Port Dot3RGB and Dot3RGBA from zmm-progress in a less broken way

src/video_core/pica.h

@@ -262,6 +262,8 @@ struct Regs {
             Lerp            = 4,
             Subtract        = 5,
 
+            Dot3RGB         = 6,
+            Dot3RGBA        = 7,
             MultiplyThenAdd = 8,
             AddThenMultiply = 9,
         };

src/video_core/rasterizer.cpp

@@ -502,6 +502,20 @@ static void ProcessTriangleInternal(const VertexShader::OutputVertex& v0,
                         return result.Cast<u8>();
                     }
 
+                    // This computes the dot product of the RGB components according to GLES spec.
+                    // Dot3RGB just stores the dot product in the RGB components.
+                    // Dot3RGBA also stores it in the alpha component.
+                    case Operation::Dot3RGB:
+                    case Operation::Dot3RGBA:
+                    {
+                        std::array<int, 3> a = {input[0].r(), input[0].g(), input[0].b()};
+                        std::array<int, 3> b = {input[1].r(), input[1].g(), input[1].b()};
+
+                        auto res = std::inner_product(a, a.end(), b, 0);
+                        
+                        return {res, res, res};
+                    }
+
                     case Operation::MultiplyThenAdd:
                     {
                         auto result = (input[0] * input[1] + 255 * input[2].Cast<int>()) / 255;
@@ -545,6 +559,16 @@ static void ProcessTriangleInternal(const VertexShader::OutputVertex& v0,
                     case Operation::Subtract:
                         return std::max(0, (int)input[0] - (int)input[1]);
 
+                    case Operation::Dot3RGBA:
+                    {
+                        std::array<int, 3> a = {input[0].r(), input[0].g(), input[0].b()};
+                        std::array<int, 3> b = {input[1].r(), input[1].g(), input[1].b()};
+
+                        auto res = std::inner_product(a, a.end(), b, 0);
+                        
+                        return res;
+                    }
+
                     case Operation::MultiplyThenAdd:
                         return std::min(255, (input[0] * input[1] + 255 * input[2]) / 255);