Backported fixes that we already know are going to be necessary for full Unity support:

- Removed empty parantheses on GroupMemoryBarrier macro, which confuse the 2020.1 shader compiler
- Added padding field to cbFSR2 to make it exactly 128 bytes in size, which gives correct buffer alignment on iOS Metal
- Changed auto-exposure reset threshold value to 1e4f, as part of a fix for black screen flashes in OpenGL Core on Nvidia GPUs
- Clamp luma to >= 0 in auto-exposure to fix artifacting in OpenGL Core on Nvidia GPUs
- Removed #extension directives meant for GLSL, which cause shader compiler warnings in Unity

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

@@ -32,7 +32,7 @@
 /// A define for abstracting compute memory barriers between shading languages.
 ///
 /// @ingroup HLSLCore
-#define FFXM_GROUP_MEMORY_BARRIER() GroupMemoryBarrierWithGroupSync()
+#define FFXM_GROUP_MEMORY_BARRIER GroupMemoryBarrierWithGroupSync()
 
 /// A define for abstracting compute atomic additions between shading languages.
 ///

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

@@ -79,6 +79,8 @@
         FfxFloat32    fDeltaTime;
         FfxFloat32    fDynamicResChangeFactor;
         FfxFloat32    fViewSpaceToMetersFactor;
+
+        FfxFloat32    fPadding;
     };
 
 #define FFXM_FSR2_CONSTANT_BUFFER_1_SIZE (sizeof(cbFSR2) / 4)  // Number of 32-bit values. This must be kept in sync with the cbFSR2 size.

Packages/com.unity.postprocessing@3.2.2/PostProcessing/Runtime/Effects/Upscaling/ASR/Shaders/shaders/fsr2/ffxm_fsr2_common.h

@@ -57,7 +57,7 @@ FFXM_STATIC const FfxFloat32 fAverageLanczosWeightPerFrame = 0.74f * fUpsampleLa
 FFXM_STATIC const FfxFloat32 fAccumulationMaxOnMotion = 3.0f * fUpsampleLanczosWeightScale;
 
 // Auto exposure
-FFXM_STATIC const FfxFloat32 resetAutoExposureAverageSmoothing = 1e8f;
+FFXM_STATIC const FfxFloat32 resetAutoExposureAverageSmoothing = 1e4f;
 
 // Optimizations defines
 #ifndef FFXM_OPT_USE_GATHER_OPS

Packages/com.unity.postprocessing@3.2.2/PostProcessing/Runtime/Effects/Upscaling/ASR/Shaders/shaders/fsr2/ffxm_fsr2_compute_luminance_pyramid.h

@@ -82,6 +82,7 @@ void SpdStore(FfxInt32x2 pix, FfxFloat32x4 outValue, FfxUInt32 index, FfxUInt32
             {
                 FfxFloat32 rate = 1.0f;
                 result = prev + (result - prev) * (1 - exp(-DeltaTime() * rate));
+                result = ffxMax(0.0f, result);
             }
             FfxFloat32x2 spdOutput = FfxFloat32x2(ComputeAutoExposureFromLavg(result), result);
             SPD_SetExposureBuffer(spdOutput);

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

@@ -119,12 +119,12 @@ FfxFloat32x4 SpdReduce4(FfxFloat32x4 v0, FfxFloat32x4 v1, FfxFloat32x4 v2, FfxFl
 
 //_____________________________________________________________/\_______________________________________________________________
 #if defined(FFXM_GLSL) && !defined(FFXM_SPD_NO_WAVE_OPERATIONS)
-#extension GL_KHR_shader_subgroup_quad:require
+//#extension GL_KHR_shader_subgroup_quad:require
 #endif
 
 void ffxSpdWorkgroupShuffleBarrier()
 {
-    FFXM_GROUP_MEMORY_BARRIER();
+    FFXM_GROUP_MEMORY_BARRIER;
 }
 
 // Only last active workgroup should proceed
@@ -578,7 +578,7 @@ void SpdDownsample(FfxUInt32x2 workGroupID, FfxUInt32 localInvocationIndex, FfxU
 #if FFXM_HALF
 
 #if defined(FFXM_GLSL)
-#extension GL_EXT_shader_subgroup_extended_types_float16:require
+//#extension GL_EXT_shader_subgroup_extended_types_float16:require
 #endif
 
 FfxFloat16x4 SpdReduceQuadH(FfxFloat16x4 v)