De-tile GOB when DMA copying from block linear to pitch kind memory regions (#3207)

* De-tile GOB when DMA copying from block linear to pitch kind memory regions

* XML docs + nits

* Remove using

* No flush for regular buffer copies

* Add back ulong casts, fix regression due to oversight

ARMeilleure/Instructions/InstEmitMemoryEx.cs

@@ -130,11 +130,6 @@ namespace ARMeilleure.Instructions
             bool ordered   = (accType & AccessType.Ordered)   != 0;
             bool exclusive = (accType & AccessType.Exclusive) != 0;
 
-            if (ordered)
-            {
-                EmitBarrier(context);
-            }
-
             Operand address = context.Copy(GetIntOrSP(context, op.Rn));
 
             Operand t = GetIntOrZR(context, op.Rt);
@@ -163,6 +158,11 @@ namespace ARMeilleure.Instructions
             {
                 EmitStoreExclusive(context, address, t, exclusive, op.Size, op.Rs, a32: false);
             }
+
+            if (ordered)
+            {
+                EmitBarrier(context);
+            }
         }
 
         private static void EmitBarrier(ArmEmitterContext context)

ARMeilleure/Instructions/InstEmitMemoryEx32.cs

@@ -146,13 +146,13 @@ namespace ARMeilleure.Instructions
             var exclusive = (accType & AccessType.Exclusive) != 0;
             var ordered = (accType & AccessType.Ordered) != 0;
 
-            if (ordered)
-            {
-                EmitBarrier(context);
-            }
-
             if ((accType & AccessType.Load) != 0)
             {
+                if (ordered)
+                {
+                    EmitBarrier(context);
+                }
+
                 if (size == DWordSizeLog2)
                 {
                     // Keep loads atomic - make the call to get the whole region and then decompose it into parts
@@ -219,6 +219,11 @@ namespace ARMeilleure.Instructions
                     Operand value = context.ZeroExtend32(OperandType.I64, GetIntA32(context, op.Rt));
                     EmitStoreExclusive(context, address, value, exclusive, size, op.Rd, a32: true);
                 }
+
+                if (ordered)
+                {
+                    EmitBarrier(context);
+                }
             }
         }
 

ARMeilleure/Translation/PTC/Ptc.cs

@@ -27,7 +27,7 @@ namespace ARMeilleure.Translation.PTC
         private const string OuterHeaderMagicString = "PTCohd\0\0";
         private const string InnerHeaderMagicString = "PTCihd\0\0";
 
-        private const uint InternalVersion = 3179; //! To be incremented manually for each change to the ARMeilleure project.
+        private const uint InternalVersion = 3193; //! To be incremented manually for each change to the ARMeilleure project.
 
         private const string ActualDir = "0";
         private const string BackupDir = "1";

Ryujinx.Common/Logging/LogClass.cs

@@ -47,6 +47,7 @@ namespace Ryujinx.Common.Logging
         ServiceNim,
         ServiceNs,
         ServiceNsd,
+        ServiceNtc,
         ServiceNv,
         ServiceOlsc,
         ServicePctl,

Ryujinx.Graphics.GAL/Format.cs

@@ -52,7 +52,7 @@ namespace Ryujinx.Graphics.GAL
         R32G32B32A32Sint,
         S8Uint,
         D16Unorm,
-        D24X8Unorm,
+        S8UintD24Unorm,
         D32Float,
         D24UnormS8Uint,
         D32FloatS8Uint,
@@ -266,7 +266,7 @@ namespace Ryujinx.Graphics.GAL
             {
                 case Format.D16Unorm:
                 case Format.D24UnormS8Uint:
-                case Format.D24X8Unorm:
+                case Format.S8UintD24Unorm:
                 case Format.D32Float:
                 case Format.D32FloatS8Uint:
                 case Format.S8Uint:

Ryujinx.Graphics.Gpu/Engine/Dma/DmaClass.cs

@@ -1,7 +1,7 @@
 using Ryujinx.Common;
-using Ryujinx.Common.Logging;
 using Ryujinx.Graphics.Device;
 using Ryujinx.Graphics.Gpu.Engine.Threed;
+using Ryujinx.Graphics.Gpu.Memory;
 using Ryujinx.Graphics.Texture;
 using System;
 using System.Collections.Generic;
@@ -330,11 +330,95 @@ namespace Ryujinx.Graphics.Gpu.Engine.Dma
                 {
                     // TODO: Implement remap functionality.
                     // Buffer to buffer copy.
-                    memoryManager.Physical.BufferCache.CopyBuffer(memoryManager, srcGpuVa, dstGpuVa, size);
+
+                    bool srcIsPitchKind = memoryManager.GetKind(srcGpuVa).IsPitch();
+                    bool dstIsPitchKind = memoryManager.GetKind(dstGpuVa).IsPitch();
+
+                    if (!srcIsPitchKind && dstIsPitchKind)
+                    {
+                        CopyGobBlockLinearToLinear(memoryManager, srcGpuVa, dstGpuVa, size);
+                    }
+                    else if (srcIsPitchKind && !dstIsPitchKind)
+                    {
+                        CopyGobLinearToBlockLinear(memoryManager, srcGpuVa, dstGpuVa, size);
+                    }
+                    else
+                    {
+                        memoryManager.Physical.BufferCache.CopyBuffer(memoryManager, srcGpuVa, dstGpuVa, size);
+                    }
                 }
             }
         }
 
+        /// <summary>
+        /// Copies block linear data with block linear GOBs to a block linear destination with linear GOBs.
+        /// </summary>
+        /// <param name="memoryManager">GPU memory manager</param>
+        /// <param name="srcGpuVa">Source GPU virtual address</param>
+        /// <param name="dstGpuVa">Destination GPU virtual address</param>
+        /// <param name="size">Size in bytes of the copy</param>
+        private static void CopyGobBlockLinearToLinear(MemoryManager memoryManager, ulong srcGpuVa, ulong dstGpuVa, ulong size)
+        {
+            if (((srcGpuVa | dstGpuVa | size) & 0xf) == 0)
+            {
+                for (ulong offset = 0; offset < size; offset += 16)
+                {
+                    Vector128<byte> data = memoryManager.Read<Vector128<byte>>(ConvertGobLinearToBlockLinearAddress(srcGpuVa + offset), true);
+                    memoryManager.Write(dstGpuVa + offset, data);
+                }
+            }
+            else
+            {
+                for (ulong offset = 0; offset < size; offset++)
+                {
+                    byte data = memoryManager.Read<byte>(ConvertGobLinearToBlockLinearAddress(srcGpuVa + offset), true);
+                    memoryManager.Write(dstGpuVa + offset, data);
+                }
+            }
+        }
+
+        /// <summary>
+        /// Copies block linear data with linear GOBs to a block linear destination with block linear GOBs.
+        /// </summary>
+        /// <param name="memoryManager">GPU memory manager</param>
+        /// <param name="srcGpuVa">Source GPU virtual address</param>
+        /// <param name="dstGpuVa">Destination GPU virtual address</param>
+        /// <param name="size">Size in bytes of the copy</param>
+        private static void CopyGobLinearToBlockLinear(MemoryManager memoryManager, ulong srcGpuVa, ulong dstGpuVa, ulong size)
+        {
+            if (((srcGpuVa | dstGpuVa | size) & 0xf) == 0)
+            {
+                for (ulong offset = 0; offset < size; offset += 16)
+                {
+                    Vector128<byte> data = memoryManager.Read<Vector128<byte>>(srcGpuVa + offset, true);
+                    memoryManager.Write(ConvertGobLinearToBlockLinearAddress(dstGpuVa + offset), data);
+                }
+            }
+            else
+            {
+                for (ulong offset = 0; offset < size; offset++)
+                {
+                    byte data = memoryManager.Read<byte>(srcGpuVa + offset, true);
+                    memoryManager.Write(ConvertGobLinearToBlockLinearAddress(dstGpuVa + offset), data);
+                }
+            }
+        }
+
+        /// <summary>
+        /// Calculates the GOB block linear address from a linear address.
+        /// </summary>
+        /// <param name="address">Linear address</param>
+        /// <returns>Block linear address</returns>
+        private static ulong ConvertGobLinearToBlockLinearAddress(ulong address)
+        {
+            // y2 y1 y0 x5 x4 x3 x2 x1 x0 -> x5 y2 y1 x4 y0 x3 x2 x1 x0
+            return (address & ~0x1f0UL) |
+                ((address & 0x40) >> 2) |
+                ((address & 0x10) << 1) |
+                ((address & 0x180) >> 1) |
+                ((address & 0x20) << 3);
+        }
+
         /// <summary>
         /// Performs a buffer to buffer, or buffer to texture copy, then optionally releases a semaphore.
         /// </summary>

Ryujinx.Graphics.Gpu/Engine/Types/ZetaFormat.cs

@@ -28,13 +28,13 @@ namespace Ryujinx.Graphics.Gpu.Engine.Types
         {
             return format switch
             {
-                ZetaFormat.D32Float          => new FormatInfo(Format.D32Float,          1, 1, 4,  1),
-                ZetaFormat.D16Unorm          => new FormatInfo(Format.D16Unorm,          1, 1, 2,  1),
-                ZetaFormat.D24UnormS8Uint    => new FormatInfo(Format.D24UnormS8Uint,    1, 1, 4,  2),
-                ZetaFormat.D24Unorm          => new FormatInfo(Format.D24UnormS8Uint,    1, 1, 4,  1),
-                ZetaFormat.S8UintD24Unorm    => new FormatInfo(Format.D24UnormS8Uint,    1, 1, 4,  2),
-                ZetaFormat.S8Uint            => new FormatInfo(Format.S8Uint,            1, 1, 1,  1),
-                ZetaFormat.D32FloatS8Uint    => new FormatInfo(Format.D32FloatS8Uint,    1, 1, 8,  2),
+                ZetaFormat.D32Float          => new FormatInfo(Format.D32Float,       1, 1, 4, 1),
+                ZetaFormat.D16Unorm          => new FormatInfo(Format.D16Unorm,       1, 1, 2, 1),
+                ZetaFormat.D24UnormS8Uint    => new FormatInfo(Format.D24UnormS8Uint, 1, 1, 4, 2),
+                ZetaFormat.D24Unorm          => new FormatInfo(Format.D24UnormS8Uint, 1, 1, 4, 1),
+                ZetaFormat.S8UintD24Unorm    => new FormatInfo(Format.S8UintD24Unorm, 1, 1, 4, 2),
+                ZetaFormat.S8Uint            => new FormatInfo(Format.S8Uint,         1, 1, 1, 1),
+                ZetaFormat.D32FloatS8Uint    => new FormatInfo(Format.D32FloatS8Uint, 1, 1, 8, 2),
                 _                            => FormatInfo.Default
             };
         }

Ryujinx.Graphics.Gpu/Image/FormatTable.cs

@@ -55,6 +55,7 @@ namespace Ryujinx.Graphics.Gpu.Image
             { 0x24a0e, new FormatInfo(Format.D24UnormS8Uint,    1,  1,  4,  2) },
             { 0x24a29, new FormatInfo(Format.D24UnormS8Uint,    1,  1,  4,  2) },
             { 0x48a29, new FormatInfo(Format.D24UnormS8Uint,    1,  1,  4,  2) },
+            { 0x4912b, new FormatInfo(Format.S8UintD24Unorm,    1,  1,  4,  2) },
             { 0x25385, new FormatInfo(Format.D32FloatS8Uint,    1,  1,  8,  2) },
             { 0x253b0, new FormatInfo(Format.D32FloatS8Uint,    1,  1,  8,  2) },
             { 0xa4908, new FormatInfo(Format.R8G8B8A8Srgb,      1,  1,  4,  4) },

Ryujinx.Graphics.Gpu/Image/TextureCache.cs

@@ -7,7 +7,6 @@ using Ryujinx.Graphics.Gpu.Engine.Types;
 using Ryujinx.Graphics.Gpu.Image;
 using Ryujinx.Graphics.Gpu.Memory;
 using Ryujinx.Graphics.Texture;
-using Ryujinx.Memory;
 using Ryujinx.Memory.Range;
 using System;
 using System.Collections.Generic;

Ryujinx.Graphics.Gpu/Image/TextureCompatibility.cs

@@ -203,7 +203,7 @@ namespace Ryujinx.Graphics.Gpu.Image
                 }
 
                 if ((lhs.FormatInfo.Format == Format.D24UnormS8Uint ||
-                     lhs.FormatInfo.Format == Format.D24X8Unorm) && rhs.FormatInfo.Format == Format.B8G8R8A8Unorm)
+                     lhs.FormatInfo.Format == Format.S8UintD24Unorm) && rhs.FormatInfo.Format == Format.B8G8R8A8Unorm)
                 {
                     return TextureMatchQuality.FormatAlias;
                 }

Ryujinx.Graphics.Gpu/Image/TexturePool.cs

@@ -362,7 +362,7 @@ namespace Ryujinx.Graphics.Gpu.Image
                 return DepthStencilMode.Depth;
             }
 
-            if (format == Format.D24X8Unorm || format == Format.D24UnormS8Uint)
+            if (format == Format.D24UnormS8Uint)
             {
                 return component == SwizzleComponent.Red
                     ? DepthStencilMode.Stencil

Ryujinx.Graphics.Gpu/Memory/MemoryManager.cs

@@ -264,7 +264,8 @@ namespace Ryujinx.Graphics.Gpu.Memory
         /// <param name="pa">CPU virtual address to map into</param>
         /// <param name="va">GPU virtual address to be mapped</param>
         /// <param name="size">Size in bytes of the mapping</param>
-        public void Map(ulong pa, ulong va, ulong size)
+        /// <param name="kind">Kind of the resource located at the mapping</param>
+        public void Map(ulong pa, ulong va, ulong size, PteKind kind)
         {
             lock (_pageTable)
             {
@@ -272,7 +273,7 @@ namespace Ryujinx.Graphics.Gpu.Memory
 
                 for (ulong offset = 0; offset < size; offset += PageSize)
                 {
-                    SetPte(va + offset, pa + offset);
+                    SetPte(va + offset, PackPte(pa + offset, kind));
                 }
             }
         }
@@ -462,14 +463,37 @@ namespace Ryujinx.Graphics.Gpu.Memory
                 return PteUnmapped;
             }
 
-            ulong baseAddress = GetPte(va);
+            ulong pte = GetPte(va);
 
-            if (baseAddress == PteUnmapped)
+            if (pte == PteUnmapped)
             {
                 return PteUnmapped;
             }
 
-            return baseAddress + (va & PageMask);
+            return UnpackPaFromPte(pte) + (va & PageMask);
+        }
+
+        /// <summary>
+        /// Gets the kind of a given memory page.
+        /// This might indicate the type of resource that can be allocated on the page, and also texture tiling.
+        /// </summary>
+        /// <param name="va">GPU virtual address</param>
+        /// <returns>Kind of the memory page</returns>
+        public PteKind GetKind(ulong va)
+        {
+            if (!ValidateAddress(va))
+            {
+                return PteKind.Invalid;
+            }
+
+            ulong pte = GetPte(va);
+
+            if (pte == PteUnmapped)
+            {
+                return PteKind.Invalid;
+            }
+
+            return UnpackKindFromPte(pte);
         }
 
         /// <summary>
@@ -512,5 +536,36 @@ namespace Ryujinx.Graphics.Gpu.Memory
 
             _pageTable[l0][l1] = pte;
         }
+
+        /// <summary>
+        /// Creates a page table entry from a physical address and kind.
+        /// </summary>
+        /// <param name="pa">Physical address</param>
+        /// <param name="kind">Kind</param>
+        /// <returns>Page table entry</returns>
+        private static ulong PackPte(ulong pa, PteKind kind)
+        {
+            return pa | ((ulong)kind << 56);
+        }
+
+        /// <summary>
+        /// Unpacks kind from a page table entry.
+        /// </summary>
+        /// <param name="pte">Page table entry</param>
+        /// <returns>Kind</returns>
+        private static PteKind UnpackKindFromPte(ulong pte)
+        {
+            return (PteKind)(pte >> 56);
+        }
+
+        /// <summary>
+        /// Unpacks physical address from a page table entry.
+        /// </summary>
+        /// <param name="pte">Page table entry</param>
+        /// <returns>Physical address</returns>
+        private static ulong UnpackPaFromPte(ulong pte)
+        {
+            return pte & 0xffffffffffffffUL;
+        }
     }
 }

Ryujinx.Graphics.Gpu/Memory/PteKind.cs

@@ -0,0 +1,268 @@
+namespace Ryujinx.Graphics.Gpu.Memory
+{
+    /// <summary>
+    /// Kind of the resource at the given memory mapping.
+    /// </summary>
+    public enum PteKind : byte
+    {
+        Invalid = 0xff,
+        Pitch = 0x00,
+        Z16 = 0x01,
+        Z162C = 0x02,
+        Z16MS22C = 0x03,
+        Z16MS42C = 0x04,
+        Z16MS82C = 0x05,
+        Z16MS162C = 0x06,
+        Z162Z = 0x07,
+        Z16MS22Z = 0x08,
+        Z16MS42Z = 0x09,
+        Z16MS82Z = 0x0a,
+        Z16MS162Z = 0x0b,
+        Z162CZ = 0x36,
+        Z16MS22CZ = 0x37,
+        Z16MS42CZ = 0x38,
+        Z16MS82CZ = 0x39,
+        Z16MS162CZ = 0x5f,
+        Z164CZ = 0x0c,
+        Z16MS24CZ = 0x0d,
+        Z16MS44CZ = 0x0e,
+        Z16MS84CZ = 0x0f,
+        Z16MS164CZ = 0x10,
+        S8Z24 = 0x11,
+        S8Z241Z = 0x12,
+        S8Z24MS21Z = 0x13,
+        S8Z24MS41Z = 0x14,
+        S8Z24MS81Z = 0x15,
+        S8Z24MS161Z = 0x16,
+        S8Z242CZ = 0x17,
+        S8Z24MS22CZ = 0x18,
+        S8Z24MS42CZ = 0x19,
+        S8Z24MS82CZ = 0x1a,
+        S8Z24MS162CZ = 0x1b,
+        S8Z242CS = 0x1c,
+        S8Z24MS22CS = 0x1d,
+        S8Z24MS42CS = 0x1e,
+        S8Z24MS82CS = 0x1f,
+        S8Z24MS162CS = 0x20,
+        S8Z244CSZV = 0x21,
+        S8Z24MS24CSZV = 0x22,
+        S8Z24MS44CSZV = 0x23,
+        S8Z24MS84CSZV = 0x24,
+        S8Z24MS164CSZV = 0x25,
+        V8Z24MS4VC12 = 0x26,
+        V8Z24MS4VC4 = 0x27,
+        V8Z24MS8VC8 = 0x28,
+        V8Z24MS8VC24 = 0x29,
+        V8Z24MS4VC121ZV = 0x2e,
+        V8Z24MS4VC41ZV = 0x2f,
+        V8Z24MS8VC81ZV = 0x30,
+        V8Z24MS8VC241ZV = 0x31,
+        V8Z24MS4VC122CS = 0x32,
+        V8Z24MS4VC42CS = 0x33,
+        V8Z24MS8VC82CS = 0x34,
+        V8Z24MS8VC242CS = 0x35,
+        V8Z24MS4VC122CZV = 0x3a,
+        V8Z24MS4VC42CZV = 0x3b,
+        V8Z24MS8VC82CZV = 0x3c,
+        V8Z24MS8VC242CZV = 0x3d,
+        V8Z24MS4VC122ZV = 0x3e,
+        V8Z24MS4VC42ZV = 0x3f,
+        V8Z24MS8VC82ZV = 0x40,
+        V8Z24MS8VC242ZV = 0x41,
+        V8Z24MS4VC124CSZV = 0x42,
+        V8Z24MS4VC44CSZV = 0x43,
+        V8Z24MS8VC84CSZV = 0x44,
+        V8Z24MS8VC244CSZV = 0x45,
+        Z24S8 = 0x46,
+        Z24S81Z = 0x47,
+        Z24S8MS21Z = 0x48,
+        Z24S8MS41Z = 0x49,
+        Z24S8MS81Z = 0x4a,
+        Z24S8MS161Z = 0x4b,
+        Z24S82CS = 0x4c,
+        Z24S8MS22CS = 0x4d,
+        Z24S8MS42CS = 0x4e,
+        Z24S8MS82CS = 0x4f,
+        Z24S8MS162CS = 0x50,
+        Z24S82CZ = 0x51,
+        Z24S8MS22CZ = 0x52,
+        Z24S8MS42CZ = 0x53,
+        Z24S8MS82CZ = 0x54,
+        Z24S8MS162CZ = 0x55,
+        Z24S84CSZV = 0x56,
+        Z24S8MS24CSZV = 0x57,
+        Z24S8MS44CSZV = 0x58,
+        Z24S8MS84CSZV = 0x59,
+        Z24S8MS164CSZV = 0x5a,
+        Z24V8MS4VC12 = 0x5b,
+        Z24V8MS4VC4 = 0x5c,
+        Z24V8MS8VC8 = 0x5d,
+        Z24V8MS8VC24 = 0x5e,
+        YUVB8C12Y = 0x60,
+        YUVB8C22Y = 0x61,
+        YUVB10C12Y = 0x62,
+        YUVB10C22Y = 0x6b,
+        YUVB12C12Y = 0x6c,
+        YUVB12C22Y = 0x6d,
+        Z24V8MS4VC121ZV = 0x63,
+        Z24V8MS4VC41ZV = 0x64,
+        Z24V8MS8VC81ZV = 0x65,
+        Z24V8MS8VC241ZV = 0x66,
+        Z24V8MS4VC122CS = 0x67,
+        Z24V8MS4VC42CS = 0x68,
+        Z24V8MS8VC82CS = 0x69,
+        Z24V8MS8VC242CS = 0x6a,
+        Z24V8MS4VC122CZV = 0x6f,
+        Z24V8MS4VC42CZV = 0x70,
+        Z24V8MS8VC82CZV = 0x71,
+        Z24V8MS8VC242CZV = 0x72,
+        Z24V8MS4VC122ZV = 0x73,
+        Z24V8MS4VC42ZV = 0x74,
+        Z24V8MS8VC82ZV = 0x75,
+        Z24V8MS8VC242ZV = 0x76,
+        Z24V8MS4VC124CSZV = 0x77,
+        Z24V8MS4VC44CSZV = 0x78,
+        Z24V8MS8VC84CSZV = 0x79,
+        Z24V8MS8VC244CSZV = 0x7a,
+        ZF32 = 0x7b,
+        ZF321Z = 0x7c,
+        ZF32MS21Z = 0x7d,
+        ZF32MS41Z = 0x7e,
+        ZF32MS81Z = 0x7f,
+        ZF32MS161Z = 0x80,
+        ZF322CS = 0x81,
+        ZF32MS22CS = 0x82,
+        ZF32MS42CS = 0x83,
+        ZF32MS82CS = 0x84,
+        ZF32MS162CS = 0x85,
+        ZF322CZ = 0x86,
+        ZF32MS22CZ = 0x87,
+        ZF32MS42CZ = 0x88,
+        ZF32MS82CZ = 0x89,
+        ZF32MS162CZ = 0x8a,
+        X8Z24X16V8S8MS4VC12 = 0x8b,
+        X8Z24X16V8S8MS4VC4 = 0x8c,
+        X8Z24X16V8S8MS8VC8 = 0x8d,
+        X8Z24X16V8S8MS8VC24 = 0x8e,
+        X8Z24X16V8S8MS4VC121CS = 0x8f,
+        X8Z24X16V8S8MS4VC41CS = 0x90,
+        X8Z24X16V8S8MS8VC81CS = 0x91,
+        X8Z24X16V8S8MS8VC241CS = 0x92,
+        X8Z24X16V8S8MS4VC121ZV = 0x97,
+        X8Z24X16V8S8MS4VC41ZV = 0x98,
+        X8Z24X16V8S8MS8VC81ZV = 0x99,
+        X8Z24X16V8S8MS8VC241ZV = 0x9a,
+        X8Z24X16V8S8MS4VC121CZV = 0x9b,
+        X8Z24X16V8S8MS4VC41CZV = 0x9c,
+        X8Z24X16V8S8MS8VC81CZV = 0x9d,
+        X8Z24X16V8S8MS8VC241CZV = 0x9e,
+        X8Z24X16V8S8MS4VC122CS = 0x9f,
+        X8Z24X16V8S8MS4VC42CS = 0xa0,
+        X8Z24X16V8S8MS8VC82CS = 0xa1,
+        X8Z24X16V8S8MS8VC242CS = 0xa2,
+        X8Z24X16V8S8MS4VC122CSZV = 0xa3,
+        X8Z24X16V8S8MS4VC42CSZV = 0xa4,
+        X8Z24X16V8S8MS8VC82CSZV = 0xa5,
+        X8Z24X16V8S8MS8VC242CSZV = 0xa6,
+        ZF32X16V8S8MS4VC12 = 0xa7,
+        ZF32X16V8S8MS4VC4 = 0xa8,
+        ZF32X16V8S8MS8VC8 = 0xa9,
+        ZF32X16V8S8MS8VC24 = 0xaa,
+        ZF32X16V8S8MS4VC121CS = 0xab,
+        ZF32X16V8S8MS4VC41CS = 0xac,
+        ZF32X16V8S8MS8VC81CS = 0xad,
+        ZF32X16V8S8MS8VC241CS = 0xae,
+        ZF32X16V8S8MS4VC121ZV = 0xb3,
+        ZF32X16V8S8MS4VC41ZV = 0xb4,
+        ZF32X16V8S8MS8VC81ZV = 0xb5,
+        ZF32X16V8S8MS8VC241ZV = 0xb6,
+        ZF32X16V8S8MS4VC121CZV = 0xb7,
+        ZF32X16V8S8MS4VC41CZV = 0xb8,
+        ZF32X16V8S8MS8VC81CZV = 0xb9,
+        ZF32X16V8S8MS8VC241CZV = 0xba,
+        ZF32X16V8S8MS4VC122CS = 0xbb,
+        ZF32X16V8S8MS4VC42CS = 0xbc,
+        ZF32X16V8S8MS8VC82CS = 0xbd,
+        ZF32X16V8S8MS8VC242CS = 0xbe,
+        ZF32X16V8S8MS4VC122CSZV = 0xbf,
+        ZF32X16V8S8MS4VC42CSZV = 0xc0,
+        ZF32X16V8S8MS8VC82CSZV = 0xc1,
+        ZF32X16V8S8MS8VC242CSZV = 0xc2,
+        ZF32X24S8 = 0xc3,
+        ZF32X24S81CS = 0xc4,
+        ZF32X24S8MS21CS = 0xc5,
+        ZF32X24S8MS41CS = 0xc6,
+        ZF32X24S8MS81CS = 0xc7,
+        ZF32X24S8MS161CS = 0xc8,
+        ZF32X24S82CSZV = 0xce,
+        ZF32X24S8MS22CSZV = 0xcf,
+        ZF32X24S8MS42CSZV = 0xd0,
+        ZF32X24S8MS82CSZV = 0xd1,
+        ZF32X24S8MS162CSZV = 0xd2,
+        ZF32X24S82CS = 0xd3,
+        ZF32X24S8MS22CS = 0xd4,
+        ZF32X24S8MS42CS = 0xd5,
+        ZF32X24S8MS82CS = 0xd6,
+        ZF32X24S8MS162CS = 0xd7,
+        S8 = 0x2a,
+        S82S = 0x2b,
+        Generic16Bx2 = 0xfe,
+        C322C = 0xd8,
+        C322CBR = 0xd9,
+        C322CBA = 0xda,
+        C322CRA = 0xdb,
+        C322BRA = 0xdc,
+        C32MS22C = 0xdd,
+        C32MS22CBR = 0xde,
+        C32MS24CBRA = 0xcc,
+        C32MS42C = 0xdf,
+        C32MS42CBR = 0xe0,
+        C32MS42CBA = 0xe1,
+        C32MS42CRA = 0xe2,
+        C32MS42BRA = 0xe3,
+        C32MS44CBRA = 0x2c,
+        C32MS8MS162C = 0xe4,
+        C32MS8MS162CRA = 0xe5,
+        C642C = 0xe6,
+        C642CBR = 0xe7,
+        C642CBA = 0xe8,
+        C642CRA = 0xe9,
+        C642BRA = 0xea,
+        C64MS22C = 0xeb,
+        C64MS22CBR = 0xec,
+        C64MS24CBRA = 0xcd,
+        C64MS42C = 0xed,
+        C64MS42CBR = 0xee,
+        C64MS42CBA = 0xef,
+        C64MS42CRA = 0xf0,
+        C64MS42BRA = 0xf1,
+        C64MS44CBRA = 0x2d,
+        C64MS8MS162C = 0xf2,
+        C64MS8MS162CRA = 0xf3,
+        C1282C = 0xf4,
+        C1282CR = 0xf5,
+        C128MS22C = 0xf6,
+        C128MS22CR = 0xf7,
+        C128MS42C = 0xf8,
+        C128MS42CR = 0xf9,
+        C128MS8MS162C = 0xfa,
+        C128MS8MS162CR = 0xfb,
+        X8C24 = 0xfc,
+        PitchNoSwizzle = 0xfd,
+        SmSkedMessage = 0xca,
+        SmHostMessage = 0xcb
+    }
+
+    static class PteKindExtensions
+    {
+        /// <summary>
+        /// Checks if the kind is pitch.
+        /// </summary>
+        /// <param name="kind">Kind to check</param>
+        /// <returns>True if pitch, false otherwise</returns>
+        public static bool IsPitch(this PteKind kind)
+        {
+            return kind == PteKind.Pitch || kind == PteKind.PitchNoSwizzle;
+        }
+    }
+}

Ryujinx.Graphics.Gpu/Memory/ResourceName.cs

@@ -1,13 +0,0 @@
-namespace Ryujinx.Graphics.Gpu.Memory
-{
-    /// <summary>
-    /// Name of a GPU resource.
-    /// </summary>
-    public enum ResourceName
-    {
-        Buffer,
-        Texture,
-        TexturePool,
-        SamplerPool
-    }
-}

Ryujinx.Graphics.Nvdec/Image/SurfaceReader.cs

@@ -16,7 +16,7 @@ namespace Ryujinx.Graphics.Nvdec.Image
             int width = surface.Width;
             int height = surface.Height;
             int stride = surface.Stride;
- 
+
             ReadOnlySpan<byte> luma = gmm.DeviceGetSpan(lumaOffset, GetBlockLinearSize(width, height, 1));
 
             ReadLuma(surface.YPlane.AsSpan(), luma, stride, width, height);

Ryujinx.Graphics.OpenGL/FormatTable.cs

@@ -6,15 +6,15 @@ namespace Ryujinx.Graphics.OpenGL
 {
     struct FormatTable
     {
-        private static FormatInfo[] Table;
-        private static SizedInternalFormat[] TableImage;
+        private static FormatInfo[] _table;
+        private static SizedInternalFormat[] _tableImage;
 
         static FormatTable()
         {
             int tableSize = Enum.GetNames<Format>().Length;
 
-            Table = new FormatInfo[tableSize];
-            TableImage = new SizedInternalFormat[tableSize];
+            _table = new FormatInfo[tableSize];
+            _tableImage = new SizedInternalFormat[tableSize];
 
             Add(Format.R8Unorm,             new FormatInfo(1, true,  false, All.R8,                PixelFormat.Red,            PixelType.UnsignedByte));
             Add(Format.R8Snorm,             new FormatInfo(1, true,  false, All.R8Snorm,           PixelFormat.Red,            PixelType.Byte));
@@ -66,7 +66,7 @@ namespace Ryujinx.Graphics.OpenGL
             Add(Format.R32G32B32A32Sint,    new FormatInfo(4, false, false, All.Rgba32i,           PixelFormat.RgbaInteger,    PixelType.Int));
             Add(Format.S8Uint,              new FormatInfo(1, false, false, All.StencilIndex8,     PixelFormat.StencilIndex,   PixelType.UnsignedByte));
             Add(Format.D16Unorm,            new FormatInfo(1, false, false, All.DepthComponent16,  PixelFormat.DepthComponent, PixelType.UnsignedShort));
-            Add(Format.D24X8Unorm,          new FormatInfo(1, false, false, All.DepthComponent24,  PixelFormat.DepthComponent, PixelType.UnsignedInt));
+            Add(Format.S8UintD24Unorm,      new FormatInfo(1, false, false, All.Depth24Stencil8,   PixelFormat.DepthStencil,   PixelType.UnsignedInt248));
             Add(Format.D32Float,            new FormatInfo(1, false, false, All.DepthComponent32f, PixelFormat.DepthComponent, PixelType.Float));
             Add(Format.D24UnormS8Uint,      new FormatInfo(1, false, false, All.Depth24Stencil8,   PixelFormat.DepthStencil,   PixelType.UnsignedInt248));
             Add(Format.D32FloatS8Uint,      new FormatInfo(1, false, false, All.Depth32fStencil8,  PixelFormat.DepthStencil,   PixelType.Float32UnsignedInt248Rev));
@@ -218,22 +218,22 @@ namespace Ryujinx.Graphics.OpenGL
 
         private static void Add(Format format, FormatInfo info)
         {
-            Table[(int)format] = info;
+            _table[(int)format] = info;
         }
 
         private static void Add(Format format, SizedInternalFormat sif)
         {
-            TableImage[(int)format] = sif;
+            _tableImage[(int)format] = sif;
         }
 
         public static FormatInfo GetFormatInfo(Format format)
         {
-            return Table[(int)format];
+            return _table[(int)format];
         }
 
         public static SizedInternalFormat GetImageFormat(Format format)
         {
-            return TableImage[(int)format];
+            return _tableImage[(int)format];
         }
     }
 }

Ryujinx.Graphics.OpenGL/Framebuffer.cs

@@ -127,14 +127,13 @@ namespace Ryujinx.Graphics.OpenGL
         private static bool IsPackedDepthStencilFormat(Format format)
         {
             return format == Format.D24UnormS8Uint ||
-                   format == Format.D32FloatS8Uint;
+                   format == Format.D32FloatS8Uint ||
+                   format == Format.S8UintD24Unorm;
         }
 
         private static bool IsDepthOnlyFormat(Format format)
         {
-            return format == Format.D16Unorm ||
-                   format == Format.D24X8Unorm ||
-                   format == Format.D32Float;
+            return format == Format.D16Unorm || format == Format.D32Float;
         }
 
         public void Dispose()

Ryujinx.Graphics.OpenGL/Image/FormatConverter.cs

@@ -0,0 +1,149 @@
+using System;
+using System.Numerics;
+using System.Runtime.InteropServices;
+using System.Runtime.Intrinsics;
+using System.Runtime.Intrinsics.X86;
+
+namespace Ryujinx.Graphics.OpenGL.Image
+{
+    static class FormatConverter
+    {
+        public unsafe static byte[] ConvertS8D24ToD24S8(ReadOnlySpan<byte> data)
+        {
+            byte[] output = new byte[data.Length];
+
+            int start = 0;
+
+            if (Avx2.IsSupported)
+            {
+                var mask = Vector256.Create(
+                    (byte)3, (byte)0, (byte)1, (byte)2,
+                    (byte)7, (byte)4, (byte)5, (byte)6,
+                    (byte)11, (byte)8, (byte)9, (byte)10,
+                    (byte)15, (byte)12, (byte)13, (byte)14,
+                    (byte)19, (byte)16, (byte)17, (byte)18,
+                    (byte)23, (byte)20, (byte)21, (byte)22,
+                    (byte)27, (byte)24, (byte)25, (byte)26,
+                    (byte)31, (byte)28, (byte)29, (byte)30);
+
+                int sizeAligned = data.Length & ~31;
+
+                fixed (byte* pInput = data, pOutput = output)
+                {
+                    for (uint i = 0; i < sizeAligned; i += 32)
+                    {
+                        var dataVec = Avx.LoadVector256(pInput + i);
+
+                        dataVec = Avx2.Shuffle(dataVec, mask);
+
+                        Avx.Store(pOutput + i, dataVec);
+                    }
+                }
+
+                start = sizeAligned;
+            }
+            else if (Ssse3.IsSupported)
+            {
+                var mask = Vector128.Create(
+                    (byte)3, (byte)0, (byte)1, (byte)2,
+                    (byte)7, (byte)4, (byte)5, (byte)6,
+                    (byte)11, (byte)8, (byte)9, (byte)10,
+                    (byte)15, (byte)12, (byte)13, (byte)14);
+
+                int sizeAligned = data.Length & ~15;
+
+                fixed (byte* pInput = data, pOutput = output)
+                {
+                    for (uint i = 0; i < sizeAligned; i += 16)
+                    {
+                        var dataVec = Sse2.LoadVector128(pInput + i);
+
+                        dataVec = Ssse3.Shuffle(dataVec, mask);
+
+                        Sse2.Store(pOutput + i, dataVec);
+                    }
+                }
+
+                start = sizeAligned;
+            }
+
+            var outSpan = MemoryMarshal.Cast<byte, uint>(output);
+            var dataSpan = MemoryMarshal.Cast<byte, uint>(data);
+            for (int i = start / sizeof(uint); i < dataSpan.Length; i++)
+            {
+                outSpan[i] = BitOperations.RotateLeft(dataSpan[i], 8);
+            }
+
+            return output;
+        }
+
+        public unsafe static byte[] ConvertD24S8ToS8D24(ReadOnlySpan<byte> data)
+        {
+            byte[] output = new byte[data.Length];
+
+            int start = 0;
+
+            if (Avx2.IsSupported)
+            {
+                var mask = Vector256.Create(
+                    (byte)1, (byte)2, (byte)3, (byte)0,
+                    (byte)5, (byte)6, (byte)7, (byte)4,
+                    (byte)9, (byte)10, (byte)11, (byte)8,
+                    (byte)13, (byte)14, (byte)15, (byte)12,
+                    (byte)17, (byte)18, (byte)19, (byte)16,
+                    (byte)21, (byte)22, (byte)23, (byte)20,
+                    (byte)25, (byte)26, (byte)27, (byte)24,
+                    (byte)29, (byte)30, (byte)31, (byte)28);
+
+                int sizeAligned = data.Length & ~31;
+
+                fixed (byte* pInput = data, pOutput = output)
+                {
+                    for (uint i = 0; i < sizeAligned; i += 32)
+                    {
+                        var dataVec = Avx.LoadVector256(pInput + i);
+
+                        dataVec = Avx2.Shuffle(dataVec, mask);
+
+                        Avx.Store(pOutput + i, dataVec);
+                    }
+                }
+
+                start = sizeAligned;
+            }
+            else if (Ssse3.IsSupported)
+            {
+                var mask = Vector128.Create(
+                    (byte)1, (byte)2, (byte)3, (byte)0,
+                    (byte)5, (byte)6, (byte)7, (byte)4,
+                    (byte)9, (byte)10, (byte)11, (byte)8,
+                    (byte)13, (byte)14, (byte)15, (byte)12);
+
+                int sizeAligned = data.Length & ~15;
+
+                fixed (byte* pInput = data, pOutput = output)
+                {
+                    for (uint i = 0; i < sizeAligned; i += 16)
+                    {
+                        var dataVec = Sse2.LoadVector128(pInput + i);
+
+                        dataVec = Ssse3.Shuffle(dataVec, mask);
+
+                        Sse2.Store(pOutput + i, dataVec);
+                    }
+                }
+
+                start = sizeAligned;
+            }
+
+            var outSpan = MemoryMarshal.Cast<byte, uint>(output);
+            var dataSpan = MemoryMarshal.Cast<byte, uint>(data);
+            for (int i = start / sizeof(uint); i < dataSpan.Length; i++)
+            {
+                outSpan[i] = BitOperations.RotateRight(dataSpan[i], 8);
+            }
+
+            return output;
+        }
+    }
+}

Ryujinx.Graphics.OpenGL/Image/TextureCopy.cs

@@ -291,7 +291,7 @@ namespace Ryujinx.Graphics.OpenGL.Image
 
         private static ClearBufferMask GetMask(Format format)
         {
-            if (format == Format.D24UnormS8Uint || format == Format.D32FloatS8Uint)
+            if (format == Format.D24UnormS8Uint || format == Format.D32FloatS8Uint || format == Format.S8UintD24Unorm)
             {
                 return ClearBufferMask.DepthBufferBit | ClearBufferMask.StencilBufferBit;
             }
@@ -311,9 +311,7 @@ namespace Ryujinx.Graphics.OpenGL.Image
 
         private static bool IsDepthOnly(Format format)
         {
-            return format == Format.D16Unorm   ||
-                   format == Format.D24X8Unorm ||
-                   format == Format.D32Float;
+            return format == Format.D16Unorm || format == Format.D32Float;
         }
 
         public TextureView BgraSwap(TextureView from)

Ryujinx.Graphics.OpenGL/Image/TextureView.cs

@@ -140,9 +140,11 @@ namespace Ryujinx.Graphics.OpenGL.Image
                 size += Info.GetMipSize(level);
             }
 
+            ReadOnlySpan<byte> data;
+
             if (HwCapabilities.UsePersistentBufferForFlush)
             {
-                return _renderer.PersistentBuffers.Default.GetTextureData(this, size);
+                data = _renderer.PersistentBuffers.Default.GetTextureData(this, size);
             }
             else
             {
@@ -150,8 +152,15 @@ namespace Ryujinx.Graphics.OpenGL.Image
 
                 WriteTo(target);
 
-                return new ReadOnlySpan<byte>(target.ToPointer(), size);
+                data = new ReadOnlySpan<byte>(target.ToPointer(), size);
             }
+
+            if (Format == Format.S8UintD24Unorm)
+            {
+                data = FormatConverter.ConvertD24S8ToS8D24(data);
+            }
+
+            return data;
         }
 
         public unsafe ReadOnlySpan<byte> GetData(int layer, int level)
@@ -285,6 +294,11 @@ namespace Ryujinx.Graphics.OpenGL.Image
 
         public void SetData(ReadOnlySpan<byte> data)
         {
+            if (Format == Format.S8UintD24Unorm)
+            {
+                data = FormatConverter.ConvertS8D24ToD24S8(data);
+            }
+
             unsafe
             {
                 fixed (byte* ptr = data)
@@ -296,6 +310,11 @@ namespace Ryujinx.Graphics.OpenGL.Image
 
         public void SetData(ReadOnlySpan<byte> data, int layer, int level)
         {
+            if (Format == Format.S8UintD24Unorm)
+            {
+                data = FormatConverter.ConvertS8D24ToD24S8(data);
+            }
+
             unsafe
             {
                 fixed (byte* ptr = data)

Ryujinx.Graphics.OpenGL/Pipeline.cs

@@ -1478,6 +1478,11 @@ namespace Ryujinx.Graphics.OpenGL
             _currentComponentMasks |= componentMaskAtIndex;
         }
 
+        public void RestoreClipControl()
+        {
+            GL.ClipControl(_clipOrigin, _clipDepthMode);
+        }
+
         public void RestoreScissor0Enable()
         {
             if ((_scissorEnables & 1u) != 0)
@@ -1494,6 +1499,11 @@ namespace Ryujinx.Graphics.OpenGL
             }
         }
 
+        public void RestoreViewport0()
+        {
+            GL.ViewportArray(0, 1, _viewportArray);
+        }
+
         public bool TryHostConditionalRendering(ICounterEvent value, ulong compare, bool isEqual)
         {
             if (value is CounterQueueEvent)

Ryujinx.Graphics.OpenGL/Window.cs

@@ -27,11 +27,12 @@ namespace Ryujinx.Graphics.OpenGL
         {
             GL.Disable(EnableCap.FramebufferSrgb);
 
-            CopyTextureToFrameBufferRGB(0, GetCopyFramebufferHandleLazy(), (TextureView)texture, crop);
+            CopyTextureToFrameBufferRGB(0, GetCopyFramebufferHandleLazy(), (TextureView)texture, crop, swapBuffersCallback);
 
             GL.Enable(EnableCap.FramebufferSrgb);
 
-            swapBuffersCallback();
+            // Restore unpack alignment to 4, as performance overlays such as RTSS may change this to load their resources.
+            GL.PixelStore(PixelStoreParameter.UnpackAlignment, 4);
         }
 
         public void SetSize(int width, int height)
@@ -40,7 +41,7 @@ namespace Ryujinx.Graphics.OpenGL
             _height = height;
         }
 
-        private void CopyTextureToFrameBufferRGB(int drawFramebuffer, int readFramebuffer, TextureView view, ImageCrop crop)
+        private void CopyTextureToFrameBufferRGB(int drawFramebuffer, int readFramebuffer, TextureView view, ImageCrop crop, Action swapBuffersCallback)
         {
             (int oldDrawFramebufferHandle, int oldReadFramebufferHandle) = ((Pipeline)_renderer.Pipeline).GetBoundFramebuffers();
 
@@ -139,11 +140,20 @@ namespace Ryujinx.Graphics.OpenGL
                 ((Pipeline)_renderer.Pipeline).RestoreComponentMask(i);
             }
 
+            // Set clip control, viewport and the framebuffer to the output to placate overlays and OBS capture.
+            GL.ClipControl(ClipOrigin.LowerLeft, ClipDepthMode.NegativeOneToOne);
+            GL.Viewport(0, 0, _width, _height);
+            GL.BindFramebuffer(FramebufferTarget.Framebuffer, drawFramebuffer);
+
+            swapBuffersCallback();
+
             GL.BindFramebuffer(FramebufferTarget.ReadFramebuffer, oldReadFramebufferHandle);
             GL.BindFramebuffer(FramebufferTarget.DrawFramebuffer, oldDrawFramebufferHandle);
 
+            ((Pipeline)_renderer.Pipeline).RestoreClipControl();
             ((Pipeline)_renderer.Pipeline).RestoreScissor0Enable();
             ((Pipeline)_renderer.Pipeline).RestoreRasterizerDiscard();
+            ((Pipeline)_renderer.Pipeline).RestoreViewport0();
 
             if (viewConverted != view)
             {

Ryujinx.HLE/HOS/Services/Nifm/StaticService/IGeneralService.cs

@@ -14,6 +14,9 @@ namespace Ryujinx.HLE.HOS.Services.Nifm.StaticService
     {
         private GeneralServiceDetail _generalServiceDetail;
 
+        private IPInterfaceProperties _targetPropertiesCache = null;
+        private UnicastIPAddressInformation _targetAddressInfoCache = null;
+
         public IGeneralService()
         {
             _generalServiceDetail = new GeneralServiceDetail
@@ -22,6 +25,8 @@ namespace Ryujinx.HLE.HOS.Services.Nifm.StaticService
                 IsAnyInternetRequestAccepted = true // NOTE: Why not accept any internet request?
             };
 
+            NetworkChange.NetworkAddressChanged += new NetworkAddressChangedEventHandler(LocalInterfaceCacheHandler);
+
             GeneralServiceManager.Add(_generalServiceDetail);
         }
 
@@ -165,6 +170,11 @@ namespace Ryujinx.HLE.HOS.Services.Nifm.StaticService
                 return (null, null);
             }
 
+            if (_targetPropertiesCache != null && _targetAddressInfoCache != null)
+            {
+                return (_targetPropertiesCache, _targetAddressInfoCache);
+            }
+
             IPInterfaceProperties       targetProperties  = null;
             UnicastIPAddressInformation targetAddressInfo = null;
 
@@ -194,13 +204,26 @@ namespace Ryujinx.HLE.HOS.Services.Nifm.StaticService
                 }
             }
 
+            _targetPropertiesCache  = targetProperties;
+            _targetAddressInfoCache = targetAddressInfo;
+
             return (targetProperties, targetAddressInfo);
         }
 
+        private void LocalInterfaceCacheHandler(object sender, EventArgs e)
+        {
+            Logger.Info?.Print(LogClass.ServiceNifm, $"NetworkAddress changed, invalidating cached data.");
+
+            _targetPropertiesCache  = null;
+            _targetAddressInfoCache = null;
+        }
+
         protected override void Dispose(bool isDisposing)
         {
             if (isDisposing)
             {
+                NetworkChange.NetworkAddressChanged -= LocalInterfaceCacheHandler;
+
                 GeneralServiceManager.Remove(_generalServiceDetail.ClientId);
             }
         }

Ryujinx.HLE/HOS/Services/Nim/Ntc/IStaticService.cs

@@ -1,8 +1,24 @@
-namespace Ryujinx.HLE.HOS.Services.Nim.Ntc
+using Ryujinx.Common.Logging;
+using Ryujinx.HLE.HOS.Services.Nim.Ntc.StaticService;
+
+namespace Ryujinx.HLE.HOS.Services.Nim.Ntc
 {
     [Service("ntc")]
     class IStaticService : IpcService
     {
         public IStaticService(ServiceCtx context) { }
+
+        [CommandHipc(0)]
+        // OpenEnsureNetworkClockAvailabilityService(u64) -> object<nn::ntc::detail::service::IEnsureNetworkClockAvailabilityService>
+        public ResultCode CreateAsyncInterface(ServiceCtx context)
+        {
+            ulong unknown = context.RequestData.ReadUInt64();
+
+            MakeObject(context, new IEnsureNetworkClockAvailabilityService(context));
+
+            Logger.Stub?.PrintStub(LogClass.ServiceNtc, new { unknown });
+
+            return ResultCode.Success;
+        }
     }
 }

Ryujinx.HLE/HOS/Services/Nim/Ntc/StaticService/IEnsureNetworkClockAvailabilityService.cs

@@ -0,0 +1,77 @@
+using Ryujinx.Common.Logging;
+using Ryujinx.HLE.HOS.Ipc;
+using Ryujinx.HLE.HOS.Kernel.Common;
+using Ryujinx.HLE.HOS.Kernel.Threading;
+using System;
+
+namespace Ryujinx.HLE.HOS.Services.Nim.Ntc.StaticService
+{
+    class IEnsureNetworkClockAvailabilityService : IpcService 
+    {
+        private KEvent     _finishNotificationEvent;
+        private ResultCode _taskResultCode;
+
+        public IEnsureNetworkClockAvailabilityService(ServiceCtx context)
+        {
+            _finishNotificationEvent = new KEvent(context.Device.System.KernelContext);
+            _taskResultCode          = ResultCode.Success;
+
+            // NOTE: The service starts a thread that polls Nintendo NTP server and syncs the time with it.
+            //       Additionnally it gets and uses some settings too:
+            //       autonomic_correction_interval_seconds, autonomic_correction_failed_retry_interval_seconds,
+            //       autonomic_correction_immediate_try_count_max, autonomic_correction_immediate_try_interval_milliseconds
+        }
+
+        [CommandHipc(0)]
+        // StartTask()
+        public ResultCode StartTask(ServiceCtx context)
+        {
+            if (!context.Device.Configuration.EnableInternetAccess)
+            {
+                return (ResultCode)Time.ResultCode.NetworkTimeNotAvailable;
+            }
+
+            // NOTE: Since we don't support the Nintendo NTP server, we can signal the event now to confirm the update task is done.
+            _finishNotificationEvent.ReadableEvent.Signal();
+
+            Logger.Stub?.PrintStub(LogClass.ServiceNtc);
+
+            return ResultCode.Success;
+        }
+
+        [CommandHipc(1)]
+        // GetFinishNotificationEvent() -> handle<copy>
+        public ResultCode GetFinishNotificationEvent(ServiceCtx context)
+        {
+            if (context.Process.HandleTable.GenerateHandle(_finishNotificationEvent.ReadableEvent, out int finishNotificationEventHandle) != KernelResult.Success)
+            {
+                throw new InvalidOperationException("Out of handles!");
+            }
+
+            context.Response.HandleDesc = IpcHandleDesc.MakeCopy(finishNotificationEventHandle);
+
+            return ResultCode.Success;
+        }
+
+        [CommandHipc(2)]
+        // GetResult()
+        public ResultCode GetResult(ServiceCtx context)
+        {
+            return _taskResultCode;
+        }
+
+        [CommandHipc(3)]
+        // Cancel()
+        public ResultCode Cancel(ServiceCtx context)
+        {
+            // NOTE: The update task should be canceled here.
+            _finishNotificationEvent.ReadableEvent.Signal();
+
+            _taskResultCode = (ResultCode)Time.ResultCode.NetworkTimeTaskCanceled;
+
+            Logger.Stub?.PrintStub(LogClass.ServiceNtc);
+
+            return ResultCode.Success;
+        }
+    }
+}

Ryujinx.HLE/HOS/Services/Nv/NvDrvServices/NvHostAsGpu/NvHostAsGpuDeviceFile.cs

@@ -232,7 +232,7 @@ namespace Ryujinx.HLE.HOS.Services.Nv.NvDrvServices.NvHostAsGpu
                         ulong virtualAddress = arguments.Offset + arguments.BufferOffset;
 
                         physicalAddress += arguments.BufferOffset;
-                        _asContext.Gmm.Map(physicalAddress, virtualAddress, arguments.MappingSize);
+                        _asContext.Gmm.Map(physicalAddress, virtualAddress, arguments.MappingSize, (PteKind)arguments.Kind);
 
                         return NvInternalResult.Success;
                     }
@@ -282,7 +282,7 @@ namespace Ryujinx.HLE.HOS.Services.Nv.NvDrvServices.NvHostAsGpu
                 {
                     if (_asContext.ValidateFixedBuffer(arguments.Offset, size, pageSize))
                     {
-                        _asContext.Gmm.Map(physicalAddress, arguments.Offset, size);
+                        _asContext.Gmm.Map(physicalAddress, arguments.Offset, size, (PteKind)arguments.Kind);
                     }
                     else
                     {
@@ -301,7 +301,7 @@ namespace Ryujinx.HLE.HOS.Services.Nv.NvDrvServices.NvHostAsGpu
                         _memoryAllocator.AllocateRange(va, size, freeAddressStartPosition);
                     }
 
-                    _asContext.Gmm.Map(physicalAddress, va, size);
+                    _asContext.Gmm.Map(physicalAddress, va, size, (PteKind)arguments.Kind);
                     arguments.Offset = va;
                 }
 
@@ -366,26 +366,30 @@ namespace Ryujinx.HLE.HOS.Services.Nv.NvDrvServices.NvHostAsGpu
 
             for (int index = 0; index < arguments.Length; index++)
             {
-                ulong mapOffs = (ulong)arguments[index].MapOffset << 16;
-                ulong gpuVa = (ulong)arguments[index].GpuOffset << 16;
-                ulong size = (ulong)arguments[index].Pages << 16;
+                ref RemapArguments argument = ref arguments[index];
+                ulong gpuVa = (ulong)argument.GpuOffset << 16;
+                ulong size = (ulong)argument.Pages << 16;
+                int nvmapHandle = argument.NvMapHandle;
 
-                if (arguments[index].NvMapHandle == 0)
+                if (nvmapHandle == 0)
                 {
                     gmm.Unmap(gpuVa, size);
                 }
                 else
                 {
-                    NvMapHandle map = NvMapDeviceFile.GetMapFromHandle(Owner, arguments[index].NvMapHandle);
+                    ulong mapOffs = (ulong)argument.MapOffset << 16;
+                    PteKind kind = (PteKind)argument.Kind;
+
+                    NvMapHandle map = NvMapDeviceFile.GetMapFromHandle(Owner, nvmapHandle);
 
                     if (map == null)
                     {
-                        Logger.Warning?.Print(LogClass.ServiceNv, $"Invalid NvMap handle 0x{arguments[index].NvMapHandle:x8}!");
+                        Logger.Warning?.Print(LogClass.ServiceNv, $"Invalid NvMap handle 0x{nvmapHandle:x8}!");
 
                         return NvInternalResult.InvalidInput;
                     }
 
-                    gmm.Map(mapOffs + map.Address, gpuVa, size);
+                    gmm.Map(mapOffs + map.Address, gpuVa, size, kind);
                 }
             }
 

Ryujinx.HLE/HOS/Services/Nv/NvDrvServices/NvHostChannel/NvHostChannelDeviceFile.cs

@@ -251,7 +251,7 @@ namespace Ryujinx.HLE.HOS.Services.Nv.NvDrvServices.NvHostChannel
                         if (va != NvMemoryAllocator.PteUnmapped && va <= uint.MaxValue && (va + (uint)map.Size) <= uint.MaxValue)
                         {
                             _host1xContext.MemoryAllocator.AllocateRange(va, (uint)map.Size, freeAddressStartPosition);
-                            _host1xContext.Smmu.Map(map.Address, va, (uint)map.Size);
+                            _host1xContext.Smmu.Map(map.Address, va, (uint)map.Size, PteKind.Pitch); // FIXME: This should not use the GMMU.
                             map.DmaMapAddress = va;
                         }
                         else

Ryujinx.HLE/HOS/Services/Olsc/IOlscServiceForApplication.cs

@@ -45,7 +45,7 @@ namespace Ryujinx.HLE.HOS.Services.Olsc
                 return ResultCode.NullArgument;
             }
 
-            if (_saveDataBackupSettingDatabase[userId])
+            if (_saveDataBackupSettingDatabase.TryGetValue(userId, out bool enabled) && enabled)
             {
                 context.ResponseData.Write((byte)1); // TODO: Determine value.
             }

Ryujinx.HLE/HOS/Services/Time/ResultCode.cs

@@ -7,16 +7,18 @@
 
         Success = 0,
 
-        TimeServiceNotInitialized = (0   << ErrorCodeShift) | ModuleId,
-        PermissionDenied          = (1   << ErrorCodeShift) | ModuleId,
-        TimeMismatch              = (102 << ErrorCodeShift) | ModuleId,
-        UninitializedClock        = (103 << ErrorCodeShift) | ModuleId,
-        TimeNotFound              = (200 << ErrorCodeShift) | ModuleId,
-        Overflow                  = (201 << ErrorCodeShift) | ModuleId,
-        LocationNameTooLong       = (801 << ErrorCodeShift) | ModuleId,
-        OutOfRange                = (902 << ErrorCodeShift) | ModuleId,
-        TimeZoneConversionFailed  = (903 << ErrorCodeShift) | ModuleId,
-        TimeZoneNotFound          = (989 << ErrorCodeShift) | ModuleId,
-        NotImplemented            = (990 << ErrorCodeShift) | ModuleId,
+        TimeServiceNotInitialized = (0    << ErrorCodeShift) | ModuleId,
+        PermissionDenied          = (1    << ErrorCodeShift) | ModuleId,
+        TimeMismatch              = (102  << ErrorCodeShift) | ModuleId,
+        UninitializedClock        = (103  << ErrorCodeShift) | ModuleId,
+        TimeNotFound              = (200  << ErrorCodeShift) | ModuleId,
+        Overflow                  = (201  << ErrorCodeShift) | ModuleId,
+        LocationNameTooLong       = (801  << ErrorCodeShift) | ModuleId,
+        OutOfRange                = (902  << ErrorCodeShift) | ModuleId,
+        TimeZoneConversionFailed  = (903  << ErrorCodeShift) | ModuleId,
+        TimeZoneNotFound          = (989  << ErrorCodeShift) | ModuleId,
+        NotImplemented            = (990  << ErrorCodeShift) | ModuleId,
+        NetworkTimeNotAvailable   = (1000 << ErrorCodeShift) | ModuleId,
+        NetworkTimeTaskCanceled   = (1003 << ErrorCodeShift) | ModuleId,
     }
 }