Support NVDEC H264 interlaced video decoding and VIC deinterlacing (#3225)

* Support NVDEC H264 interlaced video decoding and VIC deinterlacing

* Remove unused code

ARMeilleure/Decoders/Decoder.cs

@@ -121,7 +121,7 @@ namespace ARMeilleure.Decoders
                             currBlock.Branch = GetBlock((ulong)op.Immediate);
                         }
 
-                        if (!IsUnconditionalBranch(lastOp) || isCall)
+                        if (isCall || !(IsUnconditionalBranch(lastOp) || IsTrap(lastOp)))
                         {
                             currBlock.Next = GetBlock(currBlock.EndAddress);
                         }
@@ -263,6 +263,11 @@ namespace ARMeilleure.Decoders
             // so we must consider such operations as a branch in potential aswell.
             if (opCode is IOpCode32Alu opAlu && opAlu.Rd == RegisterAlias.Aarch32Pc)
             {
+                if (opCode is OpCodeT32)
+                {
+                    return opCode.Instruction.Name != InstName.Tst && opCode.Instruction.Name != InstName.Teq &&
+                           opCode.Instruction.Name != InstName.Cmp && opCode.Instruction.Name != InstName.Cmn;
+                }
                 return true;
             }
 
@@ -324,9 +329,13 @@ namespace ARMeilleure.Decoders
         }
 
         private static bool IsException(OpCode opCode)
+        {
+            return IsTrap(opCode) || opCode.Instruction.Name == InstName.Svc;
+        }
+
+        private static bool IsTrap(OpCode opCode)
         {
             return opCode.Instruction.Name == InstName.Brk ||
-                   opCode.Instruction.Name == InstName.Svc ||
                    opCode.Instruction.Name == InstName.Trap ||
                    opCode.Instruction.Name == InstName.Und;
         }

ARMeilleure/Decoders/OpCode32.cs

@@ -13,11 +13,25 @@ namespace ARMeilleure.Decoders
             Cond = (Condition)((uint)opCode >> 28);
         }
 
+        public bool IsThumb()
+        {
+            return this is OpCodeT16 || this is OpCodeT32;
+        }
+
         public uint GetPc()
         {
             // Due to backwards compatibility and legacy behavior of ARMv4 CPUs pipeline,
             // the PC actually points 2 instructions ahead.
-            return (uint)Address + (uint)OpCodeSizeInBytes * 2;
+            if (IsThumb())
+            {
+                // PC is ahead by 4 in thumb mode whether or not the current instruction
+                // is 16 or 32 bit.
+                return (uint)Address + 4u;
+            }
+            else
+            {
+                return (uint)Address + 8u;
+            }
         }
     }
 }

ARMeilleure/Decoders/OpCodeT16BImm11.cs

@@ -1,6 +1,6 @@
 namespace ARMeilleure.Decoders
 {
-    class OpCodeT16BImm11 : OpCode32, IOpCode32BImm
+    class OpCodeT16BImm11 : OpCodeT16, IOpCode32BImm
     {
         public long Immediate { get; }
 

ARMeilleure/Decoders/OpCodeT16BImm8.cs

@@ -1,6 +1,6 @@
 namespace ARMeilleure.Decoders
 {
-    class OpCodeT16BImm8 : OpCode32, IOpCode32BImm
+    class OpCodeT16BImm8 : OpCodeT16, IOpCode32BImm
     {
         public long Immediate { get; }
 

ARMeilleure/Decoders/OpCodeT32.cs

@@ -0,0 +1,14 @@
+namespace ARMeilleure.Decoders
+{
+    class OpCodeT32 : OpCode32
+    {
+        public new static OpCode Create(InstDescriptor inst, ulong address, int opCode) => new OpCodeT32(inst, address, opCode);
+
+        public OpCodeT32(InstDescriptor inst, ulong address, int opCode) : base(inst, address, opCode)
+        {
+            Cond = Condition.Al;
+
+            OpCodeSizeInBytes = 4;
+        }
+    }
+}

ARMeilleure/Decoders/OpCodeT32Alu.cs

@@ -0,0 +1,20 @@
+namespace ARMeilleure.Decoders
+{
+    class OpCodeT32Alu : OpCodeT32, IOpCode32Alu
+    {
+        public int Rd { get; }
+        public int Rn { get; }
+
+        public bool? SetFlags { get; }
+
+        public new static OpCode Create(InstDescriptor inst, ulong address, int opCode) => new OpCodeT32Alu(inst, address, opCode);
+
+        public OpCodeT32Alu(InstDescriptor inst, ulong address, int opCode) : base(inst, address, opCode)
+        {
+            Rd = (opCode >> 8) & 0xf;
+            Rn = (opCode >> 16) & 0xf;
+
+            SetFlags = ((opCode >> 20) & 1) != 0;
+        }
+    }
+}

ARMeilleure/Decoders/OpCodeT32AluImm.cs

@@ -0,0 +1,38 @@
+using ARMeilleure.Common;
+using System.Runtime.Intrinsics;
+
+namespace ARMeilleure.Decoders
+{
+    class OpCodeT32AluImm : OpCodeT32Alu, IOpCode32AluImm
+    {
+        public int Immediate { get; }
+
+        public bool IsRotated { get; }
+
+        public new static OpCode Create(InstDescriptor inst, ulong address, int opCode) => new OpCodeT32AluImm(inst, address, opCode);
+
+        private static readonly Vector128<int> _factor = Vector128.Create(1, 0x00010001, 0x01000100, 0x01010101);
+
+        public OpCodeT32AluImm(InstDescriptor inst, ulong address, int opCode) : base(inst, address, opCode)
+        {
+            int imm8 = (opCode >> 0) & 0xff;
+            int imm3 = (opCode >> 12) & 7;
+            int imm1 = (opCode >> 26) & 1;
+
+            int imm12 = imm8 | (imm3 << 8) | (imm1 << 11);
+
+            if ((imm12 >> 10) == 0)
+            {
+                Immediate = imm8 * _factor.GetElement((imm12 >> 8) & 3);
+                IsRotated = false;
+            }
+            else
+            {
+                int shift = imm12 >> 7;
+
+                Immediate = BitUtils.RotateRight(0x80 | (imm12 & 0x7f), shift, 32);
+                IsRotated = shift != 0;
+            }
+        }
+    }
+}

ARMeilleure/Decoders/OpCodeT32AluRsImm.cs

@@ -0,0 +1,20 @@
+namespace ARMeilleure.Decoders
+{
+    class OpCodeT32AluRsImm : OpCodeT32Alu, IOpCode32AluRsImm
+    {
+        public int Rm        { get; }
+        public int Immediate { get; }
+
+        public ShiftType ShiftType { get; }
+
+        public new static OpCode Create(InstDescriptor inst, ulong address, int opCode) => new OpCodeT32AluRsImm(inst, address, opCode);
+
+        public OpCodeT32AluRsImm(InstDescriptor inst, ulong address, int opCode) : base(inst, address, opCode)
+        {
+            Rm        = (opCode >> 0) & 0xf;
+            Immediate = ((opCode >> 6) & 3) | ((opCode >> 10) & 0x1c);
+
+            ShiftType = (ShiftType)((opCode >> 4) & 3);
+        }
+    }
+}

ARMeilleure/Decoders/OpCodeT32BImm20.cs

@@ -0,0 +1,29 @@
+using ARMeilleure.Instructions;
+
+namespace ARMeilleure.Decoders
+{
+    class OpCodeT32BImm20 : OpCodeT32, IOpCode32BImm
+    {
+        public long Immediate { get; }
+
+        public new static OpCode Create(InstDescriptor inst, ulong address, int opCode) => new OpCodeT32BImm20(inst, address, opCode);
+
+        public OpCodeT32BImm20(InstDescriptor inst, ulong address, int opCode) : base(inst, address, opCode)
+        {
+            uint pc = GetPc();
+
+            int imm11 = (opCode >> 0) & 0x7ff;
+            int j2 = (opCode >> 11) & 1;
+            int j1 = (opCode >> 13) & 1;
+            int imm6 = (opCode >> 16) & 0x3f;
+            int s = (opCode >> 26) & 1;
+
+            int imm32 = imm11 | (imm6 << 11) | (j1 << 17) | (j2 << 18) | (s << 19);
+            imm32 = (imm32 << 13) >> 12;
+
+            Immediate = pc + imm32;
+
+            Cond = (Condition)((opCode >> 22) & 0xf);
+        }
+    }
+}

ARMeilleure/Decoders/OpCodeT32BImm24.cs

@@ -0,0 +1,35 @@
+using ARMeilleure.Instructions;
+
+namespace ARMeilleure.Decoders
+{
+    class OpCodeT32BImm24 : OpCodeT32, IOpCode32BImm
+    {
+        public long Immediate { get; }
+
+        public new static OpCode Create(InstDescriptor inst, ulong address, int opCode) => new OpCodeT32BImm24(inst, address, opCode);
+
+        public OpCodeT32BImm24(InstDescriptor inst, ulong address, int opCode) : base(inst, address, opCode)
+        {
+            uint pc = GetPc();
+
+            if (inst.Name == InstName.Blx)
+            {
+                pc &= ~3u;
+            }
+
+            int imm11 = (opCode >> 0) & 0x7ff;
+            int j2 = (opCode >> 11) & 1;
+            int j1 = (opCode >> 13) & 1;
+            int imm10 = (opCode >> 16) & 0x3ff;
+            int s = (opCode >> 26) & 1;
+
+            int i1 = j1 ^ s ^ 1;
+            int i2 = j2 ^ s ^ 1;
+
+            int imm32 = imm11 | (imm10 << 11) | (i2 << 21) | (i1 << 22) | (s << 23);
+            imm32 = (imm32 << 9) >> 8;
+
+            Immediate = pc + imm32;
+        }
+    }
+}

ARMeilleure/Decoders/OpCodeTable.cs

@@ -1,6 +1,7 @@
 using ARMeilleure.Instructions;
 using System;
 using System.Collections.Generic;
+using System.Numerics;
 
 namespace ARMeilleure.Decoders
 {
@@ -972,8 +973,7 @@ namespace ARMeilleure.Decoders
             SetA32("111100111x11<<10xxxx00011xx0xxxx", InstName.Vzip,     InstEmit32.Vzip,     OpCode32SimdCmpZ.Create);
 #endregion
 
-#region "OpCode Table (AArch32, T16/T32)"
-            // T16
+#region "OpCode Table (AArch32, T16)"
             SetT16("000<<xxxxxxxxxxx", InstName.Mov,    InstEmit32.Mov,     OpCodeT16ShiftImm.Create);
             SetT16("0001100xxxxxxxxx", InstName.Add,    InstEmit32.Add,     OpCodeT16AddSubReg.Create);
             SetT16("0001101xxxxxxxxx", InstName.Sub,    InstEmit32.Sub,     OpCodeT16AddSubReg.Create);
@@ -1045,6 +1045,46 @@ namespace ARMeilleure.Decoders
             SetT16("11100xxxxxxxxxxx", InstName.B,      InstEmit32.B,       OpCodeT16BImm11.Create);
 #endregion
 
+#region "OpCode Table (AArch32, T32)"
+            // Base
+            SetT32("11101011010xxxxx0xxxxxxxxxxxxxxx", InstName.Adc,      InstEmit32.Adc,      OpCodeT32AluRsImm.Create);
+            SetT32("11110x01010xxxxx0xxxxxxxxxxxxxxx", InstName.Adc,      InstEmit32.Adc,      OpCodeT32AluImm.Create);
+            SetT32("11101011000<xxxx0xxx<<<<xxxxxxxx", InstName.Add,      InstEmit32.Add,      OpCodeT32AluRsImm.Create);
+            SetT32("11110x01000<xxxx0xxx<<<<xxxxxxxx", InstName.Add,      InstEmit32.Add,      OpCodeT32AluImm.Create);
+            SetT32("11101010000<xxxx0xxx<<<<xxxxxxxx", InstName.And,      InstEmit32.And,      OpCodeT32AluRsImm.Create);
+            SetT32("11110x00000<xxxx0xxx<<<<xxxxxxxx", InstName.And,      InstEmit32.And,      OpCodeT32AluImm.Create);
+            SetT32("11110x<<<xxxxxxx10x0xxxxxxxxxxxx", InstName.B,        InstEmit32.B,        OpCodeT32BImm20.Create);
+            SetT32("11110xxxxxxxxxxx10x1xxxxxxxxxxxx", InstName.B,        InstEmit32.B,        OpCodeT32BImm24.Create);
+            SetT32("11101010001xxxxx0xxxxxxxxxxxxxxx", InstName.Bic,      InstEmit32.Bic,      OpCodeT32AluRsImm.Create);
+            SetT32("11110x00001xxxxx0xxxxxxxxxxxxxxx", InstName.Bic,      InstEmit32.Bic,      OpCodeT32AluImm.Create);
+            SetT32("11110xxxxxxxxxxx11x1xxxxxxxxxxxx", InstName.Bl,       InstEmit32.Bl,       OpCodeT32BImm24.Create);
+            SetT32("11110xxxxxxxxxxx11x0xxxxxxxxxxx0", InstName.Blx,      InstEmit32.Blx,      OpCodeT32BImm24.Create);
+            SetT32("111010110001xxxx0xxx1111xxxxxxxx", InstName.Cmn,      InstEmit32.Cmn,      OpCodeT32AluRsImm.Create);
+            SetT32("11110x010001xxxx0xxx1111xxxxxxxx", InstName.Cmn,      InstEmit32.Cmn,      OpCodeT32AluImm.Create);
+            SetT32("111010111011xxxx0xxx1111xxxxxxxx", InstName.Cmp,      InstEmit32.Cmp,      OpCodeT32AluRsImm.Create);
+            SetT32("11110x011011xxxx0xxx1111xxxxxxxx", InstName.Cmp,      InstEmit32.Cmp,      OpCodeT32AluImm.Create);
+            SetT32("11101010100<xxxx0xxx<<<<xxxxxxxx", InstName.Eor,      InstEmit32.Eor,      OpCodeT32AluRsImm.Create);
+            SetT32("11110x00100<xxxx0xxx<<<<xxxxxxxx", InstName.Eor,      InstEmit32.Eor,      OpCodeT32AluImm.Create);
+            SetT32("11101010010x11110xxxxxxxxxxxxxxx", InstName.Mov,      InstEmit32.Mov,      OpCodeT32AluRsImm.Create);
+            SetT32("11110x00010x11110xxxxxxxxxxxxxxx", InstName.Mov,      InstEmit32.Mov,      OpCodeT32AluImm.Create);
+            SetT32("11101010011x11110xxxxxxxxxxxxxxx", InstName.Mvn,      InstEmit32.Mvn,      OpCodeT32AluRsImm.Create);
+            SetT32("11110x00011x11110xxxxxxxxxxxxxxx", InstName.Mvn,      InstEmit32.Mvn,      OpCodeT32AluImm.Create);
+            SetT32("11101010011x<<<<0xxxxxxxxxxxxxxx", InstName.Orn,      InstEmit32.Orn,      OpCodeT32AluRsImm.Create);
+            SetT32("11110x00011x<<<<0xxxxxxxxxxxxxxx", InstName.Orn,      InstEmit32.Orn,      OpCodeT32AluImm.Create);
+            SetT32("11101010010x<<<<0xxxxxxxxxxxxxxx", InstName.Orr,      InstEmit32.Orr,      OpCodeT32AluRsImm.Create);
+            SetT32("11110x00010x<<<<0xxxxxxxxxxxxxxx", InstName.Orr,      InstEmit32.Orr,      OpCodeT32AluImm.Create);
+            SetT32("11101011110xxxxx0xxxxxxxxxxxxxxx", InstName.Rsb,      InstEmit32.Rsb,      OpCodeT32AluRsImm.Create);
+            SetT32("11110x01110xxxxx0xxxxxxxxxxxxxxx", InstName.Rsb,      InstEmit32.Rsb,      OpCodeT32AluImm.Create);
+            SetT32("11101011011xxxxx0xxxxxxxxxxxxxxx", InstName.Sbc,      InstEmit32.Sbc,      OpCodeT32AluRsImm.Create);
+            SetT32("11110x01011xxxxx0xxxxxxxxxxxxxxx", InstName.Sbc,      InstEmit32.Sbc,      OpCodeT32AluImm.Create);
+            SetT32("11101011101<xxxx0xxx<<<<xxxxxxxx", InstName.Sub,      InstEmit32.Sub,      OpCodeT32AluRsImm.Create);
+            SetT32("11110x01101<xxxx0xxx<<<<xxxxxxxx", InstName.Sub,      InstEmit32.Sub,      OpCodeT32AluImm.Create);
+            SetT32("111010101001xxxx0xxx1111xxxxxxxx", InstName.Teq,      InstEmit32.Teq,      OpCodeT32AluRsImm.Create);
+            SetT32("11110x001001xxxx0xxx1111xxxxxxxx", InstName.Teq,      InstEmit32.Teq,      OpCodeT32AluImm.Create);
+            SetT32("111010100001xxxx0xxx1111xxxxxxxx", InstName.Tst,      InstEmit32.Tst,      OpCodeT32AluRsImm.Create);
+            SetT32("11110x000001xxxx0xxx1111xxxxxxxx", InstName.Tst,      InstEmit32.Tst,      OpCodeT32AluImm.Create);
+#endregion
+
             FillFastLookupTable(InstA32FastLookup, AllInstA32, ToFastLookupIndexA);
             FillFastLookupTable(InstT32FastLookup, AllInstT32, ToFastLookupIndexT);
             FillFastLookupTable(InstA64FastLookup, AllInstA64, ToFastLookupIndexA);
@@ -1092,8 +1132,11 @@ namespace ARMeilleure.Decoders
 
         private static void SetT32(string encoding, InstName name, InstEmitter emitter, MakeOp makeOp)
         {
-            encoding = encoding.Substring(16) + encoding.Substring(0, 16);
-            Set(encoding, AllInstT32, new InstDescriptor(name, emitter), makeOp);
+            string reversedEncoding = encoding.Substring(16) + encoding.Substring(0, 16);
+            MakeOp reversedMakeOp =
+                (InstDescriptor inst, ulong address, int opCode)
+                    => makeOp(inst, address, (int)BitOperations.RotateRight((uint)opCode, 16));
+            Set(reversedEncoding, AllInstT32, new InstDescriptor(name, emitter), reversedMakeOp);
         }
 
         private static void SetA64(string encoding, InstName name, InstEmitter emitter, MakeOp makeOp)

ARMeilleure/Instructions/InstEmitAlu32.cs

@@ -244,6 +244,23 @@ namespace ARMeilleure.Instructions
             EmitAluStore(context, res);
         }
 
+        public static void Orn(ArmEmitterContext context)
+        {
+            IOpCode32Alu op = (IOpCode32Alu)context.CurrOp;
+
+            Operand n = GetAluN(context);
+            Operand m = GetAluM(context);
+
+            Operand res = context.BitwiseOr(n, context.BitwiseNot(m));
+
+            if (ShouldSetFlags(context))
+            {
+                EmitNZFlagsCheck(context, res);
+            }
+
+            EmitAluStore(context, res);
+        }
+
         public static void Pkh(ArmEmitterContext context)
         {
             OpCode32AluRsImm op = (OpCode32AluRsImm)context.CurrOp;

ARMeilleure/Instructions/InstEmitAluHelper.cs

@@ -128,7 +128,7 @@ namespace ARMeilleure.Instructions
         {
             Debug.Assert(value.Type == OperandType.I32);
 
-            if (IsThumb(context.CurrOp))
+            if (((OpCode32)context.CurrOp).IsThumb())
             {
                 bool isReturn = IsA32Return(context);
                 if (!isReturn)
@@ -197,7 +197,7 @@ namespace ARMeilleure.Instructions
                 // ARM32.
                 case IOpCode32AluImm op:
                 {
-                    if (ShouldSetFlags(context) && op.IsRotated)
+                    if (ShouldSetFlags(context) && op.IsRotated && setCarry)
                     {
                         SetFlag(context, PState.CFlag, Const((uint)op.Immediate >> 31));
                     }

ARMeilleure/Instructions/InstEmitException.cs

@@ -9,18 +9,25 @@ namespace ARMeilleure.Instructions
     {
         public static void Brk(ArmEmitterContext context)
         {
-            EmitExceptionCall(context, nameof(NativeInterface.Break));
+            OpCodeException op = (OpCodeException)context.CurrOp;
+
+            string name = nameof(NativeInterface.Break);
+
+            context.StoreToContext();
+
+            context.Call(typeof(NativeInterface).GetMethod(name), Const(op.Address), Const(op.Id));
+
+            context.LoadFromContext();
+
+            context.Return(Const(op.Address));
         }
 
         public static void Svc(ArmEmitterContext context)
-        {
-            EmitExceptionCall(context, nameof(NativeInterface.SupervisorCall));
-        }
-
-        private static void EmitExceptionCall(ArmEmitterContext context, string name)
         {
             OpCodeException op = (OpCodeException)context.CurrOp;
 
+            string name = nameof(NativeInterface.SupervisorCall);
+
             context.StoreToContext();
 
             context.Call(typeof(NativeInterface).GetMethod(name), Const(op.Address), Const(op.Id));
@@ -41,6 +48,8 @@ namespace ARMeilleure.Instructions
             context.Call(typeof(NativeInterface).GetMethod(name), Const(op.Address), Const(op.RawOpCode));
 
             context.LoadFromContext();
+
+            context.Return(Const(op.Address));
         }
     }
 }

ARMeilleure/Instructions/InstEmitException32.cs

@@ -9,19 +9,11 @@ namespace ARMeilleure.Instructions
     static partial class InstEmit32
     {
         public static void Svc(ArmEmitterContext context)
-        {
-            EmitExceptionCall(context, nameof(NativeInterface.SupervisorCall));
-        }
-
-        public static void Trap(ArmEmitterContext context)
-        {
-            EmitExceptionCall(context, nameof(NativeInterface.Break));
-        }
-
-        private static void EmitExceptionCall(ArmEmitterContext context, string name)
         {
             IOpCode32Exception op = (IOpCode32Exception)context.CurrOp;
 
+            string name = nameof(NativeInterface.SupervisorCall);
+
             context.StoreToContext();
 
             context.Call(typeof(NativeInterface).GetMethod(name), Const(((IOpCode)op).Address), Const(op.Id));
@@ -30,5 +22,20 @@ namespace ARMeilleure.Instructions
 
             Translator.EmitSynchronization(context);
         }
+
+        public static void Trap(ArmEmitterContext context)
+        {
+            IOpCode32Exception op = (IOpCode32Exception)context.CurrOp;
+
+            string name = nameof(NativeInterface.Break);
+
+            context.StoreToContext();
+
+            context.Call(typeof(NativeInterface).GetMethod(name), Const(((IOpCode)op).Address), Const(op.Id));
+
+            context.LoadFromContext();
+
+            context.Return(Const(context.CurrOp.Address));
+        }
     }
 }

ARMeilleure/Instructions/InstEmitFlow32.cs

@@ -34,7 +34,7 @@ namespace ARMeilleure.Instructions
 
             uint pc = op.GetPc();
 
-            bool isThumb = IsThumb(context.CurrOp);
+            bool isThumb = ((OpCode32)context.CurrOp).IsThumb();
 
             uint currentPc = isThumb
                 ? pc | 1
@@ -61,7 +61,7 @@ namespace ARMeilleure.Instructions
             Operand addr = context.Copy(GetIntA32(context, op.Rm));
             Operand bitOne = context.BitwiseAnd(addr, Const(1));
 
-            bool isThumb = IsThumb(context.CurrOp);
+            bool isThumb = ((OpCode32)context.CurrOp).IsThumb();
 
             uint currentPc = isThumb
                 ? (pc - 2) | 1

ARMeilleure/Instructions/InstEmitHelper.cs

@@ -10,11 +10,6 @@ namespace ARMeilleure.Instructions
 {
     static class InstEmitHelper
     {
-        public static bool IsThumb(OpCode op)
-        {
-            return op is OpCodeT16;
-        }
-
         public static Operand GetExtendedM(ArmEmitterContext context, int rm, IntType type)
         {
             Operand value = GetIntOrZR(context, rm);

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/State/ExecutionContext.cs

@@ -43,6 +43,12 @@ namespace ARMeilleure.State
         public long TpidrEl0 { get; set; }
         public long Tpidr    { get; set; }
 
+        public uint Pstate
+        {
+            get => _nativeContext.GetPstate();
+            set => _nativeContext.SetPstate(value);
+        }
+
         public FPCR Fpcr { get; set; }
         public FPSR Fpsr { get; set; }
         public FPCR StandardFpcrValue => (Fpcr & (FPCR.Ahp)) | FPCR.Dn | FPCR.Fz;

ARMeilleure/State/NativeContext.cs

@@ -95,6 +95,25 @@ namespace ARMeilleure.State
             GetStorage().Flags[(int)flag] = value ? 1u : 0u;
         }
 
+        public unsafe uint GetPstate()
+        {
+            uint value = 0;
+            for (int flag = 0; flag < RegisterConsts.FlagsCount; flag++)
+            {
+                value |= GetStorage().Flags[flag] != 0 ? 1u << flag : 0u;
+            }
+            return value;
+        }
+
+        public unsafe void SetPstate(uint value)
+        {
+            for (int flag = 0; flag < RegisterConsts.FlagsCount; flag++)
+            {
+                uint bit = 1u << flag;
+                GetStorage().Flags[flag] = (value & bit) == bit ? 1u : 0u;
+            }
+        }
+
         public unsafe bool GetFPStateFlag(FPState flag)
         {
             if ((uint)flag >= RegisterConsts.FpFlagsCount)

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 = 3138; //! 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.Audio/Renderer/Server/Performance/PerformanceManagerGeneric.cs

@@ -149,11 +149,21 @@ namespace Ryujinx.Audio.Renderer.Server.Performance
 
                 Span<byte> targetSpan = performanceOutput.Slice(nextOffset);
 
+                // NOTE: We check for the space for two headers for the final blank header.
+                int requiredSpace = Unsafe.SizeOf<THeader>() + Unsafe.SizeOf<TEntry>() * inputHeader.GetEntryCount()
+                                                             + Unsafe.SizeOf<TEntryDetail>() * inputHeader.GetEntryDetailCount()
+                                                             + Unsafe.SizeOf<THeader>();
+
+                if (targetSpan.Length < requiredSpace)
+                {
+                    break;
+                }
+
                 ref THeader outputHeader = ref MemoryMarshal.Cast<byte, THeader>(targetSpan)[0];
 
                 nextOffset += Unsafe.SizeOf<THeader>();
 
-                Span<TEntry> outputEntries = MemoryMarshal.Cast<byte, TEntry>(targetSpan.Slice(nextOffset));
+                Span<TEntry> outputEntries = MemoryMarshal.Cast<byte, TEntry>(performanceOutput.Slice(nextOffset));
 
                 int totalProcessingTime = 0;
 
@@ -175,7 +185,7 @@ namespace Ryujinx.Audio.Renderer.Server.Performance
                     }
                 }
 
-                Span<TEntryDetail> outputEntriesDetail = MemoryMarshal.Cast<byte, TEntryDetail>(targetSpan.Slice(nextOffset));
+                Span<TEntryDetail> outputEntriesDetail = MemoryMarshal.Cast<byte, TEntryDetail>(performanceOutput.Slice(nextOffset));
 
                 int effectiveEntryDetailCount = 0;
 

Ryujinx.Common/Configuration/AppDataManager.cs

@@ -34,6 +34,7 @@ namespace Ryujinx.Common.Configuration
         private const string DefaultModsDir = "mods";
 
         public static string CustomModsPath { get; set; }
+        public static string CustomSdModsPath {get; set; }
         public static string CustomNandPath { get; set; } // TODO: Actually implement this into VFS
         public static string CustomSdCardPath { get; set; } // TODO: Actually implement this into VFS
 
@@ -84,6 +85,7 @@ namespace Ryujinx.Common.Configuration
             Directory.CreateDirectory(KeysDirPath = Path.Combine(BaseDirPath, KeysDir));
         }
 
-        public static string GetModsPath() => CustomModsPath ?? Directory.CreateDirectory(Path.Combine(BaseDirPath, DefaultModsDir)).FullName;
+        public static string GetModsPath()   => CustomModsPath ?? Directory.CreateDirectory(Path.Combine(BaseDirPath, DefaultModsDir)).FullName;
+        public static string GetSdModsPath() => CustomSdModsPath ?? Directory.CreateDirectory(Path.Combine(BaseDirPath, DefaultSdcardDir, "atmosphere")).FullName;
     }
 }

Ryujinx.Common/Configuration/Hid/Controller/JoyconConfigControllerStick.cs

@@ -5,6 +5,7 @@
         public Stick Joystick { get; set; }
         public bool InvertStickX { get; set; }
         public bool InvertStickY { get; set; }
+        public bool Rotate90CW { get; set; }
         public Button StickButton { get; set; }
     }
 }

Ryujinx.Common/Logging/LogClass.cs

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

Ryujinx.Common/ReleaseInformations.cs

@@ -1,10 +1,14 @@
-using System.Reflection;
+using Ryujinx.Common.Configuration;
+using System;
+using System.Reflection;
 
 namespace Ryujinx.Common
 {
     // DO NOT EDIT, filled by CI
     public static class ReleaseInformations
     {
+        private const string FlatHubChannelOwner = "flathub";
+
         public static string BuildVersion = "%%RYUJINX_BUILD_VERSION%%";
         public static string BuildGitHash = "%%RYUJINX_BUILD_GIT_HASH%%";
         public static string ReleaseChannelName = "%%RYUJINX_TARGET_RELEASE_CHANNEL_NAME%%";
@@ -19,6 +23,11 @@ namespace Ryujinx.Common
                    !ReleaseChannelRepo.StartsWith("%%");
         }
 
+        public static bool IsFlatHubBuild()
+        {
+            return IsValid() && ReleaseChannelOwner.Equals(FlatHubChannelOwner);
+        }
+
         public static string GetVersion()
         {
             if (IsValid())
@@ -30,5 +39,15 @@ namespace Ryujinx.Common
                 return Assembly.GetEntryAssembly().GetCustomAttribute<AssemblyInformationalVersionAttribute>().InformationalVersion;
             }
         }
+
+        public static string GetBaseApplicationDirectory()
+        {
+            if (IsFlatHubBuild())
+            {
+                return AppDataManager.BaseDirPath;
+            }
+
+            return AppDomain.CurrentDomain.BaseDirectory;
+        }
     }
 }

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/BufferCache.cs

@@ -17,6 +17,8 @@ namespace Ryujinx.Graphics.Gpu.Memory
         private const ulong BufferAlignmentSize = 0x1000;
         private const ulong BufferAlignmentMask = BufferAlignmentSize - 1;
 
+        private const ulong MaxDynamicGrowthSize = 0x100000;
+
         private readonly GpuContext _context;
         private readonly PhysicalMemory _physicalMemory;
 
@@ -166,10 +168,35 @@ namespace Ryujinx.Graphics.Gpu.Memory
                 // Otherwise, we must delete the overlapping buffers and create a bigger buffer
                 // that fits all the data we need. We also need to copy the contents from the
                 // old buffer(s) to the new buffer.
+
                 ulong endAddress = address + size;
 
                 if (_bufferOverlaps[0].Address > address || _bufferOverlaps[0].EndAddress < endAddress)
                 {
+                    // Check if the following conditions are met:
+                    // - We have a single overlap.
+                    // - The overlap starts at or before the requested range. That is, the overlap happens at the end.
+                    // - The size delta between the new, merged buffer and the old one is of at most 2 pages.
+                    // In this case, we attempt to extend the buffer further than the requested range,
+                    // this can potentially avoid future resizes if the application keeps using overlapping
+                    // sequential memory.
+                    // Allowing for 2 pages (rather than just one) is necessary to catch cases where the
+                    // range crosses a page, and after alignment, ends having a size of 2 pages.
+                    if (overlapsCount == 1 &&
+                        address >= _bufferOverlaps[0].Address &&
+                        endAddress - _bufferOverlaps[0].EndAddress <= BufferAlignmentSize * 2)
+                    {
+                        // Try to grow the buffer by 1.5x of its current size.
+                        // This improves performance in the cases where the buffer is resized often by small amounts.
+                        ulong existingSize = _bufferOverlaps[0].Size;
+                        ulong growthSize = (existingSize + Math.Min(existingSize >> 1, MaxDynamicGrowthSize)) & ~BufferAlignmentMask;
+
+                        size = Math.Max(size, growthSize);
+                        endAddress = address + size;
+
+                        overlapsCount = _buffers.FindOverlapsNonOverlapping(address, size, ref _bufferOverlaps);
+                    }
+
                     for (int index = 0; index < overlapsCount; index++)
                     {
                         Buffer buffer = _bufferOverlaps[index];
@@ -183,7 +210,9 @@ namespace Ryujinx.Graphics.Gpu.Memory
                         }
                     }
 
-                    Buffer newBuffer = new Buffer(_context, _physicalMemory, address, endAddress - address, _bufferOverlaps.Take(overlapsCount));
+                    ulong newSize = endAddress - address;
+
+                    Buffer newBuffer = new Buffer(_context, _physicalMemory, address, newSize, _bufferOverlaps.Take(overlapsCount));
 
                     lock (_buffers)
                     {
@@ -202,7 +231,7 @@ namespace Ryujinx.Graphics.Gpu.Memory
                         buffer.DisposeData();
                     }
 
-                    newBuffer.SynchronizeMemory(address, endAddress - address);
+                    newBuffer.SynchronizeMemory(address, newSize);
 
                     // Existing buffers were modified, we need to rebind everything.
                     NotifyBuffersModified?.Invoke();

Ryujinx.Graphics.Gpu/Memory/MemoryManager.cs

@@ -255,6 +255,49 @@ namespace Ryujinx.Graphics.Gpu.Memory
             }
         }
 
+        /// <summary>
+        /// Writes data to GPU mapped memory, stopping at the first unmapped page at the memory region, if any.
+        /// </summary>
+        /// <param name="va">GPU virtual address to write the data into</param>
+        /// <param name="data">The data to be written</param>
+        public void WriteMapped(ulong va, ReadOnlySpan<byte> data)
+        {
+            if (IsContiguous(va, data.Length))
+            {
+                Physical.Write(Translate(va), data);
+            }
+            else
+            {
+                int offset = 0, size;
+
+                if ((va & PageMask) != 0)
+                {
+                    ulong pa = Translate(va);
+
+                    size = Math.Min(data.Length, (int)PageSize - (int)(va & PageMask));
+
+                    if (pa != PteUnmapped && Physical.IsMapped(pa))
+                    {
+                        Physical.Write(pa, data.Slice(0, size));
+                    }
+
+                    offset += size;
+                }
+
+                for (; offset < data.Length; offset += size)
+                {
+                    ulong pa = Translate(va + (ulong)offset);
+
+                    size = Math.Min(data.Length - offset, (int)PageSize);
+
+                    if (pa != PteUnmapped && Physical.IsMapped(pa))
+                    {
+                        Physical.Write(pa, data.Slice(offset, size));
+                    }
+                }
+            }
+        }
+
         /// <summary>
         /// Maps a given range of pages to the specified CPU virtual address.
         /// </summary>
@@ -264,7 +307,8 @@ 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 +316,7 @@ 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 +506,37 @@ 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 +579,36 @@ 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/PhysicalMemory.cs

@@ -340,6 +340,16 @@ namespace Ryujinx.Graphics.Gpu.Memory
             return _cpuMemory.BeginSmartGranularTracking(address, size, granularity);
         }
 
+        /// <summary>
+        /// Checks if the page at a given address is mapped on CPU memory.
+        /// </summary>
+        /// <param name="address">CPU virtual address of the page to check</param>
+        /// <returns>True if mapped, false otherwise</returns>
+        public bool IsMapped(ulong address)
+        {
+            return _cpuMemory.IsMapped(address);
+        }
+
         /// <summary>
         /// Release our reference to the CPU memory manager.
         /// </summary>

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.Gpu/Shader/ShaderCache.cs

@@ -40,7 +40,7 @@ namespace Ryujinx.Graphics.Gpu.Shader
         /// <summary>
         /// Version of the codegen (to be changed when codegen or guest format change).
         /// </summary>
-        private const ulong ShaderCodeGenVersion = 3132;
+        private const ulong ShaderCodeGenVersion = 3054;
 
         // Progress reporting helpers
         private volatile int _shaderCount;

Ryujinx.Graphics.Nvdec.FFmpeg/Surface.cs

@@ -15,11 +15,13 @@ namespace Ryujinx.Graphics.Nvdec.FFmpeg
         public Plane UPlane => new Plane((IntPtr)Frame->data[1], UvStride * UvHeight);
         public Plane VPlane => new Plane((IntPtr)Frame->data[2], UvStride * UvHeight);
 
+        public FrameField Field => Frame->interlaced_frame != 0 ? FrameField.Interlaced : FrameField.Progressive;
+
         public int Width => Frame->width;
         public int Height => Frame->height;
         public int Stride => Frame->linesize[0];
-        public int UvWidth => (Frame->width + 1) >> 1;
-        public int UvHeight => (Frame->height + 1) >> 1;
+        public int UvWidth => (Width + 1) >> 1;
+        public int UvHeight => (Height + 1) >> 1;
         public int UvStride => Frame->linesize[1];
 
         public Surface(int width, int height)

Ryujinx.Graphics.Nvdec.Vp9/Dsp/InvTxfm.cs

@@ -486,8 +486,7 @@ namespace Ryujinx.Graphics.Nvdec.Vp9.Dsp
                 Idct8(tempIn, tempOut);
                 for (j = 0; j < 8; ++j)
                 {
-                    dest[j * stride + i] = ClipPixelAdd(dest[j * stride + i],
-                                                          BitUtils.RoundPowerOfTwo(tempOut[j], 5));
+                    dest[j * stride + i] = ClipPixelAdd(dest[j * stride + i], BitUtils.RoundPowerOfTwo(tempOut[j], 5));
                 }
             }
         }

Ryujinx.Graphics.Nvdec.Vp9/Types/Surface.cs

@@ -15,6 +15,8 @@ namespace Ryujinx.Graphics.Nvdec.Vp9.Types
         public unsafe Plane UPlane => new Plane((IntPtr)UBuffer.ToPointer(), UBuffer.Length);
         public unsafe Plane VPlane => new Plane((IntPtr)VBuffer.ToPointer(), VBuffer.Length);
 
+        public FrameField Field => FrameField.Progressive;
+
         public int Width { get; }
         public int Height { get; }
         public int AlignedWidth { get; }

Ryujinx.Graphics.Nvdec/H264Decoder.cs

@@ -31,7 +31,24 @@ namespace Ryujinx.Graphics.Nvdec
 
             if (decoder.Decode(ref info, outputSurface, bitstream))
             {
-                SurfaceWriter.Write(rm.Gmm, outputSurface, lumaOffset, chromaOffset);
+                if (outputSurface.Field == FrameField.Progressive)
+                {
+                    SurfaceWriter.Write(
+                        rm.Gmm,
+                        outputSurface,
+                        lumaOffset   + pictureInfo.LumaFrameOffset,
+                        chromaOffset + pictureInfo.ChromaFrameOffset);
+                }
+                else
+                {
+                    SurfaceWriter.WriteInterlaced(
+                        rm.Gmm,
+                        outputSurface,
+                        lumaOffset   + pictureInfo.LumaTopFieldOffset,
+                        chromaOffset + pictureInfo.ChromaTopFieldOffset,
+                        lumaOffset   + pictureInfo.LumaBottomFieldOffset,
+                        chromaOffset + pictureInfo.ChromaBottomFieldOffset);
+                }
             }
 
             rm.Cache.Put(outputSurface);

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.Nvdec/Image/SurfaceWriter.cs

@@ -38,6 +38,55 @@ namespace Ryujinx.Graphics.Nvdec.Image
                 surface.UvHeight);
         }
 
+        public static void WriteInterlaced(
+            MemoryManager gmm,
+            ISurface surface,
+            uint lumaTopOffset,
+            uint chromaTopOffset,
+            uint lumaBottomOffset,
+            uint chromaBottomOffset)
+        {
+            int lumaSize = GetBlockLinearSize(surface.Width, surface.Height / 2, 1);
+
+            using var lumaTop = gmm.GetWritableRegion(ExtendOffset(lumaTopOffset), lumaSize);
+            using var lumaBottom = gmm.GetWritableRegion(ExtendOffset(lumaBottomOffset), lumaSize);
+
+            WriteLuma(
+                lumaTop.Memory.Span,
+                surface.YPlane.AsSpan(),
+                surface.Stride * 2,
+                surface.Width,
+                surface.Height / 2);
+
+            WriteLuma(
+                lumaBottom.Memory.Span,
+                surface.YPlane.AsSpan().Slice(surface.Stride),
+                surface.Stride * 2,
+                surface.Width,
+                surface.Height / 2);
+
+            int chromaSize = GetBlockLinearSize(surface.UvWidth, surface.UvHeight / 2, 2);
+
+            using var chromaTop = gmm.GetWritableRegion(ExtendOffset(chromaTopOffset), chromaSize);
+            using var chromaBottom = gmm.GetWritableRegion(ExtendOffset(chromaBottomOffset), chromaSize);
+
+            WriteChroma(
+                chromaTop.Memory.Span,
+                surface.UPlane.AsSpan(),
+                surface.VPlane.AsSpan(),
+                surface.UvStride * 2,
+                surface.UvWidth,
+                surface.UvHeight / 2);
+
+            WriteChroma(
+                chromaBottom.Memory.Span,
+                surface.UPlane.AsSpan().Slice(surface.UvStride),
+                surface.VPlane.AsSpan().Slice(surface.UvStride),
+                surface.UvStride * 2,
+                surface.UvWidth,
+                surface.UvHeight / 2);
+        }
+
         private static void WriteLuma(Span<byte> dst, ReadOnlySpan<byte> src, int srcStride, int width, int height)
         {
             LayoutConverter.ConvertLinearToBlockLinear(dst, width, height, srcStride, 1, 2, src);

Ryujinx.Graphics.Nvdec/Types/H264/PictureInfo.cs

@@ -26,10 +26,10 @@ namespace Ryujinx.Graphics.Nvdec.Types.H264
         public uint Transform8x8ModeFlag;
         public uint LumaPitch;
         public uint ChromaPitch;
-        public uint LumaTopOffset;
-        public uint LumaBottomOffset;
+        public uint LumaTopFieldOffset;
+        public uint LumaBottomFieldOffset;
         public uint LumaFrameOffset;
-        public uint ChromaTopOffset;
+        public uint ChromaTopFieldOffset;
         public uint ChromaBottomFieldOffset;
         public uint ChromaFrameOffset;
         public uint HistBufferSize;

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

@@ -612,7 +612,7 @@ namespace Ryujinx.Graphics.OpenGL
                     _program?.Bind();
                     _unit0Sampler?.Bind(0);
 
-                    GL.ViewportArray(0, 1, _viewportArray);
+                    RestoreViewport0();
 
                     Enable(EnableCap.CullFace, _cullEnable);
                     Enable(EnableCap.StencilTest, _stencilTestEnable);
@@ -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 +1499,14 @@ namespace Ryujinx.Graphics.OpenGL
             }
         }
 
+        public void RestoreViewport0()
+        {
+            if (_viewportArray.Length > 0)
+            {
+                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.Graphics.Shader/Decoders/Decoder.cs

@@ -308,7 +308,8 @@ namespace Ryujinx.Graphics.Shader.Decoders
                     int attr = offset + elemIndex * 4;
                     if (attr >= AttributeConsts.UserAttributeBase && attr < AttributeConsts.UserAttributeEnd)
                     {
-                        int index = (attr - AttributeConsts.UserAttributeBase) / 16;
+                        int userAttr = attr - AttributeConsts.UserAttributeBase;
+                        int index = userAttr / 16;
 
                         if (isStore)
                         {
@@ -316,7 +317,7 @@ namespace Ryujinx.Graphics.Shader.Decoders
                         }
                         else
                         {
-                            config.SetInputUserAttribute(index, perPatch);
+                            config.SetInputUserAttribute(index, (userAttr >> 2) & 3, perPatch);
                         }
                     }
 

Ryujinx.Graphics.Shader/Translation/ShaderConfig.cs

@@ -54,6 +54,11 @@ namespace Ryujinx.Graphics.Shader.Translation
         private int _nextUsedInputAttributes;
         private int _thisUsedInputAttributes;
 
+        public UInt128 NextInputAttributesComponents { get; private set; }
+        public UInt128 ThisInputAttributesComponents { get; private set; }
+        public UInt128 NextInputAttributesPerPatchComponents { get; private set; }
+        public UInt128 ThisInputAttributesPerPatchComponents { get; private set; }
+
         private int _usedConstantBuffers;
         private int _usedStorageBuffers;
         private int _usedStorageBuffersWrite;
@@ -227,11 +232,12 @@ namespace Ryujinx.Graphics.Shader.Translation
             UsedOutputAttributes |= 1 << index;
         }
 
-        public void SetInputUserAttribute(int index, bool perPatch)
+        public void SetInputUserAttribute(int index, int component, bool perPatch)
         {
             if (perPatch)
             {
                 UsedInputAttributesPerPatch |= 1 << index;
+                ThisInputAttributesPerPatchComponents |= UInt128.Pow2(index * 4 + component);
             }
             else
             {
@@ -239,6 +245,7 @@ namespace Ryujinx.Graphics.Shader.Translation
 
                 UsedInputAttributes |= mask;
                 _thisUsedInputAttributes |= mask;
+                ThisInputAttributesComponents |= UInt128.Pow2(index * 4 + component);
             }
         }
 
@@ -256,6 +263,8 @@ namespace Ryujinx.Graphics.Shader.Translation
 
         public void MergeFromtNextStage(ShaderConfig config)
         {
+            NextInputAttributesComponents = config.ThisInputAttributesComponents;
+            NextInputAttributesPerPatchComponents = config.ThisInputAttributesPerPatchComponents;
             NextUsesFixedFuncAttributes = config.UsedFeatures.HasFlag(FeatureFlags.FixedFuncAttr);
             MergeOutputUserAttributes(config.UsedInputAttributes, config.UsedInputAttributesPerPatch);
         }
@@ -319,6 +328,7 @@ namespace Ryujinx.Graphics.Shader.Translation
         public void SetAllInputUserAttributes()
         {
             UsedInputAttributes |= Constants.AllAttributesMask;
+            ThisInputAttributesComponents |= ~UInt128.Zero >> (128 - Constants.MaxAttributes * 4);
         }
 
         public void SetAllOutputUserAttributes()

Ryujinx.Graphics.Shader/Translation/Translator.cs

@@ -214,24 +214,31 @@ namespace Ryujinx.Graphics.Shader.Translation
                 InitializeOutput(context, AttributeConsts.PositionX, perPatch: false);
             }
 
-            int usedAttributes = context.Config.UsedOutputAttributes;
-            while (usedAttributes != 0)
+            UInt128 usedAttributes = context.Config.NextInputAttributesComponents;
+            while (usedAttributes != UInt128.Zero)
             {
-                int index = BitOperations.TrailingZeroCount(usedAttributes);
+                int index = usedAttributes.TrailingZeroCount();
+                int vecIndex = index / 4;
 
-                InitializeOutput(context, AttributeConsts.UserAttributeBase + index * 16, perPatch: false);
+                usedAttributes &= ~UInt128.Pow2(index);
 
-                usedAttributes &= ~(1 << index);
+                // We don't need to initialize passthrough attributes.
+                if ((context.Config.PassthroughAttributes & (1 << vecIndex)) != 0)
+                {
+                    continue;
+                }
+
+                InitializeOutputComponent(context, AttributeConsts.UserAttributeBase + index * 4, perPatch: false);
             }
 
-            int usedAttributesPerPatch = context.Config.UsedOutputAttributesPerPatch;
-            while (usedAttributesPerPatch != 0)
+            UInt128 usedAttributesPerPatch = context.Config.NextInputAttributesPerPatchComponents;
+            while (usedAttributesPerPatch != UInt128.Zero)
             {
-                int index = BitOperations.TrailingZeroCount(usedAttributesPerPatch);
+                int index = usedAttributesPerPatch.TrailingZeroCount();
 
-                InitializeOutput(context, AttributeConsts.UserAttributeBase + index * 16, perPatch: true);
+                InitializeOutputComponent(context, AttributeConsts.UserAttributeBase + index * 4, perPatch: true);
 
-                usedAttributesPerPatch &= ~(1 << index);
+                usedAttributesPerPatch &= ~UInt128.Pow2(index);
             }
 
             if (config.NextUsesFixedFuncAttributes)
@@ -260,6 +267,12 @@ namespace Ryujinx.Graphics.Shader.Translation
             }
         }
 
+        private static void InitializeOutputComponent(EmitterContext context, int attrOffset, bool perPatch)
+        {
+            int c = (attrOffset >> 2) & 3;
+            context.Copy(perPatch ? AttributePerPatch(attrOffset) : Attribute(attrOffset), ConstF(c == 3 ? 1f : 0f));
+        }
+
         private static void EmitOps(EmitterContext context, Block block)
         {
             for (int opIndex = 0; opIndex < block.OpCodes.Count; opIndex++)

Ryujinx.Graphics.Shader/Translation/UInt128.cs

@@ -0,0 +1,112 @@
+using System;
+using System.Numerics;
+
+namespace Ryujinx.Graphics.Shader.Translation
+{
+    struct UInt128 : IEquatable<UInt128>
+    {
+        public static UInt128 Zero => new UInt128() { _v0 = 0, _v1 = 0 };
+
+        private ulong _v0;
+        private ulong _v1;
+
+        public UInt128(ulong low, ulong high)
+        {
+            _v0 = low;
+            _v1 = high;
+        }
+
+        public int TrailingZeroCount()
+        {
+            int count = BitOperations.TrailingZeroCount(_v0);
+            if (count == 64)
+            {
+                count += BitOperations.TrailingZeroCount(_v1);
+            }
+
+            return count;
+        }
+
+        public static UInt128 Pow2(int x)
+        {
+            if (x >= 64)
+            {
+                return new UInt128(0, 1UL << (x - 64));
+            }
+
+            return new UInt128(1UL << x, 0);
+        }
+
+        public static UInt128 operator ~(UInt128 x)
+        {
+            return new UInt128(~x._v0, ~x._v1);
+        }
+
+        public static UInt128 operator &(UInt128 x, UInt128 y)
+        {
+            return new UInt128(x._v0 & y._v0, x._v1 & y._v1);
+        }
+
+        public static UInt128 operator |(UInt128 x, UInt128 y)
+        {
+            return new UInt128(x._v0 | y._v0, x._v1 | y._v1);
+        }
+
+        public static UInt128 operator <<(UInt128 x, int shift)
+        {
+            if (shift == 0)
+            {
+                return new UInt128(x._v0, x._v1);
+            }
+            else if (shift >= 64)
+            {
+                return new UInt128(0, x._v0 << (shift - 64));
+            }
+
+            ulong shiftOut = x._v0 >> (64 - shift);
+
+            return new UInt128(x._v0 << shift, (x._v1 << shift) | shiftOut);
+        }
+
+        public static UInt128 operator >>(UInt128 x, int shift)
+        {
+            if (shift == 0)
+            {
+                return new UInt128(x._v0, x._v1);
+            }
+            else if (shift >= 64)
+            {
+                return new UInt128(x._v1 >> (shift - 64), 0);
+            }
+
+            ulong shiftOut = x._v1 & ((1UL << shift) - 1);
+
+            return new UInt128((x._v0 >> shift) | (shiftOut << (64 - shift)), x._v1 >> shift);
+        }
+
+        public static bool operator ==(UInt128 x, UInt128 y)
+        {
+            return x.Equals(y);
+        }
+
+        public static bool operator !=(UInt128 x, UInt128 y)
+        {
+            return !x.Equals(y);
+        }
+
+        public override bool Equals(object obj)
+        {
+            return obj is UInt128 other && Equals(other);
+        }
+
+        public bool Equals(UInt128 other)
+        {
+            return _v0 == other._v0 && _v1 == other._v1;
+        }
+
+        public override int GetHashCode()
+        {
+            return HashCode.Combine(_v0, _v1);
+        }
+    }
+}

Ryujinx.Graphics.Vic/Blender.cs

@@ -48,38 +48,10 @@ namespace Ryujinx.Graphics.Vic
 
             int one = 1 << (mtx.MatrixRShift + 8);
 
-
-            // NOTE: This is buggy on .NET 5.0.100, we use a workaround for now (see https://github.com/dotnet/runtime/issues/44704)
-            // TODO: Uncomment this when fixed.
-            //Vector128<int> col1 = Vector128.Create(mtx.MatrixCoeff00, mtx.MatrixCoeff10, mtx.MatrixCoeff20, 0);
-            //Vector128<int> col2 = Vector128.Create(mtx.MatrixCoeff01, mtx.MatrixCoeff11, mtx.MatrixCoeff21, 0);
-            //Vector128<int> col3 = Vector128.Create(mtx.MatrixCoeff02, mtx.MatrixCoeff12, mtx.MatrixCoeff22, one);
-            //Vector128<int> col4 = Vector128.Create(mtx.MatrixCoeff03, mtx.MatrixCoeff13, mtx.MatrixCoeff23, 0);
-
-            Vector128<int> col1 = new Vector128<int>();
-            Vector128<int> col2 = new Vector128<int>();
-            Vector128<int> col3 = new Vector128<int>();
-            Vector128<int> col4 = new Vector128<int>();
-
-            col1 = Sse41.Insert(col1, mtx.MatrixCoeff00, 0);
-            col1 = Sse41.Insert(col1, mtx.MatrixCoeff10, 1);
-            col1 = Sse41.Insert(col1, mtx.MatrixCoeff20, 2);
-            col1 = Sse41.Insert(col1, 0, 3);
-
-            col2 = Sse41.Insert(col2, mtx.MatrixCoeff01, 0);
-            col2 = Sse41.Insert(col2, mtx.MatrixCoeff11, 1);
-            col2 = Sse41.Insert(col2, mtx.MatrixCoeff21, 2);
-            col2 = Sse41.Insert(col2, 0, 3);
-
-            col3 = Sse41.Insert(col3, mtx.MatrixCoeff02, 0);
-            col3 = Sse41.Insert(col3, mtx.MatrixCoeff12, 1);
-            col3 = Sse41.Insert(col3, mtx.MatrixCoeff22, 2);
-            col3 = Sse41.Insert(col3, one, 3);
-
-            col4 = Sse41.Insert(col4, mtx.MatrixCoeff03, 0);
-            col4 = Sse41.Insert(col4, mtx.MatrixCoeff13, 1);
-            col4 = Sse41.Insert(col4, mtx.MatrixCoeff23, 2);
-            col4 = Sse41.Insert(col4, 0, 3);
+            Vector128<int> col1 = Vector128.Create(mtx.MatrixCoeff00, mtx.MatrixCoeff10, mtx.MatrixCoeff20, 0);
+            Vector128<int> col2 = Vector128.Create(mtx.MatrixCoeff01, mtx.MatrixCoeff11, mtx.MatrixCoeff21, 0);
+            Vector128<int> col3 = Vector128.Create(mtx.MatrixCoeff02, mtx.MatrixCoeff12, mtx.MatrixCoeff22, one);
+            Vector128<int> col4 = Vector128.Create(mtx.MatrixCoeff03, mtx.MatrixCoeff13, mtx.MatrixCoeff23, 0);
 
             Vector128<int> rShift = Vector128.CreateScalar(mtx.MatrixRShift);
             Vector128<ushort> clMin = Vector128.Create((ushort)slot.SlotConfig.SoftClampLow);

Ryujinx.Graphics.Vic/Image/BufferPool.cs

@@ -14,7 +14,7 @@ namespace Ryujinx.Graphics.Vic.Image
         /// If the required buffer is larger than this, it won't be
         /// added to the pool to avoid long term high memory usage.
         /// </summary>
-        private const int MaxBufferSize = 2048 * 1280;
+        private const int MaxBufferSize = 2048 * 2048;
 
         private struct PoolItem
         {

Ryujinx.Graphics.Vic/Image/InputSurface.cs

@@ -2,16 +2,85 @@
 
 namespace Ryujinx.Graphics.Vic.Image
 {
+    ref struct RentedBuffer
+    {
+        public static RentedBuffer Empty => new RentedBuffer(Span<byte>.Empty, -1);
+
+        public Span<byte> Data;
+        public int Index;
+
+        public RentedBuffer(Span<byte> data, int index)
+        {
+            Data = data;
+            Index = index;
+        }
+
+        public void Return(BufferPool<byte> pool)
+        {
+            if (Index != -1)
+            {
+                pool.Return(Index);
+            }
+        }
+    }
+
     ref struct InputSurface
     {
         public ReadOnlySpan<byte> Buffer0;
         public ReadOnlySpan<byte> Buffer1;
         public ReadOnlySpan<byte> Buffer2;
 
+        public int Buffer0Index;
+        public int Buffer1Index;
+        public int Buffer2Index;
+
         public int Width;
         public int Height;
 
         public int UvWidth;
         public int UvHeight;
+
+        public void Initialize()
+        {
+            Buffer0Index = -1;
+            Buffer1Index = -1;
+            Buffer2Index = -1;
+        }
+
+        public void SetBuffer0(RentedBuffer buffer)
+        {
+            Buffer0 = buffer.Data;
+            Buffer0Index = buffer.Index;
+        }
+
+        public void SetBuffer1(RentedBuffer buffer)
+        {
+            Buffer1 = buffer.Data;
+            Buffer1Index = buffer.Index;
+        }
+
+        public void SetBuffer2(RentedBuffer buffer)
+        {
+            Buffer2 = buffer.Data;
+            Buffer2Index = buffer.Index;
+        }
+
+        public void Return(BufferPool<byte> pool)
+        {
+            if (Buffer0Index != -1)
+            {
+                pool.Return(Buffer0Index);
+            }
+
+            if (Buffer1Index != -1)
+            {
+                pool.Return(Buffer1Index);
+            }
+
+            if (Buffer2Index != -1)
+            {
+                pool.Return(Buffer2Index);
+            }
+        }
     }
 }

Ryujinx.Graphics.Vic/Image/SurfaceReader.cs

@@ -1,5 +1,5 @@
 using Ryujinx.Common.Logging;
-using Ryujinx.Graphics.Gpu.Memory;
+using Ryujinx.Common.Memory;
 using Ryujinx.Graphics.Texture;
 using Ryujinx.Graphics.Vic.Types;
 using System;
@@ -12,24 +12,32 @@ namespace Ryujinx.Graphics.Vic.Image
 {
     static class SurfaceReader
     {
-        public static Surface Read(ResourceManager rm, ref SlotSurfaceConfig config, ref PlaneOffsets offsets)
+        public static Surface Read(
+            ResourceManager rm,
+            ref SlotConfig config,
+            ref SlotSurfaceConfig surfaceConfig,
+            ref Array8<PlaneOffsets> offsets)
         {
-            switch (config.SlotPixelFormat)
+            switch (surfaceConfig.SlotPixelFormat)
             {
-                case PixelFormat.Y8___V8U8_N420: return ReadNv12(rm, ref config, ref offsets);
+                case PixelFormat.Y8___V8U8_N420: return ReadNv12(rm, ref config, ref surfaceConfig, ref offsets);
             }
 
-            Logger.Error?.Print(LogClass.Vic, $"Unsupported pixel format \"{config.SlotPixelFormat}\".");
+            Logger.Error?.Print(LogClass.Vic, $"Unsupported pixel format \"{surfaceConfig.SlotPixelFormat}\".");
 
-            int lw = config.SlotLumaWidth + 1;
-            int lh = config.SlotLumaHeight + 1;
+            int lw = surfaceConfig.SlotLumaWidth + 1;
+            int lh = surfaceConfig.SlotLumaHeight + 1;
 
             return new Surface(rm.SurfacePool, lw, lh);
         }
 
-        private unsafe static Surface ReadNv12(ResourceManager rm, ref SlotSurfaceConfig config, ref PlaneOffsets offsets)
+        private unsafe static Surface ReadNv12(
+            ResourceManager rm,
+            ref SlotConfig config,
+            ref SlotSurfaceConfig surfaceConfig,
+            ref Array8<PlaneOffsets> offsets)
         {
-            InputSurface input = ReadSurface(rm.Gmm, ref config, ref offsets, 1, 2);
+            InputSurface input = ReadSurface(rm, ref config, ref surfaceConfig, ref offsets, 1, 2);
 
             int width = input.Width;
             int height = input.Height;
@@ -160,6 +168,8 @@ namespace Ryujinx.Graphics.Vic.Image
                 }
             }
 
+            input.Return(rm.BufferPool);
+
             return output;
         }
 
@@ -170,84 +180,227 @@ namespace Ryujinx.Graphics.Vic.Image
         }
 
         private static InputSurface ReadSurface(
-            MemoryManager gmm,
-            ref SlotSurfaceConfig config,
-            ref PlaneOffsets offsets,
+            ResourceManager rm,
+            ref SlotConfig config,
+            ref SlotSurfaceConfig surfaceConfig,
+            ref Array8<PlaneOffsets> offsets,
             int bytesPerPixel,
             int planes)
         {
             InputSurface surface = new InputSurface();
 
-            int gobBlocksInY = 1 << config.SlotBlkHeight;
+            surface.Initialize();
 
-            bool linear = config.SlotBlkKind == 0;
+            int gobBlocksInY = 1 << surfaceConfig.SlotBlkHeight;
 
-            int lw = config.SlotLumaWidth + 1;
-            int lh = config.SlotLumaHeight + 1;
+            bool linear = surfaceConfig.SlotBlkKind == 0;
 
-            int cw = config.SlotChromaWidth + 1;
-            int ch = config.SlotChromaHeight + 1;
+            int lw = surfaceConfig.SlotLumaWidth + 1;
+            int lh = surfaceConfig.SlotLumaHeight + 1;
+
+            int cw = surfaceConfig.SlotChromaWidth + 1;
+            int ch = surfaceConfig.SlotChromaHeight + 1;
+
+            // Interlaced inputs have double the height when deinterlaced.
+            int heightShift = config.FrameFormat.IsField() ? 1 : 0;
 
             surface.Width = lw;
-            surface.Height = lh;
+            surface.Height = lh << heightShift;
             surface.UvWidth = cw;
-            surface.UvHeight = ch;
+            surface.UvHeight = ch << heightShift;
 
             if (planes > 0)
             {
-                surface.Buffer0 = ReadBuffer(gmm, offsets.LumaOffset, linear, lw, lh, bytesPerPixel, gobBlocksInY);
+                surface.SetBuffer0(ReadBuffer(rm, ref config, ref offsets, linear, 0, lw, lh, bytesPerPixel, gobBlocksInY));
             }
 
             if (planes > 1)
             {
-                surface.Buffer1 = ReadBuffer(gmm, offsets.ChromaUOffset, linear, cw, ch, planes == 2 ? 2 : 1, gobBlocksInY);
+                surface.SetBuffer1(ReadBuffer(rm, ref config, ref offsets, linear, 1, cw, ch, planes == 2 ? 2 : 1, gobBlocksInY));
             }
 
             if (planes > 2)
             {
-                surface.Buffer2 = ReadBuffer(gmm, offsets.ChromaVOffset, linear, cw, ch, 1, gobBlocksInY);
+                surface.SetBuffer2(ReadBuffer(rm, ref config, ref offsets, linear, 2, cw, ch, 1, gobBlocksInY));
             }
 
             return surface;
         }
 
-        private static ReadOnlySpan<byte> ReadBuffer(
-            MemoryManager gmm,
-            uint offset,
+        private static RentedBuffer ReadBuffer(
+            ResourceManager rm,
+            ref SlotConfig config,
+            ref Array8<PlaneOffsets> offsets,
             bool linear,
+            int plane,
+            int width,
+            int height,
+            int bytesPerPixel,
+            int gobBlocksInY)
+        {
+            FrameFormat frameFormat = config.FrameFormat;
+            bool isLuma = plane == 0;
+            bool isField = frameFormat.IsField();
+            bool isTopField = frameFormat.IsTopField(isLuma);
+            int stride = GetPitch(width, bytesPerPixel);
+            uint offset = GetOffset(ref offsets[0], plane);
+
+            int dstStart = 0;
+            int dstStride = stride;
+
+            if (isField)
+            {
+                dstStart = isTopField ? 0 : stride;
+                dstStride = stride * 2;
+            }
+
+            RentedBuffer buffer;
+
+            if (linear)
+            {
+                buffer = ReadBufferLinear(rm, offset, width, height, dstStart, dstStride, bytesPerPixel);
+            }
+            else
+            {
+                buffer = ReadBufferBlockLinear(rm, offset, width, height, dstStart, dstStride, bytesPerPixel, gobBlocksInY);
+            }
+
+            if (isField || frameFormat.IsInterlaced())
+            {
+                RentedBuffer prevBuffer = RentedBuffer.Empty;
+                RentedBuffer nextBuffer = RentedBuffer.Empty;
+
+                if (config.PrevFieldEnable)
+                {
+                    prevBuffer = ReadBufferNoDeinterlace(rm, ref offsets[1], linear, plane, width, height, bytesPerPixel, gobBlocksInY);
+                }
+
+                if (config.NextFieldEnable)
+                {
+                    nextBuffer = ReadBufferNoDeinterlace(rm, ref offsets[2], linear, plane, width, height, bytesPerPixel, gobBlocksInY);
+                }
+
+                int w = width * bytesPerPixel;
+
+                switch (config.DeinterlaceMode)
+                {
+                    case DeinterlaceMode.Weave:
+                        Scaler.DeinterlaceWeave(buffer.Data, prevBuffer.Data, w, stride, isTopField);
+                        break;
+                    case DeinterlaceMode.BobField:
+                        Scaler.DeinterlaceBob(buffer.Data, w, stride, isTopField);
+                        break;
+                    case DeinterlaceMode.Bob:
+                        bool isCurrentTop = isLuma ? config.IsEven : config.ChromaEven;
+                        Scaler.DeinterlaceBob(buffer.Data, w, stride, isCurrentTop ^ frameFormat.IsInterlacedBottomFirst());
+                        break;
+                    case DeinterlaceMode.NewBob:
+                    case DeinterlaceMode.Disi1:
+                        Scaler.DeinterlaceMotionAdaptive(buffer.Data, prevBuffer.Data, nextBuffer.Data, w, stride, isTopField);
+                        break;
+                    case DeinterlaceMode.WeaveLumaBobFieldChroma:
+                        if (isLuma)
+                        {
+                            Scaler.DeinterlaceWeave(buffer.Data, prevBuffer.Data, w, stride, isTopField);
+                        }
+                        else
+                        {
+                            Scaler.DeinterlaceBob(buffer.Data, w, stride, isTopField);
+                        }
+                        break;
+                    default:
+                        Logger.Error?.Print(LogClass.Vic, $"Unsupported deinterlace mode \"{config.DeinterlaceMode}\".");
+                        break;
+                }
+
+                prevBuffer.Return(rm.BufferPool);
+                nextBuffer.Return(rm.BufferPool);
+            }
+
+            return buffer;
+        }
+
+        private static uint GetOffset(ref PlaneOffsets offsets, int plane)
+        {
+            return plane switch
+            {
+                0 => offsets.LumaOffset,
+                1 => offsets.ChromaUOffset,
+                2 => offsets.ChromaVOffset,
+                _ => throw new ArgumentOutOfRangeException(nameof(plane))
+            };
+        }
+
+        private static RentedBuffer ReadBufferNoDeinterlace(
+            ResourceManager rm,
+            ref PlaneOffsets offsets,
+            bool linear,
+            int plane,
             int width,
             int height,
             int bytesPerPixel,
             int gobBlocksInY)
         {
             int stride = GetPitch(width, bytesPerPixel);
+            uint offset = GetOffset(ref offsets, plane);
 
             if (linear)
             {
-                return gmm.GetSpan(ExtendOffset(offset), stride * height);
+                return ReadBufferLinear(rm, offset, width, height, 0, stride, bytesPerPixel);
             }
 
-            return ReadBuffer(gmm, offset, width, height, stride, bytesPerPixel, gobBlocksInY);
+            return ReadBufferBlockLinear(rm, offset, width, height, 0, stride, bytesPerPixel, gobBlocksInY);
         }
 
-        private static ReadOnlySpan<byte> ReadBuffer(
-            MemoryManager gmm,
+        private static RentedBuffer ReadBufferLinear(
+            ResourceManager rm,
             uint offset,
             int width,
             int height,
+            int dstStart,
+            int dstStride,
+            int bytesPerPixel)
+        {
+            int srcStride = GetPitch(width, bytesPerPixel);
+            int inSize = srcStride * height;
+
+            ReadOnlySpan<byte> src = rm.Gmm.GetSpan(ExtendOffset(offset), inSize);
+
+            int outSize = dstStride * height;
+            int bufferIndex = rm.BufferPool.RentMinimum(outSize, out byte[] buffer);
+            Span<byte> dst = buffer;
+            dst = dst.Slice(0, outSize);
+
+            for (int y = 0; y < height; y++)
+            {
+                src.Slice(y * srcStride, srcStride).CopyTo(dst.Slice(dstStart + y * dstStride, srcStride));
+            }
+
+            return new RentedBuffer(dst, bufferIndex);
+        }
+
+        private static RentedBuffer ReadBufferBlockLinear(
+            ResourceManager rm,
+            uint offset,
+            int width,
+            int height,
+            int dstStart,
             int dstStride,
             int bytesPerPixel,
             int gobBlocksInY)
         {
             int inSize = GetBlockLinearSize(width, height, bytesPerPixel, gobBlocksInY);
 
-            ReadOnlySpan<byte> src = gmm.GetSpan(ExtendOffset(offset), inSize);
+            ReadOnlySpan<byte> src = rm.Gmm.GetSpan(ExtendOffset(offset), inSize);
 
-            Span<byte> dst = new byte[dstStride * height];
+            int outSize = dstStride * height;
+            int bufferIndex = rm.BufferPool.RentMinimum(outSize, out byte[] buffer);
+            Span<byte> dst = buffer;
+            dst = dst.Slice(0, outSize);
 
-            LayoutConverter.ConvertBlockLinearToLinear(dst, width, height, dstStride, bytesPerPixel, gobBlocksInY, src);
+            LayoutConverter.ConvertBlockLinearToLinear(dst.Slice(dstStart), width, height, dstStride, bytesPerPixel, gobBlocksInY, src);
 
-            return dst;
+            return new RentedBuffer(dst, bufferIndex);
         }
     }
 }

Ryujinx.Graphics.Vic/Image/SurfaceWriter.cs

@@ -15,7 +15,7 @@ namespace Ryujinx.Graphics.Vic.Image
             switch (config.OutPixelFormat)
             {
                 case PixelFormat.A8B8G8R8:
-                case PixelFormat.X8B8G8R8:    
+                case PixelFormat.X8B8G8R8:
                     WriteA8B8G8R8(rm, input, ref config, ref offsets);
                     break;
                 case PixelFormat.A8R8G8B8:
@@ -433,7 +433,7 @@ namespace Ryujinx.Graphics.Vic.Image
         {
             if (linear)
             {
-                rm.Gmm.Write(ExtendOffset(offset), src);
+                rm.Gmm.WriteMapped(ExtendOffset(offset), src);
                 return;
             }
 
@@ -456,7 +456,7 @@ namespace Ryujinx.Graphics.Vic.Image
 
             LayoutConverter.ConvertLinearToBlockLinear(dst, width, height, dstStride, bytesPerPixel, gobBlocksInY, src);
 
-            rm.Gmm.Write(ExtendOffset(offset), dst);
+            rm.Gmm.WriteMapped(ExtendOffset(offset), dst);
 
             rm.BufferPool.Return(dstIndex);
         }

Ryujinx.Graphics.Vic/Scaler.cs

@@ -0,0 +1,124 @@
+using System;
+using System.Runtime.Intrinsics;
+using System.Runtime.Intrinsics.X86;
+
+namespace Ryujinx.Graphics.Vic
+{
+    static class Scaler
+    {
+        public static void DeinterlaceWeave(Span<byte> data, ReadOnlySpan<byte> prevData, int width, int fieldSize, bool isTopField)
+        {
+            // Prev I    Curr I    Curr P
+            // TTTTTTTT  BBBBBBBB  TTTTTTTT
+            // --------  --------  BBBBBBBB
+
+            if (isTopField)
+            {
+                for (int offset = 0; offset < data.Length; offset += fieldSize * 2)
+                {
+                    prevData.Slice(offset >> 1, width).CopyTo(data.Slice(offset + fieldSize, width));
+                }
+            }
+            else
+            {
+                for (int offset = 0; offset < data.Length; offset += fieldSize * 2)
+                {
+                    prevData.Slice(offset >> 1, width).CopyTo(data.Slice(offset, width));
+                }
+            }
+        }
+
+        public static void DeinterlaceBob(Span<byte> data, int width, int fieldSize, bool isTopField)
+        {
+            // Curr I    Curr P
+            // TTTTTTTT  TTTTTTTT
+            // --------  TTTTTTTT
+
+            if (isTopField)
+            {
+                for (int offset = 0; offset < data.Length; offset += fieldSize * 2)
+                {
+                    data.Slice(offset, width).CopyTo(data.Slice(offset + fieldSize, width));
+                }
+            }
+            else
+            {
+                for (int offset = 0; offset < data.Length; offset += fieldSize * 2)
+                {
+                    data.Slice(offset + fieldSize, width).CopyTo(data.Slice(offset, width));
+                }
+            }
+        }
+
+        public unsafe static void DeinterlaceMotionAdaptive(
+            Span<byte> data,
+            ReadOnlySpan<byte> prevData,
+            ReadOnlySpan<byte> nextData,
+            int width,
+            int fieldSize,
+            bool isTopField)
+        {
+            // Very simple motion adaptive algorithm.
+            // If the pixel changed between previous and next frame, use Bob, otherwise use Weave.
+            //
+            // Example pseudo code:
+            // C_even = (P_even == N_even) ? P_even : C_odd
+            // Where: C is current frame, P is previous frame and N is next frame, and even/odd are the fields.
+            //
+            // Note: This does not fully match the hardware algorithm.
+            // The motion adaptive deinterlacing implemented on hardware is considerably more complex,
+            // and hard to implement accurately without proper documentation as for example, the
+            // method used for motion estimation is unknown.
+
+            int start = isTopField ? fieldSize : 0;
+            int otherFieldOffset = isTopField ? -fieldSize : fieldSize;
+
+            fixed (byte* pData = data, pPrevData = prevData, pNextData = nextData)
+            {
+                for (int offset = start; offset < data.Length; offset += fieldSize * 2)
+                {
+                    int refOffset = (offset - start) >> 1;
+                    int x = 0;
+
+                    if (Avx2.IsSupported)
+                    {
+                        for (; x < (width & ~0x1f); x += 32)
+                        {
+                            Vector256<byte> prevPixels = Avx.LoadVector256(pPrevData + refOffset + x);
+                            Vector256<byte> nextPixels = Avx.LoadVector256(pNextData + refOffset + x);
+                            Vector256<byte> bob = Avx.LoadVector256(pData + offset + otherFieldOffset + x);
+                            Vector256<byte> diff = Avx2.CompareEqual(prevPixels, nextPixels);
+                            Avx.Store(pData + offset + x, Avx2.BlendVariable(bob, prevPixels, diff));
+                        }
+                    }
+                    else if (Sse41.IsSupported)
+                    {
+                        for (; x < (width & ~0xf); x += 16)
+                        {
+                            Vector128<byte> prevPixels = Sse2.LoadVector128(pPrevData + refOffset + x);
+                            Vector128<byte> nextPixels = Sse2.LoadVector128(pNextData + refOffset + x);
+                            Vector128<byte> bob = Sse2.LoadVector128(pData + offset + otherFieldOffset + x);
+                            Vector128<byte> diff = Sse2.CompareEqual(prevPixels, nextPixels);
+                            Sse2.Store(pData + offset + x, Sse41.BlendVariable(bob, prevPixels, diff));
+                        }
+                    }
+
+                    for (; x < width; x++)
+                    {
+                        byte prevPixel = prevData[refOffset + x];
+                        byte nextPixel = nextData[refOffset + x];
+
+                        if (nextPixel != prevPixel)
+                        {
+                            data[offset + x] = data[offset + otherFieldOffset + x];
+                        }
+                        else
+                        {
+                            data[offset + x] = prevPixel;
+                        }
+                    }
+                }
+            }
+        }
+    }
+}

Ryujinx.Graphics.Vic/Types/DeinterlaceMode.cs

@@ -0,0 +1,12 @@
+namespace Ryujinx.Graphics.Vic.Types
+{
+    enum DeinterlaceMode
+    {
+        Weave,
+        BobField,
+        Bob,
+        NewBob,
+        Disi1,
+        WeaveLumaBobFieldChroma
+    }
+}

Ryujinx.Graphics.Vic/Types/FrameFormat.cs

@@ -0,0 +1,79 @@
+namespace Ryujinx.Graphics.Vic.Types
+{
+    enum FrameFormat
+    {
+        Progressive,
+        InterlacedTopFieldFirst,
+        InterlacedBottomFieldFirst,
+        TopField,
+        BottomField,
+        SubPicProgressive,
+        SubPicInterlacedTopFieldFirst,
+        SubPicInterlacedBottomFieldFirst,
+        SubPicTopField,
+        SubPicBottomField,
+        TopFieldChromaBottom,
+        BottomFieldChromaTop,
+        SubPicTopFieldChromaBottom,
+        SubPicBottomFieldChromaTop
+    }
+
+    static class FrameFormatExtensions
+    {
+        public static bool IsField(this FrameFormat frameFormat)
+        {
+            switch (frameFormat)
+            {
+                case FrameFormat.TopField:
+                case FrameFormat.BottomField:
+                case FrameFormat.SubPicTopField:
+                case FrameFormat.SubPicBottomField:
+                case FrameFormat.TopFieldChromaBottom:
+                case FrameFormat.BottomFieldChromaTop:
+                case FrameFormat.SubPicTopFieldChromaBottom:
+                case FrameFormat.SubPicBottomFieldChromaTop:
+                    return true;
+            }
+
+            return false;
+        }
+
+        public static bool IsInterlaced(this FrameFormat frameFormat)
+        {
+            switch (frameFormat)
+            {
+                case FrameFormat.InterlacedTopFieldFirst:
+                case FrameFormat.InterlacedBottomFieldFirst:
+                case FrameFormat.SubPicInterlacedTopFieldFirst:
+                case FrameFormat.SubPicInterlacedBottomFieldFirst:
+                    return true;
+            }
+
+            return false;
+        }
+
+        public static bool IsInterlacedBottomFirst(this FrameFormat frameFormat)
+        {
+            return frameFormat == FrameFormat.InterlacedBottomFieldFirst ||
+                   frameFormat == FrameFormat.SubPicInterlacedBottomFieldFirst;
+        }
+
+        public static bool IsTopField(this FrameFormat frameFormat, bool isLuma)
+        {
+            switch (frameFormat)
+            {
+                case FrameFormat.TopField:
+                case FrameFormat.SubPicTopField:
+                    return true;
+                case FrameFormat.TopFieldChromaBottom:
+                case FrameFormat.SubPicTopFieldChromaBottom:
+                    return isLuma;
+                case FrameFormat.BottomFieldChromaTop:
+                case FrameFormat.SubPicBottomFieldChromaTop:
+                    return !isLuma;
+            }
+
+            return false;
+        }
+    }
+}

Ryujinx.Graphics.Vic/Types/SlotConfig.cs

@@ -27,7 +27,7 @@
         public bool PrevMotionFieldEnable => _word0.Extract(13);
         public bool PpMotionFieldEnable => _word0.Extract(14);
         public bool CombMotionFieldEnable => _word0.Extract(15);
-        public int FrameFormat => _word0.Extract(16, 4);
+        public FrameFormat FrameFormat => (FrameFormat)_word0.Extract(16, 4);
         public int FilterLengthY => _word0.Extract(20, 2);
         public int FilterLengthX => _word0.Extract(22, 2);
         public int Panoramic => _word0.Extract(24, 12);
@@ -36,7 +36,7 @@
         public int FilterDetail => _word1.Extract(74, 10);
         public int ChromaNoise => _word1.Extract(84, 10);
         public int ChromaDetail => _word1.Extract(94, 10);
-        public int DeinterlaceMode => _word1.Extract(104, 4);
+        public DeinterlaceMode DeinterlaceMode => (DeinterlaceMode)_word1.Extract(104, 4);
         public int MotionAccumWeight => _word1.Extract(108, 3);
         public int NoiseIir => _word1.Extract(111, 11);
         public int LightLevel => _word1.Extract(122, 4);

Ryujinx.Graphics.Vic/VicDevice.cs

@@ -43,9 +43,9 @@ namespace Ryujinx.Graphics.Vic
                     continue;
                 }
 
-                var offsets = _state.State.SetSurfacexSlotx[i][0];
+                ref var offsets = ref _state.State.SetSurfacexSlotx[i];
 
-                using Surface src = SurfaceReader.Read(_rm, ref slot.SlotSurfaceConfig, ref offsets);
+                using Surface src = SurfaceReader.Read(_rm, ref slot.SlotConfig, ref slot.SlotSurfaceConfig, ref offsets);
 
                 Blender.BlendOne(output, src, ref slot);
             }

Ryujinx.Graphics.Video/FrameField.cs

@@ -0,0 +1,8 @@
+namespace Ryujinx.Graphics.Video
+{
+    public enum FrameField
+    {
+        Progressive,
+        Interlaced
+    }
+}

Ryujinx.Graphics.Video/ISurface.cs

@@ -8,6 +8,8 @@ namespace Ryujinx.Graphics.Video
         Plane UPlane { get; }
         Plane VPlane { get; }
 
+        FrameField Field { get; }
+
         int Width { get; }
         int Height { get; }
         int Stride { get; }

Ryujinx.HLE/FileSystem/Content/ContentPath.cs

@@ -1,19 +0,0 @@
-namespace Ryujinx.HLE.FileSystem.Content
-{
-    static class ContentPath
-    {
-        public const string SystemContent    = "@SystemContent";
-        public const string UserContent      = "@UserContent";
-        public const string SdCardContent    = "@SdCardContent";
-        public const string SdCard           = "@SdCard";
-        public const string CalibFile        = "@CalibFile";
-        public const string Safe             = "@Safe";
-        public const string User             = "@User";
-        public const string System           = "@System";
-        public const string Host             = "@Host";
-        public const string GamecardApp      = "@GcApp";
-        public const string GamecardContents = "@GcS00000001";
-        public const string GamecardUpdate   = "@upp";
-        public const string RegisteredUpdate = "@RegUpdate";
-    }
-}

Ryujinx.HLE/FileSystem/Content/LocationHelper.cs

@@ -1,91 +0,0 @@
-using System;
-using System.IO;
-
-using static Ryujinx.HLE.FileSystem.VirtualFileSystem;
-
-namespace Ryujinx.HLE.FileSystem.Content
-{
-    internal static class LocationHelper
-    {
-        public static string GetRealPath(VirtualFileSystem fileSystem, string switchContentPath)
-        {
-            string basePath = fileSystem.GetBasePath();
-
-            switch (switchContentPath)
-            {
-                case ContentPath.SystemContent:
-                    return Path.Combine(basePath, SystemNandPath, "Contents");
-                case ContentPath.UserContent:
-                    return Path.Combine(basePath, UserNandPath, "Contents");
-                case ContentPath.SdCardContent:
-                    return Path.Combine(fileSystem.GetSdCardPath(), "Nintendo", "Contents");
-                case ContentPath.System:
-                    return Path.Combine(basePath, SystemNandPath);
-                case ContentPath.User:
-                    return Path.Combine(basePath, UserNandPath);
-                default:
-                    throw new NotSupportedException($"Content Path `{switchContentPath}` is not supported.");
-            }
-        }
-
-        public static string GetContentPath(ContentStorageId contentStorageId)
-        {
-            switch (contentStorageId)
-            {
-                case ContentStorageId.NandSystem:
-                    return ContentPath.SystemContent;
-                case ContentStorageId.NandUser:
-                    return ContentPath.UserContent;
-                case ContentStorageId.SdCard:
-                    return ContentPath.SdCardContent;
-                default:
-                    throw new NotSupportedException($"Content Storage `{contentStorageId}` is not supported.");
-            }
-        }
-
-        public static string GetContentRoot(StorageId storageId)
-        {
-            switch (storageId)
-            {
-                case StorageId.NandSystem:
-                    return ContentPath.SystemContent;
-                case StorageId.NandUser:
-                    return ContentPath.UserContent;
-                case StorageId.SdCard:
-                    return ContentPath.SdCardContent;
-                default:
-                    throw new NotSupportedException($"Storage Id `{storageId}` is not supported.");
-            }
-        }
-
-        public static StorageId GetStorageId(string contentPathString)
-        {
-            string cleanedPath = contentPathString.Split(':')[0];
-
-            switch (cleanedPath)
-            {
-                case ContentPath.SystemContent:
-                case ContentPath.System:
-                    return StorageId.NandSystem;
-
-                case ContentPath.UserContent:
-                case ContentPath.User:
-                    return StorageId.NandUser;
-
-                case ContentPath.SdCardContent:
-                    return StorageId.SdCard;
-
-                case ContentPath.Host:
-                    return StorageId.Host;
-
-                case ContentPath.GamecardApp:
-                case ContentPath.GamecardContents:
-                case ContentPath.GamecardUpdate:
-                    return StorageId.GameCard;
-
-                default:
-                    return StorageId.None;
-            }
-        }
-    }
-}

Ryujinx.HLE/FileSystem/Content/StorageId.cs

@@ -1,9 +0,0 @@
-namespace Ryujinx.HLE.FileSystem.Content
-{
-    public enum ContentStorageId
-    {
-        NandSystem,
-        NandUser,
-        SdCard
-    }
-}

Ryujinx.HLE/FileSystem/Content/TitleType.cs

@@ -1,15 +0,0 @@
-namespace Ryujinx.HLE.FileSystem.Content
-{
-    enum TitleType
-    {
-        SystemPrograms     = 0x01,
-        SystemDataArchive  = 0x02,
-        SystemUpdate       = 0x03,
-        FirmwarePackageA   = 0x04,
-        FirmwarePackageB   = 0x05,
-        RegularApplication = 0x80,
-        Update             = 0x81,
-        AddOnContent       = 0x82,
-        DeltaTitle         = 0x83
-    }
-}

Ryujinx.HLE/FileSystem/ContentManager.cs

@@ -20,7 +20,7 @@ using System.IO.Compression;
 using System.Linq;
 using Path = System.IO.Path;
 
-namespace Ryujinx.HLE.FileSystem.Content
+namespace Ryujinx.HLE.FileSystem
 {
     public class ContentManager
     {
@@ -110,8 +110,8 @@ namespace Ryujinx.HLE.FileSystem.Content
 
                     try
                     {
-                        contentPathString   = LocationHelper.GetContentRoot(storageId);
-                        contentDirectory    = LocationHelper.GetRealPath(_virtualFileSystem, contentPathString);
+                        contentPathString   = ContentPath.GetContentPath(storageId);
+                        contentDirectory    = ContentPath.GetRealPath(_virtualFileSystem, contentPathString);
                         registeredDirectory = Path.Combine(contentDirectory, "registered");
                     }
                     catch (NotSupportedException)
@@ -367,8 +367,7 @@ namespace Ryujinx.HLE.FileSystem.Content
             {
                 LocationEntry locationEntry = GetLocation(titleId, contentType, storageId);
 
-                return locationEntry.ContentPath != null ?
-                    LocationHelper.GetStorageId(locationEntry.ContentPath) : StorageId.None;
+                return locationEntry.ContentPath != null ? ContentPath.GetStorageId(locationEntry.ContentPath) : StorageId.None;
             }
         }
 
@@ -493,8 +492,8 @@ namespace Ryujinx.HLE.FileSystem.Content
 
         public void InstallFirmware(string firmwareSource)
         {
-            string contentPathString   = LocationHelper.GetContentRoot(StorageId.NandSystem);
-            string contentDirectory    = LocationHelper.GetRealPath(_virtualFileSystem, contentPathString);
+            string contentPathString   = ContentPath.GetContentPath(StorageId.BuiltInSystem);
+            string contentDirectory    = ContentPath.GetRealPath(_virtualFileSystem, contentPathString);
             string registeredDirectory = Path.Combine(contentDirectory, "registered");
             string temporaryDirectory  = Path.Combine(contentDirectory, "temp");
 
@@ -998,9 +997,9 @@ namespace Ryujinx.HLE.FileSystem.Content
 
                     foreach (var entry in updateNcas)
                     {
-                        foreach (var nca in entry.Value)
+                        foreach (var (type, path) in entry.Value)
                         {
-                            extraNcas += nca.path + Environment.NewLine;
+                            extraNcas += path + Environment.NewLine;
                         }
                     }
 
@@ -1019,7 +1018,7 @@ namespace Ryujinx.HLE.FileSystem.Content
 
             lock (_lock)
             {
-                var locationEnties = _locationEntries[StorageId.NandSystem];
+                var locationEnties = _locationEntries[StorageId.BuiltInSystem];
 
                 foreach (var entry in locationEnties)
                 {

Ryujinx.HLE/FileSystem/ContentPath.cs

@@ -0,0 +1,82 @@
+using LibHac.Fs;
+using LibHac.Ncm;
+using Ryujinx.Common.Configuration;
+using System;
+
+using static Ryujinx.HLE.FileSystem.VirtualFileSystem;
+using Path = System.IO.Path;
+
+namespace Ryujinx.HLE.FileSystem
+{
+    internal static class ContentPath
+    {
+        public const string SystemContent    = "@SystemContent";
+        public const string UserContent      = "@UserContent";
+        public const string SdCardContent    = "@SdCardContent";
+        public const string SdCard           = "@Sdcard";
+        public const string CalibFile        = "@CalibFile";
+        public const string Safe             = "@Safe";
+        public const string User             = "@User";
+        public const string System           = "@System";
+        public const string Host             = "@Host";
+        public const string GamecardApp      = "@GcApp";
+        public const string GamecardContents = "@GcS00000001";
+        public const string GamecardUpdate   = "@upp";
+        public const string RegisteredUpdate = "@RegUpdate";
+
+        public const string Nintendo = "Nintendo";
+        public const string Contents = "Contents";
+
+        public static string GetRealPath(VirtualFileSystem fileSystem, string switchContentPath)
+        {
+            return switchContentPath switch
+            {
+                SystemContent => Path.Combine(AppDataManager.BaseDirPath, SystemNandPath, Contents),
+                UserContent   => Path.Combine(AppDataManager.BaseDirPath, UserNandPath,   Contents),
+                SdCardContent => Path.Combine(fileSystem.GetSdCardPath(), Nintendo,       Contents),
+                System        => Path.Combine(AppDataManager.BaseDirPath, SystemNandPath),
+                User          => Path.Combine(AppDataManager.BaseDirPath, UserNandPath),
+                _ => throw new NotSupportedException($"Content Path \"`{switchContentPath}`\" is not supported.")
+            };
+        }
+
+        public static string GetContentPath(ContentStorageId contentStorageId)
+        {
+            return contentStorageId switch
+            {
+                ContentStorageId.System => SystemContent,
+                ContentStorageId.User   => UserContent,
+                ContentStorageId.SdCard => SdCardContent,
+                _ => throw new NotSupportedException($"Content Storage Id \"`{contentStorageId}`\" is not supported.")
+            };
+        }
+
+        public static string GetContentPath(StorageId storageId)
+        {
+            return storageId switch
+            {
+                StorageId.BuiltInSystem => SystemContent,
+                StorageId.BuiltInUser   => UserContent,
+                StorageId.SdCard        => SdCardContent,
+                _ => throw new NotSupportedException($"Storage Id \"`{storageId}`\" is not supported.")
+            };
+        }
+
+        public static StorageId GetStorageId(string contentPathString)
+        {
+            return contentPathString.Split(':')[0] switch
+            {
+                SystemContent or
+                System         => StorageId.BuiltInSystem,
+                UserContent or
+                User           => StorageId.BuiltInUser,
+                SdCardContent  => StorageId.SdCard,
+                Host           => StorageId.Host,
+                GamecardApp or
+                GamecardContents or
+                GamecardUpdate => StorageId.GameCard,
+                _              => StorageId.None
+            };
+        }
+    }
+}

Ryujinx.HLE/FileSystem/EncryptedFileSystemCreator.cs

@@ -8,7 +8,6 @@ namespace Ryujinx.HLE.FileSystem
 {
     public class EncryptedFileSystemCreator : IEncryptedFileSystemCreator
     {
-
         public Result Create(ref SharedRef<IFileSystem> outEncryptedFileSystem,
             ref SharedRef<IFileSystem> baseFileSystem, IEncryptedFileSystemCreator.KeyId idIndex,
             in EncryptionSeed encryptionSeed)
@@ -18,10 +17,10 @@ namespace Ryujinx.HLE.FileSystem
                 return ResultFs.InvalidArgument.Log();
             }
 
-            // Todo: Reenable when AesXtsFileSystem is fixed
+            // TODO: Reenable when AesXtsFileSystem is fixed.
             outEncryptedFileSystem = SharedRef<IFileSystem>.CreateMove(ref baseFileSystem);
 
             return Result.Success;
         }
     }
-}
+}

Ryujinx.HLE/FileSystem/LocationEntry.cs

@@ -1,6 +1,6 @@
 using LibHac.FsSystem;
 
-namespace Ryujinx.HLE.FileSystem.Content
+namespace Ryujinx.HLE.FileSystem
 {
     public struct LocationEntry
     {

Ryujinx.HLE/FileSystem/SaveDataType.cs

@@ -1,12 +0,0 @@
-namespace Ryujinx.HLE.FileSystem
-{
-    enum SaveDataType : byte
-    {
-        SystemSaveData,
-        SaveData,
-        BcatDeliveryCacheStorage,
-        DeviceSaveData,
-        TemporaryStorage,
-        CacheStorage
-    }
-}

Ryujinx.HLE/FileSystem/SaveInfo.cs

@@ -1,27 +0,0 @@
-using Ryujinx.HLE.HOS.Services.Account.Acc;
-
-namespace Ryujinx.HLE.FileSystem
-{
-    struct SaveInfo
-    {
-        public ulong        TitleId      { get; private set; }
-        public long         SaveId       { get; private set; }
-        public SaveDataType SaveDataType { get; private set; }
-        public SaveSpaceId  SaveSpaceId  { get; private set; }
-        public UserId       UserId       { get; private set; }
-
-        public SaveInfo(
-            ulong        titleId,
-            long         saveId,
-            SaveDataType saveDataType,
-            SaveSpaceId  saveSpaceId,
-            UserId       userId = new UserId())
-        {
-            TitleId      = titleId;
-            SaveId       = saveId;
-            SaveDataType = saveDataType;
-            SaveSpaceId  = saveSpaceId;
-            UserId       = userId;
-        }
-    }
-}

Ryujinx.HLE/FileSystem/SaveSpaceId.cs

@@ -1,10 +0,0 @@
-namespace Ryujinx.HLE.FileSystem
-{
-    enum SaveSpaceId
-    {
-        NandSystem,
-        NandUser,
-        SdCard,
-        TemporaryStorage
-    }
-}

Ryujinx.HLE/FileSystem/StorageId.cs

@@ -1,12 +0,0 @@
-namespace Ryujinx.HLE.FileSystem
-{
-    public enum StorageId
-    {
-        None,
-        Host,
-        GameCard,
-        NandSystem,
-        NandUser,
-        SdCard
-    }
-}

Ryujinx.HLE/FileSystem/SystemVersion.cs

@@ -1,8 +1,7 @@
 using Ryujinx.HLE.Utilities;
 using System.IO;
-using System.Text;
 
-namespace Ryujinx.HLE.FileSystem.Content
+namespace Ryujinx.HLE.FileSystem
 {
     public class SystemVersion
     {

Ryujinx.HLE/FileSystem/VirtualFileSystem.cs

@@ -13,7 +13,6 @@ using LibHac.Tools.Fs;
 using LibHac.Tools.FsSystem;
 using Ryujinx.Common.Configuration;
 using Ryujinx.Common.Logging;
-using Ryujinx.HLE.FileSystem.Content;
 using Ryujinx.HLE.HOS;
 using System;
 using System.Buffers.Text;
@@ -28,20 +27,29 @@ namespace Ryujinx.HLE.FileSystem
 {
     public class VirtualFileSystem : IDisposable
     {
-        public const string NandPath = AppDataManager.DefaultNandDir;
-        public const string SdCardPath = AppDataManager.DefaultSdcardDir;
+        public static string SafeNandPath   = Path.Combine(AppDataManager.DefaultNandDir, "safe");
+        public static string SystemNandPath = Path.Combine(AppDataManager.DefaultNandDir, "system");
+        public static string UserNandPath   = Path.Combine(AppDataManager.DefaultNandDir, "user");
 
-        public static string SafeNandPath = Path.Combine(NandPath, "safe");
-        public static string SystemNandPath = Path.Combine(NandPath, "system");
-        public static string UserNandPath = Path.Combine(NandPath, "user");
+        public KeySet           KeySet    { get; private set; }
+        public EmulatedGameCard GameCard  { get; private set; }
+        public EmulatedSdCard   SdCard    { get; private set; }
+        public ModLoader        ModLoader { get; private set; }
+        public Stream           RomFs     { get; private set; }
 
         private static bool _isInitialized = false;
 
-        public KeySet KeySet { get; private set; }
-        public EmulatedGameCard GameCard { get; private set; }
-        public EmulatedSdCard SdCard { get; private set; }
+        public static VirtualFileSystem CreateInstance()
+        {
+            if (_isInitialized)
+            {
+                throw new InvalidOperationException("VirtualFileSystem can only be instantiated once!");
+            }
 
-        public ModLoader ModLoader { get; private set; }
+            _isInitialized = true;
+
+            return new VirtualFileSystem();
+        }
 
         private VirtualFileSystem()
         {
@@ -49,8 +57,6 @@ namespace Ryujinx.HLE.FileSystem
             ModLoader = new ModLoader(); // Should only be created once
         }
 
-        public Stream RomFs { get; private set; }
-
         public void LoadRomFs(string fileName)
         {
             RomFs = new FileStream(fileName, FileMode.Open, FileAccess.Read);
@@ -79,7 +85,7 @@ namespace Ryujinx.HLE.FileSystem
 
             string fullPath = Path.GetFullPath(Path.Combine(basePath, fileName));
 
-            if (!fullPath.StartsWith(GetBasePath()))
+            if (!fullPath.StartsWith(AppDataManager.BaseDirPath))
             {
                 return null;
             }
@@ -87,14 +93,8 @@ namespace Ryujinx.HLE.FileSystem
             return fullPath;
         }
 
-        internal string GetBasePath() => AppDataManager.BaseDirPath;
-        internal string GetSdCardPath() => MakeFullPath(SdCardPath);
-        public string GetNandPath() => MakeFullPath(NandPath);
-
-        public string GetFullPartitionPath(string partitionPath)
-        {
-            return MakeFullPath(partitionPath);
-        }
+        internal string GetSdCardPath() => MakeFullPath(AppDataManager.DefaultSdcardDir);
+        public string GetNandPath() => MakeFullPath(AppDataManager.DefaultNandDir);
 
         public string SwitchPathToSystemPath(string switchPath)
         {
@@ -110,7 +110,7 @@ namespace Ryujinx.HLE.FileSystem
 
         public string SystemPathToSwitchPath(string systemPath)
         {
-            string baseSystemPath = GetBasePath() + Path.DirectorySeparatorChar;
+            string baseSystemPath = AppDataManager.BaseDirPath + Path.DirectorySeparatorChar;
 
             if (systemPath.StartsWith(baseSystemPath))
             {
@@ -136,8 +136,7 @@ namespace Ryujinx.HLE.FileSystem
             switch (path)
             {
                 case ContentPath.SdCard:
-                case "@Sdcard":
-                    path = SdCardPath;
+                    path = AppDataManager.DefaultSdcardDir;
                     break;
                 case ContentPath.User:
                     path = UserNandPath;
@@ -146,7 +145,7 @@ namespace Ryujinx.HLE.FileSystem
                     path = SystemNandPath;
                     break;
                 case ContentPath.SdCardContent:
-                    path = Path.Combine(SdCardPath, "Nintendo", "Contents");
+                    path = Path.Combine(AppDataManager.DefaultSdcardDir, "Nintendo", "Contents");
                     break;
                 case ContentPath.UserContent:
                     path = Path.Combine(UserNandPath, "Contents");
@@ -156,27 +155,19 @@ namespace Ryujinx.HLE.FileSystem
                     break;
             }
 
-            string fullPath = Path.Combine(GetBasePath(), path);
+            string fullPath = Path.Combine(AppDataManager.BaseDirPath, path);
 
-            if (isDirectory)
+            if (isDirectory && !Directory.Exists(fullPath))
             {
-                if (!Directory.Exists(fullPath))
-                {
-                    Directory.CreateDirectory(fullPath);
-                }
+                Directory.CreateDirectory(fullPath);
             }
 
             return fullPath;
         }
 
-        public DriveInfo GetDrive()
-        {
-            return new DriveInfo(Path.GetPathRoot(GetBasePath()));
-        }
-
         public void InitializeFsServer(LibHac.Horizon horizon, out HorizonClient fsServerClient)
         {
-            LocalFileSystem serverBaseFs = new LocalFileSystem(GetBasePath());
+            LocalFileSystem serverBaseFs = new LocalFileSystem(AppDataManager.BaseDirPath);
 
             fsServerClient = horizon.CreatePrivilegedHorizonClient();
             var fsServer = new FileSystemServer(fsServerClient);
@@ -505,7 +496,7 @@ namespace Ryujinx.HLE.FileSystem
 
             bool canFixBySaveDataId = extraData.Attribute.StaticSaveDataId == 0 && info.StaticSaveDataId != 0;
 
-            bool hasEmptyOwnerId = extraData.OwnerId == 0 && info.Type != LibHac.Fs.SaveDataType.System;
+            bool hasEmptyOwnerId = extraData.OwnerId == 0 && info.Type != SaveDataType.System;
 
             if (!canFixByProgramId && !canFixBySaveDataId && !hasEmptyOwnerId)
             {
@@ -523,7 +514,7 @@ namespace Ryujinx.HLE.FileSystem
 
             // The rest of the extra data can't be created from the save data info.
             // On user saves the owner ID will almost certainly be the same as the program ID.
-            if (info.Type != LibHac.Fs.SaveDataType.System)
+            if (info.Type != SaveDataType.System)
             {
                 extraData.OwnerId = info.ProgramId.Value;
             }
@@ -580,11 +571,6 @@ namespace Ryujinx.HLE.FileSystem
             }
         };
 
-        public void Unload()
-        {
-            RomFs?.Dispose();
-        }
-
         public void Dispose()
         {
             Dispose(true);
@@ -594,20 +580,8 @@ namespace Ryujinx.HLE.FileSystem
         {
             if (disposing)
             {
-                Unload();
+                RomFs?.Dispose();
             }
         }
-
-        public static VirtualFileSystem CreateInstance()
-        {
-            if (_isInitialized)
-            {
-                throw new InvalidOperationException("VirtualFileSystem can only be instantiated once!");
-            }
-
-            _isInitialized = true;
-
-            return new VirtualFileSystem();
-        }
     }
 }

Ryujinx.HLE/HLEConfiguration.cs

@@ -3,7 +3,6 @@ using Ryujinx.Audio.Integration;
 using Ryujinx.Common.Configuration;
 using Ryujinx.Graphics.GAL;
 using Ryujinx.HLE.FileSystem;
-using Ryujinx.HLE.FileSystem.Content;
 using Ryujinx.HLE.HOS;
 using Ryujinx.HLE.HOS.Services.Account.Acc;
 using Ryujinx.HLE.HOS.SystemState;

Ryujinx.HLE/HOS/Applets/Error/ErrorApplet.cs

@@ -2,10 +2,10 @@
 using LibHac.Fs;
 using LibHac.Fs.Fsa;
 using LibHac.FsSystem;
+using LibHac.Ncm;
 using LibHac.Tools.FsSystem;
 using LibHac.Tools.FsSystem.NcaUtils;
 using Ryujinx.Common.Logging;
-using Ryujinx.HLE.FileSystem;
 using Ryujinx.HLE.HOS.Services.Am.AppletAE;
 using Ryujinx.HLE.HOS.SystemState;
 using System;
@@ -106,7 +106,7 @@ namespace Ryujinx.HLE.HOS.Applets.Error
 
         private string GetMessageText(uint module, uint description, string key)
         {
-            string binaryTitleContentPath = _horizon.ContentManager.GetInstalledContentPath(ErrorMessageBinaryTitleId, StorageId.NandSystem, NcaContentType.Data);
+            string binaryTitleContentPath = _horizon.ContentManager.GetInstalledContentPath(ErrorMessageBinaryTitleId, StorageId.BuiltInSystem, NcaContentType.Data);
 
             using (LibHac.Fs.IStorage ncaFileStream = new LocalStorage(_horizon.Device.FileSystem.SwitchPathToSystemPath(binaryTitleContentPath), FileAccess.Read, FileMode.Open))
             {

Ryujinx.HLE/HOS/ApplicationLoader.cs

@@ -84,7 +84,10 @@ namespace Ryujinx.HLE.HOS
 
             MetaLoader metaData = ReadNpdm(codeFs);
 
-            _device.Configuration.VirtualFileSystem.ModLoader.CollectMods(new[] { TitleId }, _device.Configuration.VirtualFileSystem.ModLoader.GetModsBasePath());
+            _device.Configuration.VirtualFileSystem.ModLoader.CollectMods(
+                new[] { TitleId }, 
+                _device.Configuration.VirtualFileSystem.ModLoader.GetModsBasePath(), 
+                _device.Configuration.VirtualFileSystem.ModLoader.GetSdModsBasePath());
 
             if (TitleId != 0)
             {
@@ -388,7 +391,10 @@ namespace Ryujinx.HLE.HOS
 
             MetaLoader metaData = ReadNpdm(codeFs);
 
-            _device.Configuration.VirtualFileSystem.ModLoader.CollectMods(_device.Configuration.ContentManager.GetAocTitleIds().Prepend(TitleId), _device.Configuration.VirtualFileSystem.ModLoader.GetModsBasePath());
+            _device.Configuration.VirtualFileSystem.ModLoader.CollectMods(
+                _device.Configuration.ContentManager.GetAocTitleIds().Prepend(TitleId), 
+                _device.Configuration.VirtualFileSystem.ModLoader.GetModsBasePath(), 
+                _device.Configuration.VirtualFileSystem.ModLoader.GetSdModsBasePath());
 
             if (controlNca != null)
             {
@@ -481,14 +487,14 @@ namespace Ryujinx.HLE.HOS
 
                 if (result.IsSuccess() && bytesRead == controlData.ByteSpan.Length)
                 {
-                    titleName = controlData.Value.Titles[(int)device.System.State.DesiredTitleLanguage].Name.ToString();
+                    titleName = controlData.Value.Title[(int)device.System.State.DesiredTitleLanguage].NameString.ToString();
 
                     if (string.IsNullOrWhiteSpace(titleName))
                     {
-                        titleName = controlData.Value.Titles.ToArray().FirstOrDefault(x => x.Name[0] != 0).Name.ToString();
+                        titleName = controlData.Value.Title.ItemsRo.ToArray().FirstOrDefault(x => x.Name[0] != 0).NameString.ToString();
                     }
 
-                    displayVersion = controlData.Value.DisplayVersion.ToString();
+                    displayVersion = controlData.Value.DisplayVersionString.ToString();
                 }
             }
             else
@@ -615,20 +621,20 @@ namespace Ryujinx.HLE.HOS
 
                                 ref ApplicationControlProperty nacp = ref ControlData.Value;
 
-                                programInfo.Name = nacp.Titles[(int)_device.System.State.DesiredTitleLanguage].Name.ToString();
+                                programInfo.Name = nacp.Title[(int)_device.System.State.DesiredTitleLanguage].NameString.ToString();
 
                                 if (string.IsNullOrWhiteSpace(programInfo.Name))
                                 {
-                                    programInfo.Name = nacp.Titles.ToArray().FirstOrDefault(x => x.Name[0] != 0).Name.ToString();
+                                    programInfo.Name = nacp.Title.ItemsRo.ToArray().FirstOrDefault(x => x.Name[0] != 0).NameString.ToString();
                                 }
 
                                 if (nacp.PresenceGroupId != 0)
                                 {
                                     programInfo.ProgramId = nacp.PresenceGroupId;
                                 }
-                                else if (nacp.SaveDataOwnerId.Value != 0)
+                                else if (nacp.SaveDataOwnerId != 0)
                                 {
-                                    programInfo.ProgramId = nacp.SaveDataOwnerId.Value;
+                                    programInfo.ProgramId = nacp.SaveDataOwnerId;
                                 }
                                 else if (nacp.AddOnContentBaseId != 0)
                                 {
@@ -776,14 +782,14 @@ namespace Ryujinx.HLE.HOS
                 // The set sizes don't actually matter as long as they're non-zero because we use directory savedata.
                 control.UserAccountSaveDataSize = 0x4000;
                 control.UserAccountSaveDataJournalSize = 0x4000;
-                control.SaveDataOwnerId = applicationId;
+                control.SaveDataOwnerId = applicationId.Value;
 
                 Logger.Warning?.Print(LogClass.Application,
                     "No control file was found for this game. Using a dummy one instead. This may cause inaccuracies in some games.");
             }
 
             HorizonClient hos = _device.System.LibHacHorizonManager.RyujinxClient;
-            Result resultCode = hos.Fs.EnsureApplicationCacheStorage(out _, out _, applicationId, ref control);
+            Result resultCode = hos.Fs.EnsureApplicationCacheStorage(out _, out _, applicationId, in control);
 
             if (resultCode.IsFailure())
             {
@@ -792,7 +798,7 @@ namespace Ryujinx.HLE.HOS
                 return resultCode;
             }
 
-            resultCode = EnsureApplicationSaveData(hos.Fs, out _, applicationId, ref control, ref user);
+            resultCode = hos.Fs.EnsureApplicationSaveData(out _, applicationId, in control, in user);
 
             if (resultCode.IsFailure())
             {

Ryujinx.HLE/HOS/Horizon.cs

@@ -10,7 +10,7 @@ using Ryujinx.Audio.Integration;
 using Ryujinx.Audio.Output;
 using Ryujinx.Audio.Renderer.Device;
 using Ryujinx.Audio.Renderer.Server;
-using Ryujinx.HLE.FileSystem.Content;
+using Ryujinx.HLE.FileSystem;
 using Ryujinx.HLE.HOS.Kernel;
 using Ryujinx.HLE.HOS.Kernel.Memory;
 using Ryujinx.HLE.HOS.Kernel.Process;
@@ -238,6 +238,7 @@ namespace Ryujinx.HLE.HOS
             SurfaceFlinger = new SurfaceFlinger(device);
 
             InitializeAudioRenderer();
+            InitializeServices();
         }
 
         private void InitializeAudioRenderer()
@@ -288,7 +289,7 @@ namespace Ryujinx.HLE.HOS
             AudioManager.Start();
         }
 
-        public void InitializeServices()
+        private void InitializeServices()
         {
             SmServer = new ServerBase(KernelContext, "SmServer", () => new IUserInterface(KernelContext));
 
@@ -466,7 +467,7 @@ namespace Ryujinx.HLE.HOS
 
                 AudioRendererManager.Dispose();
 
-                LibHacHorizonManager.AmClient.Fs.UnregisterProgram(LibHacHorizonManager.ApplicationClient.Os.GetCurrentProcessId().Value);
+                LibHacHorizonManager.PmClient.Fs.UnregisterProgram(LibHacHorizonManager.ApplicationClient.Os.GetCurrentProcessId().Value).ThrowIfFailure();
                 
                 KernelContext.Dispose();
             }

Ryujinx.HLE/HOS/Kernel/Process/KProcess.cs

@@ -751,7 +751,8 @@ namespace Ryujinx.HLE.HOS.Kernel.Process
         {
             KThread currentThread = KernelStatic.GetCurrentThread();
 
-            if (currentThread.Owner != null &&
+            if (currentThread.Context.Running &&
+                currentThread.Owner != null &&
                 currentThread.GetUserDisableCount() != 0 &&
                 currentThread.Owner.PinnedThreads[currentThread.CurrentCore] == null)
             {

Ryujinx.HLE/HOS/Kernel/Threading/KThread.cs

@@ -658,10 +658,7 @@ namespace Ryujinx.HLE.HOS.Kernel.Threading
 
         private static uint GetPsr(ARMeilleure.State.ExecutionContext context)
         {
-            return (context.GetPstateFlag(ARMeilleure.State.PState.NFlag) ? (1U << (int)ARMeilleure.State.PState.NFlag) : 0U) |
-                   (context.GetPstateFlag(ARMeilleure.State.PState.ZFlag) ? (1U << (int)ARMeilleure.State.PState.ZFlag) : 0U) |
-                   (context.GetPstateFlag(ARMeilleure.State.PState.CFlag) ? (1U << (int)ARMeilleure.State.PState.CFlag) : 0U) |
-                   (context.GetPstateFlag(ARMeilleure.State.PState.VFlag) ? (1U << (int)ARMeilleure.State.PState.VFlag) : 0U);
+            return context.Pstate & 0xFF0FFE20;
         }
 
         private ThreadContext GetCurrentContext()
@@ -1371,7 +1368,7 @@ namespace Ryujinx.HLE.HOS.Kernel.Threading
 
             PreferredCore = _originalPreferredCore;
             AffinityMask = _originalAffinityMask;
-            
+
             if (AffinityMask != affinityMask)
             {
                 if ((AffinityMask & 1UL << ActiveCore) != 0)

Ryujinx.HLE/HOS/LibHacHorizonManager.cs

@@ -24,6 +24,7 @@ namespace Ryujinx.HLE.HOS
         public HorizonClient BcatClient        { get; private set; }
         public HorizonClient FsClient          { get; private set; }
         public HorizonClient NsClient          { get; private set; }
+        public HorizonClient PmClient          { get; private set; }
         public HorizonClient SdbClient         { get; private set; }
 
         private SharedRef<LibHacIReader> _arpIReader;
@@ -65,6 +66,7 @@ namespace Ryujinx.HLE.HOS
 
         public void InitializeSystemClients()
         {
+            PmClient      = Server.CreatePrivilegedHorizonClient();
             AccountClient = Server.CreateHorizonClient(new ProgramLocation(SystemProgramId.Account, StorageId.BuiltInSystem), AccountFsPermissions);
             AmClient      = Server.CreateHorizonClient(new ProgramLocation(SystemProgramId.Am,      StorageId.BuiltInSystem), AmFsPermissions);
             NsClient      = Server.CreateHorizonClient(new ProgramLocation(SystemProgramId.Ns,      StorageId.BuiltInSystem), NsFsPermissions);

Ryujinx.HLE/HOS/ModLoader.cs

@@ -136,7 +136,8 @@ namespace Ryujinx.HLE.HOS
 
         private static bool StrEquals(string s1, string s2) => string.Equals(s1, s2, StringComparison.OrdinalIgnoreCase);
 
-        public string GetModsBasePath() => EnsureBaseDirStructure(AppDataManager.GetModsPath());
+        public string GetModsBasePath()   => EnsureBaseDirStructure(AppDataManager.GetModsPath());
+        public string GetSdModsBasePath() => EnsureBaseDirStructure(AppDataManager.GetSdModsPath());
 
         private string EnsureBaseDirStructure(string modsBasePath)
         {
@@ -695,7 +696,7 @@ namespace Ryujinx.HLE.HOS
 
             var buildIds = programs.Select(p => p switch
             {
-                NsoExecutable nso => BitConverter.ToString(nso.BuildId.Bytes.ToArray()).Replace("-", "").TrimEnd('0'),
+                NsoExecutable nso => BitConverter.ToString(nso.BuildId.ItemsRo.ToArray()).Replace("-", "").TrimEnd('0'),
                 NroExecutable nro => BitConverter.ToString(nro.Header.BuildId).Replace("-", "").TrimEnd('0'),
                 _ => string.Empty
             }).ToList();

Ryujinx.HLE/HOS/ProgramLoader.cs

@@ -160,7 +160,7 @@ namespace Ryujinx.HLE.HOS
 
             var buildIds = executables.Select(e => (e switch
             {
-                NsoExecutable nso => BitConverter.ToString(nso.BuildId.Bytes.ToArray()),
+                NsoExecutable nso => BitConverter.ToString(nso.BuildId.ItemsRo.ToArray()),
                 NroExecutable nro => BitConverter.ToString(nro.Header.BuildId),
                 _ => ""
             }).Replace("-", "").ToUpper());

Ryujinx.HLE/HOS/Services/Account/Acc/AccountManager.cs

@@ -3,6 +3,7 @@ using LibHac.Common;
 using LibHac.Fs;
 using LibHac.Fs.Shim;
 using Ryujinx.Common;
+using Ryujinx.Common.Logging;
 using System;
 using System.Collections.Concurrent;
 using System.Collections.Generic;
@@ -25,7 +26,7 @@ namespace Ryujinx.HLE.HOS.Services.Account.Acc
 
         public UserProfile LastOpenedUser { get; private set; }
 
-        public AccountManager(HorizonClient horizonClient)
+        public AccountManager(HorizonClient horizonClient, string initialProfileName = null)
         {
             _horizonClient = horizonClient;
 
@@ -43,7 +44,14 @@ namespace Ryujinx.HLE.HOS.Services.Account.Acc
             }
             else
             {
-                OpenUser(_accountSaveDataManager.LastOpened);
+                UserId commandLineUserProfileOverride = default; 
+                if (!string.IsNullOrEmpty(initialProfileName))
+                { 
+                    commandLineUserProfileOverride = _profiles.Values.FirstOrDefault(x => x.Name == initialProfileName)?.UserId ?? default;
+                    if (commandLineUserProfileOverride.IsNull)
+                        Logger.Warning?.Print(LogClass.Application, $"The command line specified profile named '{initialProfileName}' was not found");
+                }
+                OpenUser(commandLineUserProfileOverride.IsNull ? _accountSaveDataManager.LastOpened : commandLineUserProfileOverride);
             }
         }
 
@@ -168,8 +176,8 @@ namespace Ryujinx.HLE.HOS.Services.Account.Acc
 
         private void DeleteSaveData(UserId userId)
         {
-            SaveDataFilter saveDataFilter = new SaveDataFilter();
-            saveDataFilter.SetUserId(new LibHac.Fs.UserId((ulong)userId.High, (ulong)userId.Low));
+            var saveDataFilter = SaveDataFilter.Make(programId: default, saveType: default,
+                new LibHac.Fs.UserId((ulong)userId.High, (ulong)userId.Low), saveDataId: default, index: default);
 
             using var saveDataIterator = new UniqueRef<SaveDataIterator>();
 

Ryujinx.HLE/HOS/Services/Account/Acc/IAccountServiceForApplication.cs

@@ -169,7 +169,7 @@ namespace Ryujinx.HLE.HOS.Services.Account.Acc
             // TODO: Account actually calls nn::arp::detail::IReader::GetApplicationControlProperty() with the current Pid and store the result (NACP file) internally.
             //       But since we use LibHac and we load one Application at a time, it's not necessary.
 
-            context.ResponseData.Write(context.Device.Application.ControlData.Value.UserAccountSwitchLock);
+            context.ResponseData.Write((byte)context.Device.Application.ControlData.Value.UserAccountSwitchLock);
 
             Logger.Stub?.PrintStub(LogClass.ServiceAcc);
 

Ryujinx.HLE/HOS/Services/Am/AppletOE/ApplicationProxyService/ApplicationProxy/IApplicationFunctions.cs

@@ -131,7 +131,7 @@ namespace Ryujinx.HLE.HOS.Services.Am.AppletOE.ApplicationProxyService.Applicati
             }
 
             HorizonClient hos = context.Device.System.LibHacHorizonManager.AmClient;
-            Result result = EnsureApplicationSaveData(hos.Fs, out long requiredSize, applicationId, ref control, ref userId);
+            Result result = hos.Fs.EnsureApplicationSaveData(out long requiredSize, applicationId, in control, in userId);
 
             context.ResponseData.Write(requiredSize);
 
@@ -148,7 +148,7 @@ namespace Ryujinx.HLE.HOS.Services.Am.AppletOE.ApplicationProxyService.Applicati
             // TODO: When above calls are implemented, switch to using ns:am
 
             long desiredLanguageCode = context.Device.System.State.DesiredLanguageCode;
-            int  supportedLanguages  = (int)context.Device.Application.ControlData.Value.SupportedLanguages;
+            int  supportedLanguages  = (int)context.Device.Application.ControlData.Value.SupportedLanguageFlag;
             int  firstSupported      = BitOperations.TrailingZeroCount(supportedLanguages);
 
             if (firstSupported > (int)SystemState.TitleLanguage.BrazilianPortuguese)
@@ -190,7 +190,6 @@ namespace Ryujinx.HLE.HOS.Services.Am.AppletOE.ApplicationProxyService.Applicati
         // GetDisplayVersion() -> nn::oe::DisplayVersion
         public ResultCode GetDisplayVersion(ServiceCtx context)
         {
-            // This should work as DisplayVersion U8Span always gives a 0x10 size byte array.
             // If an NACP isn't found, the buffer will be all '\0' which seems to be the correct implementation.
             context.ResponseData.Write(context.Device.Application.ControlData.Value.DisplayVersion);
 
@@ -252,7 +251,7 @@ namespace Ryujinx.HLE.HOS.Services.Am.AppletOE.ApplicationProxyService.Applicati
             BlitStruct<ApplicationControlProperty> controlHolder = context.Device.Application.ControlData;
 
             Result result = _horizon.Fs.CreateApplicationCacheStorage(out long requiredSize,
-                out CacheStorageTargetMedia storageTarget, applicationId, ref controlHolder.Value, index, saveSize,
+                out CacheStorageTargetMedia storageTarget, applicationId, in controlHolder.Value, index, saveSize,
                 journalSize);
 
             if (result.IsFailure()) return (ResultCode)result.Value;

Ryujinx.HLE/HOS/Services/Arp/ApplicationLaunchProperty.cs

@@ -1,4 +1,4 @@
-using Ryujinx.HLE.FileSystem;
+using LibHac.Ncm;
 
 namespace Ryujinx.HLE.HOS.Services.Arp
 {
@@ -20,7 +20,7 @@ namespace Ryujinx.HLE.HOS.Services.Arp
                 {
                     TitleId             = 0x00,
                     Version             = 0x00,
-                    BaseGameStorageId   = (byte)StorageId.NandSystem,
+                    BaseGameStorageId   = (byte)StorageId.BuiltInSystem,
                     UpdateGameStorageId = (byte)StorageId.None
                 };
             }
@@ -35,7 +35,7 @@ namespace Ryujinx.HLE.HOS.Services.Arp
             {
                 TitleId             = context.Device.Application.TitleId,
                 Version             = 0x00,
-                BaseGameStorageId   = (byte)StorageId.NandSystem,
+                BaseGameStorageId   = (byte)StorageId.BuiltInSystem,
                 UpdateGameStorageId = (byte)StorageId.None
             };
         }

Ryujinx.HLE/HOS/Services/Arp/LibHacIReader.cs

@@ -16,7 +16,7 @@ namespace Ryujinx.HLE.HOS.Services.Arp
         {
             launchProperty = new LibHac.Arp.ApplicationLaunchProperty
             {
-                BaseStorageId = StorageId.BuiltInUser,
+                StorageId = StorageId.BuiltInUser,
                 ApplicationId = ApplicationId
             };
 
@@ -29,7 +29,7 @@ namespace Ryujinx.HLE.HOS.Services.Arp
         {
             launchProperty = new LibHac.Arp.ApplicationLaunchProperty
             {
-                BaseStorageId = StorageId.BuiltInUser,
+                StorageId = StorageId.BuiltInUser,
                 ApplicationId = applicationId
             };
 

Ryujinx.HLE/HOS/Services/Fs/FileSystemProxy/FileSystemProxyHelper.cs

@@ -132,22 +132,10 @@ namespace Ryujinx.HLE.HOS.Services.Fs.FileSystemProxy
             }
         }
 
-        public static Result ReadFsPath(out FsPath path, ServiceCtx context, int index = 0)
-        {
-            ulong position = context.Request.PtrBuff[index].Position;
-            ulong size     = context.Request.PtrBuff[index].Size;
-
-            byte[] pathBytes = new byte[size];
-
-            context.Memory.Read(position, pathBytes);
-
-            return FsPath.FromSpan(out path, pathBytes);
-        }
-
         public static ref readonly FspPath GetFspPath(ServiceCtx context, int index = 0)
         {
-            ulong position = (ulong)context.Request.PtrBuff[index].Position;
-            ulong size = (ulong)context.Request.PtrBuff[index].Size;
+            ulong position = context.Request.PtrBuff[index].Position;
+            ulong size = context.Request.PtrBuff[index].Size;
 
             ReadOnlySpan<byte> buffer = context.Memory.GetSpan(position, (int)size);
             ReadOnlySpan<FspPath> fspBuffer = MemoryMarshal.Cast<byte, FspPath>(buffer);
@@ -157,8 +145,8 @@ namespace Ryujinx.HLE.HOS.Services.Fs.FileSystemProxy
 
         public static ref readonly LibHac.FsSrv.Sf.Path GetSfPath(ServiceCtx context, int index = 0)
         {
-            ulong position = (ulong)context.Request.PtrBuff[index].Position;
-            ulong size = (ulong)context.Request.PtrBuff[index].Size;
+            ulong position = context.Request.PtrBuff[index].Position;
+            ulong size = context.Request.PtrBuff[index].Size;
 
             ReadOnlySpan<byte> buffer = context.Memory.GetSpan(position, (int)size);
             ReadOnlySpan<LibHac.FsSrv.Sf.Path> pathBuffer = MemoryMarshal.Cast<byte, LibHac.FsSrv.Sf.Path>(buffer);

Ryujinx.HLE/HOS/Services/Fs/IDeviceOperator.cs

@@ -1,6 +1,6 @@
 using LibHac;
 using LibHac.Common;
-using LibHac.FsSrv;
+using LibHac.Fs;
 
 namespace Ryujinx.HLE.HOS.Services.Fs
 {