer needed as we now use a global
  `g_data.N` for that.


Closure sampling with upstread weight eval is now supported if the engine needs
it.

This also makes all the material GLSL sources use `GPUSource` for better
debugging experience. The `GPUFunction` parsing now happens in `GPUSource`
creation.

The whole `GPUCodegen` now uses the `ShaderCreateInfo` and is object type
agnostic. Is has also been rewritten in C++.

This patch changes a view behavior for EEVEE:
- Mix shader node factor imput is now clamped.
- Tangent Vector displacement behavior is now matching cycles.
- The chosen BSDF used for SSR might change.
- Hair shading may have very small changes on very large hairs when using hair
  polygon stripes.
- ShaderToRGB node will remove any SSR and SSS form a shader.
- SSS radius input now is no longer a scaling factor but defines an average
  radius. The SSS kernel "shape" (radii) are still defined by the socket default
  values.

Appart from the listed changes no other regressions are expected.

1 files changed, 173 insertions, 0 deletions

diff --git a/source/blender/gpu/shaders/gpu_shader_codegen_lib.glsl b/source/blender/gpu/shaders/gpu_shader_codegen_lib.glsl
index 193a4190cbf..5c97eada77d 100644
--- a/source/blender/gpu/shaders/gpu_shader_codegen_lib.glsl
+++ b/source/blender/gpu/shaders/gpu_shader_codegen_lib.glsl
@@ -95,3 +95,176 @@ vec4 tangent_get(vec4 attr, mat3 normalmat)
 }
 
 #endif
+
+/* Assumes GPU_VEC4 is color data. So converting to luminance like cycles. */
+#define float_from_vec4(v) dot(v.rgb, vec3(0.2126, 0.7152, 0.0722))
+#define float_from_vec3(v) avg(v.rgb)
+#define float_from_vec2(v) v.r
+
+#define vec2_from_vec4(v) vec2(avg(v.rgb), v.a)
+#define vec2_from_vec3(v) vec2(avg(v.rgb), 1.0)
+#define vec2_from_float(v) vec2(v)
+
+#define vec3_from_vec4(v) v.rgb
+#define vec3_from_vec2(v) v.rrr
+#define vec3_from_float(v) vec3(v)
+
+#define vec4_from_vec3(v) vec4(v, 1.0)
+#define vec4_from_vec2(v) v.rrrg
+#define vec4_from_float(v) vec4(vec3(v), 1.0)
+
+/* TODO: Move to shader_shared. */
+#define RAY_TYPE_CAMERA 0
+#define RAY_TYPE_SHADOW 1
+#define RAY_TYPE_DIFFUSE 2
+#define RAY_TYPE_GLOSSY 3
+
+#ifdef GPU_FRAGMENT_SHADER
+#  define FrontFacing gl_FrontFacing
+#else
+#  define FrontFacing true
+#endif
+
+struct ClosureDiffuse {
+  float weight;
+  vec3 color;
+  vec3 N;
+  vec3 sss_radius;
+  uint sss_id;
+};
+
+struct ClosureTranslucent {
+  float weight;
+  vec3 color;
+  vec3 N;
+};
+
+struct ClosureReflection {
+  float weight;
+  vec3 color;
+  vec3 N;
+  float roughness;
+};
+
+struct ClosureRefraction {
+  float weight;
+  vec3 color;
+  vec3 N;
+  float roughness;
+  float ior;
+};
+
+struct ClosureHair {
+  float weight;
+  vec3 color;
+  float offset;
+  vec2 roughness;
+  vec3 T;
+};
+
+struct ClosureVolumeScatter {
+  float weight;
+  vec3 scattering;
+  float anisotropy;
+};
+
+struct ClosureVolumeAbsorption {
+  float weight;
+  vec3 absorption;
+};
+
+struct ClosureEmission {
+  float weight;
+  vec3 emission;
+};
+
+struct ClosureTransparency {
+  float weight;
+  vec3 transmittance;
+  float holdout;
+};
+
+struct GlobalData {
+  /** World position. */
+  vec3 P;
+  /** Surface Normal. */
+  vec3 N;
+  /** Geometric Normal. */
+  vec3 Ng;
+  /** Surface default Tangent. */
+  vec3 T;
+  /** Barycentric coordinates. */
+  vec2 barycentric_coords;
+  vec3 barycentric_dists;
+  /** Ray properties (approximation). */
+  int ray_type;
+  float ray_depth;
+  float ray_length;
+  /** Hair time along hair length. 0 at base 1 at tip. */
+  float hair_time;
+  /** Hair time along width of the hair. */
+  float hair_time_width;
+  /** Hair thickness in world space. */
+  float hair_thickness;
+  /** Index of the strand for per strand effects. */
+  int hair_strand_id;
+  /** Is hair. */
+  bool is_strand;
+};
+
+GlobalData g_data;
+
+#ifndef GPU_FRAGMENT_SHADER
+/* Stubs. */
+vec3 dF_impl(vec3 v)
+{
+  return vec3(0.0);
+}
+
+void dF_branch(float fn, out vec2 result)
+{
+  result = vec2(0.0);
+}
+
+#elif 0 /* TODO(@fclem): User Option? */
+/* Fast derivatives */
+vec3 dF_impl(vec3 v)
+{
+  return vec3(0.0);
+}
+
+void dF_branch(float fn, out vec2 result)
+{
+  result.x = DFDX_SIGN * dFdx(fn);
+  result.y = DFDY_SIGN * dFdy(fn);
+}
+
+#else
+/* Precise derivatives */
+int g_derivative_flag = 0;
+
+vec3 dF_impl(vec3 v)
+{
+  if (g_derivative_flag > 0) {
+    return DFDX_SIGN * dFdx(v);
+  }
+  else if (g_derivative_flag < 0) {
+    return DFDY_SIGN * dFdy(v);
+  }
+  return vec3(0.0);
+}
+
+#  define dF_branch(fn, result) \
+    if (true) { \
+      g_derivative_flag = 1; \
+      result.x = (fn); \
+      g_derivative_flag = -1; \
+      result.y = (fn); \
+      g_derivative_flag = 0; \
+      result -= vec2((fn)); \
+    }
+
+#endif
+
+/* TODO(fclem): Remove. */
+#define CODEGEN_LIB