Merge branch 'asr-console' into armasr

# Conflicts:
#	Packages/com.unity.postprocessing@3.2.2/PostProcessing/Runtime/Effects/Upscaling/ASR/Shaders/ffxm_fsr2_common.cginc

Packages/com.unity.postprocessing@3.2.2/PostProcessing/Runtime/Effects/Upscaling/ASR/Runtime/Asr.cs

@@ -44,7 +44,7 @@ namespace ArmASR
             flags |= InitializationFlags.EnableDebugChecking;
 #endif
             
-            Debug.Log($"Setting up ASR with render size: {maxRenderSize.x}x{maxRenderSize.y}, display size: {displaySize.x}x{displaySize.y}, flags: {flags}");
+            Debug.Log($"Setting up ASR {variant} with render size: {maxRenderSize.x}x{maxRenderSize.y}, display size: {displaySize.x}x{displaySize.y}, flags: {flags}");
             
             var contextDescription = new ContextDescription
             {

Packages/com.unity.postprocessing@3.2.2/PostProcessing/Runtime/Effects/Upscaling/ASR/Runtime/AsrContext.cs

@@ -96,7 +96,7 @@ namespace ArmASR
             _lockPass = new AsrLockPass(_contextDescription, _resources, _upscalerConstantsBuffer);
             _accumulatePass = new AsrAccumulatePass(_contextDescription, _resources, _upscalerConstantsBuffer);
             _sharpenPass = new AsrSharpenPass(_contextDescription, _resources, _upscalerConstantsBuffer, _rcasConstantsBuffer);
-            _generateReactivePass = new AsrGenerateReactivePass(_contextDescription, _resources, _generateReactiveConstantsBuffer);
+            _generateReactivePass = new AsrGenerateReactivePass(_contextDescription, _resources, _upscalerConstantsBuffer, _generateReactiveConstantsBuffer);
         }
         
         public void Destroy()
@@ -198,6 +198,9 @@ namespace ArmASR
             commandBuffer.SetRenderTarget(AsrShaderIDs.UavReconstructedPrevNearestDepth);
             commandBuffer.ClearRenderTarget(false, true, depthInverted ? Color.clear : Color.white);
             
+            commandBuffer.SetRenderTarget(AsrShaderIDs.UavNewLocks);
+            commandBuffer.ClearRenderTarget(false, true, Color.clear);
+            
             // Auto exposure
             SetupSpdConstants(dispatchParams, out var dispatchThreadGroupCount);
             

Packages/com.unity.postprocessing@3.2.2/PostProcessing/Runtime/Effects/Upscaling/ASR/Runtime/AsrPass.cs

@@ -215,8 +215,8 @@ namespace ArmASR
             commandBuffer.SetGlobalTexture(AsrShaderIDs.SrvDilatedDepth, AsrShaderIDs.UavDilatedDepth);
             commandBuffer.SetGlobalTexture(AsrShaderIDs.SrvPrevDilatedMotionVectors, Resources.DilatedMotionVectors[frameIndex ^ 1]);
 
-            _mrt[0] = AsrShaderIDs.RtDilatedReactiveMasks;  // fDilatedReactiveMasks
-            _mrt[1] = AsrShaderIDs.RtPreparedInputColor;    // fTonemapped
+            _mrt[0] = AsrShaderIDs.RtPreparedInputColor;    // fTonemapped
+            _mrt[1] = AsrShaderIDs.RtDilatedReactiveMasks;  // fDilatedReactiveMasks
             
             FragmentProperties.SetConstantBuffer(AsrShaderIDs.CbFsr2, Constants, 0, Constants.stride);
             BlitFragment(commandBuffer, _mrt);
@@ -277,9 +277,7 @@ namespace ArmASR
             commandBuffer.SetGlobalTexture(AsrShaderIDs.SrvAutoExposure, Resources.AutoExposure[frameIndex]);
             commandBuffer.SetGlobalTexture(AsrShaderIDs.SrvLumaHistory, Resources.LumaHistory[frameIndex ^ 1]);
             commandBuffer.SetGlobalTexture(AsrShaderIDs.SrvInternalTemporalReactive, Resources.InternalReactive[frameIndex ^ 1]);
-            
-            // UAV binding in fragment shader, index needs to match the register binding in HLSL
-            commandBuffer.SetRandomWriteTarget(4, AsrShaderIDs.UavNewLocks);
+            commandBuffer.SetGlobalTexture(AsrShaderIDs.SrvNewLocks, AsrShaderIDs.UavNewLocks);
 
             if (ContextDescription.Variant == Asr.Variant.Quality)
             {
@@ -329,8 +327,8 @@ namespace ArmASR
     {
         private readonly ComputeBuffer _generateReactiveConstants;
 
-        public AsrGenerateReactivePass(Asr.ContextDescription contextDescription, AsrResources resources, ComputeBuffer generateReactiveConstants)
-            : base(contextDescription, resources, null)
+        public AsrGenerateReactivePass(Asr.ContextDescription contextDescription, AsrResources resources, ComputeBuffer constants, ComputeBuffer generateReactiveConstants)
+            : base(contextDescription, resources, constants)
         {
             _generateReactiveConstants = generateReactiveConstants;
             
@@ -348,6 +346,7 @@ namespace ArmASR
             commandBuffer.SetGlobalResource(AsrShaderIDs.SrvOpaqueOnly, dispatchParams.ColorOpaqueOnly);
             commandBuffer.SetGlobalResource(AsrShaderIDs.SrvInputColor, dispatchParams.ColorPreUpscale);
             
+            FragmentProperties.SetConstantBuffer(AsrShaderIDs.CbFsr2, Constants, 0, Constants.stride);
             FragmentProperties.SetConstantBuffer(AsrShaderIDs.CbGenReactive, _generateReactiveConstants, 0, _generateReactiveConstants.stride);
             BlitFragment(commandBuffer, dispatchParams.OutReactive.RenderTarget);
             

Packages/com.unity.postprocessing@3.2.2/PostProcessing/Runtime/Effects/Upscaling/ASR/Shaders/ffxm_fsr2_common.cginc

@@ -1,4 +1,4 @@
-#pragma warning(disable: 3078)      // Loop control variable conflicts
+#pragma warning(disable: 3078)      // Loop control variable conflicts
 #pragma warning(disable: 3203)      // Signed/unsigned mismatch
 #pragma warning(disable: 3205)      // Conversion from larger type to smaller, possible loss of data
 #pragma warning(disable: 3556)      // Integer divides might be much slower, try using uints if possible
@@ -6,7 +6,7 @@
 // Doesn't work for:
 // - DX11: forces use of DXC which is not supported
 // - XB1: native FP16 seems to be unsupported
-#if defined(SHADER_API_VULKAN)
+#if defined(SHADER_API_VULKAN) || defined(SHADER_API_PSSL)
 //#pragma require Native16Bit   // *sigh* Unity STILL ignores the #if guard when compiling for standalone
 #endif
 
@@ -38,3 +38,15 @@
 #define InterlockedMin(dest, val)           { (dest) = min((dest), (val)); }
 #define InterlockedMax(dest, val)           { (dest) = max((dest), (val)); }
 #endif
+
+// PSSL uses different semantics and doesn't support certain type qualifiers 
+#if defined(SHADER_API_PSSL)
+#define SV_VERTEXID                 S_VERTEX_ID
+#define SV_POSITION                 S_POSITION
+#define SV_TARGET0                  S_TARGET_OUTPUT0
+#define SV_TARGET1                  S_TARGET_OUTPUT1
+#define SV_TARGET2                  S_TARGET_OUTPUT2
+#define SV_TARGET3                  S_TARGET_OUTPUT3
+#define unorm
+#define globallycoherent
+#endif

Packages/com.unity.postprocessing@3.2.2/PostProcessing/Runtime/Effects/Upscaling/ASR/Shaders/shaders/ffxm_common_types.h

@@ -472,10 +472,7 @@ typedef FfxUInt32x4     Prefix##_U4;
 
 #if defined(SHADER_API_PSSL)
 
-#define unorm
-#define globallycoherent
-
-#if FFX_HALF
+#if FFXM_HALF
 
 #define  FFXM_MIN16_F  half
 #define  FFXM_MIN16_F2 half2
@@ -507,7 +504,7 @@ typedef FfxUInt32x4     Prefix##_U4;
 #define FFXM_16BIT_U3 ushort3
 #define FFXM_16BIT_U4 ushort4
 
-#else // FFX_HALF
+#else // FFXM_HALF
 
 #define  FFXM_MIN16_F  float
 #define  FFXM_MIN16_F2 float2
@@ -539,7 +536,7 @@ typedef FfxUInt32x4     Prefix##_U4;
 #define FFXM_16BIT_U3 uint3
 #define FFXM_16BIT_U4 uint4
 
-#endif // FFX_HALF
+#endif // FFXM_HALF
 
 #endif  // #if defined(SHADER_API_PSSL)
 

Packages/com.unity.postprocessing@3.2.2/PostProcessing/Runtime/Effects/Upscaling/ASR/Shaders/shaders/ffxm_fsr2_accumulate_pass_fs.hlsl

@@ -35,8 +35,7 @@
 #define FSR2_BIND_SRV_AUTO_EXPOSURE                          9
 #define FSR2_BIND_SRV_LUMA_HISTORY                           10
 #define FSR2_BIND_SRV_TEMPORAL_REACTIVE                      11
-
-#define FSR2_BIND_UAV_NEW_LOCKS                              4
+#define FSR2_BIND_SRV_NEW_LOCKS                              12
 
 #define FSR2_BIND_CB_FSR2                                    0
 
@@ -56,6 +55,12 @@
 #include "fsr2/ffxm_fsr2_reproject.h"
 #include "fsr2/ffxm_fsr2_accumulate.h"
 
+#if defined(SHADER_API_PSSL)
+#pragma PSSL_target_output_format(target 0 FMT_FP16_ABGR)
+#pragma PSSL_target_output_format(target 1 FMT_FP16_ABGR)
+#pragma PSSL_target_output_format(target 2 FMT_FP16_ABGR)
+#endif
+
 struct AccumulateOutputsFS
 {
 #if !FFXM_SHADER_QUALITY_BALANCED_OR_PERFORMANCE
@@ -63,14 +68,14 @@ struct AccumulateOutputsFS
     FfxFloat32x2 fLockStatus        : SV_TARGET1;
     FfxFloat32x4 fLumaHistory       : SV_TARGET2;
 #if FFXM_FSR2_OPTION_APPLY_SHARPENING == 0
-    FfxFloat32x3 fColor             : SV_TARGET3;
+    FfxFloat32x4 fColor             : SV_TARGET3;
 #endif
 #else // FFXM_SHADER_QUALITY_BALANCED_OR_PERFORMANCE
-    FfxFloat32x3 fUpscaledColor     : SV_TARGET0;
+    FfxFloat32x4 fUpscaledColor     : SV_TARGET0;
     FfxFloat32 fTemporalReactive    : SV_TARGET1;
     FfxFloat32x2 fLockStatus        : SV_TARGET2;
 #if FFXM_FSR2_OPTION_APPLY_SHARPENING == 0
-    FfxFloat32x3 fColor             : SV_TARGET3;
+    FfxFloat32x4 fColor             : SV_TARGET3;
 #endif
 #endif
 };
@@ -84,12 +89,12 @@ AccumulateOutputsFS main(float4 SvPosition : SV_POSITION)
     output.fColorAndWeight = result.fColorAndWeight;
     output.fLumaHistory = result.fLumaHistory;
 #else
-    output.fUpscaledColor = result.fUpscaledColor;
+    output.fUpscaledColor = FfxFloat32x4(result.fUpscaledColor, 1.0f);
     output.fTemporalReactive = result.fTemporalReactive;
 #endif
     output.fLockStatus = result.fLockStatus;
 #if FFXM_FSR2_OPTION_APPLY_SHARPENING == 0
-    output.fColor = result.fColor;
+    output.fColor = FfxFloat32x4(result.fColor, 1.0f);
 #endif
     return output;
 }

Packages/com.unity.postprocessing@3.2.2/PostProcessing/Runtime/Effects/Upscaling/ASR/Shaders/shaders/ffxm_fsr2_autogen_reactive_pass_fs.hlsl

@@ -36,6 +36,10 @@
 #include "fsr2/ffxm_fsr2_callbacks_hlsl.h"
 #include "fsr2/ffxm_fsr2_common.h"
 
+#if defined(SHADER_API_PSSL)
+#pragma PSSL_target_output_format(default FMT_FP16_ABGR)
+#endif
+
 struct GenReactiveMaskOutputs
 {
     FfxFloat32 fReactiveMask : SV_TARGET0;

Packages/com.unity.postprocessing@3.2.2/PostProcessing/Runtime/Effects/Upscaling/ASR/Shaders/shaders/ffxm_fsr2_depth_clip_pass_fs.hlsl

@@ -45,10 +45,15 @@
 #include "fsr2/ffxm_fsr2_sample.h"
 #include "fsr2/ffxm_fsr2_depth_clip.h"
 
+#if defined(SHADER_API_PSSL)
+#pragma PSSL_target_output_format(target 0 FMT_FP16_ABGR)
+#pragma PSSL_target_output_format(target 1 FMT_FP16_ABGR)
+#endif
+
 struct DepthClipOutputsFS
 {
-    FfxFloat32x2 fDilatedReactiveMasks    : SV_TARGET0;
-    FfxFloat32x4 fTonemapped              : SV_TARGET1;
+    FfxFloat32x4 fTonemapped              : SV_TARGET0;
+    FfxFloat32x2 fDilatedReactiveMasks    : SV_TARGET1;
 };
 
 DepthClipOutputsFS main(float4 SvPosition : SV_POSITION)
@@ -56,7 +61,7 @@ DepthClipOutputsFS main(float4 SvPosition : SV_POSITION)
     uint2 uPixelCoord = uint2(SvPosition.xy);
     DepthClipOutputs result = DepthClip(uPixelCoord);
     DepthClipOutputsFS output = (DepthClipOutputsFS)0;
-    output.fDilatedReactiveMasks = result.fDilatedReactiveMasks;
     output.fTonemapped = result.fTonemapped;
+    output.fDilatedReactiveMasks = result.fDilatedReactiveMasks;
     return output;
 }

Packages/com.unity.postprocessing@3.2.2/PostProcessing/Runtime/Effects/Upscaling/ASR/Shaders/shaders/ffxm_fsr2_rcas_pass_fs.hlsl

@@ -39,7 +39,7 @@
 
 struct RCASOutputsFS
 {
-    FfxFloat32x3 fUpscaledColor    : SV_TARGET0;
+    FfxFloat32x4 fUpscaledColor    : SV_TARGET0;
 };
 
 RCASOutputsFS main(float4 SvPosition : SV_POSITION)
@@ -47,6 +47,6 @@ RCASOutputsFS main(float4 SvPosition : SV_POSITION)
     uint2 uPixelCoord = uint2(SvPosition.xy);
     RCASOutputs result = RCAS(uPixelCoord);
     RCASOutputsFS output = (RCASOutputsFS)0;
-    output.fUpscaledColor = result.fUpscaledColor;
+    output.fUpscaledColor = FfxFloat32x4(result.fUpscaledColor, 1.0f);
     return output;
 }

Packages/com.unity.postprocessing@3.2.2/PostProcessing/Runtime/Effects/Upscaling/ASR/Shaders/shaders/ffxm_fsr2_reconstruct_previous_depth_pass_fs.hlsl

@@ -41,6 +41,13 @@
 #include "fsr2/ffxm_fsr2_sample.h"
 #include "fsr2/ffxm_fsr2_reconstruct_dilated_velocity_and_previous_depth.h"
 
+#if defined(SHADER_API_PSSL)
+#pragma PSSL_target_output_format(target 0 FMT_32_R)
+#pragma PSSL_target_output_format(target 1 FMT_FP16_ABGR)
+#pragma PSSL_target_output_format(target 2 FMT_32_R)
+#pragma PSSL_target_output_format(target 3 FMT_32_R)
+#endif
+
 struct ReconstructPrevDepthOutputsFS
 {
     FfxFloat32 fDepth           : SV_TARGET0;

Packages/com.unity.postprocessing@3.2.2/PostProcessing/Runtime/Effects/Upscaling/ASR/Shaders/shaders/fsr2/ffxm_fsr2_accumulate.h

@@ -372,7 +372,6 @@ AccumulateOutputs Accumulate(FfxInt32x2 iPxHrPos)
 #if FFXM_FSR2_OPTION_APPLY_SHARPENING == 0
     results.fColor = fHistoryColor;
 #endif
-    StoreNewLocks(iPxHrPos, 0);
 
     return results;
 }

Packages/com.unity.postprocessing@3.2.2/PostProcessing/Runtime/Effects/Upscaling/ASR/Shaders/shaders/fsr2/ffxm_fsr2_callbacks_hlsl.h

@@ -379,7 +379,11 @@ SamplerState s_LinearClamp : register(s1);
 
     // UAV declarations
     #if defined FSR2_BIND_UAV_RECONSTRUCTED_PREV_NEAREST_DEPTH
+        #if defined(SHADER_API_PSSL) && defined(SHADER_STAGE_FRAGMENT)
+        RWTexture2D<FfxUInt32> rw_reconstructed_previous_nearest_depth;     // Need to use implicit register binding for random write targets on PS4/5
+        #else
         RWTexture2D<FfxUInt32> rw_reconstructed_previous_nearest_depth : FFXM_FSR2_DECLARE_UAV(FSR2_BIND_UAV_RECONSTRUCTED_PREV_NEAREST_DEPTH);
+        #endif
     #endif
     #if defined FSR2_BIND_UAV_DILATED_MOTION_VECTORS
         RWTexture2D<FfxFloat32x2> rw_dilated_motion_vectors : FFXM_FSR2_DECLARE_UAV(FSR2_BIND_UAV_DILATED_MOTION_VECTORS);
@@ -397,7 +401,11 @@ SamplerState s_LinearClamp : register(s1);
         RWTexture2D<FfxFloat32> rw_lock_input_luma : FFXM_FSR2_DECLARE_UAV(FSR2_BIND_UAV_LOCK_INPUT_LUMA);
     #endif
     #if defined FSR2_BIND_UAV_NEW_LOCKS
+        #if defined(SHADER_API_PSSL) && defined(SHADER_STAGE_FRAGMENT)
+        RWTexture2D<unorm FfxFloat32> rw_new_locks;     // Need to use implicit register binding for random write targets on PS4/5
+        #else
         RWTexture2D<unorm FfxFloat32> rw_new_locks : FFXM_FSR2_DECLARE_UAV(FSR2_BIND_UAV_NEW_LOCKS);
+        #endif
     #endif
     #if defined FSR2_BIND_UAV_PREPARED_INPUT_COLOR
         RWTexture2D<FfxFloat32x4> rw_prepared_input_color : FFXM_FSR2_DECLARE_UAV(FSR2_BIND_UAV_PREPARED_INPUT_COLOR);

Packages/com.unity.postprocessing@3.2.2/PostProcessing/Runtime/Effects/Upscaling/ASR/Shaders/shaders/fsr2/ffxm_fsr2_depth_clip.h

@@ -27,8 +27,8 @@
 
 struct DepthClipOutputs
 {
-    FfxFloat32x2 fDilatedReactiveMasks;
     FfxFloat32x4 fTonemapped;
+    FfxFloat32x2 fDilatedReactiveMasks;
 };
 
 FFXM_STATIC const FfxFloat32 DepthClipBaseScale = 4.0f;

Packages/com.unity.postprocessing@3.2.2/PostProcessing/Runtime/Effects/Upscaling/ASR/Shaders/shaders/fsr2/ffxm_fsr2_reproject.h

@@ -331,7 +331,7 @@ void ReprojectHistoryColor(const AccumulationPassCommonParams params, FFXM_PARAM
 LockState ReprojectHistoryLockStatus(const AccumulationPassCommonParams params, FFXM_PARAMETER_OUT FfxFloat32x2 fReprojectedLockStatus)
 {
     LockState state = { FFXM_FALSE, FFXM_FALSE };
-    const FfxFloat32 fNewLockIntensity = LoadRwNewLocks(params.iPxHrPos);
+    const FfxFloat32 fNewLockIntensity = LoadNewLocks(params.iPxHrPos);
     state.NewLock = fNewLockIntensity > (127.0f / 255.0f);
 
     FfxFloat32 fInPlaceLockLifetime = state.NewLock ? fNewLockIntensity : 0;
@@ -368,7 +368,7 @@ void ReprojectHistoryColor(const AccumulationPassCommonParams params, FFXM_PARAM
 LockState ReprojectHistoryLockStatus(const AccumulationPassCommonParams params, FFXM_PARAMETER_OUT FfxFloat16x2 fReprojectedLockStatus)
 {
     LockState state = { FFXM_FALSE, FFXM_FALSE };
-    const FfxFloat16 fNewLockIntensity = FfxFloat16(LoadRwNewLocks(params.iPxHrPos));
+    const FfxFloat16 fNewLockIntensity = FfxFloat16(LoadNewLocks(params.iPxHrPos));
     state.NewLock = fNewLockIntensity > (127.0f / 255.0f);
 
     FfxFloat16 fInPlaceLockLifetime = state.NewLock ? fNewLockIntensity : FfxFloat16(0);