LibGL+LibGPU+LibSoftGPU: Implement matrix stack per texture unit

Each texture unit now has its own texture transformation matrix stack.
Introduce a new texture unit configuration that is synced when changed.
Because we're no longer passing a silly `Vector` when drawing each
primitive, this results in a slightly improved frames per second :^)

Userland/Libraries/LibGPU/TextureUnitConfiguration.h

@@ -0,0 +1,30 @@
+/*
+ * Copyright (c) 2022, Jelle Raaijmakers <jelle@gmta.nl>
+ *
+ * SPDX-License-Identifier: BSD-2-Clause
+ */
+
+#pragma once
+
+#include <LibGPU/Enums.h>
+#include <LibGfx/Matrix4x4.h>
+#include <LibGfx/Vector4.h>
+
+namespace GPU {
+
+typedef u8 TextureUnitIndex;
+
+struct TexCoordGeneration {
+    bool enabled;
+    TexCoordGenerationMode mode;
+    FloatVector4 coefficients;
+};
+
+struct TextureUnitConfiguration {
+    bool enabled { false };
+    FloatMatrix4x4 transformation_matrix { FloatMatrix4x4::identity() };
+    u8 tex_coord_generation_enabled;
+    TexCoordGeneration tex_coord_generation[4];
+};
+
+}