diff --git a/src/video_core/renderer_opengl/gl_texture_cache.cpp b/src/video_core/renderer_opengl/gl_texture_cache.cpp
index bddb15cb11..71f6888c6c 100644
--- a/src/video_core/renderer_opengl/gl_texture_cache.cpp
+++ b/src/video_core/renderer_opengl/gl_texture_cache.cpp
@@ -450,7 +450,7 @@ Surface TextureCacheOpenGL::CreateSurface(GPUVAddr gpu_addr, const SurfaceParams
     return std::make_shared<CachedSurface>(gpu_addr, params);
 }
 
-void TextureCacheOpenGL::ImageCopy(Surface src_surface, Surface dst_surface,
+void TextureCacheOpenGL::ImageCopy(Surface& src_surface, Surface& dst_surface,
                                    const VideoCommon::CopyParams& copy_params) {
     if (!support_info.depth_color_image_copies) {
         const auto& src_params = src_surface->GetSurfaceParams();
@@ -471,7 +471,7 @@ void TextureCacheOpenGL::ImageCopy(Surface src_surface, Surface dst_surface,
                        copy_params.depth);
 }
 
-void TextureCacheOpenGL::ImageBlit(View src_view, View dst_view,
+void TextureCacheOpenGL::ImageBlit(View& src_view, View& dst_view,
                                    const Tegra::Engines::Fermi2D::Config& copy_config) {
     const auto& src_params{src_view->GetSurfaceParams()};
     const auto& dst_params{dst_view->GetSurfaceParams()};
@@ -528,7 +528,7 @@ void TextureCacheOpenGL::ImageBlit(View src_view, View dst_view,
                       is_linear ? GL_LINEAR : GL_NEAREST);
 }
 
-void TextureCacheOpenGL::BufferCopy(Surface src_surface, Surface dst_surface) {
+void TextureCacheOpenGL::BufferCopy(Surface& src_surface, Surface& dst_surface) {
     const auto& src_params = src_surface->GetSurfaceParams();
     const auto& dst_params = dst_surface->GetSurfaceParams();
 
diff --git a/src/video_core/renderer_opengl/gl_texture_cache.h b/src/video_core/renderer_opengl/gl_texture_cache.h
index f514f137c0..dda3bf715e 100644
--- a/src/video_core/renderer_opengl/gl_texture_cache.h
+++ b/src/video_core/renderer_opengl/gl_texture_cache.h
@@ -137,13 +137,13 @@ public:
 protected:
     Surface CreateSurface(GPUVAddr gpu_addr, const SurfaceParams& params) override;
 
-    void ImageCopy(Surface src_surface, Surface dst_surface,
+    void ImageCopy(Surface& src_surface, Surface& dst_surface,
                    const VideoCommon::CopyParams& copy_params) override;
 
-    void ImageBlit(View src_view, View dst_view,
+    void ImageBlit(View& src_view, View& dst_view,
                    const Tegra::Engines::Fermi2D::Config& copy_config) override;
 
-    void BufferCopy(Surface src_surface, Surface dst_surface) override;
+    void BufferCopy(Surface& src_surface, Surface& dst_surface) override;
 
 private:
     GLuint FetchPBO(std::size_t buffer_size);
diff --git a/src/video_core/texture_cache/texture_cache.h b/src/video_core/texture_cache/texture_cache.h
index e0d0e1f700..9511683576 100644
--- a/src/video_core/texture_cache/texture_cache.h
+++ b/src/video_core/texture_cache/texture_cache.h
@@ -234,15 +234,15 @@ protected:
 
     virtual TSurface CreateSurface(GPUVAddr gpu_addr, const SurfaceParams& params) = 0;
 
-    virtual void ImageCopy(TSurface src_surface, TSurface dst_surface,
+    virtual void ImageCopy(TSurface& src_surface, TSurface& dst_surface,
                            const CopyParams& copy_params) = 0;
 
-    virtual void ImageBlit(TView src_view, TView dst_view,
+    virtual void ImageBlit(TView& src_view, TView& dst_view,
                            const Tegra::Engines::Fermi2D::Config& copy_config) = 0;
 
     // Depending on the backend, a buffer copy can be slow as it means deoptimizing the texture
     // and reading it from a sepparate buffer.
-    virtual void BufferCopy(TSurface src_surface, TSurface dst_surface) = 0;
+    virtual void BufferCopy(TSurface& src_surface, TSurface& dst_surface) = 0;
 
     void Register(TSurface surface) {
         std::lock_guard lock{mutex};
@@ -516,8 +516,9 @@ private:
         // Step 1
         // Check Level 1 Cache for a fast structural match. If candidate surface
         // matches at certain level we are pretty much done.
-        if (l1_cache.count(cache_addr) > 0) {
-            TSurface current_surface = l1_cache[cache_addr];
+        auto iter = l1_cache.find(cache_addr);
+        if (iter != l1_cache.end()) {
+            TSurface& current_surface = iter->second;
             auto topological_result = current_surface->MatchesTopology(params);
             if (topological_result != MatchTopologyResult::FullMatch) {
                 std::vector<TSurface> overlaps{current_surface};
@@ -526,7 +527,6 @@ private:
             }
             MatchStructureResult s_result = current_surface->MatchesStructure(params);
             if (s_result != MatchStructureResult::None &&
-                current_surface->GetGpuAddr() == gpu_addr &&
                 (params.target != SurfaceTarget::Texture3D ||
                  current_surface->MatchTarget(params.target))) {
                 if (s_result == MatchStructureResult::FullMatch) {