MSAA works with clustered light list generation

Changed clustered kernel string generation
LightLoop settings disables FPTL with MSAA
Disable SSS HTILE allocation if SSS is diabled (fails with MSAA)

I want to make the config more dynamic, then...I can start re-adding passes!

ScriptableRenderPipeline/HDRenderPipeline/HDRP/HDRenderPipeline.cs

         bool m_FullScreenDebugPushed;
 
         // temp hack
-        public const int kMsaaSamplesFixed = 1;
+        public const int kMsaaSamplesFixed = 4;
 
         public HDRenderPipeline(HDRenderPipelineAsset asset)
         {
             if (resolutionChanged || m_CameraDepthStencilBuffer == null)
             {
                 CreateDepthStencilBuffer(hdCamera);
-                m_SSSBufferManager.Resize(hdCamera);
+
+                if (m_FrameSettings.enableSubsurfaceScattering)
+                {
+                    // TODO: The R8 target doesn't allocate with MSAA support...
+                    m_SSSBufferManager.Resize(hdCamera);
+                }
             }
 
             if (resolutionChanged || m_LightLoop.NeedResize())

ScriptableRenderPipeline/HDRenderPipeline/HDRP/Lighting/LightLoop/LightLoop.cs

 
             // Cluster
             {
-                var kernelName = m_FrameSettings.lightLoopSettings.enableBigTilePrepass ? (k_UseDepthBuffer ? "TileLightListGen_DepthRT_SrcBigTile" : "TileLightListGen_NoDepthRT_SrcBigTile") : (k_UseDepthBuffer ? "TileLightListGen_DepthRT" : "TileLightListGen_NoDepthRT");
+                //var kernelName = m_FrameSettings.lightLoopSettings.enableBigTilePrepass ? (k_UseDepthBuffer ? "TileLightListGen_DepthRT_SrcBigTile" : "TileLightListGen_NoDepthRT_SrcBigTile") : (k_UseDepthBuffer ? "TileLightListGen_DepthRT" : "TileLightListGen_NoDepthRT");
+
+                var kernelName = "TileLightListGen";
+                if (k_UseDepthBuffer)
+                {
+                    kernelName += "_DepthRT";
+
+                    // if MSAA is actually active, use the MSAA path
+                    if (m_FrameSettings.enableMSAA && 
+                        (HDRenderPipeline.kMsaaSamplesFixed > 1))
+                        kernelName += "_MSAA";
+                }
+                else
+                {
+                    kernelName += "_NoDepthRT";
+                }
+                if (m_FrameSettings.lightLoopSettings.enableBigTilePrepass)
+                    kernelName += "_SrcBigTile";
                 s_GenListPerVoxelKernel = buildPerVoxelLightListShader.FindKernel(kernelName);
             }
 

ScriptableRenderPipeline/HDRenderPipeline/HDRP/Lighting/LightLoop/LightLoopSettings.cs

             // Deferred opaque are always using Fptl. Forward opaque can use Fptl or Cluster, transparent use cluster.
             // When MSAA is enabled we disable Fptl as it become expensive compare to cluster
             // In HD, MSAA is only supported for forward only rendering, no MSAA in deferred mode (for code complexity reasons)
-            aggregate.enableFptlForForwardOpaque = aggregate.enableFptlForForwardOpaque && aggregateFrameSettings.enableMSAA;
+            //aggregate.enableFptlForForwardOpaque = aggregate.enableFptlForForwardOpaque && aggregateFrameSettings.enableMSAA;
+            aggregate.enableFptlForForwardOpaque = aggregate.enableFptlForForwardOpaque && !aggregateFrameSettings.enableMSAA;
             // If Deferred, enable Fptl. If we are forward renderer only and not using Fptl for forward opaque, disable Fptl
             aggregate.isFptlEnabled = !aggregateFrameSettings.enableForwardRenderingOnly || aggregate.enableFptlForForwardOpaque;
         }