Fix WMO wall collision, normal mapping, POM backfill, and M2/WMO rendering performance

- Fix MOPY flag check (0x08 not 0x01) for proper wall collision detection
- Cap MAX_PUSH to PLAYER_RADIUS to prevent gradual clip-through
- Fix WMO doodad quaternion component ordering (X/Y swap)
- Linear normal map strength blend in shader for smooth slider control
- Enable shadow sampling for interior WMO groups (covered outdoor areas)
- Backfill deferred normal/height maps after streaming with descriptor rebind
- M2: prepareRender only iterates animated instances, bone dirty flag
- M2: remove worker thread VMA allocation, skip unready bone instances
- WMO: persistent visibility vectors, sequential culling
- Add FSR EASU/RCAS shaders

include/rendering/wmo_renderer.hpp

@@ -148,6 +148,8 @@ public:
      * @param perFrameSet Per-frame descriptor set (set 0)
      * @param camera Camera for frustum culling
      */
+    /** Pre-update mutable state (frame ID, material UBOs) on main thread before parallel render. */
+    void prepareRender();
     void render(VkCommandBuffer cmd, VkDescriptorSet perFrameSet, const Camera& camera);
 
     /**
@@ -332,6 +334,9 @@ public:
     // Defer normal/height map generation during streaming to avoid CPU stalls
     void setDeferNormalMaps(bool defer) { deferNormalMaps_ = defer; }
 
+    // Generate normal/height maps for cached textures that were loaded while deferred
+    void backfillNormalMaps();
+
 private:
     // WMO material UBO — matches WMOMaterial in wmo.frag.glsl
     struct WMOMaterialUBO {
@@ -720,6 +725,8 @@ private:
         uint32_t distanceCulled = 0;
     };
     std::vector<std::future<void>> cullFutures_;
+    std::vector<size_t> visibleInstances_;      // reused per frame
+    std::vector<InstanceDrawList> drawLists_;    // reused per frame
 
     // Collision query profiling (per frame).
     mutable double queryTimeMs = 0.0;