Implement lazy flush-on-read for Buffers (SSBO/Copy) (#1790)

* Initial implementation of buffer flush (VERY WIP)

* Host shaders need to be rebuilt for the SSBO write flag.

* New approach with reserved regions and gl sync

* Fix a ton of buffer issues.

* Remove unused buffer unmapped behaviour

* Revert "Remove unused buffer unmapped behaviour"

This reverts commit f1700e52fb8760180ac5e0987a07d409d1e70ece.

* Delete modified ranges on unmap

Fixes potential crashes in Super Smash Bros, where a previously modified range could lie on either side of an unmap.

* Cache some more delegates.

* Dispose Sync on Close

* Also create host sync for GPFifo syncpoint increment.

* Copy buffer optimization, add docs

* Fix race condition with OpenGL Sync

* Enable read tracking on CommandBuffer, insert syncpoint on WaitForIdle

* Performance: Only flush individual pages of SSBO at a time

This avoids flushing large amounts of data when only a small amount is actually used.

* Signal Modified rather than flushing after clear

* Fix some docs and code style.

* Introduce a new test for tracking memory protection.

Sucessfully demonstrates that the bug causing write protection to be cleared by a read action has been fixed. (these tests fail on master)

* Address Comments

* Add host sync for SetReference

This ensures that any indirect draws will correctly flush any related buffer data written before them. Fixes some flashing and misplaced world geometry in MH rise.

* Make PageAlign static

* Re-enable read tracking, for reads.

5 files changed, 18 insertions, 3 deletions

diff --git a/Ryujinx.Graphics.Gpu/Engine/Compute.cs b/Ryujinx.Graphics.Gpu/Engine/Compute.cs
index fd3114a7..c7e059ba 100644
--- a/Ryujinx.Graphics.Gpu/Engine/Compute.cs
+++ b/Ryujinx.Graphics.Gpu/Engine/Compute.cs
@@ -97,7 +97,7 @@ namespace Ryujinx.Graphics.Gpu.Engine
 
                 SbDescriptor sbDescriptor = _context.PhysicalMemory.Read<SbDescriptor>(sbDescAddress);
 
-                BufferManager.SetComputeStorageBuffer(sb.Slot, sbDescriptor.PackAddress(), (uint)sbDescriptor.Size);
+                BufferManager.SetComputeStorageBuffer(sb.Slot, sbDescriptor.PackAddress(), (uint)sbDescriptor.Size, sb.Flags);
             }
 
             BufferManager.SetComputeStorageBufferBindings(info.SBuffers);
diff --git a/Ryujinx.Graphics.Gpu/Engine/GPFifo/GPFifoClass.cs b/Ryujinx.Graphics.Gpu/Engine/GPFifo/GPFifoClass.cs
index 0e87aa3d..84d35350 100644
--- a/Ryujinx.Graphics.Gpu/Engine/GPFifo/GPFifoClass.cs
+++ b/Ryujinx.Graphics.Gpu/Engine/GPFifo/GPFifoClass.cs
@@ -39,6 +39,7 @@ namespace Ryujinx.Graphics.Gpu.Engine.GPFifo
                 { nameof(GPFifoClassState.Semaphored), new RwCallback(Semaphored, null) },
                 { nameof(GPFifoClassState.Syncpointb), new RwCallback(Syncpointb, null) },
                 { nameof(GPFifoClassState.WaitForIdle), new RwCallback(WaitForIdle, null) },
+                { nameof(GPFifoClassState.SetReference), new RwCallback(SetReference, null) },
                 { nameof(GPFifoClassState.LoadMmeInstructionRam), new RwCallback(LoadMmeInstructionRam, null) },
                 { nameof(GPFifoClassState.LoadMmeStartAddressRam), new RwCallback(LoadMmeStartAddressRam, null) },
                 { nameof(GPFifoClassState.SetMmeShadowRamControl), new RwCallback(SetMmeShadowRamControl, null) }
@@ -136,6 +137,7 @@ namespace Ryujinx.Graphics.Gpu.Engine.GPFifo
             }
             else if (operation == SyncpointbOperation.Incr)
             {
+                _context.CreateHostSyncIfNeeded();
                 _context.Synchronization.IncrementSyncpoint(syncpointId);
             }
 
@@ -150,6 +152,17 @@ namespace Ryujinx.Graphics.Gpu.Engine.GPFifo
         {
             _context.Methods.PerformDeferredDraws();
             _context.Renderer.Pipeline.Barrier();
+
+            _context.CreateHostSyncIfNeeded();
+        }
+
+        /// <summary>
+        /// Used as an indirect data barrier on NVN. When used, access to previously written data must be coherent.
+        /// </summary>
+        /// <param name="argument">Method call argument</param>
+        public void SetReference(int argument)
+        {
+            _context.CreateHostSyncIfNeeded();
         }
 
         /// <summary>
diff --git a/Ryujinx.Graphics.Gpu/Engine/GPFifo/GPFifoDevice.cs b/Ryujinx.Graphics.Gpu/Engine/GPFifo/GPFifoDevice.cs
index 25614a13..d0fcf142 100644
--- a/Ryujinx.Graphics.Gpu/Engine/GPFifo/GPFifoDevice.cs
+++ b/Ryujinx.Graphics.Gpu/Engine/GPFifo/GPFifoDevice.cs
@@ -52,7 +52,7 @@ namespace Ryujinx.Graphics.Gpu.Engine.GPFifo
             {
                 if (Words == null)
                 {
-                    Words = MemoryMarshal.Cast<byte, int>(context.MemoryManager.GetSpan(EntryAddress, (int)EntryCount * 4)).ToArray();
+                    Words = MemoryMarshal.Cast<byte, int>(context.MemoryManager.GetSpan(EntryAddress, (int)EntryCount * 4, true)).ToArray();
                 }
             }
         }
diff --git a/Ryujinx.Graphics.Gpu/Engine/MethodIncrementSyncpoint.cs b/Ryujinx.Graphics.Gpu/Engine/MethodIncrementSyncpoint.cs
index 8fcfb900..9c22275d 100644
--- a/Ryujinx.Graphics.Gpu/Engine/MethodIncrementSyncpoint.cs
+++ b/Ryujinx.Graphics.Gpu/Engine/MethodIncrementSyncpoint.cs
@@ -13,6 +13,7 @@ namespace Ryujinx.Graphics.Gpu.Engine
         {
             uint syncpointId = (uint)(argument) & 0xFFFF;
 
+            _context.CreateHostSyncIfNeeded();
             _context.Renderer.UpdateCounters(); // Poll the query counters, the game may want an updated result.
             _context.Synchronization.IncrementSyncpoint(syncpointId);
         }
diff --git a/Ryujinx.Graphics.Gpu/Engine/Methods.cs b/Ryujinx.Graphics.Gpu/Engine/Methods.cs
index 9f27aec2..d6bd5110 100644
--- a/Ryujinx.Graphics.Gpu/Engine/Methods.cs
+++ b/Ryujinx.Graphics.Gpu/Engine/Methods.cs
@@ -61,6 +61,7 @@ namespace Ryujinx.Graphics.Gpu.Engine
 
             context.MemoryManager.MemoryUnmapped += _counterCache.MemoryUnmappedHandler;
             context.MemoryManager.MemoryUnmapped += TextureManager.MemoryUnmappedHandler;
+            context.MemoryManager.MemoryUnmapped += BufferManager.MemoryUnmappedHandler;
         }
 
         /// <summary>
@@ -333,7 +334,7 @@ namespace Ryujinx.Graphics.Gpu.Engine
 
                     SbDescriptor sbDescriptor = _context.PhysicalMemory.Read<SbDescriptor>(sbDescAddress);
 
-                    BufferManager.SetGraphicsStorageBuffer(stage, sb.Slot, sbDescriptor.PackAddress(), (uint)sbDescriptor.Size);
+                    BufferManager.SetGraphicsStorageBuffer(stage, sb.Slot, sbDescriptor.PackAddress(), (uint)sbDescriptor.Size, sb.Flags);
                 }
             }
         }